Converting galactose into the rare sugar talose with cellobiose 2-epimerase as biocatalyst

Article abstract of DOI:10.3390/molecules23102519

Cellobiose 2-epimerase from Rhodothermus marinus (RmCE) reversibly converts a glucose residue to a mannose residue at the reducing end of β-1,4-linked oligosaccharides. In this study, the monosaccharide specificity of RmCE has been mapped and the synthesis of D-talose from D-galactose was discovered, a reaction not yet known to occur in nature. Moreover, the conversion is industrially relevant, as talose and its derivatives have been reported to possess important antimicrobial and anti-inflammatory properties. As the enzyme also catalyzes the keto-aldo isomerization of galactose to tagatose as a minor side reaction, the purity of talose was found to decrease over time. After process optimization, 23 g/L of talose could be obtained with a product purity of 86% and a yield of 8.5% (starting from 4 g (24 mmol) of galactose). However, higher purities and concentrations can be reached by decreasing and increasing the reaction time, respectively. In addition, two engineering attempts have also been performed. First, a mutant library of RmCE was created to try and increase the activity on monosaccharide substrates. Next, two residues from RmCE were introduced in the cellobiose 2-epimerase from Caldicellulosiruptor saccharolyticus (CsCE) (S99M/Q371F), increasing the kcat twofold.

Full text of DOI:10.3390/molecules23102519

molecules
Article
Converting Galactose into the Rare Sugar Talose with
Cellobiose 2-Epimerase as Biocatalyst
Stevie Van Overtveldt, Ophelia Gevaert, Martijn Cherlet, Koen Beerens and Tom Desmet *
Centre for Synthetic Biology, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000 Gent,
Belgium; stevie.vanovertveldt@ugent.be (S.V.O.); ophelia.gevaert@ugent.be (O.G.);
martijn.cherlet@telenet.be (M.C.); koen.beerens@ugent.be (K.B.)
* Correspondence: tom.desmet@ugent.be; Tel.: +32-9264-9920
ꢀꢁꢂꢀꢃꢄꢅꢆꢇ
Academic Editors: Giorgia Oliviero and Nicola Borbone
Received: 17 September 2018; Accepted: 29 September 2018; Published: 1 October 2018
ꢀꢁꢂꢃꢄꢅꢆ
Abstract: Cellobiose 2-epimerase from Rhodothermus marinus (RmCE) reversibly converts a glucose
residue to a mannose residue at the reducing end of β-1,4-linked oligosaccharides. In this study,
the monosaccharide specificity of RmCE has been mapped and the synthesis of D-talose from
D-galactose was discovered, a reaction not yet known to occur in nature. Moreover, the conversion
is industrially relevant, as talose and its derivatives have been reported to possess important
antimicrobial and anti-inflammatory properties. As the enzyme also catalyzes the keto-aldo
isomerization of galactose to tagatose as a minor side reaction, the purity of talose was found
to decrease over time. After process optimization, 23 g/L of talose could be obtained with a product
purity of 86% and a yield of 8.5% (starting from 4 g (24 mmol) of galactose). However, higher purities
and concentrations can be reached by decreasing and increasing the reaction time, respectively.
In addition, two engineering attempts have also been performed. First, a mutant library of RmCE
was created to try and increase the activity on monosaccharide substrates. Next, two residues from
RmCE were introduced in the cellobiose 2-epimerase from Caldicellulosiruptor saccharolyticus (CsCE)
(S99M/Q371F), increasing the kcat twofold.
Keywords: biocatalysis; cellobiose 2-epimerase; rare sugars; talose; enzyme engineering
1. Introduction
Carbohydrate epimerases are a subclass of isomerases capable of inverting the configuration of
hydroxyl groups of both substituted (phosphate and nucleotide diphosphate) and unsubstituted
sugars [
by the growing number of characterized enzymes and industrial applications [
3-epimerase, for example, plays an important role in the production of D-psicose (showing potential as
no-calorie sweetener and antidiabetic [ ]) from sucrose where it is combined with two other enzymes
in a one-pot cascade reaction [ ]. Another example is cellobiose 2-epimerase, which catalyzes the
1]. During recent years, more attention has been drawn towards the latter, evidenced
2–7]. D-tagatose
8
9
interconversion of glucose to mannose at the reducing end of disaccharides (e.g., cellobiose and lactose)
and oligosaccharides (e.g., cellotriose and cellotetraose) [10]. Common applications are the production
of epilactose starting from milk ultrafiltrate containing lactose or the production of lactulose from whey
powder. Epilactose is regarded as a new prebiotic and lactulose is a potential low-calorie sweetener
that could substitute for lactose in the food industry [11,12].
Even though cellobiose 2-epimerase enzymes have a rather relaxed substrate specificity, their
activity is typically limited to oligosaccharides [
saccharolyticus (CsCE) has been shown to also accept monosaccharides as substrate, a feature that was
exploited for the production of mannose from glucose [ ]. However, similar enzymes might still exist
as not all current representatives have been screened for their monosaccharide activity and, therefore,
5]. Until today, only the enzyme from Caldicellulosiruptor
4
Molecules 2018, 23, 2519; doi:10.3390/molecules23102519
www.mdpi.com/journal/molecules

Molecules 2018, 23, 2519
2 of 15
it is possible that interesting epimerase reactions might have gone unnoticed. This paper focuses on
the cellobiose 2-epimerase from Rhodothermus marinus, which has already been characterized regarding
activity on oligosaccharides and has displayed the ability to produce epil◦actose from lactose efficiently.
Furthermore, the enzyme was shown to be stable at temperatures up to 80 C, making it highly attractive
for potential production processes [
with a particular emphasis on the C2 epimerization of galactose to talose, a rare sugar with potential
applications in the pharmaceutical industry [13 16]. It can, for example, serve as a precursor molecule
5]. In this research, the substrate specificity of RmCE was explored
–
for several antibiotics of which the best example is Caminoside A, an antimicrobial compound isolated
from the sponge Caminus sphaeroconia, containing a 6-deoxytalose group [13]. Moreover, the O2 and
O3-methylated forms have been shown to be submillimolar inhibitors of galactose-binding galectin-4 and
galectin-8, proteins which are involved in inflammation and cancer [17].
Currently, talose can only be produced by a rather inefficient chemical process, requiring either
high amounts of dibutyltin oxide [18] or five consecutive chemical steps and a lot of solvents [19].
Therefore, an efficient biochemical route would provide a clear added value to the production of talose.
2. Results
2.1. Substrate Specificity Analysis of Rhodothermus marinus (RmCE) Revealed One-Step Conversion from
D-Galactose to D-Talose
The specificity of the target cellobiose 2-epimerase was evaluated on a range of substrates,
comprising D-hexoses, D-pentoses and an L-pentose (Table 1). The analysis revealed that RmCE is
able to accept various monosaccharides, with a specific activity that is about 1% of that on cellobiose.
Interestingly, the activity on galactose was found to be similar to that of mannose, which can be
regarded as the naturally preferred monosaccharide. However, further kinetic analysis showed that
the affinity for galactose is considerably lower (Km of 1.04 and 0.45 M, respectively) whereas the kcat is
roughly the same (5 s−1) on both of these substrates (Table 2). While the mannose to glucose reaction
was already described for the CsCE enzyme, it is the first time that a one-enzyme conversion from
galactose to the rare sugar talose is demonstrated.
Table 1. Substrate specificity of Rhodothermus marinus (RmCE). Specific activities were determined on
◦
200 mM substrate at pH 6.3 and 70 C. All tests were performed in triplicate and the standard deviation
is reported for every substrate. No activity could be detected on L-arabinose.
RmCE Specific Activity
CsCE Specific Activity
Substrate
Product
(mU/mg)
(mU/mg) 1
87,500 2
1130 ± 25
440 ± 24
1020 ± 83
38 ± 5
18 ± 0.2
/
Cellobiose
D-mannose
D-glucose
D-galactose
D-xylose
4-O-
β
-D-glucopyranosyl-D-mannose
3600 ± 2.1
150 ± 0.5
35 ± 0.5
/
40 ± 0.7
60 ± 0.3
/
D-glucose
D-mannose
D-talose
D-lyxose
D-xylose
D-lyxose
L-arabinose
L-ribose
1
2
Data represented as described in Park et al [4]. Data obtained from Ojima et al [5].
Table 2. Kinetic parameters of cellobiose 2-epimerase from R. marinus for D-mannose and D-galactose.
Tests were performed at pH 6.3 and 70 ◦C. Data represent the mean of three experiments with
standard error.
−1
−1
)
Substrate
Km (M)
kcat (s
5.4 ± 1
5.0 ± 0.3
)
k
cat/Km (s−1/M
D-mannose
D-galactose
0.45 ± 0.18
1.04 ± 0.13
12.0 ± 4.4
4.8 ± 0.7

Molecules 2018, 23, 2519
3 of 15
2.2. Production of D-Talose
During prolonged incubations with galactose, the keto-aldo isomer tagatose (Figure 1) also started
to appear (Figure 2), as was observed during the production of mannose from glucose by CsCE [ ].
4
As a consequence, the purity of the produced talose slowly decreased during the reaction, with the final
ratio of galactose, tagatose and talose being 53:28:19 at equilibrium, demonstrating that galactose is the
most stable isomer. Interestingly, the talose equilibrium was reached first, while the concentration of
tagatose kept increasing until about 28% yield was reached. The presence of the isomerization product
tagatose was also spotted during the kinetic analysis (see previous paragraph) but was neglibly low
and, therefore, not incorporated in the calculations.
O
O
H
HO
HO
H
OH
HO
HO
HO
H
H
H
H
H
H
OH
OH
OH
OH
Galactose
Talose
OH
O
HO
HO
H
H
H
OH
OH
Tagatose
Figure 1. Galactose to talose conversion by RmCE also generates the isomerization product tagatose.
1600
Galactose
Talose
Tagatose
1400
1200
1000
800
400
200
0
0
20
125
Time (h)
Figure 2. Time-course representation of galactose to talose/tagatose conversion. The reaction was
performed with 1.6 M of galactose at pH 6.3 and monitored for 27 h at 70 ◦C. During the first hours,
talose was being formed time-dependently, reaching a maximum conversion of about 20% after
overnight incubation. Formation of tagatose seemed to lag in the beginning of the reaction but
increases linearly after one hour. The tagatose equilibrium was only reached after 120 h, during which
extra enzyme was added to speed up the reaction. Data represent the means of three experiments with
standard deviation (until 18 h).
In order to find the optimal conditions for the production of talose, various substrate
concentrations were evaluated and the product purity (expressed as the amount of talose to the
total amount of talose and tagatose) was monitored over time (Table S1). High substrate concentrations

Molecules 2018, 23, 2519
4 of 15
were found to yield the highest purity during longer incubation times, as it takes longer to reach the
talose equilibrium. Moreover, these substrate concentrations are the most attractive as they also give
rise to the highest product concentrations.
When focusing on the reaction profile of 1.6 M galactose, it becomes clear that purity is >99%
during the first hour and then drops during the next hours. However, the total amount of talose keeps
increasing during the course of the reaction. As a consequence, the optimal reaction time is dependent
on the specific wishes of talose production. If a very pure product is desired for pharmaceutical
applications, a shorter incubation time is preferred. In contrast, if purity is less important, the reaction
time could be extended to maximize product yield (Table 3).
Table 3. Time-course representation of galactose to talose conversion depicting product purity, yield
and concentration. Tests were performed at a substrate concentration of 1.6 M and with an enzyme
concentration of 0.3 mg/mL. Data represent the means of three experiments with standard deviation.
Time (h)
Purity (%)
Yield (%)
D-talose (mM)
0.5
1
1.5
3
4.5
7
18
>99
>99
0.50 ± 0.30
1.90 ± 0.16
2.68 ± 0.23
5.43 ± 0.08
8.97 ± 0.66
12.01 ± 0.82
19.28 ± 1.28
8.07 ± 4.85
30.36 ± 2.49
42.81 ± 3.76
86.90 ± 1.36
143.50 ± 10.50
192.14 ± 13.18
308.40 ± 20.44
97.01 ± 2.59
93.65 ± 1.92
92.47 ± 1.26
91.22 ± 0.86
89.78 ± 1.53
Next, the production of talose was performed at a larger scale, starting from 288 g/L of galactose.
This generated 24.3 g/L of talose after 4.5 h, which corresponds to a molar yield of about 8.5%.
The relatively low degree of conversion is mainly caused by the superior stability of galactose,
as reflected by these compounds’ equilibrium ratio (see above). However, the remaining galactose
could simply be removed by a yeast treatment with Kluyveromyces marxianus (Figure S1) [20
This selective digestion of a 2-epimer by micro-organisms has already been demonstrated in the
production of N-glycolylneuraminic acid [22 23]. After 96 hours incubation, less than 1% of galactose
,21].
,
was still present, and 23 g/L (1.92 mmol) of talose could be recovered at a purity of 86% (containing
12% of tagatose and trace amounts of galactose and the yeast metabolite glycerol) (Figure S2).
2.3. Engineering of RmCE to Increase Activity on Monosaccharide Substrates
Given the clear application potential of the R. marinus cellobiose 2-epimerase, the opportunity
to increase the affinity towards monosaccharides by enzyme engineering was explored. Mannose to
glucose conversion was used as case study, because the detection of glucose can be performed in more
high-throughput manner by making use of the glucose oxidase-peroxidase (GOD-POD) assay [24].
The active site was explored and four positions appeared interesting for mutagenesis: Tyr-307 and
Trp-321 had no important function in catalysis but were nonetheless oriented towards the substrate;
Tyr-124 already makes a hydrogen bond with the O2 of mannose but might still be improved; Trp-385
displays a stacking interaction with the non-reducing end of cellobiose, but since this part is not present
when mannose is supplied as substrate, this provides an opportunity to create a new interaction
(Figure 3A). Simultaneously saturating all four positions would leave us with the impossible task of
screening roughly half a million mutants (3
considered, as it is able to seriously diminish library size by suggesting mutations that increase affinity
between enzyme and substrate [26–28].
×
204) [25]. Therefore, the Rosetta software suite was

Molecules 2018, 23, 2519
5 of 15
A
Y124
Y307
W 385
W 321
Shortsim ulation
Long sim ulation
B
Y124
Y307
W 321
W 385
Y124
Y307
W 321
W 385
Figure 3. (A) Active site representation of RmCE (PDB code: 3WKG) displaying the four residues (Tyr-124,
Tyr-307, Trp-321, Trp-385) that were targeted for mutagenesis [10]. The substrate glucosylmannose is
colored by element cyan (carbon in cyan, oxygen in red) while the residues were colored by element green
(carbon in green, oxygen in red, nitrogen in blue). (B) Sequence motifs for both the short simulation and
long simulation mutant library. Motifs were created by the online software Weblogo [29].
Two separate Rosetta analyses were performed, the only difference being the starting structures.
They both originated from snapshots of molecular dynamics simulations with a duration of 0.1 ns
and 5 ns, respectively. The corresponding mutant libraries were then designed based on amino
acids occurring in more than 5% of the sequences on a given position (Figure S3). Strikingly, some
differences occur between the motives (Figure 3B), evidencing that different simulation times can
provide alternative amino acids on positions of interest. However, neither of these libraries delivered
improved mutants. The percentage of inactive mutants was comparable between both libraries.
2.4. Introducing Two RmCE Residues in Caldicellulosiruptor saccharolyticus (CsCE) Led to A Twofold
kcat Increase
As mentioned earlier, the only other cellobiose 2-epimerase showing significant activity towards
monosaccharides is CsCE. When comparing the substrate specificities of these enzymes, it becomes
clear that some differences occur: CsCE displays a higher activity towards D-xylose and D-lyxose and
RmCE is active on galactose, an activity not yet reported for CsCE. Moreover, the kcat of RmCE for
mannose is significantly higher than that of CsCE (5.4 and 0.7 s−1, respectively) [4,5].
Analyzing the crystal structures of RmCE and CsCE revealed that one of the common first shell
residues is oriented differently: in RmCE, Trp-385 (in CsCE Trp-372) is closer to the non-reducing
end of cellobiose. This could be explained by the presence of the bulky amino acid Phe-384, altering
the angle of Trp-385 towards the glucose moiety. This in turn seems correlated to the neighbouring
amino acid Met-109, orienting Phe-384 (Figure 4). To validate whether the orientation of Trp-385 had
any influence on the activity, a CsCE double mutant S99M/Q371F was created that displayed two

Molecules 2018, 23, 2519
6 of 15
times higher specific activity towards mannose than the wild-type CsCE. Kinetic analysis showed
that this could be attributed to an increase in kcat of the mutant (7.5 versus 3.8 s−1), while the Km was
not significantly affected (Table 4). This can be explained by the higher kcat value of wild-type RmCE,
which apparently can be transferred towards CsCE by mutating these two residues. However, it is not
clear whether other residues can still increase the kcat of CsCE: according to our measurements the kcat
of CsCE is only slightly lower than RmCE, but literature reports a significantly lower kcat (0.7 s−1) and
K
m (52 mM) for CsCE [4,5].
W 372
W 385
S 99
F384
M 109
Q 371
Figure 4. The catalytic Trp residue in the active sites of RmCE (carbon in green, nitrogen in blue, sulfur
in yellow) (PDB code: 3WKG) and CsCE (carbon in yellow, nitrogen in blue, oxygen in red) (PDB code:
4Z4J) has a slightly different orientation towards the glucose moiety of glucosylmannose (carbon in
cyan, oxygen in red). In the active site of RmCE, Trp-385 is oriented closer to glucose, probably due to
the accompanying resides Phe-384 and Met-109.
Table 4. Kinetic parameters of cellobiose 2-epimerase from C. saccharolyticus for D-mannose, both for
◦
wild-type and double mutant S99M/Q371F. Tests were performed at pH 7.5 and 70 C. Data represent
the mean of three experiments with standard error.
−1
)
CsCE
Km (mM)
kcat (s
WT
S99M/Q371F
412 ± 91
430 ± 94
3.8 ± 0.2
7.5 ± 0.5
3. Discussion
Most cellobiose 2-epimerase enzymes are characterized by a broad substrate specificity meaning
that they can be used for the production of several important molecules (epilactose, mannose,
lactulose) from their cheaper counterparts [4,11,12,30–32]. The specificity of cellobiose 2-epimerase
from Rhodothermus marinus was explored further and revealed side-activity on several monosaccharides
comprising the conversion of D-glucose to D-mannose (comparable to CsCE) and D-galactose to D-talose
(a new reaction). As a consequence, the previous assumption that CE enzymes are only active on
β-1,4-substituted sugars might be questioned and more cellobiose 2-epimerases might still exist in
nature that exhibit monosaccharide activity. In order to further clarify the activity of CE enzymes on
monosaccharides, galactose was docked in the active site of RmCE and polar contacts were evaluated
(Figure S4). Interestingly, the residues responsible for binding of glucose (reducing end of cellobiose)
are the same as for galactose, so no clear determinant for monosaccharide activity could be discovered.
While all CE enzymes are probably able to recognize monosaccharides in their subsite, it remains
unclear why only some representatives display significant activity. Moreover, the monosaccharide

Molecules 2018, 23, 2519
7 of 15
specificity among these representatives differs, as evidenced by the activity on galactose by RmCE
which was not present in the case of CsCE.
Interestingly, the kcat of the galactose to talose reaction turned out to be relatively high, even
though it is only considered a minor activity. The combination of a medium-high activity, a stable
enzyme and a cheap substrate led us to believe that a viable production process was possible.
Moreover, the product talose, its derivatives and glycoconjugates are valuable molecules with several
important industrial applications, like anti-tumor [15] and antimicrobial [33,34] activities as well as
recrystallization-inhibition (RI) activity [35]. The only problem in the entire process was the by-product
tagatose that was being formed through the isomerization side reaction of the enzyme, similar to
the formation of fructose in the production of mannose by CsCE [4]. Therefore, reaction and process
optimization were performed to chart the purity and yield of the product by varying substrate
concentration and reaction time. This revealed that product purity remained >99% at the beginning of
the reaction, but decreased steadily afterwards. In contrast, the product yield kept increasing during
the course of the reaction, reaching a maximum yield of about 20%. Upscaling of the reaction was
performed and resulted in 23 g/L of D-talose with a purity of 86% and a yield of 8.5%. Depending on
the application, a trade-off should be considered between purity and product yield, decreasing reaction
time if the former is favored and increasing the reaction time if the latter is preferred. Alternatively,
either preparative high-performance liquid chromatography (HPLC) [36
chromatography might be used to separate monosaccharides, as established by the production of
psicose from sucrose with a three-enzyme cascade reaction [ ]. A more elegant solution might also be
,37] or simulated moving bed
9
provided by adding an additional enzyme to the reaction mixture, creating a coupled process. This has
already been successfully demonstrated in the production of N-acetylneuraminic acid, the starting
material for anti-influenza agents such as zanamivir and inavir, using an N-acyl-D-glucosamine
2-epimerase [38].
In an attempt to increase the activity on monosaccharide substrates, two mutant libraries were
constructed and in total 2700 mutants were screened for increased activity. Unfortunately, no improved
mutants could be detected, indicating that amino acid positions Tyr-124, Tyr-307, Trp-321 and Trp-385
are not ideal hotspots for gaining kcat improvements. Targeting other positions, possibly further away
from the active site, could prove to be the key to unlocking higher activities on mannose, galactose and
other monosaccharide substrates.
Literature and our own results demonstrated that the kcat of RmCE is higher than that of CsCE.
As CsCE already holds an application in the production of glucose from mannose, it was decided to try
and transfer this property from RmCE to CsCE [4]. Variant S99M/Q371F was able to accomplish this
by doubling the kcat while more or less retaining the same Km value. This shift in activity can probably
be attributed to the different orientation of Trp-385 (Trp-372 in CsCE) effectuated by the presence of
Phe-384 (Gln-371 in CsCE) and Met-109 (Ser-99 in CsCE).
In conclusion, this paper presents two enzymes that can be applied for industrial production
processes: the wild-type RmCE which can efficiently produce D-talose from the cheap substrate
D-galactose on the one hand; and a mutant CsCE with a higher kcat for the production of D-mannose
on the other hand.
4. Materials and Methods
4.1. Gene Cloning of Cellobiose 2-Epimerase
Cellobiose 2-epimerase from Rhodothermus marinus (Uniprot: F8WRK9) was ordered at Genscript
and supplied in a pUC57 vector containing a C-terminal His-tag. Next, it was cloned in a pET21a(+)
vector acquired from Novagen by making use of the Golden Gate protocol by Engler et al. [39].
The backbone was amplified with forward (50-GGT CTT CGG TCT CCG TGG CTC GCA TCA TCA
TCA T-30) and reverse primers (50-GGT CTT CGG0TCT CCA TCC CGT ATA TCT CCT TCT TAA
0
AGT-3 ) and the gene was amplified with forward (5 -GGT CTT CGG TCT CCG GAT GAG CAC CGA

Molecules 2018, 23, 2519
8 of 15
AAC GAT TCC G-30) and reverse primers (50-GGT CTT CGG TCT CCC CAC GGG AGC GAA CGT
GTT-30). Then, the resulting plasmid was transformed in E. coli BL21(DE3) and plated on lysogeny
broth (LB) agar containing 100 µg/mL of ampicillin. An ampicillin-resistant colony was selected and
plasmid DNA from the transformant was isolated with a plasmid purification kit (Analytik Jena, Jena,
Germany). Afterwards, the plasmid was checked by sequencing (LGC Genomic sequencing service in
Berlin, Germany).
4.2. Enzyme Production and Purification
The recombinant E. coli cells for the expression of RmCE were cultivated in 500 mL of terrific
broth (TB) medium in a 2 L flask containing 100 µg/mL of ampicillin. When the optical density of the
cells reached 0.6, isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of
0.1 mM and the culture was incubated shaking at 200 rpm at 30 ◦C for 16 h. The produced biomass
was harvested by centrifugation at 4 ◦C for 20 min at 4500 rpm and the resulting pellet was frozen at
◦
−
20 C for at least 16 h. For enzyme extraction and purification, cell pellets were thawed and dissolved
in 10 mL lysis buffer (300 mM NaCl, 10 mM imidazole, 0.1 mM PMSF and 50 mM sodium phosphate
buffer; pH 7.4) supplemented with lysozyme at a final concentration of 1 mg/mL. The resuspended
cells were incubated on ice for 30 min and sonicated two times for 4 min (Branson Sonifier 250, level 3,
50% duty cycle). Next, cell lysate was centrifuged two times for 15 min at 14,000 rpm.
Proteins from the soluble supernatant fraction were purified by Ni-NTA chromatography as
described by the supplier (Thermo Scientific, Waltham, MA, USA), after which the buffer was
exchanged to 50 mM MOPS buffer (pH 6.3) in an Amicon®Ultra centricon with a 30 kDa cut-off
(Merck Millipore, Darmstadt, Germany). The protein content was determined by measuring the
absorbance at 280 nm with the NanoDrop2000 Spectrophotometer (Thermo Scientific). The extinction
coefficient and molecular weight of His-tagged RmCE were calculated using the ProtParam tool on
the ExPASy server (http://web.expasy.org/protparam/). Molecular weight and purity of the protein
were verified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; 12% gel).
4.3. Substrate Specificity Analysis
The specificity of RmCE towards different monosaccharides was determined in MOPS buffer
(pH 6.3) and with 200 mM of substrate (D-glucose, D-mannose, D-galactose, D-xylose, D-lyxose,
◦
L-arabinose) at an enzyme concentration of 0.3 mg/mL. All reactions were monitored at 70 C
and 5 time points were taken during 24 h. Enzyme inactivation was achieved by transferring the
sample (5
high-performance anion exchange chromatography–pulsed amperometric detection (HPAEC–PAD)
using the Dionex ICS-3000 system (Thermo Fischer Scientific) (CarboPac PA20 column-3 150 mm)
µL) in 100 mM of NaOH (90 µL). Conversion of substrate to product was evaluated by
×
as described by Verhaeghe et al. but with a 15-min isocratic method of 20 mM NaOH [40].
The monosaccharides were quantified by determining conversion based on the peak areas. Product
concentration was then plotted in function of time and a linear correlation was fitted representing
the activity of the enzyme (in
µmol/min). Finally, specific activities were calculated by dividing the
increase in product (units) by the enzyme concentration (mg). One unit (U) of RmCE activity was
defined as the amount of enzyme required to produce 1
70 ◦C and pH 6.3.
µmol of product from substrate per minute at
4.4. Kinetic Parameters of RmCE
The kinetic parameters for mannose were determined with a discontinuous coupled assay using
glucose oxidase and peroxidase (GOD-POD) as described by Aerts et al. [41]. Since the assay can, to
a small extent, also detect mannose, different standard series were created mimicking the formation
of glucose in different starting mannose concentrations (Appendix A). Reactions were performed in
100 mM MOPS (pH 6.3) and monitored at 70 ◦C during 90 min. Five time points were taken for each

Molecules 2018, 23, 2519
9 of 15
mannose concentration (50, 100, 200, 300, 400, 750, 1000 mM) and activities were determined in the
same way as in Section 4.3.
Since the GOD-POD method was not applicable for detection of talose, the Dionex ICS-3000 system
(Thermo Fischer Scientific) (CarboPac PA20 column-3
and its products tagatose and talose. Samples were diluted to a concentration of approximately
100 M and analysed at a flow rate of 0.5 mL/min. Separation of galactose, tagatose and talose was
×
150 mm) was used for detection of galactose
µ
accomplished by isocratic elution with 8 mM of NaOH for 12 min, followed by a linear gradient for
13 min from 8 mM to 45 mM of NaOH. Afterwards, initial conditions were restored in 1 min and
maintained for 10 min. Enzyme reactions were performed in 100 mM MOPS (pH 6.3) and monitored at
70 ◦C during 90 min. Six time points were taken for each galactose concentration (200, 400, 800, 1200,
1600 mM) and activities were determined in the same way as in Section 4.3. The kinetic parameters,
including the Michaelis–Menten constant (Km) and turnover number (kcat), were determined using
Michaelis–Menten plots created in SigmaPlot (Systat Software, San Jose, CA, USA).
4.5. Talose Production
First of all, two parameters (substrate concentration and incubation time) were optimized to
maximize yield and purity. Unless otherwise stated, reactions were performed at 70 ◦C in 100 mM
MOPS (pH 6.3) with an RmCE concentration of 0.3 mg/mL. To evaluate the effect of the substrate
on the purity of the galactose/tagatose/talose mixture, five different galactose concentrations (200,
400, 800, 1200, 1600 mM) were tested by collecting seven time points during 18 h of reaction.
Subsequently, product purity ratios at each time point were determined more precisely by performing
the reaction at 1.6 M of galactose in triplicate and were calculated with the formula (Tal)/((Tal)+(Tag)).
In addition, yield was determined by the formula (Tal)/((Gal)+(Tal)+(Tag)) and displayed in percentage.
Conversion of galactose to talose and tagatose was evaluated by HPAEC–PAD using the same method
and settings as described in Section 4.4.
Next, a large-scale reaction was perfomed to evaluate whether it was possible to get larger
amounts of product with acceptable purity. 15 mL of reaction mixture containing 4.3 g (24 mmol)◦of
galactose, 0.3 mg/mL of His-tag purified RmCE and 100 mM of MOPS (pH 6.3) was incubated at 70 C
◦
and the reaction was stopped after 4.5 h by incubation for 10 min at 95 C. Protein debris was removed
by centrifugation for 20 min at 18,000 rpm and filtration through a polyethersulfone membrane with a
cut-off of 0.2 µm.
To remove galactose from the mixture, Kluyveromyces marxianus was grown i◦n 50 mL of medium
(yeast extract, 10 g/L; peptone, 20 g/L; galactose, 20g/L) at 200 rpm and 30 C until at least an
OD600 of 6 was reached. This particular strain has the advantage of consuming galactose whereas a
typical strain like Saccharomyces cerevisiae would not [20,21]. Moreover, the glycerol production is very
low, even after long incubation periods. The yeast was pregrown with galactose as primary carbon
source instead of glucose allowing the metabolism to already adapt towards galactose consumption.
This speeds up the subsequent process of removing galactose in the reaction mixture. Then, the cell
suspension was centrifuged for 20 min at 4500 rpm and the yeast cells were washed with MilliQ water
(ELGA LabWater, High Wycombe, UK) once and afterwards resuspended in 5 mL of 100 mM MOPS
(pH 7). Half of this suspension was added to the filtrated reaction mixture (diluted until a theoretical
◦
concentration of 20 g/L monosaccharides) and incubated for 96 h at 30 C and 200 rpm and centrifuged
◦
for 20 min at 4500 rpm. Finally, the supernatant was evaporated for 3 h and saved at
of the final product was evaluated by analysis with Dionex as described in Section 4.4. The amount of
talose was estimated by a standard curve with talose obtained from Carbosynth (Compton, UK).
−
20 C. The purity
4.6. Developing and Screening Two Mutant Libraries to Increase Activity on Monosaccharide Substrates
Exploration of the active site of RmCE, based on the crystal structure (PDB code: 3WKG), revealed
four interesting positions for mutagenesis: Tyr-124, Tyr-307, Trp-321 and Trp-385. Since complete

Molecules 2018, 23, 2519
10 of 15
saturation of these positions is not possible, as this would deliver too big a library, the Rosetta suite
was considered as alternative. The full procedure can be retrieved in Appendix B.
4.6.1. Generation of Mutant Libraries
Two mutant libraries (short_sim and long_sim) were created by circular polymerase extension
cloning (CPEC) [42] with different amino acids on the four targeted positions. Four prefragmens
were generated starting from the pET21a-RmCE WT construct using the primers from Appendix C.1.
This step was necessary to ensure that the distribution of the introduced amino acids was equal and not
favored towards the wild-type amino acid. The actual CPEC fragments were constructed by performing
an additional polymerase chain reaction (PCR) with the prefragments and the degenerated primers from
Appendixs C.2 and C.3. These primers were designed to have an overhang that displays an overlap with
a melting temperature of 63–64 ◦C which is important to increase the efficiency of the CPEC reaction.
Afterwards, a DpnI digest for 3 h at 37 ◦C was performed on all PCR-fragments. After pooling of the
fragments in the final PCR reaction, the resulting plasmid libraries of both short_sim and long_sim
were transformed in electrocompetent E. coli BL21(DE3) cells. Next, the transformation mixture was
transferred to LB medium containing 100 µg/mL ampicillin and plasmid DNA was isolated with a
plasmid purification kit (Analytik Jena). Afterwards, the distribution of the incorporated nucleotides
on each of the four codon positions was checked by sequencing (LGC Genomic sequencing service in
Berlin, Germany).
4.6.2. Screening of Mutant Libraries
Library plasmids were again transformed in electrocompetent E. coli BL21(DE3) cells and plated
on LB medium containing 100
µ
g/mL ampicillin. Colonies from this plate were transferred to
L LB medium complemented with 100 g/mL ampicillin. After 16 h
L from each well was transferred to a new microtiterplate
microtiterplates containing 175
µ
µ
◦
incubation at 37 C and 250 rpm, 10
µ
containing 170 µL TB medium complemented with 100 µg/mL ampicillin. After growth for two hours
at 37 ◦C and 250 rpm the mutants were induced with 0.1 mM IPTG. Then, after incubation for 16 h
at 30 ◦C and 250 rpm, cells were harvested by centrifugation for 20 min at 4500 rpm. Pellets were
◦
frozen at
−
20 C for at least 1 h and resuspended in 175
µ
L lysis buffer (50 mM MOPS (pH 6.3),
1 mg/mL lysozym, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM ethylenediaminetetraacetic
acid (EDTA), 4 mM MgCl2 and 50 mM NaCl). Cell lysis went on for 30 min at 37 ◦C. Subsequently,
25 µL of a mannose stock solution was added to a final concentration of 100 mM. After 1h a sample
from each well was taken and analyzed by the GOD-POD assay. Finally, the absorbance in the wells
containing the mutants was compared to the mean wild-type absorbance. A mutant enzyme was
considered a hit if the following condition was fulfilled: absorbance (mutant) > mean absorbance
(WT) + mean absorbance (WT)
×
CV, where CV is the coefficient of variation. The CV was calculated
by taking the standard deviation of the absorbance of 48 microtiter plate wells with wild-type and
dividing by the mean signal of these wells. To obtain 95% coverage of each library, 21 plates were
screened for short_sim and six plates were screened for long_sim.
4.7. Cloning of CsCE and S99M/Q371F Variant, Enzyme Expression, Purification and Activity Tests
Cellobiose 2-epimerase from Caldicellulosiruptor saccharolyticus (Uniprot: A4XGA6) was ordered and
subcloned in the pET21a(+) vector 0at GeneArt. The S99M/Q371F variant was created starting from this
0
wild-type plasmid using forward (5 -GCCTGTTTTGGATGGTTAGCCATAAAGG-3 ) and reverse primers
(50-GGGCATTTCCAAAATTCAACAATCGGCAT-30) with the method devoped by Sanchis et al. [43].
Expression and purification of the CsCE wild-type and variant enzyme was performed as described for
RmCE in Section 4.2. Activity tests were performed with 200 mM of mannose and 1 mg/mL enzyme
in 100 mM of MOPS pH 7.5. Reactions were incubated at 70 ◦C and six time points were taken during
20 min. Enzyme inactivation was achieved by transferring the sample (5
µL) in 100 mM of NaOH (95 µL).
Samples were properly diluted and analyzed by the GOD-POD assay (Section 4.4).

Molecules 2018, 23, 2519
11 of 15
Supplementary Materials: The following are available online. Table S1: Substrate concentration optimization
of the RmCE reaction with galactose performed at pH 6.3 and 70 ◦C; Figure S1: Time-course representation of
Kluyveromyces marxianus treatment of the upscaled RmCE reaction for talose production; Figure S2: Dionex profile
of the final dried mixture containing glycerol (Gly), galactose (Gal), tagatose (Tag) and talose (Tal); Figure S3:
Overview of the introduced amino acids on each position in both libraries. Figure S4: Active site representation of
RmCE (PDB code: 3WKG) containing docked galactose (element = blue) as substrate.
Author Contributions: S.V.O., T.D., O.G. and K.B. conceived and designed the experiments. S.V.O. and M.C.
performed and analyzed the biological experiments. S.V.O. was responsible for writing and original draft
preparation. T.D., O.G. and KB were responsible for reviewing and editing the manuscript. T.D. and K.B.
supervised the project.
Funding: This research was funded by the Fund for Scientific Research Flanders (FWO-Vlaanderen) grant number
FWO.3F0.2016.0056.01.
Conflicts of Interest: The authors declare no conflict of interest.
Appendix A.
Table A1. Standard series employed during the GOD-POD assays. The ratios represent the amount of
mannose and glucose present in the different standards (mm mannose:mm glucose).
50 mM Std
100 mM Std
200 mM Std
300 mM Std
400 mM Std
750 mM Std
1 M Std
50:0
49:1
48:2
47:3
46:4
45:5
100:0
98:2
96:4
94:6
92:8
90:10
200:0
198:2
195:5
192:8
190:10
180:20
300:0
295:5
290:10
280:20
275:25
270:30
400:0
395:5
390:10
380:20
370:30
360:40
750:0
740:10
730:20
710:40
700:50
690:60
1000:0
990:10
980:20
970:30
950:50
930:70
Appendix B. Employing Rosetta EnzDes to Determine Useful Amino Acid Substitutions on
Target Positions
The RmCE structure (3WKG) was subjected to a molecular dynamics simulation to generate
starting structures for the Rosetta suite (all steps were performed with the Yasara Structure software).
Two separate simulations of different simulation times (0.1 ns and 5 ns) were executed and five
snapshots were taken at regular time intervals. First of all, the structure was prepared for the molecular
dynamics simulation by employing the ‘clean structure’ function. Then, a simulation cell 5 Å around
all atoms was defined and this box was filled with water atoms and, accordingly, neutralized by ions
if necessary. Subsequently, an energy minimization was performed until the mean velocity dropped
below 3000 m/s after which ‘simulated annealing’ mode was used until temperature dropped below
30 K. Finally, the actual simulation was performed with the ‘rescale velocity’ function and following
characteristics: temperature was constant at 298 K, pressure control was enabled and solvent molecules
remained at a constant water density/pressure.
Ligand input files (consisting of a pdb and a .params for each conformer) and constraint files were
constructed. The catalytic constraints are specific parameters (like distance and angle) between atoms
of the catalytic residues and ligand atoms and represent the borders within which decent catalysis can
still occur. Four residues (Tyr-124, Tyr-307, Trp-321 and Trp-385) were defined as designable, while all
other residues within 7 Å of the ligand were defined as repackable.
The Rosetta EnzDes suite [26,44] was run 500 times per starting structure, generating 2500 possibly
improved structures for both the short and the long simulation. All these structures were ranked by
Rosetta based on the ∆∆G value representing the difference in binding energy between the ligand
and the designed protein against the wild type. Furthermore, structures displaying a constraint value,
representing the fit of the ligand to the catalytic residues, higher than 30 were omitted from further
analysis. Finally, the introduced amino acids on each designable position were evaluated and amino
acids occurring in more than 5% of the remaining sequencing were considered in the generation of a
mutant library.

Molecules 2018, 23, 2519
12 of 15
Appendix C.
Appendix C.1. Primers for the Amplification of Prefragments
Primer Name
Primer
RmCE_bib_A125_Fw
RmCE_bib_Y306_Rv
RmCE_bib_W322_Fw
RmCE_bib_F384_Rv
RmCE_bib_K386_Fw
RmCE_bib_V123_Rv
GCCCAGAGTTTTGCTATC
ATACAGAGAGCCATCCG
TGGCCGCAAGCCGAA
GAAATCGACTTTGTCATCCG
AAAGGTCCGTATCATAATGGC
AACATGTTTACGGGTGTC
Appendix C.2. Primers for the Long Library Fragments
Primer Name
Primer
RmCE_bib_Y124Y_Fw
RmCE_bib_Y124P_Fw
RmCE_bib_Y124V_Fw
RmCE_bib_Y307YH_Rv
RmCE_bib_Y307K_Rv
RmCE_bib_W321WR_Fw
RmCE_bib_W321A_Fw
RmCE_bib_W321M_Fw
RmCE_bib_W385WR_Rv
RmCE_bib_W385T_Rv
RmCE_bib_W385YD_Rv
CCGCTGGACACCCGTAAACATGTTTATGCCCAGAGTTTTGCTATC
CCGCTGGACACCCGTAAACATGTTCCGGCCCAGAGTTTTGCTATC
CCGCTGGACACCCGTAAACATGTTGTGGCCCAGAGTTTTGCTATC
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCATRATACAGAGAGCCATCCG
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCTTTATACAGAGAGCCATCCG
GTGAACAGGGCCATCTGGATACCGACCGTCACYGGTGGCCGCAAGCCGAA
GTGAACAGGGCCATCTGGATACCGACCGTCACGCGTGGCCGCAAGCCGAA
GTGAACAGGGCCATCTGGATACCGACCGTCACATGTGGCCGCAAGCCGAA
CGCGGCCATTATGATACGGACCTTTCCRGAAATCGACTTTGTCATCCG
CGCGGCCATTATGATACGGACCTTTGGTGAAATCGACTTTGTCATCCG
CGCGGCCATTATGATACGGACCTTTATMGAAATCGACTTTGTCATCCG
Appendix C.3. Primers for the Short Library Fragments
Primer Name
Primer
RmCE_bib_Y124YH_Fw
RmCE_bib_Y124A_Fw
RmCE_bib_Y124L_Fw
RmCE_bib_Y307YH_Rv
RmCE_bib_Y307K_Rv
RmCE_bib_Y307R_Rv
RmCE_bib_Y307A_Rv
RmCE_bib_W321A_Fw
RmCE_bib_W321M_Fw
RmCE_bib_W321YH_Fw
RmCE_bib_W321W_Fw
RmCE_bib_W385KE_Rv
RmCE_bib_W385WR_Rv
RmCE_bib_W385T_Rv
RmCE_bib_W385YD_Rv
CCGCTGGACACCCGTAAACATGTTYATGCCCAGAGTTTTGCTATC
CCGCTGGACACCCGTAAACATGTTGCGGCCCAGAGTTTTGCTATC
CCGCTGGACACCCGTAAACATGTTCTGGCCCAGAGTTTTGCTATC
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCATRATACAGAGAGCCATCCG
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCTTTATACAGAGAGCCATCCG
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCGCGATACAGAGAGCCATCCG
GTCGGTATCCAGATGGCCCTGTTCACCAATTTCCGCATACAGAGAGCCATCCG
GTGAACAGGGCCATCTGGATACCGACCGTCACGCGTGGCCGCAAGCCGAA
GTGAACAGGGCCATCTGGATACCGACCGTCACATGTGGCCGCAAGCCGAA
GTGAACAGGGCCATCTGGATACCGACCGTCACYATTGGCCGCAAGCCGAA
GTGAACAGGGCCATCTGGATACCGACCGTCACTGGTGGCCGCAAGCCGAA
CGCGGCCATTATGATACGGACCTTTTTYGAAATCGACTTTGTCATCCG
CGCGGCCATTATGATACGGACCTTTCCRGAAATCGACTTTGTCATCCG
CGCGGCCATTATGATACGGACCTTTGGTGAAATCGACTTTGTCATCCG
CGCGGCCATTATGATACGGACCTTTATMGAAATCGACTTTGTCATCCG
References
1.
2.
3.
Van Overtveldt, S.; Verhaeghe, T.; Joosten, H.-J.; van den Bergh, T.; Beerens, K.; Desmet, T. A structural
classification of carbohydrate epimerases: From mechanistic insights to practical applications. Biotechnol. Adv.
2015, 33, 1814–1828. [CrossRef] [PubMed]
Kim, J.E.; Kim, Y.S.; Kang, L.W.; Oh, D.K. Characterization of a recombinant cellobiose 2-epimerase
from Dictyoglomus turgidum that epimerizes and isomerizes beta-1,4- and alfa-1,4-gluco-oligosaccharides.
Biotechnol. Lett. 2012, 34, 2061–2068. [CrossRef] [PubMed]
Uechi, K.; Takata, G.; Fukai, Y.; Yoshihara, A.; Morimoto, K. Gene cloning and characterization of L-ribulose
3-epimerase from Mesorhizobium loti and its application to rare sugar production. Biosci. Biotechnol. Biochem.
2013, 77, 511–515. [CrossRef] [PubMed]

Molecules 2018, 23, 2519
13 of 15
4.
5.
6.
Park, C.-S.; Kim, J.-E.; Choi, J.-G.; Oh, D.-K. Characterization of a recombinant cellobiose 2-epimerase
from Caldicellulosiruptor saccharolyticus and its application in the production of mannose from glucose.
Appl. Microbiol. Biotechnol. 2011, 92, 1187–1196. [CrossRef] [PubMed]
Ojima, T.; Saburi, W.; Sato, H.; Yamamoto, T.; Mori, H.; Matsui, H. Biochemical characterization of a
thermophilic cellobiose 2-epimerase from a thermohalophilic bacterium, Rhodothermus marinus JCM9785.
Biosci. Biotechnol. Biochem. 2011, 75, 2162–2168. [CrossRef] [PubMed]
Zhang, W.; Zhang, T.; Jiang, B.; Mu, W. Biochemical characterization of a d-psicose 3-epimerase from
Treponema primitia ZAS-1 and its application on enzymatic production of d-psicose. J. Sci. Food Agric. 2015
96, 49–56. [CrossRef] [PubMed]
,
7.
8.
9.
Beerens, K.; Desmet, T.; Soetaert, W. Enzymes for the biocatalytic production of rare sugars. J. Ind. Microbiol.
Biotechnol. 2012, 39, 823–834. [CrossRef] [PubMed]
Baek, S.H.; Park, S.J.; Lee, H.G. D-Psicose, a Sweet Monosaccharide, Ameliorate Hyperglycemia, and
Dyslipidemia in C57BL/6J db/db Mice. J. Food Sci. 2010, 75, 6–10. [CrossRef] [PubMed]
Wagner, N.; Bosshart, A.; Failmezger, J.; Bechtold, M.; Panke, S. A Separation-Integrated Cascade Reaction to
Overcome Thermodynamic Limitations in Rare-Sugar Synthesis. Angew. Chem. Int. Ed. 2015, 127. [CrossRef]
10. Fujiwara, T.; Saburi, W.; Matsui, H.; Mori, H.; Yao, M. Structural insights into the epimerization of β-1,4-linked
oligosaccharides catalyzed by cellobiose 2-epimerase, the sole enzyme epimerizing non-anomeric hydroxyl
groups of unmodified sugars. J. Biol. Chem. 2014, 289, 3405–3415. [CrossRef] [PubMed]
11. Krewinkel, M.; Kaiser, J.; Merz, M.; Rentschler, E.; Kuschel, B.; Hinrichs, J.; Fischer, L. Novel cellobiose
2-epimerases for the production of epilactose from milk ultrafiltrate containing lactose. J. Dairy Sci. 2015, 98,
3665–3678. [CrossRef] [PubMed]
12. Wu, L.; Xu, C.; Li, S.; Liang, J.; Xu, H.; Xu, Z. Efficient production of lactulose from whey powder by cellobiose
2-epimerase in an enzymatic membrane bioreactor. Bioresour. Technol. 2017, 233, 305–312. [CrossRef]
[PubMed]
13. Linington, R.G.; Robertson, M.; Gauthier, A.; Finlay, B.B.; MacMillan, J.B.; Molinski, T.F.; Van Soest, R.;
Andersen, R.J. Caminosides B-D, antimicrobial glycolipids isolated from the marine sponge Caminus
sphaeroconia. J. Nat. Prod. 2006, 69, 173–177. [CrossRef] [PubMed]
14. Haukaas, M.H.; O’Doherty, G.A. Enantioselective synthesis of N-Cbz-protected 6-amino-6-deoxymannose,
-talose, and -gulose. Org. Lett. 2001, 3, 3899–3902. [CrossRef] [PubMed]
15. Sznaidman, M.L.; Hecht, S.M. Studies on the total synthesis of tallysomycin. Synthesis of the threonylbithiazole
moiety containing a structurally unique glycosylcarbinolamide. Org. Lett. 2001, 3, 2811–2814. [CrossRef]
[PubMed]
16. Beerens, K.; Van Overtveldt, S.; Desmet, T. The “epimerring” highlights the potential of carbohydrate
epimerases for rare sugar production. Biocatal. Biotransform. 2017, 35, 230–237. [CrossRef]
17. Öberg, C.T.; Blanchard, H.; Leffler, H.; Nilsson, U.J. Protein subtype-targeting through ligand epimerization:
Talose-selectivity of galectin-4 and galectin-8. Bioorg. Med. Chem. Lett. 2008, 18, 3691–3694. [CrossRef]
[PubMed]
18. Hodosi, G.; Kovácˇ, P. Epimerization of carbohydrates via stannylene acetals. A practical synthesis of D-Talose.
J. Carbohydr. Chem. 1998, 17, 557–565. [CrossRef]
19. Xiao, H.; Wang, G.; Wang, P.; Li, Y. Convenient synthesis of D-talose from D-galactose. Chinese J. Chem. 2010
28, 1229–1232. [CrossRef]
,
20. Fonseca, G.G.; De Carvalho, N.M.B.; Gombert, A.K. Growth of the yeast Kluyveromyces marxianus CBS 6556
on different sugar combinations as sole carbon and energy source. Appl. Microbiol. Biotechnol. 2013, 97,
5055–5067. [CrossRef] [PubMed]
21. Beniwal, A.; Saini, P.; Kokkiligadda, A.; Vij, S. Physiological growth and galactose utilization by dairy yeast
Kluyveromyces marxianus in mixed sugars and whey during fermentation. 3 Biotech 2017, 7. [CrossRef]
[PubMed]
22. Kuboki, A.; Okazaki, H.; Sugai, T.; Ohta, H. An expeditious route to N-glycolylneuraminic acid based on
enzyme-catalyzed reaction. Tetrahedron 1997, 53, 2387–2400. [CrossRef]
23. Arbor, A.; Rackham, T.; Rackham, H. Glucosamine metabolism I. N-acetylglucosamine deacetylase. J. Biol. Chem.
1957, 226, 115–124.
24. Trinder, P. Determination of blood glucose using an oxidase-peroxidase system with a non-carcinogenic
chromogen. J. Clin. Pathol. 1969, 22, 158–161. [CrossRef] [PubMed]

Molecules 2018, 23, 2519
14 of 15
25. Reetz, M.T.; Kahakeaw, D.; Lohmer, R. Addressing the numbers problem in directed evolution. Chembiochem
2008, 9, 1797–1804. [CrossRef] [PubMed]
26. Röthlisberger, D.; Khersonsky, O.; Wollacott, A.M.; Jiang, L.; DeChancie, J.; Betker, J.; Gallaher, J.L.;
Althoff, E.A; Zanghellini, A.; Dym, O.; et al. Kemp elimination catalysts by computational enzyme design.
Nature 2008, 453, 190. [CrossRef] [PubMed]
27. Zanghellini, A.; Jiang, L.I.N.; Wollacott, A.M.; Cheng, G.; Meiler, J.; Althoff, E.A.; Ro, D. New algorithms
and an in silico benchmark for computational enzyme design. Protein Sci. 2006, 15, 2785–2794. [CrossRef]
[PubMed]
28. Jiang, L.; Althoff, E.A.; Clemente, F.R.; Doyle, L.; Zanghellini, A.; Gallaher, J.L.; Betker, J.L.; Iii, C.F.B.;
Hilvert, D.; Houk, K.N.; et al. De novo computational design of retro-aldo enzymes. Science 2008, 319,
1387–1391. [CrossRef] [PubMed]
29. Crooks, G.; Hon, G.; Chandonia, J.; Brenner, S. WebLogo: A sequence logo generator. Genome Res 2004, 14,
1188–1190. [CrossRef] [PubMed]
30. Mu, W.; Li, Q.; Fan, C.; Zhou, C.; Jiang, B. Recent advances on physiological functions and biotechnological
production of epilactose. Appl. Microbiol. Biotechnol. 2013, 97, 1821–1827. [CrossRef] [PubMed]
31. Ito, S.; Taguchi, H.; Hamada, S.; Kawauchi, S.; Ito, H.; Senoura, T.; Watanabe, J.; Nishimukai, M.; Ito, S.;
Matsui, H. Enzymatic properties of cellobiose 2-epimerase from Ruminococcus albus and the synthesis of rare
oligosaccharides by the enzyme. Appl. Microbiol. Biotechnol. 2008, 79, 433–441. [CrossRef] [PubMed]
32. Chen, Q.; Levin, R.; Zhang, W.; Zhang, T.; Jiang, B.; Stressler, T.; Fischer, L.; Mu, W. Characterisation of a
novel cellobiose 2-epimerase from thermophilic Caldicellulosiruptor obsidiansis for lactulose production. J. Sci.
Food Agric. 2017, 97, 3095–3105. [CrossRef] [PubMed]
33. De Castro, C.; Gargiulo, V.; Lanzetta, R.; Parrilli, M. Agrobacterium rubiT DSM 6772 produces a lipophilic
polysaccharide capsule whose degree of acetylation is growth modulated. Biomacromolecules 2007, 8,
1047–1051. [CrossRef] [PubMed]
34. Haukaas, M.H.; O’Doherty, G.A. Synthesis of D- and L-Deoxymannojirimycin via an Asymmetric
Aminohydroxylation of Vinylfuran. Org. Lett. 2001, 3, 401–404. [CrossRef] [PubMed]
35. Czechura, P.; Tam, R.Y.; Dimitrijevic, E.; Murphy, A.V.; Ben, R.N. The importance of hydration for inhibiting
ice recrystallization with C-linked antifreeze glycoproteins. J. Am. Chem. Soc. 2008, 130, 2928–2929. [CrossRef]
[PubMed]
36. Hicks, K.B. Preparative HPLC of Carbohydrates. J. Chromatogr. Libr. 1995, 58, 361–387. [CrossRef]
37. Stead, P. Isolation by preparative HPLC. Methods Biotechnol. 1998, 4, 165–208.
38. Maru, I.; Ohnishi, J.; Ohta, Y.; Tsukada, Y. Simple and large-scale production of N-acetylneuraminic acid from
N-acetyl-D-glucosamine and pyruvate using N-acyl-dglucosamine 2-epimerase and n-acetylneuraminate
lyase. Carbohydr. Res. 1998, 306, 575–578. [CrossRef]
39. Engler, C.; Kandzia, R.; Marillonnet, S. A one pot, one step, precision cloning method with high throughput
capability. PLoS ONE 2008, 3. [CrossRef] [PubMed]
40. Verhaeghe, T.; Aerts, D.; Diricks, M.; Soetaert, W.; Desmet, T. The quest for a thermostable sucrose
phosphorylase reveals sucrose 6’-phosphate phosphorylase as a novel specificity. Appl. Microbiol. Biotechnol.
2014, 98, 7027–7037. [CrossRef] [PubMed]
41. Aerts, D.; Verhaeghe, T.F.; Roman, B.I.; Stevens, C.V; Desmet, T.; Soetaert, W. Transglucosylation potential
of six sucrose phosphorylases toward different classes of acceptors. Carbohydr. Res. 2011, 346, 1860–1867.
[CrossRef] [PubMed]
42. Quan, J.; Tian, J. PROTOCOL Circular polymerase extension cloning for high- throughput cloning of complex
and combinatorial DNA libraries. Nat. Protoc. 2011, 6, 242–251. [CrossRef] [PubMed]
43. Sanchis, J.; Fernández, L.; Carballeira, J.D.; Drone, J.; Gumulya, Y.; Höbenreich, H.; Kahakeaw, D.; Kille, S.;
Lohmer, R.; Peyralans, J.J.-P.; et al. Improved PCR method for the creation of saturation mutagenesis libraries
in directed evolution: Application to difficult-to-amplify templates. Appl. Microbiol. Biotechnol. 2008, 81,
387–397. [CrossRef] [PubMed]

Molecules 2018, 23, 2519
15 of 15
44. Richter, F.; Leaver-Fay, A.; Khare, S.D.; Bjelic, S.; Baker, D. De novo enzyme design using Rosetta3. PLoS ONE
2011, 6, e19230. [CrossRef] [PubMed]
Sample Availability: Samples of the compounds are available from the authors.
©
2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).