APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Aug. 2008, p. 5031–5037
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Vol. 74, No. 16
Engineering Pseudomonas putida S12 for Efficient Utilization of
D-Xylose and L-Arabinose?
Jean-Paul Meijnen,1,2,3* Johannes H. de Winde,2,3and Harald J. Ruijssenaars1,2
TNO-Quality of Life, Business Unit Food and Biotechnology Innovations, Julianalaan 67, 2628 BC Delft, The Netherlands1;
B-Basic, Julianalaan 67, 2628 BC Delft, The Netherlands2; and Department of Biotechnology, Delft University of
Technology, Julianalaan 67, 2628 BC Delft, The Netherlands3
Received 23 April 2008/Accepted 19 June 2008
The solvent-tolerant bacterium Pseudomonas putida S12 was engineered to utilize xylose as a substrate by
expressing xylose isomerase (XylA) and xylulokinase (XylB) from Escherichia coli. The initial yield on xylose
was low (9% [g CDW g substrate?1], where CDW is cell dry weight), and the growth rate was poor (0.01 h?1).
The main cause of the low yield was the oxidation of xylose into the dead-end product xylonate by endogenous
glucose dehydrogenase (Gcd). Subjecting the XylAB-expressing P. putida S12 to laboratory evolution yielded a
strain that efficiently utilized xylose (yield, 52% [g CDW g xylose?1]) at a considerably improved growth rate
(0.35 h?1). The high yield could be attributed in part to Gcd inactivity, whereas the improved growth rate may
be connected to alterations in the primary metabolism. Surprisingly, without any further engineering, the
evolved D-xylose-utilizing strain metabolized L-arabinose as efficiently as D-xylose. Furthermore, despite the
loss of Gcd activity, the ability to utilize glucose was not affected. Thus, a P. putida S12-derived strain was
obtained that efficiently utilizes the three main sugars present in lignocellulosic hydrolysate: glucose, xylose,
and arabinose. This strain will form the basis for a platform host for the efficient production of biochemicals
from renewable feedstock.
The increasing price of oil and imminent shortage of fossil
fuels raise the necessity for the development of alternative
technologies for the production of petrochemicals. The use of
lignocellulosic biomass as feedstock for the chemical industry
is a promising alternative that is being studied widely and
extensively. Ethanol and other biochemicals are currently pro-
duced from glucose by organisms such as Zymomonas mobilis
and Saccharomyces cerevisiae. Although glucose is the primary
sugar in lignocellulosic biomass, a considerable fraction con-
sists of xylose and arabinose, which can make up to 25% of the
total sugar amount (14). Therefore, expanding the substrate
range of whole-cell biocatalysts with these pentose sugars will
greatly contribute to the economic feasibility of biochemical
production from renewable feedstock.
Several approaches have been used to achieve the utilization
of pentose sugars by whole-cell biocatalysts. Expressing xylose
isomerase and/or xylulokinase, encoded by, respectively, xylA
and xylB, has proven to be a successful strategy to enable
phosphorylative growth on xylose (11, 12, 31). Also, genes
encoding xylose reductase and xylitol dehydrogenase have
been employed, especially for engineering yeast cells (6). Still,
problems like redox imbalance or an incomplete pentose phos-
phate (PP) pathway have been encountered, hampering effi-
cient xylose utilization (6, 31). In addition to xylose utilization,
microorganisms have been engineered to utilize arabinose,
e.g., by expressing the AraBAD pathway from Escherichia coli
(10) or Lactobacillus plantarum (30). Ultimately, microorgan-
isms should be engineered to efficiently and concomitantly
utilize glucose, xylose, and arabinose to attain cost-effective
production of biochemicals.
Our laboratory is developing Pseudomonas putida S12 as a
platform for the production of chemicals from renewable feed-
stock via central metabolites as the precursor (17, 18, 26, 29).
P. putida S12 is exceptionally tolerant to organic solvents (1),
which makes this strain an excellent host for the production of
chemicals that are generally toxic to other bacterial cells, such
as substituted aromatic compounds. For these compounds, the
use of mixtures of hexoses and pentoses as substrate may offer
a specific advantage, as they are derived from the aromatic
amino acids L-phenylalanine and L-tyrosine (17, 18, 26, 29).
The key precursors for the aromatic amino acids are phospho-
enol pyruvate and erythrose-4-phosphate, which are respec-
tively derived efficiently from hexoses (via the Entner-Doudor-
off pathway) and pentoses (via the PP pathway). The aim of
this study was to construct a P. putida S12 strain that is capable
of utilizing glucose, xylose, and arabinose to serve as an opti-
mized host strain for efficient, green production of chemicals
from renewable lignocellulose-derived feedstock.
MATERIALS AND METHODS
Culture conditions. The strains and plasmids used in this study are presented
in Table 1. The media used were Luria broth (22) and a phosphate-buffered
mineral salts medium described previously (8). In the mineral salts medium, 10
mM glucose (MMG), 12 mM xylose (MMX), or 12 mM arabinose (MMA) was
used as a sole carbon source unless stated otherwise. Antibiotics were added as
required, in the following concentrations: ampicillin, 100 ?g ml?1(E. coli);
gentamicin, 30 ?g ml?1for Luria broth or 10 ?g ml?1for mineral salts medium;
and kanamycin 50 ?g ml?1. Shake-flask experiments were performed in Boston
bottles containing 20 ml MMG, MMX, or MMA in a horizontally shaking
incubator at 30°C for P. putida S12 or 37°C for E. coli. For E. coli, 26.5 mg liter?1
thiamine was added to the mineral salts medium.
DNA techniques. Genomic DNA was isolated by using a DNeasy tissue kit
(Qiagen). Plasmid DNA was isolated with a QIAprep spin miniprep kit (Qiagen).
DNA concentrations were measured with an ND-1000 spectrophotometer
* Corresponding author. Mailing address: TNO-Quality of Life,
Julianalaan 67, 2628 BC Delft, The Netherlands. Phone: (31) 15-
2789871. Fax: (31) 15-2782355. E-mail: firstname.lastname@example.org.
?Published ahead of print on 27 June 2008.
(Nanodrop). Agarose-trapped DNA fragments were isolated with a QIAEXII
gel extraction kit (Qiagen). PCRs were performed with Accuprime Pfx polymer-
ase (Invitrogen) according to the manufacturer’s instructions. Plasmid DNA was
introduced into electrocompetent cells by using a Gene Pulser electroporation
device. DNA sequencing reactions were performed by MWG Biotech AG.
qPCR. The mRNA levels of the glucose dehydrogenase (Gcd) gene gcd were
analyzed by quantitative PCR (qPCR). Total RNA extractions were performed
with an RNeasy kit (Qiagen). qPCR was performed with oligonucleotide primers
9 and 10 (Table 2) using mRNA of mid-log-phase samples of batch cultures with
a spectrofluorimetric thermal cycler (iCycler thermal cycler equipped with opti-
cal module; Bio-Rad) using IQ Sybr green supermix (Bio-Rad) according to the
Construction of the expression vector pJTmcs. Expression vector pJTmcs was
constructed by using pJTTmcs, formerly named pTac (18) (Table 1), as the
backbone. PCR on pJTTmcs was carried out to amplify the plasmid fragment
containing the tac promoter site, omitting the tac ribosomal binding site (RBS).
The PCR product was cloned into vector pJTTmcs using restriction sites KpnI
and ScaI. The resulting expression vector, pJTmcs, has the same characteristics
as pJTTmcs but contains no tac RBS.
Construction of recombinant plasmids. XylAB was amplified by PCR using
genomic DNA from E. coli DH5? as the template and oligonucleotide primers
1 and 2 (Table 2). The resulting 2.8-kb DNA fragment was ligated into vector
pJTmcs using the restriction sites KpnI and NotI. The resulting plasmid was
The suicide vector pJQ200SK (20) was used to construct a gene replacement
plasmid for the gcd gene as described below. Primers 3 to 6 (Table 2) were used
to amplify 1,158 bp of the 5? end (gcd1) and 951 bp of the 3? end (gcd2) of the gcd
gene. The kanamycin resistance gene, flanked by loxP recombination sites, was
amplified by using primers 7 and 8 on pSK-kanalox (unpublished data) as the
template. pJQ200SK was digested by using restriction enzymes NotI and BamHI,
and gcd1and gcd2were digested with NotI/XbaI and XbaI/BamHI, respectively.
The three resulting fragments were ligated in vector pJQ200SK to yield vector
pJQgcd. pJQgcd was linearized with XbaI, and the cohesive ends were dephos-
phorylated. The loxP-kanaR-loxP fragment, digested with XbaI, was cloned into
the linearized pJQgcd, yielding pJQgcd::kana. This vector was introduced into
wild-type P. putida S12, and transformants were selected for kanamycin resis-
tance. Double-crossover mutants were selected for kanamycin resistance (Kmr)
and gentamicin sensitivity (Gms). The replacement of the native gcd by the
TABLE 1. Strains and plasmids used in this study
Strain or plasmid Characteristic(s)a
Source or reference
E. coli DH5?
P. putida S12
P. putida S12xylAB
P. putida S12xylAB_FGH
P. putida S12xylAB2
P. putida S12xylAB2c
P. putida S12?gcd
P. putida S12?gcd_xylAB
P. putida S12araFGH
P. putida S12 containing plasmid pJTxylAB
P. putida S12 containing plasmid pJTxylAB_FGH
P. putida S12 containing plasmid pJTxylAB, evolved to efficient pentose utilizer
P. putida S12xylAB2 cured from pJTxylAB
P. putida S12 glucose dehydrogenase knockout
P. putida S12 glucose dehydrogenase knockout containing plasmid pJTxylAB
P. putida S12 containing plasmid pBTaraFGH
P. putida S12 containing plasmids pJTxylAB and pBTaraFGH
Hartmans et al. (9)
pJNTmcs(t)AprGmr; basic expression vector derived from plasmid pJWB1 (28) containing the
salicylate-inducible promoter nagR-nagAa
AprGmr; expression vector containing the constitutive tac promoter, derived from
AprGmr; expression vector containing the constitutive tac promoter without the tac
Cmr; expression vector containing the constitutive tac promoter without the tac RBS
pJTmcs containing the xylAB genes from E. coli DH5?
pJTmcs containing the xylAB and xylFGH genes from E. coli DH5?
pBTmcs containing the araFGH genes from E. coli DH5?
P15A ori sacB RP4 Gmr(pBluescriptSK); suicide vector
pJQ200SK containing a Kmr-disrupted copy of the gcd gene
Quandt et al. (20)
aApr, ampicillin resistance; Gmr, gentamicin resistance; Cmr, chloramphenicol resistance; Kmr, kanamycin resistance; ori, origin of replication; sacB, gene producing
TABLE 2. Oligonucleotide primers used in this study
PrimerTarget Sequence (5? 3 3?)a
xylA from E. coli
xylB from E. coli
5? End of gcd
Positions 1154–1134 in gcd
Positions 1261–1281 in gcd
3? End of gcd
5? End of loxP-Kmr-loxP
3? End of loxP-Kmr-loxP
Positions 2088–2106 in gcd
Positions 2266–2285 in gcd
Positions 588–608 in gcd
Positions 397–376 upstream of gcd
KpnI cohesive end
NotI cohesive end
NotI cohesive end
XbaI cohesive end
XbaI cohesive end
BamHI cohesive end
XbaI cohesive end
aThe restriction sites used for cloning are underlined.
5032MEIJNEN ET AL.APPL. ENVIRON. MICROBIOL.
KanaR-disrupted gcd was confirmed by colony PCR using primers 3 and 6 and
sequence analysis of the disrupted gene. Plasmid pJTNcre (unpublished data)
was introduced to cure P. putida S12?gcd from KanaR. This plasmid encodes the
Cre recombinase that catalyzes the site-specific recombination at the loxP target
sites by which the DNA fragment enclosed by the two loxP sites is removed, in
this case KanaR (23, 25). Vector pJTNcre was removed by overnight culturing in
nonselective Luria broth.
Analytical methods. Optical densities were measured at 600 nm using a Bio-
wave cell density meter (WPA Ltd.). An optical density of 1.0 corresponds with
a cell dry weight (CDW) of 465 mg liter?1. Sugars and organic acids were
analyzed by ion chromatography (Dionex ICS3000 system) using a CarboPac
PA20 column with 10 mM NaOH as the eluent for sugars or an IonPac ICE AS6
column with 0.4 mM heptafluorobutyric acid as the eluent for organic acids.
D-Xylulose-5-phosphate (Xu5P) was assessed by using a transketolase activity
Enzyme activity assays. Cell extracts for enzyme assays were prepared by
sonication of 5 ml of concentrated cell suspensions (0.9 g liter?1CDW in 50 mM
Tris-HCl buffer, pH 7.5) from overnight cultures. After centrifugation, superna-
tants were desalted by using PD-10 desalting columns (GE Healthcare). The
resulting cell extracts were used for enzyme assays.
The activity of xylose isomerase (XylA) was determined as described by Gao
et al. (5). In the assay, xylose isomerase activity is coupled to NADH consump-
tion via sorbitol dehydrogenase. The assay was performed at 30°C in a total
volume of 1 ml. The assay mixture contained 50 mM Tris-HCl buffer (pH 7.5),
10 mM MgSO4, 1 mM triethanolamine, 0.5 U sorbitol dehydrogenase, 0.2 mM
NADH, and cell extract. The reaction was started by adding xylose to a final
concentration of 50 mM.
The xylulokinase activity was determined as described by Eliasson et al. (3).
Xylulokinase activity is coupled to the consumption of NADH in the reduction
of pyruvate to lactate by lactate dehydrogenase. The assay was performed at
30°C, in a total volume of 1 ml. The reaction mixture contained 50 mM Tris-HCl
buffer (pH 7.5), 2.0 mM MgCl2, 2.0 mM ATP, 0.2 mM phosphoenolpyruvate, 10
U pyruvate kinase, 10 U lactate dehydrogenase, 0.2 mM NADH, and cell extract.
The reaction was started by adding xylulose to a final concentration of 10 mM.
The transketolase activity was measured to demonstrate D-xylulose-5-phos-
phate formation from L-arabinose or D-xylose. Transketolase couples D-xylulose-
5-phosphate to D-ribulose-5-phosphate to form glyceraldehyde-3-phosphate and
sedoheptulose-7-phosphate. The transketolase reaction is coupled to NADH
consumption via glyceraldehyde-phosphate dehydrogenase. The assay was per-
formed at 30°C, in a total volume of 1 ml. The reaction mixture contained 216
mM glycylglycine buffer (pH 7.7), 1.7 mM D-ribose-5-phosphate, 0.002% (wt/vol)
cocarboxylase, 0.14 mM NADH, 15 mM MgCl2, 2.0 mM ATP, 20 U ?-glyceral-
dehyde-phosphate dehydrogenase–triosephosphate isomerase (based on triose-
phosphate isomerase units), 0.05 U transketolase, and cell extract. The reaction
was started by adding xylose or arabinose to a final concentration of 50 mM.
The pyrroloquinoline quinone (PQQ)-dependent Gcd activity was determined
as described by Liu et al. (16). The activity of Gcd was determined spectropho-
tometrically by measuring the decrease in the absorbance of 2,6-dichloropheno-
lindophenol (DCPIP) at 600 nm. The assay was performed at 30°C, in a total
volume of 1 ml. The reaction mixture contained 50 mM Tris-HCl buffer (pH 7.5),
15 mM NH4Cl, 80 ?M DCPIP, 1 ?M KCN, 0.33 mM phenazine methosulfate,
and cell extract. The reaction was started by adding glucose to a final concen-
tration of 1 mM.
For calculations of enzyme activities, the following molar extinction coeffi-
cients were used: 6.22 mM?1cm?1for NADH and 19 mM?1cm?1for reduced
DCPIP. One unit is defined as the amount of enzyme that oxidizes 1 ?mol of
substrate per minute in the coupled assays described above.
Cloning and functional expression of xylAB in P. putida S12.
Genes xylA and xylB, part of the xyl operon of E. coli, were
cloned into expression vector pJTmcs under the transcriptional
control of the constitutive tac promoter. The resulting vector,
pJTxylAB, was introduced into P. putida S12, yielding P. putida
S12xylAB. The results of enzyme assays confirmed that xylose
isomerase and xylulokinase were expressed as functional en-
zymes. The specific activities were 34 U g?1protein for xylose
isomerase and 134 U g?1protein for xylulokinase. The overall
activity of XylAB was 22 U g?1.
After demonstrating that XylAB were functionally ex-
pressed, P. putida S12xylAB was tested for its ability to utilize
xylose as a carbon source. When strain S12xylAB was inocu-
lated into xylose-containing medium, growth was observed.
However, the biomass yield was low (9% [g CDW g sub-
strate?1]) (Fig. 1A) compared to the biomass yield on glucose
(typically 55%). The growth rate was also much lower on xylose
than on glucose (0.01 h?1versus 0.5 h?1). The same charac-
teristics were found for strain S12xylAB_FGH, indicating that
xylose transport was not limiting xylose utilization (Table 3). It
FIG. 1. (A) Growth of wild-type P. putida S12 (triangles) and P.
putida S12xylAB (diamonds) cells in mineral salts medium with xylose
as the sole carbon source. Data points are the averages of the results
of duplicate measurements. Error bars represent the maximum devi-
ations from the averages. (B) Xylonate (squares) and xylose (circles)
concentrations in P. putida S12xylAB culture on MMX. P. putida
S12xylAB cells were grown in mineral salts medium with xylose as the
sole carbon source. CDW concentrations are presented by diamonds.
Data points are the averages of the results of duplicate measurements.
Error bars represent the maximum deviations of the averages. L, liter.
TABLE 3. Overview of growth characteristics of pentose-utilizing
P. putida S12-derived strainsa
(%; g CDW g
(%; g CDW g
aNG, no growth; ND, not determined.
VOL. 74, 2008 ENGINEERING P. PUTIDA TO UTILIZE XYLOSE AND ARABINOSE5033
was observed previously that P. putida S12 oxidizes xylose to
xylonate, a reaction shown to be catalyzed in other P. putida
strains by PQQ-dependent Gcd (7). Also in xylose-grown P.
putida S12xylAB cultures, 81% of the initial amount of xylose
was oxidized to xylonate (Fig. 1B), rendering a large part of the
xylose unavailable for growth. When the biomass formation in
such cultures was related to the amount of xylose that was not
oxidized, an apparent yield of 47% (g CDW g substrate?1) was
Optimizing xylose utilization by laboratory evolution. To
optimize the biomass yield and growth rate on xylose, a labo-
ratory evolution approach similar to that described by Kuyper
et al. was applied (13). Two parallel cultures were maintained
and repeatedly transferred to fresh minimal medium with xy-
lose. Throughout, the best performing of the two parallel cul-
tures was used as inoculum for the next transfer while the other
was discarded. Three stages could be discriminated in the
evolutionary process (Fig. 2). In the first stage, transformant
S12xylAB strains were selected for increased biomass yield.
When the culture entered the stationary phase, the culture
with the highest biomass yield was selected for further evolu-
tion. After approximately 20 transfers, the biomass yield sta-
bilized at 52%, comparable to the yield on glucose (55%).
Although the growth rate increased with increasing biomass
yield, it stabilized when the maximum yield had been achieved,
at only 0.05 h?1. Since this was an order of magnitude lower
than the growth rate on glucose, the evolutionary approach was
continued to select for a strain with a higher growth rate.
Instead of transferring the culture with the highest biomass
yield, the faster-growing cultures were selected and reinocu-
lated into fresh medium. After 10 transfers, a strain was ob-
tained that exhibited a significantly higher growth rate of 0.35
h?1. At this stage of the evolutionary procedure, the cultures
were found to lyse upon entering the stationary phase. Evolu-
tion was therefore prolonged to select for a strain that was less
prone to lysis. With every transfer, the susceptibility to lysis
decreased, ultimately leading to a nonlysing and efficiently
xylose-utilizing P. putida S12-derived strain that was desig-
Characterization of P. putida S12xylAB2. The increased bio-
mass yield on xylose after laboratory evolution provided a
strong indication that xylose oxidation was affected in P. putida
S12xylAB2. Indeed, no xylonate formation was observed dur-
ing growth on xylose, and also, glucose was no longer oxidized
to its corresponding aldonic acids. This suggested that Gcd had
become inactive during the evolutionary procedure, which was
confirmed by the results of Gcd assays. Sequence analysis of
the gcd gene showed no mutations, indicating that the absence
of active Gcd was caused at a different level. qPCR was used to
determine the gcd transcript levels. The mRNA concentration
in P. putida S12 cells was 7.00 ? 1.1 ng ?l?1(average ?
standard deviation of the results from three independent ex-
ponentially growing cultures); that in P. putida S12?gcd cells
was 10.62 ? 2.0 ng ?l?1; and that in P. putida S12xylAB2 cells
was 9.10 ? 0.8 ng ?l?1. Surprisingly, the gcd mRNA level in
S12xylAB2 was increased in comparison to that of the wild-
type S12 strain, eliminating impaired transcription of gcd as an
explanation for the absence of Gcd activity. Also, inefficient
translation of gcd mRNA is unlikely to have caused the abol-
ishment of Gcd activity as no mutations were found in the gcd
RBS. Therefore, it is proposed that the inactivation of Gcd
results from a yet-unidentified posttranslational event.
In order to investigate whether the improved growth char-
acteristics of the evolved strain on xylose could be attributed to
the absence of Gcd activity, the gcd gene was disrupted in
wild-type P. putida S12. The resulting strain, P. putida S12?gcd,
was transformed with plasmid pJTxylAB to enable xylose uti-
lization. Like P. putida S12xylAB2, strain S12?gcd_xylAB uti-
lized both glucose and xylose as the sole carbon source without
formation of the associated aldonates. Also in strain
S12?gcd_xylAB, in which 107 bp were removed from gcd, the
gcd mRNA level was increased with respect to the level in the
wild-type S12 strain (see above). Compared to the yield of
S12?gcd_xylAB was considerably improved (44% [g CDW g
xylose?1]), but the growth rate was equally low (0.01 h?1).
Thus, the absence of Gcd activity explains only part of the
improved xylose utilization. Mutations in the xylAB genes of
strain S12xylAB2 resulting in a higher xylose conversion rate
could be excluded by sequence analysis, which was confirmed
by the results of XylAB activity measurements (not shown).
Utilization of arabinose and mixtures of glucose and pen-
toses by P. putida S12xylAB2. The evolved strain S12xylAB2
was able to efficiently utilize xylose, as well as glucose, despite
the apparent loss of a key enzyme activity for direct oxidative
glucose metabolism. For optimal utilization of lignocellulose-
derived feedstock, L-arabinose should also be metabolized in
addition to glucose and D-xylose. Although strain S12xylAB2
was not specifically engineered for arabinose utilization, the
introduced xylose metabolic enzymes may show nonspecific
activity toward L-arabinose, a C4 epimer of D-xylose (19, 21).
The ability of strain S12xylAB2 to utilize L-arabinose was
therefore assessed by growth in mineral salts medium contain-
ing 12 mM arabinose. Surprisingly, L-arabinose was utilized as
a carbon source at an efficiency identical to that of growth on
xylose, with a biomass yield of 52% (g CDW g arabinose?1)
and a maximum growth rate of 0.35 h?1.
Strain S12xylAB2 lost the ability to utilize arabinose when
cured from pJTxylAB (Table 3), demonstrating that the xylAB
genes introduced for xylose consumption were also essential for
arabinose consumption. Also, the evolutionary procedure appar-
ently made a key contribution to efficient arabinose consumption
FIG. 2. Laboratory evolution of P. putida S12xylAB. Transformant
cells were repeatedly transferred into fresh xylose-containing medium
in order to optimize biomass yield (diamonds) and growth rate
(squares) on MMX.
5034 MEIJNEN ET AL.APPL. ENVIRON. MICROBIOL.
since the nonevolved strain S12xylAB did not utilize arabinose
(Table 3). The results of previous work demonstrated that P.
putida S12 expressing the AraBAD pathway that converts L-ar-
abinose into D-xylulose-5-phosphate did not grow on arabinose
unless the high-efficiency arabinose transporter AraFGH was co-
expressed (unpublished data). The coexpression of AraFGH and
strain, whereas the introduction of araFGH alone did not estab-
lish growth on arabinose (Table 3). These results suggest that the
evolutionary procedure improved arabinose transport efficiency.
The effect of Gcd having become inactive probably plays a minor
role, as wild-type P. putida S12 oxidizes arabinose only to a very
limited extent (not shown).
The involvement of XylA and XylB in L-arabinose metabolism
was further confirmed by the results of enzyme measurements
(Table 4). NADH consumption was observed when L-arabinose
was added as the substrate instead of D-xylose when assaying
xylose isomerase and xylulokinase in the cell extract of strain
S12xylAB2. In addition, D-xylulose-5-phosphate from L-arabinose
could be detected with the transketolase assay (Table 4). Since
the product of L-arabinose formed by XylAB is expected to be
L-ribulose-5-phosphate, the formation of D-xylulose-5-phosphate
suggests the presence of a C4 epimerase activity that remains to
be identified. The observation that the nonevolved strain
S12xylAB_araFGH also utilizes L-arabinose suggests that the C4
epimerase activity is endogenous to P. putida S12 and not the
result of the evolutionary procedure.
Finally, the growth of strain S12xylAB2 on mixtures of sugars
was investigated. Cells were inoculated into mineral salts medium
containing glucose and xylose (MMGX); glucose and arabinose
(MMGA); or glucose, xylose, and arabinose (MMGXA). Timed
samples were drawn and analyzed for CDW and sugar content.
The results show that all sugars in the tested combinations are
consumed (Fig. 3). A diauxic shift was observed in all cultures:
only after glucose was depleted were the pentoses consumed.
When both xylose and arabinose were present in addition to
glucose, the pentose sugars were utilized simultaneously when
cells were deprived of glucose.
A P. putida S12 strain was constructed that efficiently utilizes
D-xylose, as well as L-arabinose. The expression of xylose
isomerase and xylulokinase is essential for the utilization of
both pentoses, but the subsequent laboratory evolution is key
to the efficiency with which these pentoses are metabolized.
The improved yield on xylose attained by the evolved strain
could largely be attributed to Gcd having become inactive in
the evolved strain, preventing xylose “loss” as a result of oxi-
dation to xylonate. Although targeted disruption of the gcd
gene in wild-type P. putida S12 resulted in an improved bio-
mass yield on xylose, this strategy did not result in an improved
growth rate, indicating that other changes occurred in the
evolved strain. It may be speculated that mutations occurred
that affected the metabolic fluxes through the PP pathway,
which is the expected route by which xylose is metabolized. In
FIG. 3. Growth of P. putida S12xylAB2 (CDW is represented by
squares) and consumption of glucose (diamonds), xylose (triangles),
and arabinose (circles). P. putida S12xylAB2 was inoculated into min-
eral salts medium containing glucose and xylose (A); glucose and
arabinose (B); or glucose, xylose, and arabinose (C). Data points are
the averages of the results of duplicate measurements. Error bars
represent the maximum deviations of the averages. L, liter.
TABLE 4. Key enzyme activities for pentose utilization in cell
extracts of P. putida S12xylAB2 with xylose or arabinose
Activity (U g?1) on:
XylA (pentose isomerase)
XylB (“pentulose” kinase)
XylAB (combined isomerase/kinase)
Putative C4 epimerased
aND, not determined.
bThe activity assay of XylA with L-arabinose was compromised by the low
affinity of sorbitol dehydrogenase for L-ribulose.
cThe activity of XylB was not determined separately as L-ribulose is not
dThe formation of D-xylulose-5-phosphate from L-arabinose was quantified by
measuring transketolase activity.
VOL. 74, 2008ENGINEERING P. PUTIDA TO UTILIZE XYLOSE AND ARABINOSE5035
P. putida, a complete PP pathway is present, but metabolic flux
analyses on Pseudomonas fluorescens have shown that this
pathway mainly serves to replenish biosynthetic intermediates
(4). Similarly, modest flux distributions have been demon-
strated for P. putida S12 (unpublished data). Therefore, the
function of the PP pathway may have changed from anabolic to
catabolic in the evolved P. putida S12xylAB2 strain.
The observation that D-xylose and L-arabinose are consumed
with equal efficiency, both in terms of biomass yield and spe-
cific growth rate, suggests that both pentoses are consumed via
the PP pathway. In addition, the evolved strain appears to have
acquired an efficient L-arabinose uptake system, as wild-type P.
putida S12 required the expression of both XylAB and the
AraFGH transporter for arabinose utilization. Although coex-
pressing a high-affinity xylose transporter (XylFGH from E.
coli) did not have a significant effect on xylose metabolism in
the nonevolved strain (Table 3), the possibility that improved
xylose uptake has contributed to more-efficient xylose utiliza-
tion in strain S12xylAB2 cannot be excluded. Since pentose
transporters have been shown to be promiscuous (24), it may
be hypothesized that efficient L-arabinose uptake has co-
evolved with improved D-xylose uptake in strain S12xylAB2.
At this point it is unclear how arabinose is converted into a
PP pathway intermediate (Fig. 4). The results of enzyme assays
showed that L-arabinose is a substrate for XylAB, and growth
on arabinose did not occur without the expression of XylAB.
The expected product, L-ribulose-5-phosphate, is not a central
pathway intermediate, and a C4 epimerization would be re-
quired to form D-xylulose-5-phosphate. Indeed, the formation
of D-xylulose-5-phosphate from L-arabinose was suggested by
the results of the transketolase assay, but no indications that
this strain contains an AraD homologue were found in the P.
putida S12 genome sequence (unpublished data). Therefore, it
is proposed that the endogenous ribulose-5-phosphate-3-epi-
merase shows nonspecific epimerization activity on L-ribulose-
5-phosphate, converting the molecule into D-xylulose-5-phos-
phate. Further research is ongoing to confirm this hypothesis.
Despite the loss of Gcd activity, glucose was still efficiently
used as the sole carbon source by the evolved P. putida
S12xylAB2. The glucose catabolism in P. putida operates
through the action of three simultaneous pathways that con-
verge at 6-phosphogluconate. Glucose is preferentially oxi-
dized in the periplasm by Gcd to (2-keto-)gluconate and
subsequently phosphorylated in the cytoplasm to yield 6-phos-
phogluconate (2, 15). Alternatively, glucose is imported by an
ABC-transport system, phosphorylated by glucokinase, and ox-
idized to 6-phosphogluconate (2, 15). Apparently, the evolved
xylose-utilizing strain can readily switch to this alternative
pathway for glucose oxidation without affecting the yield or
The absence of active Gcd may itself provide an explanation
for the increased gcd transcription levels observed in both the
evolved strain and the gcd knockout strain. Glucose induces
the expression of gcd (27), and with active Gcd present, glucose
is rapidly oxidized to gluconate and 2-ketogluconate, resulting
in a swift downregulation of gcd. However, without active Gcd,
glucose persists in the medium, resulting in increased levels of
gcd mRNA. So, with an intact gcd gene and associated RBS
present in the evolved strain, and apparently even increased
transcription levels, the absence of active Gcd must be attrib-
uted to some posttranslational effect. The amino acid sequence
of Gcd shows that the protein is excreted and that it contains
four transmembrane regions (not shown), which is consistent
with the periplasmic oxidation of sugars. It raises the possibility
that malfunctions appear in the Gcd translocation machinery,
leading to faulty localization of the enzyme, improper folding,
or inadequate anchoring to the inner membrane. The exact
cause of the inactivity of Gcd remains to be investigated.
In conclusion, a P. putida S12 strain was obtained that effi-
ciently utilizes the three most-abundant sugars in lignocellu-
lose, glucose, xylose, and arabinose, as sole carbon sources.
The applied evolutionary approach proved to be a powerful
method to optimize the initial inefficient xylose-utilizing strain.
Transcriptome and proteome analyses, as well as metabolic
flux analysis, are currently being performed to identify the
changes in the metabolism of the evolved xylose-utilizing
strain. The insight gained into the molecular background of the
efficient pentose utilization will be employed to incorporate
this property into optimized substitute-aromate-producing P.
putida S12-derived strains, thereby contributing to the eco-
nomical feasibility of the production of such biochemicals from
We thank Nicole van Luijk for developing the markerless gene
disruption system in P. putida S12, Hendrik Ballerstedt for performing
qPCR experiments, Maaike Westerhof for practical assistance, and Jan
Wery for helpful discussions.
This project was financially supported by The Netherlands Ministry
of Economic Affairs and the B-Basic partner organizations (www
.b-basic.nl) through B-Basic, a public-private NWO-ACTS (Advanced
Chemical Technologies for Sustainability) program.
1. de Bont, J. A. 1998. Solvent-tolerant bacteria in biocatalysis. Trends Bio-
2. del Castillo, T., J. L. Ramos, J. J. Rodrı ´guez-Herva, T. Fuhrer, U. Sauer, and
E. Duque. 2007. Convergent peripheral pathways catalyze initial glucose
catabolism in Pseudomonas putida: genomic and flux analysis. J. Bacteriol.
3. Eliasson, A., E. Boles, B. Johansson, M. O ¨sterberg, J. M. Thevelein, I.
Spencer-Martins, H. Juhnke, and B. Hahn-Ha ¨gerdal. 2000. Xylulose fer-
mentation by mutant and wild-type strains of Zygosaccharomyces and Sac-
charomyces cerevisiae. Appl. Microbiol. Biotechnol. 53:376–382.
4. Fuhrer, T., E. Fischer, and U. Sauer. 2005. Experimental identification and
FIG. 4. Utilization of D-xylose and L-arabinose may proceed via
(partly) shared pathways. Indications were found that L-arabinose can
be converted into D-xylulose-5-phosphate by the combined action of
XylAB and a yet-unidentified C4 epimerase (indicated by the question
5036MEIJNEN ET AL.APPL. ENVIRON. MICROBIOL.
quantification of glucose metabolism in seven bacterial species. J. Bacteriol.
5. Gao, Q., M. Zhang, J. D. McMillan, and D. S. Kompala. 2002. Character-
ization of heterologous and native enzyme activity profiles in metabolically
engineered Zymomonas mobilis strains during batch fermentation of glucose
and xylose mixtures. Appl. Biochem. Biotechnol. 98–100:341–355.
6. Hahn-Ha ¨gerdal, B., K. Karhumaa, M. Jeppsson, and M. F. Gorwa-Grauslund.
2007. Metabolic engineering for pentose utilization in Saccharomyces cerevisiae.
Adv. Biochem. Eng. Biotechnol. 108:147–177.
7. Hardy, G. P., M. J. Teixeira de Mattos, and O. M. Neijssel. 1993. Energy
conservation by pyrroloquinoline quinol-linked xylose oxidation in Pseudo-
monas putida NCTC 10936 during carbon-limited growth in chemostat cul-
ture. FEMS Microbiol. Lett. 107:107–110.
8. Hartmans, S., J. P. Smits, M. J. van der Werf, F. Volkering, and J. A. M. de
Bont. 1989. Metabolism of styrene oxide and 2-phenylethanol in the styrene-
degrading Xanthobacter strain 124X. Appl. Environ. Microbiol. 55:2850–
9. Hartmans, S., M. J. van der Werf, and J. A. de Bont. 1990. Bacterial
degradation of styrene involving a novel flavin adenine dinucleotide-depen-
dent styrene monooxygenase. Appl. Environ. Microbiol. 56:1347–1351.
10. Kawaguchi, H., M. Sasaki, A. A. Vertes, M. Inui, and H. Yukawa. 27 October
2007. Engineering of an L-arabinose metabolic pathway in Corynebacterium
glutamicum. Appl. Microbiol. Biotechnol. doi:10.1007/s00253-007-1244-x.
11. Kawaguchi, H., A. A. Verte `s, S. Okino, M. Inui, and H. Yukawa. 2006.
Engineering of a xylose metabolic pathway in Corynebacterium glutamicum.
Appl. Environ. Microbiol. 72:3418–3428.
12. Kuyper, M., H. R. Harhangi, A. K. Stave, A. A. Winkler, M. S. Jetten, W. T.
de Laat, J. J. den Ridder, H. J. Op den Camp, J. P. van Dijken, and J. T.
Pronk. 2003. High-level functional expression of a fungal xylose isomerase:
the key to efficient ethanolic fermentation of xylose by Saccharomyces cer-
evisiae? FEMS Yeast Res. 4:69–78.
13. Kuyper, M., M. J. Toirkens, J. A. Diderich, A. A. Winkler, J. P. van Dijken,
and J. T. Pronk. 2005. Evolutionary engineering of mixed-sugar utilization by
a xylose-fermenting Saccharomyces cerevisiae strain. FEMS Yeast Res.
14. Lee, J. 1997. Biological conversion of lignocellulosic biomass to ethanol.
J. Biotechnol. 56:1–24.
15. Lessie, T. G., and P. V. Phibbs, Jr. 1984. Alternative pathways of carbohy-
drate utilization in pseudomonads. Annu. Rev. Microbiol. 38:359–388.
16. Liu, Q., J. R. Kirchhoff, C. R. Faehnle, R. E. Viola, and R. A. Hudson. 2006.
bacterium extorquens. Protein Expr. Purif. 46:316–320.
17. Nijkamp, K., N. van Luijk, J. A. de Bont, and J. Wery. 2005. The solvent-
tolerant Pseudomonas putida S12 as host for the production of cinnamic acid
from glucose. Appl. Microbiol. Biotechnol. 69:170–177.
18. Nijkamp, K., R. G. Westerhof, H. Ballerstedt, J. A. de Bont, and J. Wery.
2007. Optimization of the solvent-tolerant Pseudomonas putida S12 as host
for the production of p-coumarate from glucose. Appl. Microbiol. Biotech-
19. Pastinen, O., K. Visuri, H. E. Schoemaker, and M. Leisola. 1999. Novel
reactions of xylose isomerase from Streptomyces rubiginosus. Enzyme Microb.
20. Quandt, J., and M. F. Hynes. 1993. Versatile suicide vectors which allow
direct selection for gene replacement in gram-negative bacteria. Gene 127:
21. Richard, P., M. H. Toivari, and M. Penttila ¨. 2000. The role of xylulokinase
in Saccharomyces cerevisiae xylulose catabolism. FEMS Microbiol. Lett. 190:
22. Sambrook, J., T. Maniatis, and E. Fritsch. 1982. Molecular cloning: a lab-
oratory manual. Cold Spring Harbor Press, Cold Spring Harbor, NY.
23. Sauer, B., and N. Henderson. 1988. Site-specific DNA recombination in
mammalian cells by the Cre recombinase of bacteriophage P1. Proc. Natl.
Acad. Sci. USA 85:5166–5170.
24. Song, S., and C. Park. 1998. Utilization of D-ribose through D-xylose trans-
porter. FEMS Microbiol. Lett. 163:255–261.
25. Sternberg, N., and D. Hamilton. 1981. Bacteriophage P1 site-specific recom-
bination. I. Recombination between loxP sites. J. Mol. Biol. 150:467–486.
26. Verhoef, S., H. J. Ruijssenaars, J. A. de Bont, and J. Wery. 2007. Biopro-
duction of p-hydroxybenzoate from renewable feedstock by solvent-tolerant
Pseudomonas putida S12. J. Biotechnol. 132:49–56.
27. Vicente, M., and J. L. Ca ´novas. 1973. Regulation of the glucolytic enzymes
in Pseudomonas putida. Arch. Microbiol. 93:53–64.
28. Wery, J., B. Hidayat, J. Kieboom, and J. A. de Bont. 2001. An insertion
sequence prepares Pseudomonas putida S12 for severe solvent stress. J. Biol.
29. Wierckx, N. J., H. Ballerstedt, J. A. de Bont, and J. Wery. 2005. Engineering
of solvent-tolerant Pseudomonas putida S12 for bioproduction of phenol
from glucose. Appl. Environ. Microbiol. 71:8221–8227.
30. Wisselink, H. W., M. J. Toirkens, M. D. Franco Berriel, A. A. Winkler, J. P.
van Dijken, J. T. Pronk, and A. J. van Maris. 2007. Engineering of Saccha-
romyces cerevisiae for efficient anaerobic alcoholic fermentation of L-arabi-
nose. Appl. Environ. Microbiol. 73:4881–4891.
31. Zhang, M., C. Eddy, K. Deanda, M. Finkelstein, and S. Picataggio. 1995.
Metabolic engineering of a pentose metabolism pathway in ethanologenic
Zymomonas mobilis. Science 267:240–243.
VOL. 74, 2008 ENGINEERING P. PUTIDA TO UTILIZE XYLOSE AND ARABINOSE5037