Directed evolution toward improved production of L-ribose from ribitol

Article abstract of DOI:10.2174/138620710791054277

Improvement of the one-step production of L-ribose from ribitol using a recombinant Escherichia coli is described. The gene encoding the enzyme mannitol-1-dehydrogenase (MDH) from Apium graveolens has previously been codon-optimized, cloned into the constitutive pZuc10 vector, and expressed in E. coli. This MDH catalyzes the NADdependent conversion of mannitol to D-mannose and has the ability to convert several polyols to their L-sugar counterparts, including ribitol to L-ribose. Here, three rounds of directed evolution using libraries generated through errorprone PCR and screened using a dinitrosalicylate reagent. Mutants were selected for improved conversion of L-ribose, and the best mutant was isolated by combining two round 2 mutations. Libraries were also selected for thermal stability and screened at increasingly higher temperatures with each round of mutagenesis. An overall 19.2-fold improvement was observed with a final conversion of 46.6 ± 1.7% and a productivity of 3.88 ± 0.14 gL^-1d^-1 in 50 mL shaken flasks at 34 °C. Further characterization of the mutants suggests that increased enzyme thermal stability and expression are responsible for the increase in L-ribose production. The mutant E. coli production strain isolated represents an improved system for largescale production of L-ribose.

Full text of DOI:10.2174/138620710791054277

3
02
Combinatorial Chemistry & High Throughput Screening, 2010, 13, 302-308
Directed Evolution Toward Improved Production of L-Ribose from Ribitol
*
Trevor N. Christ, Kara A. Deweese and Ryan D. Woodyer
8
01 W. Main Street, Room B107, Peoria, IL 61606, USA
Abstract: Improvement of the one-step production of L-ribose from ribitol using a recombinant Escherichia coli is
described. The gene encoding the enzyme mannitol-1-dehydrogenase (MDH) from Apium graveolens has previously been
codon-optimized, cloned into the constitutive pZuc10 vector, and expressed in E. coli. This MDH catalyzes the NAD-
dependent conversion of mannitol to D-mannose and has the ability to convert several polyols to their L-sugar
counterparts, including ribitol to L-ribose. Here, three rounds of directed evolution using libraries generated through error-
prone PCR and screened using a dinitrosalicylate reagent. Mutants were selected for improved conversion of L-ribose, and
the best mutant was isolated by combining two round 2 mutations. Libraries were also selected for thermal stability and
screened at increasingly higher temperatures with each round of mutagenesis. An overall 19.2-fold improvement was
-1 -1
observed with a final conversion of 46.6 ± 1.7% and a productivity of 3.88 ± 0.14 gL d in 50 mL shaken flasks at 34°C.
Further characterization of the mutants suggests that increased enzyme thermal stability and expression are responsible for
the increase in L-ribose production. The mutant E. coli production strain isolated represents an improved system for large-
scale production of L-ribose.
Keywords: L-ribose, ribitol, directed evolution, mannitol-1-dehydrogenase MDH.
INTRODUCTION
from D-sorbitol, and L-galactose from galactitol [21]. The
recombinant Escherichia coli adapted to express MDH was
employed in a 1 L bioconversion using L-ribose as a model
system. Conversion conditions were optimized to produce a
Chiral chemicals are of great importance in the areas of
drug discovery and natural product synthesis [1]. Of these
chemicals, enantiomerically pure carbohydrates are essential
building blocks for the synthesis of many chemical products
important to pharmaceuticals [2-4]. However, traditional
chemical synthesis of stereospecific carbohydrates is limited
by multistep synthesis, low enantiomeric purity, and low
yields [5, 6]. Enzyme catalyzed reactions via whole-cell
bioconversion differ in the fact that they are simple,
inexpensive, and time efficient [7], while maintaining high
purity products [8]. Since carbohydrates play such a large
role in cellular recognition, signaling, extra and intracellular
targeting, and even the development of disease states, it is
not surprising that an increasing number of drugs utilizing
sugars of the L-conformation are currently on the market or
undergoing clinical trails [2, 4, 9, 10]. Furthermore, interest
in L-sugars in the biochemical and pharmacological research
community is growing steadily [11]. While many methods of
production of L-ribose have been described [12-15], high
costs ($500/kg est.) remain inhibitory for intermediate stage
development [6]. It is evident that a scaleable and
economical process for the production of L-sugar
intermediates is of critical importance to both industry and
research.
5
5% conversion at a concentration of 100 g/L in 72 hours,
-1 -1
giving a volumetric productivity of 17.4 gL d [11].
However, this system had some important drawbacks
including the need for a temperature shift at induction
(
cooling is a significant manufacturing cost) and the
requirement for a moderately expensive exogenous induction
reagent, isopropyl β-D-1-thiogalactopyranoside (IPTG).
Furthermore, the temperature of optimum conversion was
significantly below the temperature of peak metabolic
activity for E. coli, suggesting that improvements could be
achieved in the time required for conversion.
Fig. (1).
Previous work demonstrated successful cloning,
expression, and characterization of the NAD-dependent
mannitol-1-dehydrogenase (MDH) from Apium graveolens
Directed evolution has been demonstrated as a successful
means to improve enzyme function in the areas of activity
and overall performance, enantioselectivity, substrate
specificity, and enzyme stability [23]. The use of error-prone
(
garden celery). The MDH is of particular interest because of
both its regioselectivity at the 1-carbon position as well as its
stringent stereoselectivity at the 2-carbon position [16-22].
Thus, the synthesis of various sugars in the L-conformation
is permitted including L-ribose from ribitol (Fig. 1), L-gulose
PCR as
a
method of mutagenesis and
a
unique
dinitrosalicylate reagent for identifying reducing sugars
allowed the production and screening of thousands of library
members in a quick and simple manner. This method was
applied to improve the activity of MDH towards an
unnatural substrate, ribitol, opposed to the natural substrate,
mannitol. Improved enzyme thermal stability allowed the
*
Address correspondence to this author at the 801 W. Main Street, Room
B107, Peoria, IL 61606, USA; Tel: 309-495-7318;
E-mail: rwoodyer@zuchem.com
1
386-2073/10 $55.00+.00
© 2010 Bentham Science Publishers Ltd.

Improved Production of L-Ribose from Ribitol
Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4 303
conversion to proceed at an increased temperature and the
need for an induction reagent and temperature shift were
eliminated through use of a constitutive vector.
DNS assay were streaked onto kanamycin (50 μg/mL)
containing LB-agar plates and incubated at 37°C overnight.
Two individual colonies from each plate were grown in 5
mL cultures containing LB-kanamycin (50 μg/mL) and were
used to inoculate culture tubes containing 3 mL of RM,
which were shaken at 30°C for 72 hours. A second DNS
reducing sugar assay was performed on a 20-fold dilution of
the resulting culture broth to verify the hits. The mutants
MATERIALS
Ribitol and L-ribose were purchased from CMS
Chemicals Ltd. (Oxsfordshire, UK), while oligonucleotide
primers were obtained from Integrated DNA Technologies
Coralville, IA). Kanamycin, ampicillin, NAD, IPTG,
glycerol, ZnCl 3,5-dinitrosalycylic acid (DNS) and
from tubes that produced the three darkest wells were
selected for scale up to a 50 mL conversion in RM and
grown at 30°C. Samples for HPLC analysis were prepared at
(
2
,
lysozyme were obtained from Sigma-Aldrich (St. Louis,
MO); other cell culture components were obtained from
2
4-hour time points. Confirmed positives served as a
template for subsequent rounds of mutagenesis and
screening. Temperature was 30°C in the first round of
screening, 34°C in the second round, and 37°C in the final
round. Approximately 1,200 colonies were screened in the
first round, 3,000 colonies in the second round, and 5,000
colonies in the final round. Verified positive mutations were
sequenced (University of Illinois-Urbana-Champaign
Biotechnology Center). To ensure improved conversion
results were not strain based, plasmids were purified, E. coli
were retransformed and strains were compared to original
hits on a 50 mL flask conversion.
-
Difco (Sparks, MD). E. coli EC100 (F mcrA Δ(mrr-
hsdRMS-mcrBC) φ80dlacZΔM15 ΔlacX74 recA1 endA1
-
araD139 Δ(ara, leu)7697 galU galK λ rpsL nupG) were
obtained from Epicentre (Madison, WI) and E. coli BL21
(
DE3) was purchased from Novagen (Madison, WI).
Restriction enzymes EcoRI, BamHI, HindIII, T4 DNA
ligase, pMal-C2x, and Taq Polymerase were purchased from
New England Biolabs (Ipswich, MA). Aminex HPX 87P
column (300 ꢀ 7.8 mm), de-ashing cartridge (30 ꢀ 4.6 mm),
and Carbo-P micro-guard cartridge (30 ꢀ 4.6 mm) and SDS-
PAGE gel and marker were obtained from Bio-Rad
Laboratories (Hercules, CA). Commercial plasmid
purification, PCR product purification and gel extraction kits
were purchased from Qiagen (Valencia, CA).
Combining Second Round Mutations. The mutation from
mutant 2-H5 was incorporated into the parent template 2-H6
via the PCR based “megaprimer” site-directed mutagenesis
method [26]. Briefly, a primer containing the 2-H5 mutation
was designed, 5’-CGA AAA CAT CAT CGA AAC CAT
CGG CAG C-3’ (underlined base designates the mutagenic
base), and a PCR was carried out using this and the pZuc10
reverse primer on the 2-H6-pZuc10 plasmid template. The
resulting PCR product was used as a primer in a second PCR
reaction with the pZuc10 forward primer and 2-H6-pZuc10
plasmid template. The resulting PCR product was subcloned
into pZuc10. The combined mutation was successful as
confirmed by sequencing (n=2).
METHODS
Library Preparation. An error-prone PCR was carried
out as previously described [24] on the MDH gene template
from the previously constructed pZuc10 (pTRP-MDH [11])
vector (containing tryptophan promoter, pMB1, ROP, and
MDH gene). This was performed using the forward primer
5
’-CGA ACT AGT TAA CTT TTA CGC AAG TTC-3’ and
reverse primer 5’-CCA TGG GTA AGT ATT TCC TTA
AGG ATC C-3’ (underlined bases designate BamHI
restriction site) under the following conditions: 2 min 95°C,
Cloning of Maltose Binding Protein (MBP)-MDH
Fusion. The MDH gene from successful mutants was
amplified by PCR from pZuc10 plasmids using the designed
pMal forward primer 5’- GCT ACG GGA TCC ATG GCG
AAA AGC AGC G -3’ (underlined bases designate BamHI
restriction site) and pMal reverse primer 5’– GCC TGC
AAG CTT TTA CGC GCC GAG GC –3’ (underlined bases
designate HindIII restriction site). Note that the wild-type
MDH was amplified with the reverse primer previously
described [11] since it did not contain a mutation found in
the first round of directed evolution in the 3’ primer region.
Inserts were digested with restriction enzymes BamHI and
HindIII and ligated into a similarly prepared pMal-C2x
vector using T4 DNA ligase at 16°C overnight. E. coli
EC100 competent cells were transformed with the resulting
ligation and plated on kanamycin containing LB-agar plates.
E. coli BL21 (DE3) competent cells were transformed with
purified plasmids from colonies containing MDH insert (as
confirmed by colony PCR).
2
3 cycles of 35 sec 95°C, 35 sec 55°C, 1 min 72°C, and
finally 5 min 72°C. The resulting PCR product was digested
with EcoRI and BamHI and ligated into a similarly prepared
pZuc10 vector using T4 DNA ligase at 16°C overnight. E.
coli EC100 competent cells were transformed with the
purified plasmid, plated onto kanamycin (50 μg/mL)
containing LB-agar plates and incubated overnight at 37°C.
Plasmids were purified from the resulting colonies. E. coli
Zuc174 (ꢀ(araD-araB)567, lacZ4787(ꢀ)(::rrnB-3), lacIp-
4
000(lacIQ), l-, rph-1, ꢀ(rhaD-rhaB)568, hsdR514, ꢀ(ptsH-
crr)) was transformed with the resulting purified plasmids
and the transformants were selected on kanamycin
containing LB-agar plates and incubated at 37°C.
Library Screening. Ribitol Medium (RM) was prepared
with final concentrations of 0.3 mM ZnCl , 50 g/L ribitol, 50
2
μg/mL kanamycin, and 0.1% glycerol in LB. RM (200 μL)
was added to wells of 96-well plates containing sterile 3 mm
glass beads. Colonies of the MDH gene library were picked
into the wells (three wells per plate contained
Zuc174/pZuc10 positive control) and shaken at 30°C. After
Purification of MBP-MDH. An overnight culture of
BL21 (DE3) pMAL-MBP-MDH (5 mL) was used to
inoculate 350 mL Terrific Broth containing 20 mM glucose
and 100 μg/mL ampicillin. After 2.5 h at 37°C, the
temperature was reduced to 25°C and 0.5 mM IPTG was
added. Cells were harvested 16 h post-induction by
7
2 hours, a DNS reducing sugar assay was performed [25]
on a 20-fold dilution of the culture broth. Cell cultures from
wells that appeared darker than the positive control in the

3
04 Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4
Christ et al.
centrifugation and resuspended in 50 mM HEPES (pH 7.2),
00 mM NaCl, 1 mg/mL lysozyme, and 10% glycerol (3
mL/g cells) and stored at -80°C overnight. MBP-MDH
enzymes were purified as previously described [11] except a
RESULTS AND DISCUSSION
3
Directed Evolution. After the screening of about 1,200
library members at 30°C, one mutant was successfully
isolated, 1-C1, that had a darker DNS results suggesting
more reducing sugar (L-ribose) was being produced. The first
round mutant served as a template for a second round library
and approximately 3,000 members were screened at 34°C
using the DNS reducing sugar assay. Two mutants, 2-H6 and
2-H5, were isolated. After the second round, two approaches
were taken, since there were two different mutants of similar
ability. In the first approach, the mutants were utilized as
templates for two separate mutagenesis and screening
experiments. After screening 5,000 library members from
each template at 37°C, one improved mutant was isolated
from the 2-H5 template, 3-G7. In the second approach, the
two mutants from round 2 were combined as described in the
Materials and Methods section. After combining the second
round mutations 2-H6 and 2-H5, a final library was
constructed from this template; however, after screening at
37°C, no improved mutants were identified.
2
mL column of amylose resin was used, and enzymes were
concentrated by centrifugation using a Millipore Amicon
Ultra-15 with an Ultracel membrane.
Characterization of MDH. The initial rates of reaction
were determined by monitoring the increase in absorbance of
-
1
-1
NADH (ꢀ = 6.22 mM cm ) at 340 nm for 0.6 min at 25°C
using a Varian Cary 1E spectrophotometer. Reactions were
initiated by addition of ~25 μg MBP-MDH to 600 μL of 100
mM tris buffer pH 8.5 containing substrates D-mannitol or
ribitol and NAD. One unit (U) of activity was defined as 1
μmol NADH produced/min and specific activity was defined
as 1 U/mg of total protein added. All assays were performed
in duplicate per experiment and the average values from 2 or
more separate experiments were utilized to calculate kinetic
M
constants kcat and K using least squares regression analysis
for 5 or more concentrations of each substrate. Half-life
values were found by adding enzyme that had been
incubating at 43°C to 2.5 mM NAD and 250 mM D-mannitol
at three-minute intervals. The values for T1/2 were calculated
using the first order decay T1/2= -ln(2)/k, where k is the rate
of decay as calculated by Microsoft Excel. Values were
collected in at least two or more separate experiments with
the average and standard deviation reported. Values are
expressed as the average of all data collected (unless
statistically irrelevant by Q test) with the associated standard
deviation.
Shaken Flasks. Shaken flask experiments were carried
out as previously described using medium with glycerol and
zinc for improved conversion [11] to quantitate the
improvements in production of L-ribose from ribitol. The
various mutants created by directed evolution were tested in
5
0 mL flask conversions at 34°C and compared to the wild-
type MDH with samples analyzed by HPLC. As displayed in
Table 1 and Fig. (2), a 19.2-fold overall improvement was
demonstrated at 34°C, which achieved a 46.6 ± 1.7%
conversion in the combined round 2 mutant. Additionally,
improvements in space-time yield increased from 0.203 ±
Modeling. Swiss-Model Workspace [27] was used to
generate a model of the wild type MDH and the mutant
MDH incorporating the sites of beneficial mutations. Rasmol
-
1 -1
0.004 to 3.88 ± 0.14 gL d . While the combined mutant
performed the best, the third round mutant 3-G7, performed
nearly as well. This evidence, along with the difficulty in
finding hits in later round libraries suggests that the MDH is
no longer the rate-limiting step towards L-ribose production
and that further rounds of mutagenesis would not improve
overall conversion.
[
28] was used to view and edit the models.
MDH Expression and SDS-PAGE. Culture tubes
containing 5 mL LB-kanamycin (50 μg/mL) were inoculated
with pZuc10-MDH mutants and grown to log-phase at 37°C.
Enzymes were expressed at room temperature for 4 hours
and 1 mL of culture medium was pelleted and lysed in 1 mL
Table 1. Conversion and Productivity of Mutants
1
00 mM tris buffer containing 1 mg/mL lysozyme.
Resuspended cells were stored at -20°C overnight, thawed at
room temperature, and pelleted to clear cell debris. Protein
expression was visualized by SDS-PAGE by the method of
Laemmli [29] using lysate samples normalized to 10 μg of
protein and Coomassie staining.
-
1 -1
Mutant
Conversion
STY (gL d )
Ratio to WT
WT
1-C1
2-H6
2.43 ± 0.05%
28.5 ± 0.3%
38.1 ± 0.2%
40.6 ± 1.1%
45.4 ± 1.1%
46.6 ± 1.7%
0.203 ± 0.004
2.37 ± 0.03
3.1 ± 0.2
1.0
11.7
15.7
17.8
18.7
19.2
Analysis. The HPLC separation system consisted of a
solvent delivery system (P2000 pump, Spectra-Physics, San
Jose, CA) equipped with an autosampler (717, Waters
Chromatography Division, Millipore Corp., Milford, MA), a
refractive index detector (410 differential refractometer,
Waters), and a computer software based integration system
2
-H5
3.6 ± 0.2
3-G7
3.78 ± 0.09
3.88 ± 0.14
Comb. Rnd 2
(
Chromquest 4.0, Spectra-Physics). An ion moderated
partition chromatography column (Aminex HPX-87P with
De-ashing and Carbo-P micro-guard cartridges) was utilized
at 85°C with a mobile phase of Milli-Q filtered water at a
flow rate of 0.6 mL/min. Peaks were detected by refractive
index and were identified and quantified by comparison to
retention times and areas of authentic standards (ribose and
ribitol).
Sequence Characterization. Upon sequencing, two amino
acid substitutions were observed in the round 1 mutant (1-
C1), F14I and S47C. Based on homology modeling using the
previously solved crystal structure of sinapyl alcohol
dehydrogenase [30] as a template (77% sequence identity
[
31]), mutation S47C is found in the active site of the
enzyme. In sinapyl alcohol dehydrogenase, the

Improved Production of L-Ribose from Ribitol
Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4 305
summarized in Table 2. The mannitol kcat decreased 8-fold
-
1
-1
from 4.85 ± 0.24 s in the WT to 0.59 ± 0.08 s in the
combined mutant. The decrease in activity was paralleled for
the K
M
values where a near 7-fold increase was observed
from 46 ± 1 mM in the WT to 315 ± 7 mM in the combined
mutant with mannitol as the substrate. There was no clear
trend for the K values of the NAD substrate with mannitol.
M
The values ranged from 61 ± 18 μM for the round 1 mutant
to 193 ± 154 μM for round 2-H6. Overall, the catalytic
efficiency decreased 55-fold from 0.11 to 0.002. The kinetic
data suggests that the directed evolution rounds swayed the
MDH away from mannitol as a substrate.
Interestingly, kinetic activity remained fairly constant
with ribitol as the substrate between the WT and the
combined round 2 mutant MDH. Regarding the kcat values, a
decrease in activity was observed in round 1 and round 2
mutants, but was maintained slightly above the level of the
Fig. (2).
corresponding amino acid residue is responsible for
M
WT in the combined mutant. Conversely, the K values saw
a modest decrease from the WT to the round 1 and round 2
MDH. However, the combined mutant resulted in a higher
2
+
coordinating the catalytic Zn [30]. While the oxygen of
serine has been reported as a zinc ligand [32], Cys-47 is
more commonly found in alcohol dehydrogenases as a
NADH binding ligand [33]. Mutant 2-H6 contained the
amino acid substitution H54Y, and mutant 2-H5 contained
the amino acid substitution D122E. The exchange of
phenylalanine (aromatic) to isoleucine (hydrophobic) in the
K
M
of 173 ± 39 mM compared to the WT value of 112 ± 52
mM. The kcat /K ratio showed no clear trend and varied
M
only slightly from the highest value of 0.012 in the round 1
mutant to the lowest value of 0.006 in the combined mutant.
Due to the lack of drastic changes with the substrate ribitol,
kinetics values of the MDH are not likely a significant factor
in the increase of the L-ribose production rate. The KM, NAD
values with ribitol as a substrate decreased with each round,
resulting in a near 4-fold decrease from the WT to the
combined mutant. However, kcat was not significantly
affected in the combined mutant, although some intermediate
mutants had reduced kcat values.
1
-C1 mutation and the exchange of histidine (positive
charge) to tyrosine (aromatic) are significant substitutions.
Models of the wild type MDH (A) with mutation sites in
black and the mutant MDH with the substituted amino acids
from the combined round 2 mutant in black (B) are shown in
Fig. (3). The third round mutant, 3-G7, contained the amino
acid substitutions E8V and I149V. It is curious that both
mutations resulted in the substitution of a valine group. The
exchange from glutamic acid (negative charge) to valine
residue (hydrophobic) is a considerable change, however, no
obvious loss or gain of function is expected from these
mutations.
Overall, the MDH with the best conversion, the
combined round 2 mutant, had decreased activity on
mannitol and similar activity on ribitol as compared to the
WT. The ratio of catalytic efficiencies of mannitol to ribitol
decreased from 13.8 in the WT to 0.3 in the combined
mutant; a 44-fold change.
M
Purified Protein Data. The Michaelis constant (K ) and
the turnover number (kcat ) for each purified enzyme were
determined for either ribitol or mannitol and NAD. Catalytic
efficiency for mannitol was reduced in the mutants as a
Thermal Stability. Since Apium graveolens, the parent
organism from which the MDH gene originated, has an
optimal vegetation temperature between 16°C and 21°C [34]
M
function of both lower kcat and higher K values, which is
Fig. (3).

3
06 Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4
Christ et al.
Table 2. Comparison of MDH Kinetic Values for Mannitol and Ribitol
Mannitol
Ribitol
Ratio kcat/K
M
Enzyme
-
1
k
mM)
M
K
(μM)
M, NAD
-1
k
M
k
K
cat
/
M
NAD
M, NAD (μM)
k
cat (s )
kcat/K
M
k
cat (s )
Mannitol:Ribitol
(
(mM)
K
WT
4.85 ± 0.24
1.80 ± 0.15
0.97 ± 0.41
1.30 ± 0.17
0.59 ± 0.08
46 ± 1
24 ± 4
0.11
88 ± 20
61 ± 18
0.94 ± 0.09
0.34 ± 0.01
0.66 ± 0.06
0.27 ± 0.03
1.00 ± 0.09
112 ± 52
29 ± 10
88 ± 3
0.008
0.012
0.008
0.007
0.006
167 ± 80
69 ± 47
82 ± 72
55 ± 37
43 ± 34
13.8
6.3
0.9
1.7
0.3
1
-C1
-H6
-H5
0.075
0.007
0.012
0.002
2
2
135 ± 4
110 ± 9
315 ± 7
193 ± 154
80 ± 35
39 ± 5
Comb. Rnd 2
104 ± 86
173 ± 39
it was not expected that proteins from the organism would
achieve optimal activity at higher temperatures suitable for
E. coli optimal growth (37°C). It was hypothesized that the
later round MDH mutants would be selected for increased
thermal stability as these rounds were screened at
increasingly higher temperatures. To demonstrate
improvements in the thermal stability of the enzymes, each
MDH was incubated at 43°C and added to a reaction mixture
of mannitol and NAD at incremental time points. Half-life
values were found for each enzyme and are summarized in
Table 3. An overall 12-fold improvement from the WT
MDH to the combined mutant was observed from half-lives
of 3.5 ± 0.7 min to 42 ± 18 min. Although 37°C was the final
screening temperature, L-ribose production is considerably
better at 34°C for all the MDH strains, including those
selected at 37°C. Also, the combined round 2 mutant, which
consisted of mutations selected at 34°C, was the most
thermal stable enzyme tested. The positive correlation
between improved enzyme thermal stability and L-ribose
production is a reasonable explanation for the improvements
observed in the shaken flask conversions. The increased
(Lane 5). It is expected that the vast increase in MDH as a
percentage of total soluble protein in the cell contributes to
the increased conversion rates seen in the shaken flask
experiments.
thermal stability observed may also be
a desirable
characteristic for future scale-up processing. Since cooling
expenses are significant for industrial fermentations [35],
strains that permit higher temperatures may be more
economically feasible.
Fig. (4).
Although many chemical syntheses for L-ribose and other
L-sugars have been demonstrated, such methods are not
practical for industrial production due to use of expensive
reagents and organic solvents, multi-step synthesis, and low
yields [36-39]. A biochemical method has also been
described that oxidizes ribitol to L-ribulose with Acetobacter
aceti IFO 3281 washed cells and isomerizes the L-ribulose to
L-ribose using L-ribose isomerase from Acinetobacter sp.
strain DL-28 [40]. While overall yield is around 70% with a
Table 3. Comparision of MDH Thermal Stability
Enzyme
T1/2 @ 43˚C (min)
Ratio T1/2
WT
3.5 ± 0.7
5.8 ± 1.6
16.0 ± 2.1
14.7 ± 1.8
42 ± 18
1.0
1.7
-
1 -1
1-C1
2-H6
2-H5
productivity of 11.7 gL d , the process requires two
separate biochemical steps that may prove inefficient in an
industrial setting [40].
4.6
4.2
Further applications of this work include greatly reducing
the cost of many antiviral pharmaceuticals used for treating
Comb. Rnd 2
12.0
®
®
®
hepatitis B such as Tyzeka , Clevudine , Valtorcitabine
®
and Maribavir [41, 42]. The number of people infected with
hepatitis B worldwide is 350 million [43], with highly
endemic areas including sub-Saharan Africa, the Pacific, and
Asia [44]. The development of the recombinant E. coli L-
ribose production strain could greatly reduce the cost of L-
nucleoside drugs. Decreased cost may also help to make L-
ribose available and affordable to the pharmaceutical and
biochemical research community, thus further spurring the
invention of novel and effective treatments.
SDS-PAGE Comparison of Expression. The expression
of MDH mutants and the WT subcloned in the constitutive
pZuc10 vector was analyzed by SDS-PAGE, which can be
found in Fig. (4). The MDH expressed well with the
expected monomer size of ~39 KDa (mutations did not
change MW significantly). Compared to the WT (Lane 1),
expression improved dramatically in the round 1 MDH
(
Lane 2) and remained consistent with round 2, 2-H6 and 2-
H5 (Lanes 3 and 4, respectively), and combined mutation

Improved Production of L-Ribose from Ribitol
Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4 307
It is expected that scale-up will prove to be simple and
efficient. Previously, the WT MDH as a MBP fusion protein
Wilson, R.M.; Danishefsky, S.J. Preparation and evaluation of
unimolecular pentavalent and hexavalent antigenic constructs
targeting prostate and breast cancer: a synthetic route to anticancer
vaccine candidates. J. Am. Chem. Soc., 2006, 128, 2715-2725.
Woodyer, R.D.; Wymer, N.J.; Racine, F.M.; Khan, S.N.; Saha,
in a similar E. coli host achieved 55% conversion in 72 hours
in a 1 L bioconversion [11]. Considering the vast improve-
ments made via improvements made by directed evolution
higher conversion rates and quicker bioconversion times are
anticipated. Additionally, the bioconversion can be carried
out at a higher temperature and does not require an induction
step or expensive induction reagent, further reducing costs.
[11]
B.C. Efficient production of L-ribose with
Escherichia coli biocatalyst. Appl. Environ. Microbiol., 2008, 74,
967-2975.
a recombinant
2
[
12]
Kawaguchi, T.; Hara, M.; Ueda, M. Process for producing L-
ribose. US Patent 6,348,326, Feb 19, 2002.
[13]
[14]
Seo, M.J.; An, J.; Shim, J.H.; Kim, G. One-pot inversion of D-
mannono-1, 4-lactone for the practical synthesis of L-ribose.
Tetrahedron Lett., 2003, 44, 3051-3052.
Takahashi, H.; Iwai, Y.; Hitomi, Y.; Ikegami, S. Novel synthesis of
L-ribose from D-mannono-1,4-lactone. Org. Lett., 2002, 4, 2401-
In summary, three rounds of directed evolution and a
combining mutations step were performed on the mannitol-
1
-dehydrogenase gene from Apium graveolens subcloned
2
403.
into E. coli. The best mutant, which was a combination of
[15]
Jumppanen, J.; Nurmi, J.; Pastinen, O. Process for the continuous
production of high purity L-ribose. US Patent 6,140, 498: Oct. 31,
2000.
Stoop, J.; Pharr, D. M. Effect of different carbon sources on
relative growth rate, internal carbohydrates, and mannitol 1-
oxidoreductase activity in celery suspension cultures. Plant
Physiol., 1993, 103, 1001-1008.
Stoop, J.; Pharr, D. M. Mannitol metabolism in celery stressed by
excess macronutrients. Plant Physiol., 1994, 106, 503-511.
Stoop, J.M.; Mooibroek, H. Cloning and characterization of
NADP-mannitol dehydrogenase cDNA from the button mushroom,
Agaricus bisporus, and its expression in response to NaCl stress.
Appl. Environ. Microbiol., 1998, 64, 4689-4696.
Stoop, J.M.; Pharr, D.M. Partial purification and characterization of
mannitol: mannose 1-oxidoreductase from celeriac (Apium
graveolens var. rapaceum) roots. Arch. Biochem. Biophys., 1992,
two round 2 mutations, produced L-ribose at a 46.6 ± 1.7%
-
1 -1
conversion and 3.88 ± 0.14 gL d productivity in shaken
flasks; a 19.2-fold improvement compared to the WT. While
enzyme activity for mannitol decreased substantially,
activity for ribitol did not change significantly. It is expected
that the large increase in thermal stability and expression are
responsible for the improved conversion of L-ribose from
ribitol.
[16]
[17]
[18]
ABBREVIATIONS
[
19]
MDH = Mannitol-1-Dehydrogenase
DNS
IPTG = Isopropyl β-D-1-thiogalactopyranoside
RM = Ribitol medium
= 3,5-dinitrosalycylic acid
2
98, 612-619.
[20]
[21]
[22]
Stoop, J.M.; Williamson, J.D.; Conkling, M.A.; Pharr, D.M.
Purification of NAD-dependent mannitol dehydrogenase from
celery suspension cultures. Plant Physiol., 1995, 108, 1219-1225.
Stoop, J. M.H.; Scott Chilton, W.; Mason Pharr, D. Substrate
stereospecificity of the NAD-dependent mannitol dehydrogenase
from celery. Phytochemistry, 1996, 43, 1145-1150.
Stoop, J.M.H.; Williamson, J.D.; Conkling, M.A.; Mackay, J.J.;
Pharr, D.M. Characterization of NAD-dependent mannitol
dehydrogenase from celery as affected by ions, chelators, reducing
agents and metabolites. Plant Sci., 1998, 131, 43-51.
Arnold, F.H.; Volkov, A.A. Directed evolution of biocatalysts.
Curr. Opin. Chem. Biol., 1999, 3, 54-59.
Cadwell, R. C.; Joyce, G. F. Randomization of genes by PCR
mutagenesis. PCR Meth. Appl., 1992, 2, 28-33.
Yang, J.; Fu, X.; Jia, Q.; Shen, J.; Biggins, J. B.; Jiang, J.; Zhao, J.;
Schmidt, J. J.; Wang, P. G.; Thorson, J. S. Studies on the substrate
specificity of Escherichia coli galactokinase. Org. Lett., 2003, 5,
MBP = Maltose Binding Protein
ACKNOWLEDGEMENTS
This work was supported by a small business innovative
research grant from the National Institutes of Health. Thanks
to Badal Saha and Greg Kennedy of USDA NCAUR, Peoria,
IL for HPLC analysis.
[
[
[
23]
24]
25]
REFERENCES
[1]
[2]
[3]
Schreiber, S.L. Target-oriented and diversity-oriented organic
synthesis in drug discovery. Science, 2000, 287, 1964-1969.
Koeller, K.M.; Wong, C.H. Emerging themes in medicinal
glycoscience. Nat. Biotechnol., 2000, 18, 835-841.
Griffith, B.R.; Langenhan, J.M.; Thorson, J.S. 'Sweetening' natural
products via glycorandomization. Curr. Opin. Biotechnol., 2005,
2
223-2226.
[
[
26]
27]
Sarkar, G.; Sommer, S.S. The "megaprimer" method of site-
directed mutagenesis. Biotechniques, 1990, 8, 404-407.
Arnold, K.; Bordoli, L.; Kopp, J.; Schwede, T. The SWISS-
MODEL workspace: a web-based environment for protein structure
homology modeling. Bioinformatics, 2006, 22, 195-201.
1
6, 622-630.
Allen, J.R.; Harris, C.R.; Danishefsky, S.J. Pursuit of optimal
carbohydrate-based anticancer vaccines: preparation of
[28]
[29]
[30]
Sayle, R.A.; Milner-White, E.J. RASMOL: biomolecular graphics
for all. Trends Biochem. Sci., 1995, 20, 374-376.
[
4]
5]
a
Laemmli, U.K. Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature, 1970, 227, 680-685.
Bomati, E.K.; Noel, J.P. Structural and kinetic basis for substrate
selectivity in Populus tremuloides sinapyl alcohol dehydrogenase.
Plant Cell, 2005, 17, 1598-1611.
Altschul, S.F.; Madden, T.L.; Schaffer, A.A.; Zhang, J.; Zhang, Z.;
Miller, W.; Lipman, D.J. Gapped BLAST and PSI-BLAST: a new
generation of protein database search programs. Nucleic Acids Res.,
multiantigenic unimolecular glycopeptide containing the Tn, MBr1,
y
and Lewis antigens. J. Am. Chem. Soc., 2001, 123, 1890-1897.
[
Nicolaou, K.C.; Mitchell, H.J. Adventures in carbohydrate
chemistry: new synthetic technologies, chemical synthesis,
molecular design, and chemical biology. Angew. Chem. Int. Ed.
Engl., 2001, 40, 1576–1624.
[31]
[
[
6]
7]
Rouhi, A. M. Chiral chemistry. Chem. Eng. News, 2004, 82, 47-62.
Lee, Y.J.; Kim, C.S.; Oh, D.K. Lactulose production by beta-
galactosidase in permeabilized cells of Kluyveromyces lactis. Appl.
Microbiol. Biotechnol., 2004, 64, 787-793.
1
997, 25, 3389-3402.
[
32]
33]
McCall, K.A.; Huang, C.; Fierke, C.A. Function and mechanism of
zinc metalloenzymes. J. Nutr., 2000, 130, 1437S-1446S.
Vallee, B.L.; Auld, D.S. Zinc coordination, function, and structure
of zinc enzymes and other proteins. Biochemistry, 1990, 29, 5647-
[
[
[
8]
Pometto, A.L.; Crawford, D.L. Whole-cell bioconversion of
vanillin to vanillic acid by Streptomyces viridosporus. Appl.
Environ. Microbiol., 1983, 45, 1582-1585.
Bartolozzi, A.; Seeberger, P.H. New approaches to the chemical
synthesis of bioactive oligosaccharides. Curr. Opin. Struct. Biol.,
[
5
659.
9]
[
34]
35]
Huxley, A.J.; Griffiths, M.; Levy, M. The New RHS dictionary of
gardening, macmillan: New York, 1999.
Blanch, H.W.; Clark, D.S. Biochemical Engineering. CRC Press:
USA, 1997.
2
001, 11, 587-592.
[
10]
Ragupathi, G.; Koide, F.; Livingston, P.O.; Cho, Y.S.; Endo, A.;
Wan, Q.; Spassova, M.K.; Keding, S.J.; Allen, J.; Ouerfelli, O.;

3
08 Combinatorial Chemistry & High Throughput Screening, 2010, Vol. 13, No. 4
Christ et al.
[36]
Takahashi, H.; Iwai, Y.; Hitomi, Y.; Ikegami, S. Novel synthesis of
L-ribose from D-mannono-1, 4-lactone. Org. Lett., 2002, 4, 2401-
arabinose from ribitol: a new approach. J. Biosci. Bioeng., 1999,
88, 444-448.
2
404.
[41]
[42]
Tam, R.; Averett, D.; Ramasamy, K. Monocyclic L-nucleosides,
analogs and uses thereof. US patent 6,642,206, Nov. 4, 2003.
Casey, J.L.; Korba, B.E.; Cote, P.J.; Gerin, J.L.; Tennant, B.C.;
Chu, C.K. Method of treating hepatitis delta virus infection. US
patent application. 2003/0158149 A1, Aug. 21, 2003.
Rehermann, B.; Nascimbeni, M. Immunology of hepatitis B virus
and hepatitis C virus infection. Nat. Rev. Immunol., 2005, 5, 215-
229.
[37]
Takahashi, H.; Shida, T.; Hitomi, Y.; Iwai, Y.; Miyama, N.;
Nishiyama, K.; Sawada, D.; Ikegami, S. Divergent Synthesis of L-
Sugars and L-Iminosugars from D-Sugars. Chem. Eur. J., 2006, 12,
5
868.
[
[
[
38]
39]
40]
Yun, M.; Moon, H.R.; Kim, H.O.; Choi, W.J.; Kim, Y.C.; Park,
C.S.; Jeong, L.S. A highly efficient synthesis of unnatural L-sugars
from D-ribose. Tetrahedron Lett., 2005, 46, 5903-5905.
Shi, Z.D.; Yang, B.H.; Wu, Y.L. A stereospecific synthesis of L-
deoxyribose, L-ribose and L-ribosides. Tetrahedron, 2002, 58,
[43]
[44]
Lavanchy, D. Hepatitis B virus epidemiology, disease burden,
treatment, and current and emerging prevention and control
measures. J. Viral Hepatitis, 2004, 11, 97-107.
3
287-3296.
Ahmed, Z.; Shimonishi, T.; Bhuiyan, S.H.; Utamura, M.; Takada,
G.; Izumori, K. Biochemical preparation of L-ribose and L-
Received: April 10, 2009
Revised: December 23, 2009
Accepted: December 23, 2009