• Open Access

Autohydrolysis of plant xylans by apoplastic expression of thermophilic bacterial endo-xylanases

Authors

  • Bernhard Borkhardt,

    Corresponding author
    1. Department of Genetics and Biotechnology, Faculty of Agricultural Sciences, Aarhus University, Thorvaldsensvej, Frederiksberg C, Denmark
    Search for more papers by this author
  • Jesper Harholt,

    1. Department of Plant Biology and Biotechnology, VKR Research Centre “Pro-Active Plants”, University of Copenhagen, Thorvaldsensvej, Frederiksberg C, Denmark
    Search for more papers by this author
  • Peter Ulvskov,

    1. Department of Plant Biology and Biotechnology, VKR Research Centre “Pro-Active Plants”, University of Copenhagen, Thorvaldsensvej, Frederiksberg C, Denmark
    Search for more papers by this author
  • Birgitte K. Ahring,

    1. Department of Biotechnology, University of Aalborg, Lautrupvang, Ballerup, Denmark
    Search for more papers by this author
  • Bodil Jørgensen,

    1. Department of Genetics and Biotechnology, Faculty of Agricultural Sciences, Aarhus University, Thorvaldsensvej, Frederiksberg C, Denmark
    Search for more papers by this author
  • Henrik Brinch-Pedersen

    1. Department of Genetics and Biotechnology, Faculty of Agricultural Sciences, Aarhus University, Flakkebjerg, Denmark
    Search for more papers by this author

*(fax +45 3533 2589; email b.borkhardt@dias.kvl.dk)

Summary

The genes encoding the two endo-xylanases XynA and XynB from the thermophilic bacterium Dictyoglomus thermophilum were codon optimized for expression in plants. Both xylanases were designed to be constitutively expressed under the control of the CaMV 35S promoter and targeted to the apoplast. Transient expression in tobacco and stable expression in transgenic Arabidopsis showed that both enzymes were expressed in an active form with temperature optima at 85 °C. Transgenic Arabidopsis accumulating heterologous endo-xylanases appeared phenotypically normal and were fully fertile. The highest xylanase activity in Arabidopsis was found in dry stems indicating that the enzymes were not degraded during stem senescence. High levels of enzyme activity were maintained in cell-free extracts from dry transgenic stems during incubation at 85 °C for 24 h. Analysis of cell wall polysaccharides after heat treatment of wildtype and transgenic extracts from dry stems showed a decrease in the molecular weight of xylans from transgenic stems.

Introduction

The use of plant biomass for the production of biofuel has drawn much interest as a viable alternative to fossil fuels. The technologies needed to allow this conversion process have been developed, and the major challenge now is to make this process cost-competitive with fossil fuels. Plant biomass is converted to fermentable sugars using pre-treatment processes that disrupt the lignocellulose and thus allowing the access of microbial enzymes for cellulose deconstruction. Both the pre-treatment processes and the application of commercial enzymes are rather expensive (for a review see (Sticklen, 2008)).

Plant biomass used for 2nd generation biofuel production is composed nearly solely of secondary cell walls. Cellulose, hemicellulose and lignin are the three major components of the secondary cell wall. In dicots, the major hemicellulose is normally xyloglucan, but in some tissues, the primary hemicellulose is xylan with varying side chain composition. The major hemicellulose in grasses, the main feedstock for biofuel production, is arabinoxylan or glucuronoarabinoxylan depending on tissue and ratio between primary and secondary cell wall (Ebringerova et al., 2005). The structure of xylan is a chain of β-(1,4)-linked xylose moieties to which are attached side chains, including arabinose, acetate and methyl-glucuronic acid (Carpita, 1996). The arabinose in grasses can further be substituted with the ester-linked hydroxycinnamates ferulic acid or p-coumaric acid (Carpita, 1996). The ferulic acids can cross-link increasing the MW of the xylans and decreasing the water extractability (Grabber et al., 2004). It has furthermore been hypothesised that it can function as initiation site and anchoring point for lignin, hereby linking the xylan to the lignin network (Grabber et al., 2004). In addition to the grass-specific hydroxycinnamate cross-linking, unsubstituted regions of the hemicellulose may associate to the cellulose microfibrils through hydrogen bonding.

Arabidopsis inflorescence stems develop a substantial amount of xylem and interfascicular fibres, unusually rich in xylan originating from the secondary wall (Gardner et al., 2002). The xylose content in alcohol insoluble residues from Arabidopsis stems is around 70 mol% as is also typical of many grass species of relevance to biofuel (Gomez et al., 2008). Xylan in Arabidopsis is believed mainly to consist of glucuronoxylan with α-d-glucuronic acid and/or 4-O-methyl-α-d-glucuronic acid at every 6–12 xylose residues. The xylose backbone may also be attached with L-arabinose and/or acetylated on C-2 or C-3 positions (Ebringerova and Heinze, 2000). A comparable degree of backbone substitution is found in grasses, exemplified by wheat, which has a xyl:ara ratio of app 12 in its xylan (Lequart et al., 1999). The xylan of Arabidopsis and the typical grass may thus be expected to be equally prone to degradation by xylanases, and Arabidopsis is thus useful as model for transgenic approaches to controlling xylan-mediated recalcitrance of the secondary walls.

The role of xylan in the secondary wall is not clear. However, it has been suggested that xylans coat the cellulose microfibrils (Awano et al., 2002) and may influence the helicoidal orientation of the microfibrils (Reis and Vian, 2004). Xylan contributes to the tensile strength of plant secondary walls and functions to limit the accessibility of cellulases to the cell wall thereby preventing damage to the plant surface by challenge of plant pathogens.

The recalcitrance of plant biomass to degradation is a function of cross-linking and aggregation within the cell wall. Recent advances in plant genetic engineering may reduce biomass recalcitrance by developing crop varieties that self-produce the enzymes that will disrupt the cross-links between hemicellulose and the other cell wall components or degrade the xylan backbone.

Reduced cross-linkage has been pursued in tall fescue and Italian ryegrass by targeting a fungal ferulic acid esterase (FAE) to the vacuole (Buanafina et al., 2006, 2008). FAE-expressing plants had reduced levels of cell wall esterified monomeric and dimeric ferulates and increased in vitro dry matter digestibility compared with non-transformed plants. In wheat, ubiquitous expressions of an Aspergillus niger ferulic acid esterase and a Bacillus subtilis xylanase were lethal (Harholt et al., 2010). Endosperm-specific expression caused wrinkled seeds and decreased grain weights. Clear changes in the properties of the arabinoxylan were observed linking the change in cell wall composition and architecture to the observed phenotypic changes (Harholt et al., 2010). This clearly demonstrates the problem of in planta accumulation of cell wall-degrading enzymes active at typical, ambient plant growth temperatures. Thermophilic enzymes have been suggested as a method to reduce the detrimental effects on plants as these enzymes have higher temperature optima than encountered during plant growth. Expression of thermophilic bacterial xylanases in tobacco and potato has resulted in high expression and no detrimental effects (Herbers et al., 1995, 1996; Yang et al., 2007). However, tobacco and potato contain very little xylan and the unchanged phenotype may well reflect this fact. Arabidopsis mutants with defects in glucuronoxylan biosynthesis all show severe growth defects indicating that glucuronoxylan plays an important role for cell wall integrity in Arabidopsis stems (reviewed in (York and O’ Neill, 2008)). Given the high amount of glucuronoxylan present in Arabidopsis stems and the severe effect observed in mutants defective in xylan biosynthesis, it could be anticipated that any detrimental effects of expressed xylanases would also lead to severe phenotypes.

The use of enzymes that are inactive at ambient temperatures is an alternative strategy, which, however, raises questions about remobilization of proteins from the apoplast during seed filling and degradation during senescence. In this article, we describe the expression of the two Dictyoglomus thermophilum thermophilic xylanases XynA and XynB in Arabidopsis. Accumulation and stability of the heterologous enzymes in the apoplast during stem development was studied using an apoplastic galactanase (Sorensen et al., 2000) and an apoplastic and Golgi-localized endo-arabinanase (Skjot et al., 2002; Obro et al., 2004) as reference enzymes. Finally, the modification of xylans in extracts from dry transgenic stems after heat treatment at 85 °C is reported.

Results

Two thermophilic xylanase-encoding genes, XynA and XynB, have been cloned and characterized from Dictyoglomus thermophilum. The XynA gene encodes a single-domain xylanase, which belongs to the family 10 group of xylanases (Gibbs et al., 1995). The temperature and pH optimum of the recombinant XynA protein have been determined to be 85 °C and pH 6.5, respectively. However, the enzyme is active across a broad pH range, with over 50% activity between pH 5.5 and 9.5. The Dictyoglomus thermophilum XynB gene encodes a family 11 xylanase (Morris et al., 1996). XynB is a multidomain enzyme comprising an N-terminal catalytic domain and a possible C-terminal substrate-binding domain that are separated by a short serine–glycine-rich 23-amino acid linker peptide (Morris et al., 1998). XynB has an optimum activity at pH 6.5 and temperature optima at 85 °C (Morris et al., 1996, 1998).

Genetic engineering of the genes encoding XynA and XynB for expression in plants

The codon usage of the D. thermophilum XynA and XynB genes was modified for expression in wheat and maize. The codon optimization resulted in a GC content of the modified XynA and XynB genes of 60% and 64%, respectively, compared to a GC content of the authentic bacterial XynA and XynB genes of 35% and 39%. The GC contents at the wobble base positions were increased from 26.7% and 27.2% to 96.9% and 98.3%, respectively. The coding region of the optimized Xyn A and XynB genes corresponding to amino acids 30–352 and 25–225, respectively, were amplified by PCR and cloned downstream of the coding region of the barley α-amylase signal peptide (SP) (Rogers, 1985), creating the genes SPmatXynAOpt and SPcatXynBOpt (Figure 1). The proteins encoded by these two genes will be referred to as matXynA and catXynB, respectively, when expressed in plants. Finally, the two engineered genes were cloned in between the 35S promoter and the Tnos sequence in the plant transformation vector pBI121 to allow constitutively expression of the two xylanases in plants. Using the computer programs TargetP and SignalP 3.0 (Emanuelsson et al., 2000, 2007; Bendtsen et al., 2004), both proteins are predicted to be secreted with cleavage site between amino acids 24 and 25.

Figure 1.

 Structure of the genes encoding SPmatXynAOpt and SPcatXynBOpt. The coding region of the XynAOpt gene encoding amino acids 30–352 (marked in bold and corresponding to amino acids 27–349 in SPmatXynA) was cloned downstream of the first 24 amino acids of the barley α-amylase signal peptide (marked in italic). The coding region of the XynBOpt gene encoding amino acids 25–225 (marked in bold and corresponding to amino acids 28–228 in SPcatXynB) was cloned in a similar manner. X and S represent restriction enzyme sites for XbaI and SacI, respectively, used for cloning of the two modified genes into the same restriction sites in pBI121. The arrow indicates the most likely cleavage site between amino acids 24 and 25 proposed by the SignalP 3.0 Server (Bendtsen et al., 2004). Stars show predicted N-glycosylation sites in SPcatXynB located at amino acid positions 36–38, 54–56 and 220–222 using the NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/).

Temperature optima of matXynA and catXynB in tobacco

It has previously been shown that the activity and thermal stability of expressed N-terminus truncated versions of XynB in bacteria are highly variable (Morris et al., 1998). The addition of a leader signal peptide to the truncated version of the two bacterial proteins and the presence of potential glycosylation sites in the two proteins may also negatively influence xylanase activity and stability when expressed in plants. Therefore, the matXynA and catXynB proteins were transiently expressed in tobacco as a fast screen for enzyme activity. It has been shown that expression of a Clostridium thermostable xylanase in tobacco results in a high apoplastic accumulation of the expressed xylanase without affecting plant development, making it likely that we could employ the same approach (Herbers et al., 1995). The two modified Dictyoglomus xylanases were transiently co-expressed with the p19 protein, a gene silencing suppressor from Tomato bushy stunt virus, in tobacco by agro-infiltration and xylanase activities were measured 3 days post inoculation. Determination of the temperature profile of the expressed matXynA and catXynB enzymes showed that both enzymes had activity optimum at 85 °C when assayed at pH 6.5 (Figure 2). The major difference in the temperature profile of the two enzymes is the higher relative xylanase activity of catXynB compared to matXynA over the entire temperature range. CatXynB had only very limited activity below 50 °C, a very important feature when plant cell walls should remain intact under normal growth temperatures.

Figure 2.

 Temperature profiles of matXynA and catXynB transiently expressed in tobacco leaves. Xylanase activity was measured at pH 6.5 and determined as the release of dyed fragments soluble in 74% ethanol from azo-wheat arabinoxylan during incubation with protein extract prepared from agro-infiltrated tobacco leaves (Megazyme, Ireland). Highest xylanase activity is set to 100%. Each data point represents the average xylanase activity from two biological repeats.

Expression of matXynA and catXynB in Arabidopsis

Transgenic Arabidopsis plants were generated by the flower-dip method. Visual inspections of greenhouse-grown transgenic T0 xylanase-expressing Arabidopsis plants (11 transformed with pBI121-SPmatXynAOpt and seven transformed with pBI121-SPcatXynBOpt) did not reveal any obvious phenotypic differences between the transgenic and wildtype (WT) plants. The xylanase activity in leaves and stems was determined in four transgenic Arabidopsis plants of each type (Figure 3). These lines were selected for further analysis as they were identified as single-gene insertion lines based on the segregation pattern of T2 seeds on plates containing 50 mm kanamycin.

Figure 3.

 Xylanase activity of expressed matXynA and catXynB during the development of transgenic Arabidopsis plants. Xylanase activity was determined as the release of reducing ends on 1% wheat arabinoxylan after incubation for 10 min at 85 °C with protein extract prepared from Arabidopsis leaves and stems using modified Somygoi methods (Megazyme, Ireland). Each data point was determined in triplicate and shown as the average ± standard deviation.

In contrast to WT plants, which had extremely low xylanase activity in all three tissues analysed, all transgenic lines had increased xylanase activity. In all transgenic lines, the lowest xylanase activity was found in leaves and the highest in completely dried stems. The highest xylanase activity was seen in stems of line A10, with approximately 90-fold higher activity than in WT. Moreover, long-term storage (>6 months) of dry stems at room temperature (RT) did not reduce enzyme activity.

Temperature optima and heat stability of matXynA and catXynB in Arabidopsis

Determination of the temperature profile of the expressed matXynA and catXynB enzymes in stem extracts from transgenic lines A8 and B4 assayed at pH 6.5 showed nearly the same profile of xylanase activity as observed in tobacco leaves with temperature optima at 85 °C and 80 °C for the matXynA and catXynB enzymes, respectively (Figure 4a). This analysis showed that the heterologous matXynA and catXynB enzymes had less than 4% of the maximal activity at temperatures below 30 °C and 3.6% and 8.1% of the maximal activity at 40 °C, respectively.

Figure 4.

 Temperature profile (a) and heat stability (b) of matXynA and catXynB expressed in Arabidopsis. Xylanase activity was measured at pH 6.5 and determined as the release of reducing ends on 1% wheat arabinoxylan after incubation for 10 min at 85 °C with protein extract prepared from stems using modified Somygoi methods (Megazyme, Ireland). (a) Highest xylanase activity for each enzyme is set to 100%. (b) Xylanase activity for each enzyme at time point zero is set to 100%. Each data point was determined in triplicate and shown as the average ± standard deviation.

The heat stability of the two xylanases expressed in Arabidopsis was analysed by incubation of stem extracts (pH 6.5) at 85 °C for 24 h and determining the residual xylanase activity during the incubation period (Figure 4b). After 24 h of incubation, both enzymes showed more than 50% residual activity, namely, 58% for matXynA and 50% for catXynB, respectively, which is very similar to the heat stability of the enzymes when expressed in bacteria (Gibbs et al., 1995; Morris et al., 1998).

matXynA and catXynB are both prominent proteins in dry transgenic Arabidopsis stems

The highest xylanase activity in the different transgenic Arabidopsis lines was in all cases found in dry stems. We performed a PAGE analysis of the accumulation and storage of expressed matXynA and catXynB during stem development from bolting to complete senescence of the stems (Figure 5) to better understand the biochemical events underlying the increase in xylanase activity. This analysis showed that during stem development, most of the soluble stem proteins disappear or are present at stem senescence at very low concentrations. In the two transgenic lines, but not in WT, there is a single protein that starts to accumulate approximately 4 weeks after stem bolting and is present at similar or higher amounts at all later stages of stem development, a 39 kDa protein in line A8 and a 30 kDa protein in line B4. The apparent molecular weight of 39 kDa for the accumulating protein band in line A8 is very similar to the calculated mass of 38 kDa for matXynA, but the apparent molecular weight of 30 kDa for the accumulating protein in line B4 is higher than the calculated mass of 23 kDa for catXynB.

Figure 5.

 Accumulation of matXynA and catXynB during stem development. Soluble protein extracts of define stem length were prepared 1 week (lanes 1–3), 4 weeks (lanes 4–6), 7 weeks (lanes 7–9) and 10 weeks (lanes 10–12) after bolting. WT (lanes 1, 4, 7 and 10), line A8 (lanes 2, 5, 8 and 11, and line B4 (lanes 3, 6, 9, and 12). Proteins were visualized on 10% NuPage Novex bis-Tris gels in morpholinepropanesulfonic acid and stained with PageBlue (Fermentas). Arrows indicate accumulating protein in line A8 (39 kDa) and line B4 (30 kDa).

Heat treatment of bacterially expressed XynA and XynB has been used as a simple and effective method to purify the two thermostable enzymes as most bacterial proteins are denatured by this treatment and can be removed by centrifugation (Gibbs et al., 1995; Morris et al., 1998). This observation was utilized to identify the matXynA and catXynB proteins by heating protein extracts prepared from 7-week-old stems at 85 °C for 30 min (Figure 6a).This treatment resulted in a nearly pure preparation of the 39 kDa protein band present in line A8 (Figure 6a; lanes 3 and 4) and the 30 kDa protein band present in line B4 (Figure 6a; lanes 5 and 6). To further prove that the 30 kDa protein band in line B4 was a modified version of catXynB, a Western blot analysis was performed and probed with rabbit anti-catXynB polyclonal antibodies (Figure 6b). Two protein bands reacted with the antibodies, a 24 kDa His-tagged version of catXynB purified from Escherichia coli (Figure 6b; lane 8) and the 30 kDa protein present in line B4 (Figure 6b; lane 7). It is most likely that the increased mass of catXynB is caused by glycosylation as it is predicted to have three potential N-glycosylation sites.

Figure 6.

 Identification of matXynA and catXynB in soluble stem protein extracts. (a) Heat-treated stem extracts were prepared by incubation of extracts from 7-week-old stems at 85 °C for 30 min, and denatured proteins were removed by centrifugation. Equal volumes (10 uL) of unheated and heated extracts were analysed by PAGE. Lanes 1, 3 and 5: unheated samples. Lanes 2, 4 and 6 heat-treated samples. Lanes 1 and 2: WT; lanes 3 and 4: line A8; lanes 5 and 6: line B4. Proteins were visualized on 10% NuPage Novex bis-Tris gels in morpholinepropanesulfonic acid and stained with PageBlue (Fermentas). (b) PAGE and Western blot analysis of stem extracts prepared 7 week after bolting. Lane 1: line A8; lane 2: purified his-tagged matXynA (calculated mass 39.8 kDa) from Escherichia coli; lane 3: WT; lane 4: line B4; lane 5: purified his-tagged catXynB (calculated mass 24.4 kDa) from E. coli. For Western blot analysis, the proteins in lanes 3–5 were transferred onto a PVDF membrane and the XynB protein detected by the ECL chemiluminescence detection system using rabbit anti-XynB polyclonal antibodies as the primary antibody and horse radish peroxidase-linked anti-rabbit immunoglobulin G (Dakopats, Denmark) as the secondary antibody. Lane 6: WT, lane 7: line B4 and lane 8: purified his-tagged catXynB from E. coli.

The concentration of the two xylanases in soluble extracts from dry stems was determined by PAGE analysis. Known amounts of protein extracts from the two transgenic lines were applied to the gel, and the amount of matXynA and catXynB was determined by comparison with a twofold dilution series of purified His-tagged catXynB of known concentration. With this method, the concentration of matXynA and catxynB in the soluble extracts was determined to be 14% and 3%, respectively.

The stability of the heterologous proteins in the apoplast during maturation may either be a property of the thermostable proteins, or of the apoplastic milieu. Transgenic Arabidopsis plants expressing endo-arabinanase, either apoplastic (Ara) or in the Golgi (ST-Ara), endo-galactanase (Gal) and combinations thereof (Obro et al., 2009) were monitored for enzyme accumulation and stability during stem development. Figure 7 shows that these enzymes all of which are mesophilic were stable throughout stem development. Surprisingly, also the Golgi-localized endo-arabinanase ST-Ara appears to be more stable during stem senescence than are the endogenous soluble stem proteins (Figure 7; lane 6 and 8).

Figure 7.

 Accumulation of heterologously expressed pectin-degrading enzymes during stem development in Arabidopsis. Soluble proteins were extracted in 0.1 m MES, pH 6.5, from the basal part of 6-week-old (lanes 1, 3, 5, 7 and 9) and 16-week-old (lanes 2, 4, 6, 8 and 10) stems. WT (lanes 1 and 2), apoplastic-localized endo-arabinanase (Ara) (lanes 3 and 4), Golgi-localized endo-arabinanase (ST-Ara) (lanes 5 and 6), fusion protein with Golgi-localized endo-arabinanase and apoplastic-localized endo-galactanase (ST-Ara/Gal) (lanes 7 and 8), and apoplastic-localized endo-galactanase (Gal) transgenic Arabidopsis plants (lanes 9 and 10). Note that the most prominent proteins in dry stems are the heterologously expressed proteins. Western blot analysis has shown that Ara, ST-Ara and Gal proteins are present as 32, 34 and 40 kD proteins, respectively (Obro et al., 2009).

Deposition of the heterologous expressed xylanases in the apoplast

A barley α-amylase signal peptide was included in the expression cassettes for matXynA and catXynB to direct the recombinant xylanases to the apoplast via the default secretory pathway. The immunolocalization of matXynA in transgenic stems of Arabidopsis line A10 revealed labelling surrounding the cells in the cortex and pith and a very thick and prominent labelling surrounding xylem and supporting cells in the vascular tissue (Figure 8a). Closer inspection of cells in the cortex with less elaborated secondary cell walls showed that the labelling was confined exclusively to the apoplast and in particular to the triangular space located between adjacent cells (Figure 8b–d).

Figure 8.

 Subcellular location of matXynA in stems of transgenic Arabidopsis expressing matXynA. Hand-cut sections from the basal part of 6-week-old stems of Arabidopsis were labelled with anti-XynA antibodies and secondary FITC-conjugated goat anti-rabbit IgG and analysed by confocal laser scanning microscopy. The detection limit was adjusted so that fluorescence from control sections (WT sections labelled with anti-XynA antibodies) just disappeared. (a) Overview of labelling of matXynA. Note the strong labelling in xylem and supporting cells (b) Localization of matXynA in cortex cells. Note the strong triangular labelling of cell corners. (c) Fluorescence from cells treated with propidium iodide to counter stain cell walls. (d) Combined picture obtained by merging pictures B and C showing co-localization of matXynA and the apoplast. (e) Bright-field image of the cortex cells. e; epidermis; co, cortex; if, interfascicular fibre; xy, xylem; pi, pith; Scale bar in (a), 150 μm and (b–e), 25 μm.

High temperature induction of xylanase activity and autohydrolysis of stem xylans

To determine whether the two heterologous expressed xylanases could degrade and solubilize Arabidopsis xylans, raw stem extracts prepared in 0.1 m MES buffer from dried stems of WT and transgenic lines A10 and B1 were incubated at 85 °C for 3 h. After centrifugation, the supernatant was saved. Determination of the sugar composition in the supernatant showed that the heat treatment resulted in an increased amount of water-extractable xylose in the two transgenic lines compared to WT (Table 1). Xylose is a good indicator of xylan content in mature stems as the secondary cell walls in Arabidopsis are by far the largest pool of xylan in stem cell walls (a parallel increase in glucuronic acid was also observed, corroborating that the polymer in question is indeed glucuronoxylan (data not shown)).

Table 1.   Extraction of xylans after heat treatment of dry stem extracts. Stem extracts were heated at 85 °C for 3 h, and after centrifugation, the supernatant was recovered. Amount of xylans in the supernatant is set to 100% for WT
Plant line0.1 m MES extractable
  1. Data shown are the average ± SD (n = 3). Data for the two transgenic lines are significantly different from WT at P < 0.002 based on student’s t-tests.

WT100 ± 2
A10108 ± 1
B1113 ± 2

Analysis of the size distribution of the buffer-extractable polysaccharides showed a lower amount of high molecular weight polysaccharides in the two transgenic lines when compared to WT (Figure 9). This is particular evident for transgenic line B4. As xylose alone constitutes more than 60% of the monosaccharides present in the buffer extract on a molar ratio, we interpret this difference as a difference in xylan composition. In the low molecular weight region, both transgenic lines had a higher amount of low molecular weight xylans. For line B4, this is evident for xylans smaller than approximately 40 kDa and for line A8 for xylans smaller than approximately 20 kDa.

Figure 9.

 Size distribution of water-extractable xylans after heat treatment of dry stem extracts. Stem extracts were heated at 85 °C for 3 h, and after centrifugation, the size distribution of xylans present in the supernatant was analysed by high pressure liquid chromatography.

Discussion

Pre-treatment of biomass prior to enzymatic saccharification for fermentation into biofuel comprises both heat treatment, treatment with chemicals and usually also a grinding or milling step all of which serve the purpose of opening cell structures and wetting the material, thus making the biomass accessible to the saccharification enzymes. Expressing some of the saccharification enzymes directly in the 2nd generation biofuel crop is not primarily to reduce enzyme cost, but to save energy spent on mechanical or heat mediated disintegration of the material. One of the limiting factors in the production of recombinant proteins in transgenic plants is the low level of protein accumulation. As a means to increase recombinant protein production in transgenic plants proteins has been targeted to different organelles. In Arabidopsis, xylanases have been targeted to the apoplast (Lu et al., 2004), the cytoplasm and chloroplast (Bae et al., 2008), and the cytoplasm, chloroplast and peroxisome (Hyunjong et al., 2006). From these studies, it appears that the highest xylanase activity in leaves is obtained when the recombinant xylanase is targeted to the chloroplast, and it is in line with previous studies in other organisms that the chloroplast is an efficient protein production system (Staub et al., 2000).

In our study, the highest xylanase activity was found in completely dried stems for both xylanases. Most notably, the best expressers showed a 10-fold increase in xylanase activity compared with the activity of xylanases targeted to the chloroplast (Hyunjong et al., 2006; Bae et al., 2008). To target the enzymes to the apoplast, they were fused N-terminally with the barley α-amylase signal peptide. Analysis of xylanase expression during stem development revealed that the high xylanase activity was both not only because of an accumulation of the two enzymes during stem development but also because of a general decrease in other soluble plant proteins. Xylanase activity was not markedly changed when the stems were kept at RT over a period of 6 months in the dry state. To our knowledge, this is the first report on the apoplastic targeting and storage of a recombinant protein during stem development. This suggests that the heterologous proteins are less efficiently degraded and remobilized than the endogenous enzymes. We anticipated that this would apply specifically or at least particularly to proteins stored in secondary walls. The survival of the Golgi-localized endo-arabinanase speaks against this, however.

Plant biomass will play a crucial role in future energy production, and there is a need to develop plants where biomass, especially the cell walls, can be more efficiently utilized. However, grass secondary cell walls are recalcitrant to enzymatic degradation because of, in particular, lignin and its cross-linkage to other cell wall components, particularly arabinoxylans. Reduced lignin content and cross-linkage are key targets for improving plants for biofuel production. Hydrolytic enzymes residing inside the mature straw are highly desirable because it may facilitate enzyme–substrate contact, is produced basically with no cost, can be stored until needed and supplement/reduce the addition of microbial enzymes. However, so far, in planta, expression and accumulation of cell wall-degrading enzymes has been problematic because active enzyme in the cell wall is causing plant malfunction and if stored in the vacuoles is degraded during plant maturation and senescence.

In the present study, we have solved these problems by heterologously expressing thermophilic xylanases in Arabidopsis with very low activities below 40 °C. The observed preservation of the two xylanases in dried stems is of great importance as most biomass intended for bioethanol production will be stored dried until needed for production purposes. Our finding that the two in planta expressed xylanases are highly thermostable with only a small decrease in activity for the first 6 h of heating at 85 °C and that over 50% residual activity still exist after 24 h of heating are also biochemical properties that are favourable from a production point of view.

A barley α-amylase signal peptide was included in the expression cassettes for matXynA and catXynB to ensure secretion of the heterologous proteins to the apoplast via the default secretory pathway. As the use of the same signal peptide for secretion of a Aspergillus phytase to the endosperm in wheat has shown unequivocally that the expressed phytase was deposited in the protein storage vacuole and not in the apoplast (Brinch-Pedersen et al., 2006) we analysed the subcellular location of the recombinant matXynA by immunocytochemistry. Our analysis showed that the matXynA protein was deposited in the apoplast. In cortex cells, labelling was preferentially located in the triangular space between adjacent cells while labelling was more uniform and intense in xylem and supporting cells. This labelling pattern is in line with earlier studies of protein targeting in rice and tobacco (Chen et al., 2004) where the use of two rice α-amylase signal peptides both led to a dual targeting of the cargo protein to the apoplast and chloroplast. The labelling of matXynA as well as its intensity largely mimics the labelling of xylans with LM10 which strongly labels interfascicular and xylem cells and to a lesser degree cortical and pith cells.

We prepared extracts from dry stems from WT and the two types of transgenic plants and activated the enzymes by incubation at 85 °C for 3 h to test whether the stored enzymes in the apoplast were capable of degrading xylan by autohydrolysis. Solubility of cell wall polymers are, among other factors, a function of the MW of the polymers. Polysaccharide size distribution in the aqueous extract was analysed by size exclusion chromatography. This showed a shift from high molecular weight xylans towards smaller-sized xylans in both transgenic lines when compared to the WT.

Xylanase activity was quantified as μmol reducing ends per min and per mg of protein. We also measured xylanase activity as the release of dyed xylan from AZO-xylan as a convenient and fast assay. The two xylanases show different activities on the two substrates with XynB more active on AZO-wheat and XynA more active in the reducing end assay. Analysis of the hydrolysis products produced by the action of XynA on oat spelt xylan has shown that not only it is broken down to oligomers, primarily xylotriose and xylobiose, but also that free xylose are produced (Gibbs et al., 1995). This digestion pattern of XynA may explain the presence of higher amounts of small xylans fragments in the XynA size distribution profile when compared to the XynB profile which show a marked reduction in xylan fragments larger than app 40 kDa compared to WT.

In conclusion, both enzymes evade the protein remobilization in Arabidopsis, retain their thermal stability and broad pH profile despite the different post-translational machineries operating in bacteria and plants, and they promote autolysis under conditions that are compatible with biomass pre-treatment technology, yet significant milder. Endo-xylanase has previously been mixed with commercially available cellulases and indeed increased the glucose yield (Tabka et al., 2006). It remains for further studies to ascertain whether xylanases produced in planta will have this effect and in particular under milder biomass pre-treatment conditions. This is our working hypothesis and unless any of the present observations turn out to be specific to Arabidopsis, these enzymes are attractive candidates for expression in plants to improve their value as bioenergy feedstocks.

Experimental procedures

Genetic modification of the Dictyoglomus thermophilum XynA and XynB genes for expression in plants

Codon-optimized versions of the Dictyoglomus thermophilum XynA and XynB genes (accession numbers L39866 and U76545, respectively) suitable for expression in wheat and maize were obtained from GenScript Corporation (Piscataway, NJ, USA) and named XynAOpt and XynBOpt, respectively.

Construction of plant transformation vectors

Plasmids pZmCOMT-SPmatXynAOpt and pZmCOMT-SPcatXynBOpt (to be published elsewhere) were cut with XbaI/EcoRI and the XynOpt coding region Tnos fragment was gel purified and cloned into pBI121 cut with the same two enzymes generating pBI121-SPmatXynAOpt and pBI121-SPcatXynBOpt, respectively. In pBI121-SPmatXynAOpt and pBI121-SPcatXynBOpt, expression of the bacterial proteins are controlled by the cauliflower mosaic virus 35S promoter, and the native bacterial signal sequences have been replaced by the barley a-amylase signal peptide (GENBANK accession no. K02637) (Rogers, 1985).

Bioinformatics

Targeting of SPmatXynA and SPcatXynB was predicted using TargetP (Emanuelsson et al., 2000, 2007), signal peptide or signal anchor predictions were predicted using SignalP 3.0 (Bendtsen et al., 2004), and N-glycosylation sites were predicted using NetNGlyc (http://www.cbs.dtu.dk/services/NetNGlyc/).

Expression of SPmatXynA and SPcatXynB in tobacco and Arabidopsis

The plant expression vectors pBI121-SPmatXynAOpt and pBI121-SPcatXynBOpt were electroporated (Shen and Forde, 1989) into Agrobacterium tumefaciens strain C58C1::pGV3810 (Zambryski et al., 1983). Leaves of 3-week-old wild-type tobacco (N. benthamiana) were agro infiltrated with pBI121-SPmatXynAOpt/p19, pBI121-SPcatXynBOpt/p19 and p19 alone using a 1-ml syringe with no needle attached, and the leaves were harvested 3 days post infiltration. In co-infiltration with recombinant agrobacterium containing p19, a gene silencing suppressor from Tomato bushy stunt virus (Voinnet et al., 2003), bacterial cultures were prepared separately in induction media and combined in a 1:1 ratio immediately before infiltration (Jensen et al., 2008). Arabidopsis thaliana (L.) Heynh. ecotype Col-0 were transformed by the floral dip method (Clough and Bent, 1998). Seeds were collected from Agrobacterium-treated plants, surface sterilized with 1.5% bleach and sown (ca. 500 seeds per 90 mm plate) on plates containing MS media (1 × Murashige and Skoog salts, 8 g/L agar, 1× B5 vitamins, 10.8 g/L sucrose) containing 50 mg/mL kanamycin and incubated for 48 h at 4 °C in the dark. Plates were transferred to a growth room at 21 °C with 16 h light/8 h dark and incubated for 2 weeks. Kanamycin-resistant seedlings were transplanted to peat in the greenhouse and grown at an 8-h photoperiod at 28 °C/day and 18 °C/night. To initiate bolting and synchronize stem growth, plants were shifted to a 16-h photoperiod after 8 weeks growth in the 8-h photoperiod. Xylanase measurements (see below) on several independent transformants of each kind were used to select two XynA high expressers (A8 and A10) and two XynB high expressers (B1 and B4) for further study. Members of each set of two high expressers were proven to be indistinguishable in our preliminary investigations and were thus used interchangeably in the experiments reported here.

Preparation of plant extracts

Plant material (leaves, green and dried primary stems) was frozen in liquid nitrogen and ground into a fine powder with a mortar and pestle, and resuspended in 4 volume of 100 mm MES (pH 6.5). The extract was kept on ice for 20 min and vortexed three times during this period. The extract was centrifuged at 18 000 g for 20 min at 4 °C, and the supernatant was stored at −20 °C. Heat treatment of plant extracts was performed at 85 °C for 30 min and after centrifugation at 18 000 g for 20 min at 4 °C, the supernatant was saved and stored at −20 °C. Protein concentrations were determined by the Bradford method with bovine serum albumin as standard (Bradford, 1976).

Measurements of xylanase activity

The temperature profiles of the enzymes transiently expressed in tobacco were determined as described by Megazyme (Wicklow, Ireland) with minor modifications using AZO-wheat arabinoxylan (Megazyme) as substrate. Leaf extracts were prepared in 100 mm MES pH 6.5, 150 μL pre-warmed 1% AZO-wheat arabinoxylan and 150 μL pre-warmed diluted leaf extract were mixed and incubated for 10 min at the temperatures indicated. The reaction was stopped by the addition of 1 mL 96% ethanol. The samples were left standing at RT for 15 min, and insoluble material was removed by centrifugation at 14 000 g for 10 min. The absorbance of the supernatant was measured at 590 nm.

A modified Somogyi reducing sugar assay was used for determining the

Xylanase activity using medium viscosity wheat arabinoxylan as substrate as described by Megazyme (Ireland) with xylose as the standard. Each reaction contained 5 or 10 μL of plant extract diluted in 0.1 m MES to 150 μL, diluted plant extract and 150 μL 1% wheat arabinoxylan were pre-warmed at 85 °C for 5 min before being mixed and incubated for 10 min at the temperatures indicated. One unit of xylanase activity was defined as the amount of enzyme that released 1 μmol of reducing sugar (expressed as xylose equivalent) per min at 85 °C and pH 6.5. Measurements of xylanase activity during plant development were performed by combining leaves, stems or senescent stems from five plants and preparing plant extracts as described previously. Xylanase activity was then determined in triplicate for each extract using the modified Somogyi reducing sugar assay described previously.

Production of antibodies against XynB and Western blot analysis

The coding region of XynB covering amino acids 25–224 and including the catalytic region of the enzyme was cloned in frame with an N-terminal extension containing a His-tag in the expression vector pQE-30.

To do so, the mature and catalytic part of XynBOpt was amplified by PCR using the primers:

catXynBOpt forw: GATCGGATCCCTGCAGATGcagacctccatcacc-ctcac

catXynBOpt rev: GATCGTTAACGGTACCTTActgggagaaggtgttct-gg

where the forward primer includes a BamHI (underlined) recognition site while the reverse primer contains a KpnI (underlined) site. The PCR fragments was gel purified and cloned in the pCR2.1-TOPO vector (Invitrogen) creating.

pCR2.1-TOPO-catXynBOpt. pCR2.1-TOPO-catXynBOpt was cut with BamHI and KpnI and the insert gel purified and cloned into pQE-30 cut with the same two enzymes giving rise to the plasmid pQE-30-catXynBOpt. The plasmid was transformed into E. coli M15 and E. coliSG13009 and evaluated for the expression of the xylanase. It was found that E. coliSG13009 expressed the His-tagged xylanase most effectively.

After large scale induction with 1 mm IPTG, the His-tagged catXynB was purified under native conditions using the nickel-nitrilotriacetic acid (Ni-NTA) Fast Start System kit (Qiagen, Copenhagen, Denmark), which eluted the protein in a final buffer of 50 mm Na phosphate, 300 mm NaCl and 250 mm imidazole, pH 8.0. The eluted xylanase was dialysed against 1 × PBS and used to immunize rabbits for the production of anti-XynB polyclonal antibodies.

Proteins were visualized on 10% NuPage Novex bis-Tris gels in morpholinepropanesulfonic acid buffer stained with PageBlue (Fermentas, Lithuania). For Western blot analysis, the electrophoresed proteins were transferred onto a PVDF (polyvinylidene difluoride) membrane and the XynB protein detected by using a 1 : 1000 dilution of the rabbit anti-XynB polyclonal antibodies with 1 : 1000 dilution of the horse radish peroxidase-linked anti-rabbit immunoglobulin G (Dakopats, Denmark) as the secondary antibody and chemiluminescence detection using the ECL system.

Antibody labelling of Arabidopsis stems

Hand-cut transverse sections from the stem base tissue were treated with skim milk powder/PBS buffer (50 uL, 5% skim milk powder in PBS pH 8.0) for 15 min at RT, the liquid was removed and the sections were then incubated with fivefold dilutions of anti-matXynA antibodies in skim milk powder/PBS buffer at RT for 2 h. The sections were washed with PBS three times and then incubated at RT for 2 h with 160-fold dilution of fluorescein isothiocyanate-conjugated secondary antibodies (Sigma-Aldrich Corp., St Louis, USA) in skim milk powder/PBS buffer. From that time point, the incubations were performed in darkness. After incubation with the secondary antibodies, the sections were washed with PBS three times. The sections were stored in darkness at 4C until the microscopic analysis. The sections were mounted in a glycerol/PBS-based antifade solution (Citifluor AF1; Agar Scientific, Stansted, England) and observed with a Leica TCS SP2/MP confocal laser scanning microscope with a 20×/1.2 numerical aperture water immersion objective. Sections were excited at 488 nm, and emissions were recovered in the interval 495–512 nm. Propidium iodide (Sigma) (100 ug/mL) was used to counter stain the cell walls, and fluorescent emissions were measured at 590–620 nm.

Analysis of autohydrolysis of xylans in plant extracts

Dry stem material was ground in liquid nitrogen with a mortar and pestle and resuspended in 300 μL of 100 mm MES (pH 6.5) per 10 mg of stem material. The suspended plant material was incubated at 85 °C for 3 h under mild shaking. After centrifugation at 18 400 g for 20 min at 4 °C, the supernatant was saved and stored at -20 °C. Vacuum-dried samples were hydrolysed in 2 m trifluoroacetic acid for 1 h at 120 °C. Trifluoroacetic acid was removed by drying under vacuum. Monosaccharide composition was subsequently determined by high performance anion exchange chromatography with pulsed amperometric detection of hydrolysed material using a PA20 column (Dionex) as described previously (Obro et al., 2004). Monosaccharide standards were from Sigma and included L-Fuc, L-Rha, L-Ara, D-Gal, D-Xyl, D-Man, D-GalUA and D-GlcUA. For verification of the response factor, a standard calibration was performed before analysis of each batch of samples. Mass distribution analysis of water-extractable polysaccharides in the supernatant was performed using size exclusion chromatography. Size exclusion chromatography was carried out on a Superose 12 column (1 × 30 cm; Amersham Pharmacia, Uppsala, Sweden) equilibrated in 0.05 m ammonium formate. Sample was applied to the column and eluted with 0.05 m ammonium formate at a flow rate of 24 mL/h. The eluent was monitored using a refractive index (RI) detector (model 131; Gilson, Middleton, USA). Equal amounts of sugars were loaded onto the column.

Acknowledgements

This project (3304-FVFP-060673) was financed by the Danish Research Council for Technology and Productivity. We wish to thank Aida Curovic, Annette Jensen and Morten Stephensen for skilful technical assistance.

Ancillary