Introduction

D-xylose, abundant in lignocellulosic biomass, is one of the most abundant carbohydrates in the biosphere. There are four routes for D-xylose metabolism in microorganisms. (1) The isomerase pathway. In numerous bacteria including Escherichia coli [3], Bacillus species [14] and Lactobacillus species [6], xylose is converted by xylose isomerase and xylulose kinase to xylulose-5-phosphate as an intermediate, which is further utilized by the pentose phosphate pathway. (2) The oxo-reductive pathway. In most fungi [12], xylose is transformed into the same intermediate xylulose-5-phosphate via xylose reductase, xylitol dehydrogenase and xylulose kinase, and then turns to the pentose phosphate pathway, as well. (3) Weimberg pathway. In Pseudomonas fragi [20], Caulobacter crecentus [18] and Halophilic Archaeon Haloferax volcanii [13], xylose is converted to α-ketoglutarate, an intermediate of the tricarboxylic acid cycle, by xylose dehydrogenase, xylonolactonase and xylonate dehydratase. (4) Dahms pathway. The Dahms pathway starts with the Weimberg pathway, but 2-keto-3 deoxy-xylonate is split by an aldolase to pyruvate and glycolaldehyde [17].

Gluconobacter oxydans, one of the most widely used biocatalysts in bioindustry and the chiral synthetic field, is able to incompletely oxidize a wide range of carbohydrates and alcohols. Most of these reactions are catalyzed by many highly active dehydrogenases located on the cell membrane, which makes G. oxydans the ideal strain for industrial microbiological production. In contrast to several well-studied carbon sources for G. oxydans, such as sorbitol, mannitol, glycerol and glucose, the mechanism of xylose metabolism pathways in G. oxydans has rarely been studied, thus construction of a xylose metabolism pathway is of significance to completing the metabolism pathway of G. oxydans, and may be of help in figuring out a different pathway from those described above. Deppenmeier and Ehrenreich [4] proposed a pathway similar to the oxo-reductive pathway for xylose metabolism in G. oxydans, as some key genes in the oxo-reductive pathway could be found in the genome of G. oxydans (Fig. 1). Buchert and Viikari [2] experimentally demonstrated that most of xylose was converted to xylonate by xylose dehydrogenase for further oxidation, however little was known about the gene involved in xylose oxidation. This study, therefore, focuses on unveiling the predominant xylose metabolic pathways in G. oxydans by genetic analysis.

Fig. 1
figure 1

Proposed scheme of xylose metabolism pathways in G. oxydans. XR, xylose reductase; XDH, xylitol dehydrogenase; XK, xylulose kinase; mGDH, membrane-bound glucose dehydrogenase; XDA, xylonate dehydratase. Definite reactions are indicated by solid arrows, predicted reactions are shown by broken arrows

Full size image

In this study, to achieve understanding of the mechanisms underlying xylose metabolism in G. oxydans, we observed that the growth profiles (growth rate, xylonate concentration and pH value in the culture broth) of the membrane-bound glucose dehydrogenase deficient strain 621H Δmgdh (the mgdh gene deletion mutant, previously constructed in our lab [22]) on xylose were significantly different from those of the wild-type G. oxydans. Therefore, the role of the membrane-bound glucose dehydrogenase in xylose metabolism was investigated. In addition, two strains deficient in xylitol dehydrogenase (XDH) and putative xylulose kinase (XK), respectively, were also constructed and the functions of both the xdh and the xk genes were investigated.

Materials and methods

Strains, plasmids, and culture conditions

The strains and plasmids used in this study are listed in Table 1. G. oxydans 621H was pre-incubated in yeast–sorbitol (Y–S) medium (yeast extract 20 g/L, sorbitol 80 g/L). To test its growth, different media including yeast–xylose (Y–X) medium (yeast extract 20 g/L, xylose 40 g/L) and yeast–xylitol (Y–XT) medium (yeast extract 20 g/L, xylitol 40 g/L) were used. All G. oxydans strains were cultivated at 30 °C, 220 rpm. E. coli strains were cultivated at 37 °C in Luria–Bertani medium (yeast extract 5 g/L, tryptone 10 g/L, NaCl 10 g/L) containing appropriate antibiotics.

Table 1 Strains and plasmids used in this study
Full size table

Whole-cell biotransformation of xylose to xylonate by G. oxydans

The different G. oxydans cells cultivated in Y–S medium in late-exponential phase were harvested by centrifugation (12,000 rpm, 10 min, 4 °C), washed three times with physiological saline (0.9 % NaCl solution), and then resuspended in 0.1 M phosphate buffer (pH 6.0). Bioconversions were carried out in triplicate in 50 mL shake flasks, containing 3.0 g DCW/L cells and 20 g/L xylose at 220 rpm and 30 °C. Samples (1 mL) were taken at suitable intervals, and the supernatants were obtained after centrifuging at 12,000 rpm and 4 °C for 10 min.

General genetic techniques

Primers used in this study are listed in Table 2. DNA manipulation was in accordance with the standard protocols. For PCRs, genomic DNA isolated from G. oxydans 621H (DSM 2343) was used as a template. LATaq DNA polymerase (TaKaRa Biotechnol., Dalian, China) was used for PCR amplification in test reactions.

Table 2 Primers used in this study
Full size table

Construction of knockout plasmids

Plasmids used in this study are summarized in Table 1. To generate a gene replacement plasmid for the inactivation of the xdh gene, the strategy developed by Derbise [5] was employed. The upstream fragment xdh1 and downstream fragment xdh2 of the xdh gene in G. oxydans [16] as well as the kanamycin resistance gene (1.2 kb) from pET28 were amplified by PCR with primers listed in Table 2. Construction steps of plasmid for disrupting the xdh gene are shown in Fig. 2. The strategy of plasmid construction for the inactivation of the xk gene was similar to that for the xdh gene except for the use of a gentamycin resistance cassette instead of a kanamycin resistance cassette.

Fig. 2
figure 2

Construction of knockout plasmid. The fragment xdh1 containing part of the upstream xdh gene was amplified with primers xdh1F-NcoI and xdh1R-KpnI-XbaI-EcoRI and cloned between the NcoI-EcoRI sites of pSUP202, resulting in plasmid pSUP202-1. The fragment Km resistance gene was amplified with primers KmF-KpnI and KmR-XbaI and cloned between KpnI and XbaI of pSUP202-1, resulting in plasmid pSUP202-2. The fragment xdh2 containing part of the downstream xdh gene was amplified with primers xdh2F-XbaI and xdh2R-EcoRI and cloned between the XbaI -EcoRI sites of pSUP202-2, resulting in plasmid pSUP202-3. The strategy for constructing the plasmid for the inactivation of the xk gene was the same as that for xdh, except that the resistance was replaced by gentamicin

Full size image

Conjugational plasmid transfer into G. oxydans

Knockout plasmid pSUP202 xdh::Km or pSUP202 xk::Gen was transferred into G. oxydans by triparental mating, using E. coli JM109 harboring the knockout plasmid as the donor strain and E. coli HB101 bearing vector pRK2013 as the helper strain. The strategy developed by Holscher and Gorisch [10] was adapted in this study. Three strains were grown to the late-exponential phase, pelleted, washed, resuspended in Y–S medium, and mixed at a 1:1:1 ratio. The mixture was spread on Y–S agar plates without antibiotics and incubated overnight at 30 °C. The grown cell patches were scraped from the plates and resuspended in sterile water, then streaked on selective Y–S agar plates containing appropriate selective antibiotic (gentamycin, kanamycin). Plates were incubated at 30 °C for 3–5 days until resistant colonies appeared. Extracting genomic DNA or plasmids of these resistant colonies, polymerase chain reaction, direct sequencing or restriction enzyme reaction were performed to remove false positives and to identify the objective mutations.

Enzyme activity assay

Xylitol dehydrogenase activity was measured spectrophotometrically at 30 °C by following the substrate-dependent formation of NADH at 340 nm [7]. The standard reaction mixture contained 100 mM Tris–HCl buffer (pH 9.5), 3 mM NAD+ and enzyme. The assay was started by addition of 100 mM xylitol.

Xylulose kinase activity was assayed by coupling the formation of ADP to the oxidation of NADH to NAD+ via pyruvate kinase and lactate dehydrogenase and monitored at 340 nm [9]. The standard reaction mixture contained 50 mM Tris buffer (pH 7.5), 10 mM MgSO4, 1.2 mM ATP, 1.2 mM phosphoenolpyruvate, 0.3 mM NADH, 2 U of pyruvate kinase, 2 U of lactate dehydrogenase, and enzyme. The assay was started by addition of 1 mM D-xylulose.

The enzyme activity assays were carried out in triplicate. One enzyme unit was defined as the amount of the enzyme required to oxidize 1 μmol of substrate in 1 min. Total protein was determined by the Lowry method [11].

Detection of growth conditions

Strains were grown in 250 mL Erlenmeyer flasks containing 50 mL of Y–X or Y–XT medium. Cell growth was monitored by measuring the optical density at 600 nm. Dry cell weight (DCW) was determined by collecting cell pellets from 100 mL of culture aliquots, washing with distilled water, and drying at 100 °C until constant weight was reached. Xylose concentration was determined by xylose assay kit (Megazyme Intl, Wicklow, Ireland) with microplate spectrofluorometer (PowerWave XS/XS2, BioTek Instruments, Winooski, VT) at 554 nm. Xylonate concentration was detected by high-pressure liquid chromatography (HPLC) on an Aminex HPX-87 Ion Exclusion column (Bio-Rad, Hercules, CA) with 5 mM sulfuric acid as a mobile phase. Every growth parameter was measured three times to get an accurate result.

Results

Growth profiles of G. oxydans 621H and 621H Δmgdh in xylose medium

In our previous study [22], we constructed a 621H Δmgdh strain that was deficient in membrane-bound glucose dehydrogenase (mGDH) to improve the growth rate and biomass yield of G. oxydans on glucose. Surprisingly, the growth of the 621H Δmgdh mutant strain on xylose was significantly inhibited compared with that of the wild-type strain (final cell density dropped from 0.648 to 0.052 g DCW/L) (Fig. 3a). In addition, when the wild-type strain was cultured in a xylose medium, the xylonate concentration in the broth gradually accumulated to a final concentration of around 9.3 g/L and the pH value of the medium dropped rapidly after the lag phase, with a final pH value of 3.5 (Fig. 3b–d). All these cultural characteristics of the wild-type strain were similar to the observations by Buchert and Viikari [2]. The only difference was that pure xylose (without glucose addition) was used to cultivate G. oxydans 621H, in comparison to the yeast–xylose medium plus a little glucose used in the above literature.

Fig. 3
figure 3

Growth profiles of G. oxydans 621H (square) and 621H Δmgdh (triangle) in xylose medium. a Growth curves in Y–X medium; b pH value; c Xylose consumption and d Xylonate concentration

Full size image

In contrast, the maximum specific growth rate of the 621H Δmgdh strain on xylose drastically decreased to below one-third of the wild-type strain (Table 3), and no obvious changes were found for the pH value and xylonate accumulation in the culture of the xylose-growing 621H Δmgdh strain. These results imply that xylonate might be an important carbon source for the growth of G. oxydans.

Table 3 Growth kinetic parameters for various G. oxydans strains with xylose or xylitol as the carbon source
Full size table

Whole-cell biotransformation of xylose to xylonate by different G. oxydans

In order to further ascertain whether the mgdh gene was the only gene involved in the bioconversion from xylose to xylonate of G. oxydans, the effects of xylose oxidation and xylonate formation by different whole-cell G. oxydans were also investigated. The different G. oxydans cells cultivated in Y–S medium in the late-exponential phase were harvested as whole-cell catalysts; after the mgdh gene was knocked out, no xylonate was produced, as shown in the HPLC graphics (Fig. 4), which implies that the bioconversion of xylose to xylonate was attributed to the mgdh gene. Therefore, in addition to the known function of oxidizing glucose, this study shows that the mgdh gene also works in the oxidization of xylose and thus plays an essential role in xylose metabolism in G. oxydans.

Fig. 4
figure 4

HPLC graphics of biotransformation of xylose to xylonate by different whole-cell G. oxydans. a Xylonate production by G. oxydans 621H; b Xylonate production by 621H Δmgdh

Full size image

Confirmation of xylitol dehydrogenase and xylulose kinase deficient strains

Gluconobacter oxydans 621H Δmgdh could still grow on xylose, albeit at a low growth rate, which indicates that there might be some other pathways for xylose metabolism in G. oxydans. It has been reported that xylose is mainly utilized by the oxo-reductive pathway in G. oxydans, and most of the genes encoding enzymes of this pathway have already been identified from the genome sequence of G. oxydans [4]. Two key enzymes of this pathway, xylitol dehydrogenase and xylulose kinase, were studied here to confirm whether the oxo-reductive pathway is responsible for xylose metabolism.

Xylitol dehydrogenase (EC 1.1.1.9) is one of the key enzymes of xylose utilization in yeast and fungi, which catalyses the reaction from xylitol to xylulose. In light of the genome sequence of G. oxydans, xylitol dehydrogenase is encoded by gox0865, whose characteristics have been fully investigated [19]. Xylulose kinase (EC 2.7.1.17) is another key enzyme for xylose utilization through the pentose phosphate (PP) pathway, which catalyses the conversion from xylulose to xylulose-5-P. Xylulose kinase has already been successfully isolated from bacteria, fungi and bovine liver. Pival et al. [15] have characterized D-xylulose kinase and recombinantly produced this protein in E. coli to improve the production of ethanol. Bu et al. [1] solubilized abundantly expressed xylulose kinase in E. coli by molecular chaperone GroEL-GroES. The gox2214 gene in G. oxydans was annotated as a putative xylulose kinase. To provide biochemical evidence for the gene function, recombinant G. oxydans XK was successfully obtained from E. coli by expressing it in a soluble form with chaperonins GroEL/GroES and purifying it using Ni–NTA affinity chromatography. Unfortunately, no xylulose kinase activity was detected for the purified enzyme (data not shown).

To confirm the function of the oxo-reductive pathway in xylose metabolism, mutant strains deficient in the gox0865 gene (xdh, encoding xylitol dehydrogenase) or the gox2214 gene (xk, encoding putative xylulose kinase) were constructed using a suicide vector pSUP202 harboring the xdh or xk gene, with an insertion of a kanamycin or gentamicin resistance cassette in the gene. The disruption was isolated as a strain with a phenotype of kanamycin and cephalosporine resistances (KmrCepr) or gentamicin and cephalosporine resistances (GenrCepr) after conjugation transfer of the suicide plasmid pSUP202 xdh::Km or pSUP202 xk::Gen into G. oxydans 621H.

The disruption and absence of the wild type gene was confirmed by PCR. In Fig. 5, a lane sized about 2.4 kb was obtained from the 621H Δxdh strain due to the insertion of the Km resistance gene, and a lane sized about 2.3 kb was gained from the 621H Δxk strain due to the insertion of the Gen resistance gene, suggesting successful knockout of the gox0865 or gox2214 gene.

Fig. 5
figure 5

PCR Confirmation of the disruption and absence of the xdh or xk gene. When xdhF-EcoRI and xdhR-HindIII were used as primers, a lane sized about 1.2 kb was obtained from the wild-type G. oxydans 621H strain, and a lane sized about 2.4 kb was obtained from the mutant 621H Δxdh strain due to the insertion of the Km resistance gene. When xkF-EcoRI and xkR-HindIII were used as primers, a lane sized about 1.6 kb was obtained from 621H, and a lane sized about 2.3 kb was obtained from the mutant 621H Δxk due to the insertion of the Gen resistance gene

Full size image

Enzyme activity detection of xylitol dehydrogenase and xylulose kinase

To provide biochemical evidence of gene functions, enzyme activities of xylitol dehydrogenase and xylulose kinase were assayed for crude cell extracts of G. oxydans 621H mutants grown on xylose. In comparison to the wild-type strain, the xylitol dehydrogenase activity of 621H Δxdh mutant strain decreased by about 75 % (Fig. 6a), suggesting that most of the intracytoplasmic xylitol dehydrogenase was encoded by gox0865. The xylulose kinase activity, however, declined by only 30 % after deletion of the xk gene (Fig. 6b), implying that genes other than gox2214 might be mainly responsible for the phosphorylation of xylulose in G. oxydans.

Fig. 6
figure 6

Enzyme activity detection of the xylitol dehydrogenase and the xylulose kinase

Full size image

Culture of the wild-type and the mutant strains in xylose or xylitol medium

The growth performances of the two mutant strains and the wild-type strain in medium with xylose or xylitol as the carbon source were further investigated to validate the roles of both genes in xylose metabolism in G. oxydans (Fig. 7).

Fig. 7
figure 7

Growth profiles of the wild-type G. oxydans 621H strain (square) and the mutant 621H Δxdh (circle) or 621H Δxk (triangle) strain in xylose or xylitol medium. a Growth profiles in xylose medium (Y–X medium); b Growth profiles in xylitol medium (Y–XT medium)

Full size image

The wild-type strain could grow on xylose, which was quite different from previous reports stating that G. oxydans could not grow on pure xylose without the addition of a small amount of glucose [2]. When the wild-type strain was cultured in a xylose medium, the pH value of the medium started to drop significantly after the lag phase, and a final pH value of 3.5 and a maximum cell density of 0.648 gDCW/L were obtained. The extremely low pH value was caused by large amount of xylonate from the direct oxidation of xylose by membrane-bound glucose dehydrogenase. When xylitol, an intermediate of xylose metabolism, was used as the carbon source, the pH value of the medium decreased slowly and reached a final value of 5.4. The growth of the wild-type strain on xylitol, however, reduced by more than 50 % in DCW compared to that on xylose, implying the presence of other xylose metabolism pathways bypassing xylose, e.g. a xylonate metabolism pathway.

Whether grown on xylose or xylitol, the deficiency of gox2214 gene caused only slight decreases in both maximum specific growth rate and maximum cell density in comparison to those of the wild-type G. oxydans (Fig. 7; Table 3). These experimental results, together with the finding that no xylulose kinase activity was detected for the purified enzyme, suggest that the gox2214 gene might not be the major gene involved in xylose metabolism, and there might be other genes encoding xylulose kinase, which was in accordance with the enzyme activity detection results of both crude and pure enzymes of xylulose kinase.

Removal of xylitol dehydrogenase, however, resulted in more than 33 % (on xylose) and 72 % (on xylitol) decreases in maximum cell density, and 16 % (on xylose) and 46 % (on xylitol) decreases in maximum specific growth rate (Table 3). These results suggest that the xdh gene investigated in this study was associated tightly with xylose utilization in G. oxydans through the oxo-reductive pathway. Moreover, the still fast growth of the 621H Δxdh strain on xylose, albeit slower than the wild-type strain, implies that there must be other pathways responsible for metabolizing the majority of xylose, which supported the above results. These observations suggest that since the gox0865 gene played a relatively major role in the xylitol metabolism, G. oxydans 621H harbored at least one additional gene responsible for the utilization of xylitol.

In summary, based on the growth profiles of the 621H Δmgdh strains, 621H Δxdh and,621H Δxk, it could be easily found that the majority of xylose was directly oxidized to xylonate for further metabolism in the cell. Only a small part of xylose was converted by xylitol dehydrogenase and xylulose kinase to obtain xylulose-5-phosphate for further assimilation via the PP pathway.

Discussion

In this study, we focused on the xylose metabolism of G. oxydans, which to date had received little study, by analyzing the effects of the mgdh, xdh and xk genes on cell growth in either a xylose or a xylitol medium. Genetic disruption of mgdh was sufficient to completely abolish xylonate formation from xylose, which, however, severely inhibited the growth of G. oxydans on xylose (Fig. 7). The whole-cell biotransformation results also confirmed that the mgdh gene was the only gene related to the reaction from xylose to xylonate. The results imply that xylonate was the main carbon source utilized by G. oxydans in vivo. Meanwhile, the genetic and biochemical characterization of xdh showed it was responsible for the intercellular xylose metabolism process via oxo-reductive pathway. From all the experiments, it could be concluded that G. oxydans utilized xylose in two different ways: one was the direct oxidation of xylose into xylonate, which could then be transported through the cytoplasmic membrane for further metabolism; the other way included xylose uptake, intracellular oxidation and further dissimilation via the oxo-reductive pathway (Fig. 1).The former was identified as the major pathway of xylose utilization by G. oxydans during typical fermentation.

A question may arise as to how in vivo xylonate transported into the cytoplasm is metabolized in G. oxydans. At present, we know of the aforementioned four routes for D-xylose metabolism in microorganisms. Among these pathways, xylonate can be generated only by the Weimberg and Dahms pathways. Due to the absence of key enzymes of phosphofructokinase and succinate dehydrogenase, the Embden-Meyerhof-Parnas (EMP) pathway and the tricarboxylic acid (TCA) cycle are not complete in G. oxydans [8]. The deficiency of the two key enzymes in the TCA cycle restrained the utilization of α-ketoglutarate, thus the Weimberg pathway [20] is not likely to contribute to the xylonate metabolism in G. oxydans. Hence, cytoplasmic xylonate can probably be utilized via the Dahms pathway [17] where pyruvate and glycolaldehyde will be generated to enter the central metabolism of G. oxydans (Fig. 1). An upcoming study will include the confirmation of the whole Dahms pathway and determination of xylose distributions between the oxo-reductive pathway and the Dahms pathway. Characterization of the key enzymes, such as D-xylonate dehydratase (catalyzing the conversion of D-xylonate to 3-deoxy-D-pentlulosonic acid) and 3-deoxy-D-pentulosonic acid aldolase (splitting 3-deoxy-D-pentlulosonic acid into pyruvate and glycolaldehyde) will be conducted to ascertain the existence of the Dahms pathway. In addition, metabolic stoichiometry and stable isotope-based flux analysis [21] are being developed to determine the contributions of individual pathways, which could therefore facilitate the engineering of G. oxydans for better growth and synthesis of useful enzymes on xylose from renewable hemicellulose biomass.