Cloning, characterization and functional analysis of novel 6-kestose exohydrolases (6-KEHs) from wheat (Triticum aestivum)


Author for correspondence: Wim Van den Ende Tel: +32 163 21952 Fax: +32 163 21967 Email:


  • • Cereals accumulate graminan-type fructans which are subject to stress-related degradation by fructan 1-exohydrolases (1-FEHs) and fructan 6-exohydrolases (6-FEHs). To find new FEH genes related to freezing tolerance, a cold-hardened wheat crown cDNA library was screened.
  • • Here we report the cloning, purification and characterization of two novel 6-kestosidase (6-KEH) isoenzymes from wheat crowns (Triticum aestivum). Functional characterization in Pichia pastoris confirmed the extreme substrate selectivity for the fructan trisaccharide 6-kestose.
  • • Northern blotting showed that 6-KEH transcripts were constantly detected at the same level from autumn to winter in crown but not in leaf tissues. Apoplastic fluid isolations and activity measurements strongly suggest that 6-KEH is localized in the apoplast.
  • • It is proposed that 6-KEHs, together with other FEHs, might be involved in the breakdown of apoplastic fructans which may fulfil a role as membrane protectors under stress. Alternatively, a role in signalling processes, or in the degradation of exogenous 6-kestose from bacterial origin, cannot be excluded.


Fructans, fructose-based oligo- and polysaccharides, are storage compounds in about 15% of flowering plant species (Hendry, 1993). They are classified into several forms depending on their glycosidic linkages. Inulin-type fructans with linearly β(2→1)-linked fructofuranosyl units occur mainly in dicot species. Levan-type fructans with β(2→6)-linked fructofuranosyl units and mixed-levan type fructans with both β(2→1)- and β(2→6)-linked fructofuranosyl units (graminans) are found in monocot species (Pollock & Cairns, 1991). Inulin biosynthesis involves two enzymes: sucrose : sucrose 1-fructosyl transferase (1-SST, EC and fructan : fructan 1-fructosyl transferase (1-FFT, EC. (Van Laere & Van den Ende, 2002). Together with 1-SST and 1-FFT, Sucrose : fructan 6-fructosyl transferase (6-SFT) determines graminan biosynthesis in cereals (Bancal et al., 1992; Sprenger et al., 1995; Kawakami & Yoshida, 2002; Kawakami et al., 2002; Nagaraj et al., 2004). Fructan : fructan 6G-fructosyl transferase (6G-FFT, Shiomi, 1981, 1989; Ritsema et al., 2003) or other 6-FT-like fructosyl transferases (Pavis et al., 2001) are probably implicated in the synthesis of fructan neoseries found in Asparagus officinalis, Allium cepa, Lolium perenne and Avena sativa (Livingston et al., 1993).

Graminan-type fructans are the predominant storage carbohydrates in vegetative tissues of temperate grasses and cereals such as wheat and barley. Wheat fructan levels might be controlled not only by fructan biosynthetic enzymes, but also by fructan exohydrolases (FEHs), releasing terminal fructoses from fructans. These are classified into two types of enzyme: 1-FEHs and 6-FEHs, based on the linkage type [β(2→1) vs β(2→6)] that is preferentially hydrolysed. 1-FEH is coexpressed with fructan biosynthetic enzymes and might act as a β(2→1) trimmer during the period of active fructan biosynthesis (Van den Ende et al., 2003a). Graminans are believed to accumulate mainly in the vacuole, and are subject to degradation by vacuolar FEHs (Wagner & Wiemken, 1986).

Similarly to raffinose oligosaccharides (Tapernoux-Luthi et al., 2004), fructans may be able to protect plants against drought and freezing (Pilon-Smits et al., 1995; Park et al., 1999; Konstantinova et al., 2002; Amiard et al., 2003; Parvanova et al., 2004), probably by stabilizing membranes (Vereyken et al., 2001; Hincha et al., 2002, 2003). Induction of 1-SST and 6-SFT, and subsequent accumulation of fructans in winter cereals hardened above freezing temperatures (first-phase cold hardening, 1-PH), was demonstrated to be essential for winter survival (Olien & Clark, 1993; Yoshida et al., 1998; Gaudet et al., 1999; Kawakami & Yoshida, 2002). In wheat, 1-PH resulted in strongly decreased FEH activities. It was suggested that the observed varietal difference in fructan accumulation under frost is affected largely by the FEH activity levels (Yukawa et al., 1995). Compared with winter wheat, spring wheat was unable to induce 1-SST activity in leaves under cold stress (Savitch et al., 2000).

A significant increase in fructan concentrations in the apoplast of oat crowns was found during second-phase cold hardening (hardening below freezing temperatures, 2-PH), suggesting the presence of two different fructan pools under freezing stress. Compared with total extracts, 2-PH apoplastic fluids showed an increased low degree of polymerization (DP)/high DP fructan and hexose/sucrose ratio, correlating well with the increased FEH and invertase activities in the apoplast (Livingston & Henson, 1998). These results were indicative for the presence of a true apoplastic FEH in cold-hardened oat. FEHs were induced in chicory and orchardgrass by freezing temperatures (Van den Ende & Van Laere, 1996; Yamamoto et al., 1999; Michiels et al., 2004), but also after defoliation (Yamamoto & Mino, 1985; Prud’homme et al., 1992; Morvan-Bertrand et al., 2001; Van den Ende et al., 2001; Asega & de Carvalho, 2004). Gibberellin and ABA might be important hormones for FEH regulation (Morvan et al., 1997; Yang et al., 2004).

So far, plant 1-FEH cDNAs have been cloned from chicory (Van den Ende et al., 2000, 2001) and wheat (Van den Ende et al., 2003a). Plant FEH enzymes are clearly different from invertases (Van Laere & Van den Ende, 2002) as they are unable to degrade sucrose (Suc). Surprisingly, it was found that 6-FEH activities occur in nonfructan plants, and a 6-FEH was recently purified and cloned from sugar beet (Van den Ende et al., 2003b). In order to find new FEH genes putatively related to freezing tolerance, a wheat 1-FEH probe was used to screen a cold-hardened wheat crown cDNA library.

Materials and Methods

Plant material

Winter wheat (Triticum aestivum L.) cv. PI 173438 was sown on 18 September in Sapporo (2002–03). In parallel wheat was grown in a field near Leuven (2001–02 and 2003–04) for sampling crowns, leaves, stems and ears (anthesis stage) for enzyme purification and measurements of enzymatic activity. Throughout October 2001 and up to February 2002, samples of leaf and crown (etiolated base of stem containing the apex) were taken at regular time intervals (9 October, 8 November, 8 December, 19 December, 10 January and 11 February at Sapporo) for Northern analysis. In parallel, wheat was sown in acid-washed vermiculite in a controlled-growth chamber (temperature 22°C; 12/12 day/night cycle, light intensity 200 µm photons m−2 s−1). Plants received a nitrogen-rich medium as described (De Roover et al., 2000). Root and leaf part samples (from 20 different plants) were pooled 3, 7, 10, 17 and 21 d after sowing (time of sampling, 4 h after light-on) for carbohydrate analysis, enzymatic activity determinations and Northern analysis. After 21 and 26 d crown tissue could be sampled as well. On day 26 it became impossible to dissect roots from individual plants and the leaves started to turn yellow. The methods of RNA isolation and Northern hybridization were as described previously (Kawakami & Yoshida, 2002).

Cloning and molecular analyses

A mixture of wheat PCR products 12.16 and 12.18 with high similarity to cell wall invertases and dicot 1-FEHs (for details see Van den Ende et al., 2003a) was used as a hybridization probe to screen a wheat cDNA library derived from T. aestivum. cv. PI 173438 (described by Kawakami & Yoshida, 2002) by plaque hybridization. Nucleotide sequences of isolated clones were determined by an automated DNA sequencer (model 373S, Applied Biosystems, Foster City, CA, USA) using a thermo sequenase dye terminator cycle sequencing kit v. 2.0 (Amersham Pharmacia Biotech, Buckinghamshire, UK). The nucleotide sequences and deduced amino acid sequences were analysed by dnasis software (Hitachi Software Engineering, Yokohama, Japan). Phylogenetic trees were created with clustal X and treeview programs (Page, 1996; Thompson et al., 1997).

Expression of recombinant proteins in methylotrophic yeast

Transformation and culturing of Pichia pastoris was carried out basically according to the instructions provided by the supplier (EasySelect Pichia Expression Kit, Invitrogen, Groningen, the Netherlands). Insertions of functional sequences that covered putative mature protein regions of 6-KEH w1 and w2 were cloned by PCR using the following primers with adapters:



(forward for 6-KEH w2), and


(reverse for both 6-KEHs). The restriction sites of PstI (6-KEH w1), EcoRI (6-KEH w1) and XbaI (both 6-KEHs) are indicated in bold in the primers. Both the amplified fragments and the pPICZα B vector (Invitrogen, Groningen, the Netherlands) were digested with their respective restriction enzymes. Subsequently the DNA fragments were further purified by using the Ultraclean DNA Purification Kit (MOBIO, Solana Beach, CA, USA). After ligation of the vector and DNA fragments, the P. pastoris strain X-33 was transformed by electroporation, using 10 µg of the PmeI-linearized vector with inserts, and transformants were selected on YPDS/Zeocin plates. A 200 ml preculture medium (BMGY, pH 6.0) was inoculated with freshly prepared single colonies and incubated for 48 h at 30°C with vigorous shaking (200 rpm). The cells were collected by centrifugation at 1500g for 5 min, transferred to 20 ml induction medium (BMMY, pH 6.0) and incubated at 28°C under aerobic conditions. 2% (w/w) methanol was added to the culture medium every day. At 96 h after induction the culture was centrifuged and proteins were precipitated in 80% ammonium-saturated citrate–phosphate buffer (10 mm final, pH 5.0) on ice for 1.5 h. Finally, centrifugation was carried out (15 000g) for 25 min at 2°C.

Purification of heterologous 6-KEHs

The precipitate was redissolved in 150 ml 50 mm sodium acetate buffer, pH 5.0. Undissolved material was spun down for 15 min at 40 000g and 4°C. The supernatant was applied to a ConA Sepharose column (25 × 100 mm). Bound proteins were eluted with 50 ml 1 m methyl α-d-mannopyranoside in 20 mm Tris–HCl buffer pH 7.5. The fractions containing enzymatic activity were applied on a Mono Q anion-exchange column (Pharmacia HR 5/5, Uppsala, Sweden), which was equilibrated with 20 mm Tris–HCl buffer pH 7.5. Proteins were eluted using a linear gradient from 0 to 0.3 m NaCl in 30 min (flow rate 1 ml min−1).

Partial purification of native 6-KEH w1

Wheat crowns (2.5 months after sowing) or ears and stems (anthesis stage) were cut into very small pieces. Of this material, 1 kg was immersed in liquid nitrogen and homogenized dry with a Waring blender. A second dry homogenization was performed until a fine powder was obtained. Finally this powder was homogenized in 1.5 l 50 mm sodium acetate buffer pH 5 containing 1 mm EDTA, 10 mm NaHSO3, 1 mm mercaptoethanol and 0.1% (w/v) Polyclar AT (Serva, Heidelberg, Germany). The homogenate was squeezed through cheesecloth. Ammonium sulfate was added to a saturation of 30% and gently stirred on ice for 30 min. After centrifugation for 20 min at 40 000g and 4°C, precipitated protein was discarded. Again ammonium sulfate was added to the supernatant to a final saturation of 80%. After a second centrifugation (20 min at 40 000g and 4°C) the precipitate was collected and redissolved in 150 ml 50 mm sodium acetate buffer pH 5.0. Undissolved material was spun down for 15 min at 40 000g and 4°C. The supernatant was applied to a ConA Sepharose column (25 × 100 mm) and eluted as described (Van den Ende et al., 1996). Active fractions were pooled, adjusted to pH 7.0 with concentrated Tris, and applied on Mono Q (Pharmacia HR 5/5) equilibrated with 20 mm Tris–HCl buffer pH 7.0 and eluted as described (Van den Ende et al., 1996). Two fractions containing 6-KEH were pooled, diluted five times with 20 mm His–HCl buffer pH 6.0, and subsequently loaded on a second Mono Q equilibrated with 20 mm His–HCl buffer pH 6.0. The fractions obtained were free from contaminating inulinase, levanase and invertase activity. Whenever possible, enzymes were kept on ice throughout the whole purification. Sodium azide (0.02%, w/v) was added to all buffers to prevent microbial growth. SDS–PAGE of the 6-KEH fractions was performed on 12.5% (w/v) polyacrylamide gels and stained with Coomassie Brilliant Blue-R250 as described (Van den Ende et al., 1996).

Q-TOF analyses on 6-KEH tryptic fragments

The SDS–PAGE protein band of 72 kDa, derived from a fraction exhibiting only 6-kestosidase activity, was subjected to mass spectrometric (MS) identification. The Coomassie BB-stained protein bands were excised, trypsinized, extracted, desalted and analysed by quadrupole/time-of-flight (Q-TOF) as described previously (Van den Ende et al., 2003a). Sequence information was derived from the MS/MS spectra with the aid of the maxent 3 (deconvoluting and de-isotoping of data) and pepseq software from the micromass biolynx software package (Micromass, Manchester, UK).

Fructan substrates

Wheat-stem low-DP fructan (anthesis) was obtained as previously described (Van den Ende et al., 2003a). 1&6-kestotetraose (bifurcose) is not commercially available. Only small amounts (not enough for kinetic analyses on native and heterologous 6-KEHs) were obtained from wheat crown tissue by preparative HPAEC (Dionex, Sunnyvale, CA, USA) and repeated manual collection. The HPAEC conversion factor of bifurcose was determined. 1,1-nystose was prepared from Neosugar P (Beghin-Meiji Industries, Paris, France) by preparative reversed-phase HPLC (Nucleosil 7 C18, 250 × 12.7 mm) with water as solvent and a flow rate of 2 ml min−1. Low molecular weight levan, 1-kestose and 6-kestose were generous gifts from Dr Iizuka (Iizuka et al., 1993). Levanbiose (F2) was purified as described by Timmermans et al. (2001). Dr Chatterton and Dr Livingston kindly provided neokestose, phlein and neokestin. The mean DP of inulin, levan and neokestin was estimated from the fructose/glucose ratio after mild acid hydrolysis in 60 mm HCl at 70°C for 75 min. Subsequently molar concentration was estimated based on this mean DP and the glucose liberation.

Carbohydrate analyses and enzyme activity determinations

To measure the activities of inulinase, levanase, invertase and 6-KEH, aliquots were incubated with 5 mm Suc, 6-kestose, commercial chicory root inulin (Sigma-Aldrich, St Louis, MO, USA) and bacterial levan (Iizuka et al., 1993), in 50 mm sodium acetate buffer pH 5.0 at 30°C. Sodium azide (0.02%, w/v) was added to all buffers to prevent microbial growth. Fructose (Fru) formation was determined by HPAEC, and proteins were determined as described (Van den Ende & Van Laere, 1996). Enzymatic activity is expressed in units (U), defined as the amount of enzyme which formed 1 µmol Fru min−1. Heterologous enzyme specificities were investigated by incubating equal U of 6-KEH w1 and w2 together with 2 mm Suc, levanbiose, 6-kestose, 1-kestose, neokestose, 1,1-nystose, bifurcose, phlein oligomers, inulin, levan and neokestin. To check for linearity of product formation, aliquots were removed after several time intervals and Fru formation was determined by HPAEC.

Apoplastic fluid extraction

Cold-hardened wheat leaves and crowns (Leuven, February 2004) were used as a starting material for apoplastic fluid isolations. Tissues were vacuum-infiltrated four times with 100 mm NaCl for 45 s and centrifuged as described (Boller & Métraux, 1988). For carbohydrate analyses, the four apoplastic fluids (termed APV1–4) were immediately heat-inactivated. APV2 and 3 were found to contain the lowest level of intracellular contaminants (using both 1-SST and PGI as intracellular markers). Therefore APV2 and 3 were combined for carbohydrate analyses (see Fig. 6) and enzyme activity measurements (see Fig. 7).

Figure 6.

HPAEC profiles of wheat crown fractions. HPAEC chromatograms of total extract (TOT) and apoplastic fluid (APO) of cold-hardened wheat (Triticum aestivum) crown and leaf tissues (harvested February 2004). A reference is included containing: G, Glc; F, Fru; S, Suc; R, raffinose; 1K, 1-kestose; 6K, 6-kestose; nK, neokestose. Other labelled peaks: N, 1,1 nystose; B or 1&6-kestotetraose, k, 6&1,1 kestopentaose; l, 1&6,1 kestopentaose; m, 1&6,6-kestopentaose. 6-kestose in the extracts is indicated by an arrow.

Figure 7.

Comparison of enzymatic activities in different fractions. (a) 6-KEH to 1-SST ratio in total extract (TOT); remaining tissue fraction (RTF); and apoplastic fluid (APO) of cold-hardened wheat (Triticum aestivum) crowns. Results are mean values (± SE) of three experiments. (b) HPAEC chromatograms showing in vitro fructan biosynthesis from 200 mm Suc by TOT, RTF and APO fractions of a cold-hardened wheat crown after 0 and 23 h incubation at 30°C. G, Glc; F, Fru; S, Suc; 1K, 1-kestose; 6K, 6-kestose, B, bifurcose or 1&6-kestotetraose.


Molecular characterization of wheat 6-KEHs

Two full-length cDNAs were obtained from a winter wheat cDNA library: 6-KEH w1 (DDBJ Accession No. AB089271) and 6-KEH w2 (AB089270). They contain one long open reading frame (ORF) encoding 589 and 587 amino acid polypeptides, respectively. These are compared with 1-FEH w2 and cell wall invertase from wheat, 1-FEH IIa from chicory, and 6-FEH from sugar beet (Fig. 1). Both 6-KEHs contain five potential N-glycosylation sites (Fig. 1) and have predicted pI of 4.91 (w1) and 4.88 (w2). The theoretical pI value matches the chromatographic behaviour of the native 6-KEH w1 protein. The cDNA-derived molecular mass of the expected mature enzymes (61 kDa) is lower than the 72 kDa estimated from SDS–PAGE, but this discrepancy can be explained by the presence of glycosyl chains. 6-KEH w1 and w2 have the typical consensus domains conserved among invertases and FEHs: the β-fructosidase motif (NDPNG); the FRDP region; and the putative catalytic site [WEC(V/P)D]. We were unable to determine the N-terminal sequence of 6-KEH w1 and w2, but an estimation was made based on the N-terminal sequences of the mature protein of chicory (Van den Ende et al., 2000, 2001) and wheat 1-FEHs (Van den Ende et al., 2003a).

Figure 1.

Multiple alignment of wheat (Triticum aestivum) 6-KEHs and some related enzymes. Sugar beet 6-FEH (Bv 6-FEH, AJ242538); wheat cell wall invertase (Ta cwINV, AF030420); chicory 1-FEH IIa (Ci 1-FEH IIa, AJ295033); wheat 6-KEH w2 (Ta 6-KEH w2, AB089270); wheat 6-KEH w1 (Ta 6-KEH w1, AB089271) and wheat 1-FEH w2 (AJ508387, Ta 1-FEH w2) are aligned. Potential glycosylation sites are underlined. The three carboxylic acids that are thought to be crucial for enzyme catalysis are in bold. Consensus line: asterisks (*) indicate identical residues; colons (:), conserved substitutions; full stops (.), semiconserved substitutions.

6-KEH w1 and 6-KEH w2 are 88% identical. They are 68% identical to wheat 1-FEH w1 and w2 (AJ516025 and AJ508387). Their homologies to wheat 1-SST (AB029888) and 6-SFT (AB029887) are much lower (35–37% identity). Their identity to wheat cell-wall invertase is 45% (AF030420), which is lower than the identity to dicot chicory 1-FEHs (48–49% identity).

Like the 1-FEHs from wheat and chicory, wheat 6-KEH w1 and w2 group together with cell-wall invertases, and not with vacuolar-type invertases and fructan biosynthetic enzymes. An unrooted radial tree of some members of cell wall-type glycosyl hydrolases is presented in Fig. 2. Four distinct groups can be discerned: the first (I) contains monocotyledonous cell-wall invertases. The second (II) contains dicotyledonous cell wall-type invertases. A third group (III) also contains dicotyledonous cell wall-type invertases and FEHs. Except for AtcwINV3 from Arabidopsis thaliana (revised nomenclature, Sherson et al., 2003) and the recently cloned 6-FEH from sugar beet (Van den Ende et al., 2003b), all members of subgroups I–III have a high pI for interaction with the cell wall. In contrast, all group IV members have an acidic pI, as usually observed for vacuolar type invertases. The fourth group contains dicotyledonous (IVa) and monocotyledonous (IVb,c) enzymes. The IVa subgroup harbours chicory 1-FEHs and AtcwINV6 from A. thaliana. It was recently demonstrated that AtcwINV3 and 6 are not invertases but are true FEHs (De Coninck et al., 2004). The IVb group contains soluble apoplastic invertase from Zea mays and some putative invertases from Oryza sativa. The IVc subgroup is quite distinct from IVb, and contains a putative rice invertase, wheat 1-FEH w1 and w2, and 6-KEH w1 and w2 described in this manuscript (arrow). It should be noted that the ‘invertase activity’ of only a limited number of enzymes (indicated by * in the tree) was tested either by protein purification and/or characterization, or by heterologous expression.

Figure 2.

Unrooted phylogenic tree of cDNA-derived amino acid sequence of plant cell wall invertases and FEHs. Four groups can be discerned. (I) Zea mays INV1, 2 and 3; two putative (question mark) and one effective Oryza sativa INV; Triticum aestivum INV. (II) Vicia faba INV1; Daucus carota INV 1, 2 and 3; Solanum tuberosum INVF; Lycopersicon esculentum INV5; Nicotiana tabacum INV; Arabidopsis thaliana INV2 and 4. (III) Beta vulgaris 6-FEH and INV; Arabidopsis thaliana INV1, 5 and 6-FEH (AtcwINV3); Vicia faba INV2; Pisum sativum INV; Chenopodium rubrum INV1; and Fragaria × ananassa putative INV. (IVa) A putative INV from Beta vulgaris; Cichorium intybus 1-FEH I, IIa and IIb and a 6&1 FEH from Arabidopsis thaliana (AtcwINV6). (IVb) Zea mays INV4; two putative and one effective INV from Oryza sativa. (IVc) 1-FEH w2 from Triticum aestivum, a putative invertase from Oryza sativa and Triticum aestivum 6-KEH w1 and w2 (arrow). Isoelectric points and accession numbers are presented between brackets. Acid isoelectric points are underlined. Scale bar indicates a distance value of 0.1. Enzymes of which the invertase functionality was confirmed by extensive protein characterization and/or heterologous expression of the cDNA are indicated with an asterisk (*). INV, cell wall invertase. Revised nomenclature was used for Arabidopsis thaliana invertases according to Sherson et al. (2003).

Functional characterization of recombinant 6-KEH w1 and w2 proteins

Heterologous expression in P. pastoris proved a valuable system to judge the main functionality of fructosyltransferases (Kawakami & Yoshida, 2002) and FEHs (Van den Ende et al., 2003b; Verhaest et al., 2004), although sometimes an additional side activity was found compared with the native enzyme (Hochstrasser et al., 1998). Table 1 shows the extreme substrate selectivity towards 6-kestose for the heterologously expressed 6-KEH w1 and w2 enzymes. Both enzymes were similar in this respect. Other 2,6-type fructans, such as levanbiose, phlein and levan, are degraded with low efficiency (1–2%), while inulin and Suc are even poorer substrates. These results demonstrate that the two enzymes are specific 6-kestose hydrolases and not general 6-FEH, 1-FEH, β-fructofuranosidase or invertase enzymes. 6-KEH w1 produces only Suc and Fru from 6-kestose (Fig. 3a). Identical profiles were observed for 6-KEH w2 (not shown). The extreme substrate selectivity was demonstrated further with total wheat graminan-type fructans as a substrate (Fig. 3b). 6-KEH w1 hydrolyses all 6-kestose (6K) within 1 h, while 6,6-kestotetraose (j), 1&6,6-kestopentaose (m), 6,6,6-kestopentaose (p) and bifurcose (b) decrease only slightly after 6 h incubation. Other graminan-type fructans are not substrates (Fig. 3b). Identical profiles were observed for 6-KEH w2 (not shown).

Table 1.  Substrate specificity of 6-kestosidase (6-KEH) w1 and w2 from Triticum aestivum
SubstrateDPRelative activity (%)
  1. Results are shown as values relative to activity with the substrate 6-kestose.

  2. nd, Not determined.

Phlein oligomers4–121.831.62
Inulin> 100.240.19
Figure 3.

Action of recombinant 6-KEHs on fructans. (a) HPAEC chromatograms of reaction mixtures of recombinant 6-KEH w1 incubated with 6-kestose after 0 and 1 h incubation at 30°C. Reference includes G (Glc), F (Fru), S (Suc), 1-kestose (1-K), 6-kestose (6-K) and 1,1-nystose (N). (b) HPAEC chromatograms of reaction mixtures of recombinant 6-KEH w1 incubated with wheat (Triticum aestivum) stem fructans and followed as a function of incubation time. G, Glc; F, Fru; S, Suc; 1K, 1-kestose; 6K, 6-kestose; N, 1,1-nystose; B, bifurcose or 1&6-kestotetraose; i, 6,1 kestotetraose; j, 6,6-kestotetraose; k, 6&1,1 kestopentaose; l, 1&6,1 kestopentaose; m, 1&6,6-kestopentaose; n, 6; 1&6-kestopentaose; o, 6; 6,1 kestopentaose; p, 6,6,6-kestopentaose. A reference with S, 1K, 6K and N is included. A slightly different HPAEC elution gradient was used compared with (a).

Properties of recombinant 6-KEH w1 and w2

The kinetic analysis of both enzymes with 6-kestose as a substrate is shown in Fig. 4a. A Km value of 29 (w1) and 28 mm (w2) was calculated (Lineweaver–Burk). In contrast to chicory 1-FEH IIa (Van den Ende et al., 2001) and wheat 1-FEH w1 and w2 (Van den Ende et al., 2003a), Suc is not a strong inhibitor of the wheat 6-KEH w1 and w2. Only a 40–60% inhibition was obtained when 100 mm Suc was added to 4 mm 6-kestose (Fig. 4b).

Figure 4.

Properties and partial purification of 6-KEH. (a) Double reciprocal plot of recombinant wheat (Triticum aestivum) 6-KEH w1 (▪) and w2 (◆) activity against the concentration of 6-kestose. The fitted linear regression line and its parameters are presented. Km values were estimated based on these parameters. (b) Percentage of inhibition of wheat 6-KEH w1 (▪) and w2 (▴) as a function of Suc concentration at a 6-kestose concentration of 4 mm. Reaction time was 10 min; reaction temperature, 30°C. Results are mean values (± SE) of three experiments. (c) Partial purification of 6-KEH w1 from wheat crowns. Activity of 6-KEH (5 mm 6-kestose, ▪); inulinase (5 mm inulin, □); levanase (5 mm levan, •); and invertase (5 mm Suc, ▴) in the fractions obtained after Mono Q anion chromatography at pH 6.0. Reaction time and temperature were 2 h and 30°C, respectively. Similar results were obtained for wheat ears and wheat stems (anthesis stage).

Partial purification and MS analysis of native 6-KEH w1

We were able to partially purify a 6-KEH activity from wheat crowns, wheat stems and ears at anthesis stage. A purification procedure based on (NH4)2·SO4 precipitation, lectin affinity chromatography (ConA), anion- and cation-exchange chromatography was designed. Despite numerous efforts, we did not succeed in obtaining a completely purified native enzyme, because of the minute amounts that were extractable from wheat tissues. Nevertheless the ‘purified’ enzyme was essentially free from invertase, inulinase and levanase activity, as demonstrated in Fig. 4c. The fraction with the highest activity was subjected to SDS–PAGE (not shown). Five bands were cut out, trypsinized in-gel, and subjected to Q-TOF MS. A 72 kDa band resulted in peptides identical to a theoretical tryptic digest of the cDNA-derived 6-KEH w1 sequence (Fig. 1) which yielded 49 peptides (designated T1–T49 from N- to C-terminus). Masses of ZipTip eluted tryptic 6-KEH peptides were compared (Table 2) with the masses of theoretical peptides (with the consideration of one possible missed cleavage site). All masses detected matched – within the acceptable mass measurement error of ±0.2 Da – with one of the theoretical fragments (Table 2). Collision-induced dissociation (CID) MS/MS analysis yielded a number of sequence tags, which proved the identity of the tryptic peptides (Table 2). No specific 6-KEH w2 or other fragments appeared, suggesting that the 72 kDa band represents the native 6-KEH w1.

Table 2.  Fragment ions detected in Q-TOF after tryptic digest of native 6-kestosidase (6-KEH) w1 from Triticum aestivum, with calculated matches to a theoretical digest of virtual cDNA derived protein, and confirmation of identity by tandem MS/MS sequencing
Observed massCharge stateCalculated mass (presumptive 6-KEH w1 ion)MS/MS sequence (from N- to C-terminus)
  1. Amino acids that are different between 6-KEH w1 and w2 are in bold and underlined.

574.42+ 524.33 [T6 + 2H]2+HQVQNIALPK
1006.62+1006.51 [T10 + 2H]2+AGNNPVIQSGVPGLNSGQFR
785.92+ 785.88 [T11 + 2H]2+DPTTGWIGPDGLWR
619.32+ 619.27 [T18 + 2H]2+DAFVPDTDDSR
286.12+ 286.16 [T19 + 2H]2+LWPR
589.32+ 589.27 [T20 + 2H]2+IDYGNFYASK
372.72+ 372.67 [T21 + 2H]2+TFFDSK
1019.92+1019.97 [T24 + 2H]2+IIWAWTTETDSSSDDIAK
577.32+ 577.28 [T25 + 2H]2+GWAGIYSFPR
512.32+ 512.24 [T44 + 2H]2+SAVESFGGGGR

6-KEH gene expression in crowns and leaves of young field-grown plants

Both novel genes were isolated from a cDNA library derived from the crown tissues of winter wheat sampled in mid-November (Kawakami & Yoshida, 2002). Fig. 5a shows the expression of 6-KEH w1 and w2 genes in field-grown wheat sampled from October to February. The typical climate at the experimental field (Sapporo, Japan) was as in a previous paper (Yoshida et al., 1998). The expression of 1-SST and 6-SFT genes has been shown previously (Kawakami & Yoshida, 2002). The temperatures decreased soon after sowing, then subzero minimum temperatures occurred from mid-November, and the wheat was covered with snow from early December. Despite the environmental changes during this period, both 6-KEH w1 and w2 genes were constitutively expressed in the crowns. In severe contrast, the genes were not expressed in photosynthetically active and exporting leaves (1–4, Fig. 5a). However, a low expression appeared in January and February leaves, when they were covered by a very thick layer of snow and both photosynthesis and growth was prevented (5–6, Fig. 5a). The roots could not be sampled because the soil was frozen from mid-November.

Figure 5.

Expression analysis of 6-KEHs and other fructan-related genes. (a) Northern analysis of transcripts of 6-KEH w1 and w2 in crowns and leaves of young field-grown and cold-hardened wheat (Triticum aestivum) plants throughout the period October 2001–February 2002 (Sapporo, Japan). Sampling dates: (1) 9 October; (2) 8 November; (3) 8 December; (4) 19 December; (5) 10 January; (6) 11 February. The 25S rRNA bands stained with ethidium bromide are shown as a loading control. (b) Northern analysis of transcripts of genes related to fructan metabolism (1-SST, 6-SFT, 1-FEH w2, 6-KEH w1 and w2) in leaves, roots and crowns of vermiculite-grown wheat seedlings at different times after sowing (3, 7, 10, 13, 17, 21 and 26 d). The 25S rRNA bands stained with ethidium bromide are shown as a loading control.

Expression of fructan metabolism genes in vermiculite-grown seedlings

Fructan concentration in wheat might be determined by the balance of both fructan biosynthetic and breakdown enzymes, and these activity levels may change under stress (Van den Ende et al., 2003a). To investigate what happens under controlled (no stress) conditions, the expression level of 1-SST, 1-FFT, 6-SFT, 1-FEH w2, 6-KEH w1 and w2 genes was followed in young seedlings grown under optimal conditions (0–26 d after sowing). Primary sink shoots (3 d after sowing) show a very strong 1-SST expression and a weaker 6-SFT expression, while the opposite is true for 3-d-old roots (Fig. 5b). Fructan biosynthetic genes were downregulated in the leaves between 3 and 17 d after sowing. A weak 1-SST expression appeared again after 21 d. During the same period, the 6-SFT expression remains at a high level in the roots, while 1-SST shows a variable pattern but increases 21 d after sowing. The novel crown tissue could be distinguished after 3 wk. 1-SST expression is stronger than 6-SFT expression in this sink tissue. Similarly to field-grown plants, 6-KEH w1 expression is specifically restricted to the crown tissue while 6-KEH w2 expression could not be detected in any case. 1-FEH w2 expression is weak but apparently constitutive.

Carbohydrates in total extracts and in apoplastic fluid

Figure 6 shows a detail of some typical total carbohydrate patterns in different fractions (total extract vs apoplast) of field-grown, cold-hardened wheat crowns and leaves. Apoplastic fluid was quickly isolated (15 min) from wheat crowns as described (Boller & Métraux, 1988) and immediately heat-inactivated after the centrifugation step. It is noteworthy that the level of 6-kestose (arrow) is relatively low compared with all other fructans, both in total extracts and in apoplastic fluids. Strikingly, 1-kestose and Suc are considerably less prominent in the apoplastic fluids of crowns and leaves. The different carbohydrate profiles observed between total and apoplastic fluid fractions could not be explained by a rapid enzymatic breakdown of Suc and 1-kestose during the short apoplastic fluid isolation procedure, as was tested during a time-dependent incubation of apoplastic proteins together with carbohydrates derived from a total extract (data not shown).

6-KEH is probably localized in the apoplast

1-SST and other fructosyl transferases are present in the vacuole (Van Laere & Van den Ende, 2002). As 6-KEH is coexpressed together with 1-SST and 6-SFT in crowns of wheat (Fig. 5b), 1-SST was used as a vacuolar marker for investigating the subcellular localization of 6-KEH. Apoplastic fluid (APO) was isolated from wheat crowns as described (Boller & Métraux, 1988). The activities of 1-SST and 6-KEH were determined in the apoplastic fluid, in the remaining tissue fraction (RTF), and in the total tissue (TOT). The 6-KEH/1-SST ratio was much higher in the APO than in the RTF and TOT fractions (Fig. 7a), indicating the presence of a substantial amount of 6-KEH activity in the apoplast. Apoplastic 1-SST activity was very low, and was within the cellular leakage level of 2–4%. During fructan biosynthesis in vitro from Suc, slightly higher 6-kestose/1-kestose and 6-kestose/1&6-kestotetraose ratios were observed for RTF compared with TOT, further suggesting the removal of a specific 6-KEH activity by the apoplastic fluid extraction (Fig. 7b).


Molecular characterization of 6-KEH w1 and w2 cDNAs

Both 1-FEH and 6-FEH are needed to break down graminan-type fructan in wheat. This paper describes the cloning and functional analysis of two wheat cDNAs encoding highly specific 6-kestosidases, termed 6-KEH w1 and w2 (Fig. 1). As observed for several other FEHs (Van den Ende et al., 2003a), the 6-KEHs show more homology to cell-wall invertases than to vacuolar invertases and fructan biosynthetic enzymes. They fall into the same subgroup (IVc) as 1-FEH from wheat, and cluster close to the 1-FEHs from chicory and a 6&1-FEH (AtcwINV6; De Coninck et al., 2004) from Arabidopsis (IVa, Fig. 2). The recently cloned 6-FEHs from sugar beet (Van den Ende et al., 2003b) and Arabidopsis (AtcwINV3, De Coninck et al., 2004) cluster in a more distinct subgroup (III, Fig. 2). These observations show that the 2,1 and 2,6 substrate specificity of FEHs is not monophyletic.

Functional analysis of wheat 6-KEHs

Plant fructan biosynthetic and breakdown enzymes, as well as invertases, differ in donor (Suc, fructan) and acceptor (Suc, fructan, water) specificities. They all belong to glycosyl hydrolase family 32 (Henrissat & Davies, 1997). Previously, the P. pastoris secretion vector system was successfully used for expressing such enzymes (see Introduction), and in all cases the characteristics of the heterologous enzymes were similar to their native counterparts, proving the validity of the system. Here we used the same system for expressing the 6-KEH w1 and w2 cDNAs. Surprisingly, the derived enzymes are highly active against 6-kestose as a substrate, while higher DP 2,6 oligo-and polymers (phlein, levan), 2,1 type fructans (1-kestose, inulin) and mixed type graminans are hardly degraded (Table 1). As expected, 6-KEH w1 and w2 showed no invertase activity, a characteristic typical for all other FEHs characterized so far (Van den Ende et al., 2003b). Longer term incubations showed some capacity to degrade 6,6-kestotetraose, 1&6,6-kestopentaose, 6,6,6-kestopentaose and 1&6-kestotetraose, but this occurred only after all 6-kestose was degraded in the reaction mixture (Fig. 3b). Although the enzyme specifically cuts between the two adjacent 2,6 bound fructoses in 6-kestose, the reducing carbohydrate levanbiose is not a good substrate (Table 1), indicating the crucial importance of the presence of a glucose (Glc) moiety or the absence of a reducing group.

In general, fructan biosynthetic and breakdown enzymes often show side activities in vitro depending on the reaction conditions (reviewed by Van Laere & Van den Ende, 2002). The high specificity of the 6-KEH enzyme towards a single fructan substrate is unique among all fructosyl transferases characterized so far. The 6-FEHs from L. perenne (Marx et al., 1997a), Dactylis glomerata (Yamamoto & Mino, 1985), sugar beet (Van den Ende et al., 2003b) and Arabidopsis (De Coninck et al., 2004) efficiently degrade low- and high-DP levans as well as neokestose and levanbiose. 1-FEHs from chicory and wheat degrade low- and high-DP inulin-type fructans and show some limited levanase side activities (De Roover et al., 1999; Van den Ende et al., 2003a).

Properties of 6-KEHs

A 6-KEH fraction was purified from wheat and was free of contaminating inulinase, levanase and invertase activities (Fig. 4c). Mass spectrometric analysis of a 72 kDa band convincingly demonstrated the link with the corresponding 6-KEH w1 cDNA (Table 2). The 72 kDa molecular mass of 6-KEH w1 was comparable to the 70 kDa molecular mass of wheat 1-FEH w1 and w2 (Van den Ende et al., 2003a); 64 kDa of 1-FEH II from Cichorium intybus (De Roover et al., 1999); 57 kDa of 6-FEH from D. glomerata (Yamamoto & Mino, 1985); 69 kDa of 6-FEH from L. perenne (Marx et al. (1997a); 62.5 kDa of 1-FEH from Hordeum vulgare (Henson & Livingston, 1998); and 75 kDa of 1-FEH from Helianthus tuberosus (Marx et al., 1997b). All these enzymes are believed to be monomers, while sugar beet 6-FEH appears to be a heterotrimer (Van den Ende et al., 2003b). Like sugar beet 6-FEH (Van den Ende et al., 2003b), and in contrast to other FEHs (Van den Ende et al., 2001, 2003a), wheat 6-KEHs are only poorly inhibited by Suc (Fig. 4).

6-KEH is probably localized in the apoplast

Like barley, wheat produces predominantly linear 2,6-type levans (stems) and branched graminans (crowns, ears and induced leaves) that are preferentially stored in the vacuole and are probably degraded by 6-FEH and 1-FEH enzymes. It has been suggested that 1-FEH w2 might be involved in 2,1 trimming of graminan-type fructans during the period of active fructan biosynthesis, in this way preventing the accumulation of higher DP inulin-type fructans in wheat stems (Van den Ende et al., 2003a). The low but constitutive expression of wheat 1-FEH w2 in young wheat plants is consistent with such a role (Fig. 5b). We became suspicious about a putative vacuolar localization of 6-KEH after we found a small but specific 6-KEH activity in wheat stems accumulating only the 6-kestose-based linear levan-type fructans (typical oligo-levan pattern as shown by Bancal et al., 1993). These initial results strongly suggested a different (subcellular) localization of fructan biosynthetic enzymes and 6-KEH.

However, the 6-KEH work was focused on cold-hardened wheat crowns, because (1) both 6-KEH activities and mRNA expression levels were found to be high in these tissues; and (2) data were available from crowns of cold-hardened oat (Livingston & Henson, 1998). These authors showed the presence of fructans and increased FEH activity in the apoplast of 2-PH cold-hardened oat. Accordingly, we performed apoplastic fluid isolations with 1-SST as a vacuolar marker, confirming an apoplastic localization of the 6-KEH enzyme in cold-hardened wheat crowns (Fig. 7a). An improved version of signalP (Bendtsen et al., 2004) predicts that 6-KEH w1 and w2 have the hydrophobic N-terminal signal sequence required for cotranslational insertion in the endoplasmic reticulum and secretion from the cell. This prediction is consistent with our experimental findings. However, it is noteworthy that the 1-FEHs from chicory and wheat also have an extracellular fate according to signalP, whereas other observations support a vacuolar localization (Wagner & Wiemken, 1986; Van den Ende et al., 2003a). Perhaps these 1-FEHs contain a vacuolar targeting signal in their prolonged C-terminal ends containing short stretches of hydrophobic amino-acids, as observed in vacuolar invertases (Unger et al., 1994).

Putative roles of 6-KEHs and other apoplastic FEHs

A role in the prevention of bacterial levan formation and infection?  FEHs fulfil different important roles in fructan plants (Van den Ende et al., 2004). Recently, we discovered that FEHs also occur in nonfructan plants. As no endogenous substrates are present in these plants, it was proposed that FEHs have a defence-related function against fructan producing pathogenic or endophytic microbes (Van den Ende et al., 2004). It could be expected that this type of FEH also occurs in fructan-producing plants. Therefore it cannot be excluded that wheat 6-KEHs are involved in the degradation of apoplastic 6-kestose of bacterial origin. Bacterial levans are produced by extracellular levansucrases, which preferentially use Suc as donor substrate and levan-type fructans (such as 6-kestose) as preferential acceptor substrate (Dedonder, 1972; Chambert et al., 1974; Han, 1990; Song & Jacques, 1999). 6-KEHs might prevent infection by the prevention of bacterial levan formation in the apoplast. This could be corroborated by investigating the resistance to bacterial infection of A. thaliana plants overexpressing wheat 6-KEH.

A role in signalling?  It is well known that invertases have the capacity to produce the three kestose trisaccharides 1-kestose, 6-kestose and neokestose in vitro from high Suc concentrations (Cairns & Ashton, 1991). It cannot be excluded that the production of these trisaccharides is of physiological importance. Interestingly, by specifically expressing yeast invertase in companion cells, it has been demonstrated that 6-kestose is specifically produced in the leaf, loaded into the phloem and transported to the developing tubers (Züther et al., 2003). It is conceivable that Suc derivatives or homologues such as 6-kestose and trehalose fulfil crucial roles in plant carbohydrate partitioning and can act as signal molecules captured by specific sensors (Van den Ende et al., 2004). Assuming such a role for 6-kestose, ideally the signal should be destroyed after the sensing process and a highly specific enzyme such as 6-KEH might realize this. Perhaps the existence of such extremely selective fructan enzyme (among many other aspecific types) is, in itself, indicative of such a role.

Degradation of endogenous apoplastic fructans   A more straightforward function for apoplastic wheat 6-KEHs is the degradation of endogenous but extracellular fructan. Compared with total extracts, the apoplastic fluid of cold-hardened wheat crowns contains lower Suc and 1-kestose levels (Fig. 6). Preliminary results show that this result can probably be explained by the presence of cell wall-bound (high pI) invertase/1-FEH activity, but it remains to be further demonstrated whether one or two different enzymes are involved. This result confirms that two fructan pools can be differentiated in cold-hardened crown tissues. It can be presumed that plants use a complex mixture of invertases and FEHs to degrade apoplastic fructans and Suc efficiently. In addition to 6-KEH, a true 6-FEH (levanase) was recently characterized and cloned from wheat inflorescenses (Van Riet and co-workers, unpublished data). While high pI invertases might become more stable by interaction with the cell wall, the mobility of low pI FEH isoforms (such as 6-KEH) might allow these enzymes to function in the vicinity of the plasma membrane, where fructans are believed to interact with membranes to stabilize them under stress (Vereyken et al., 2001; Hincha et al., 2002).

So far the origin of apoplastic fructans it is not clear. Wheat and many other species were tested in our laboratory for the presence of 6-SST or 1-SST activities in apoplastic fluids of different tissues. No evidence for fructan biosynthetic activity was detected in the apoplast (Fig. 7b), indicating that fructans are biosynthesized in the vacuole and need to be transported to the apoplast by an (until now) elusive mechanism. It is tempting to speculate that a vesicle-mediated fructan transport (exocytosis) occurs between the vacuole and the plasma membrane (Echeverria, 2000). Exocytosis of polysaccharides was recently described in Phaeocystis globosa (Chin et al., 2004), but such a process remains speculative in plants. Like Livingston & Henson (1998), we found that fructan levels in the apoplast exceed the levels that could be explained by ‘normal’ cellular leakage. Further, the presence of fructans was clearly demonstrated in the guttated liquid of oat (Livingston & Henson, 1998). These observations are consistent with transport of fructans from the vacuole to the apoplast.

The overwintering organs of winter wheat should be protected against freezing temperatures. In addition to other important adaptations such as the prevention of ice formation in the apoplast (Yeh et al., 2000), the accumulation of apoplastic fructans is probably one specific adaptation that helps plants survive freezing temperatures. Apoplastic and cell wall-bound FEHs and invertases might fulfil crucial roles in modulating the concentration, structural composition and DP of apoplastic fructans. In particular, we propose that the 6-KEHs described here catalyse the final step in graminan degradation: the specific hydrolysis of apoplastic 6-kestose to Suc and Fru. Most of the apoplastic 6-kestose probably originates from 1-FEH activity on 1&6-kestotetraose, by far the most abundant fructan in the apoplast of cold-hardened wheat crowns (Fig. 6).

It was previously shown that fructan degradation in 2-PH crowns (under snow) of freezing-tolerant wheat cultivars correlated with increased Fru levels (Yoshida et al., 1998). It can be speculated that a specific mixture of oligofructans, together with Suc and hexoses, might be optimal for membrane protection and associated freezing tolerance. Net fructan accumulation is the result of both fructan biosynthetic (FBE) and breakdown enzymes (FEH). Yukawa et al. (1995) demonstrated that fructan accumulation is associated with the FEH/FBE ratio and suggested that the varietal differences in fructan accumulation are largely affected by the level of FEH activity. Interestingly, fructan concentrations (Yoshida et al., 1998), as well as 1-SST and 6-SFT expression levels (Kawakami & Yoshida, 2002), are higher in crown tissues compared with leaf tissues. This correlates well with the observation that crown tissues (including the apex) are more tolerant than leaves to frost (Chen et al., 1983). Further research is needed to find out whether the lower freezing tolerance of leaves can be explained by (1) lower fructan concentrations; (2) the inability or reduced ability of leaves to transport fructans to the apoplast; or (3) the inability or reduced ability of leaves to partially degrade apoplastic fructans (absence of 6-KEH expression in Figs 5a). Future research on apoplastic fructan-metabolizing enzymes is needed to provide further insights in the complexity of apoplastic FEH isoforms and their putative contribution to fructan modulation and associated freezing tolerance.

Similarly to sugar beet 6-FEH, 6-KEH is apparently expressed in sink tissues such as crowns, ears and stems, but not in mature source leaves (Fig. 5a). The specific expression of 6-KEH in tissues that cannot photosynthesize, such as crown tissues and leaves under snow (Fig. 5a), suggests that these tissues may depend mainly on fructan degradation as energy source. It has previously been demonstrated that fructan breakdown is induced in leaves kept in the dark at 1°C (Yukawa et al., 1994). In contrast to other FEHs, it is tempting to speculate that the 6-KEH gene is not inhibited but rather induced by a certain Suc concentration or Suc flux. No 6-KEH expression was found in strongly growing young leaves or roots (Fig. 5b). In these tissues, however, extremely high acid invertase activities were found (not shown), keeping the actual Suc concentrations at a low level.


To date, the presence of specific 6-KEH enzymes in plants has never been reported or even suggested. Here we describe two highly similar 6-KEHs from cold-hardened wheat crowns at the DNA and protein level. Their specificity towards 6-kestose is unique among all fructan-metabolizing enzymes characterized so far. Expression analyses and localization experiments in wheat indicate that the 6-KEHs are present in the apoplastic fluid of fructan-accumulating sink tissues. Together with other apoplastic FEHs, these enzymes might control apoplastic fructan concentrations, contributing to increased frost tolerance. However, an alternative role in the degradation of exogenous 6-kestose from bacterial origin or in signalling processes cannot be excluded.


The authors wish to thank E. Nackaerts for his technical assistance. W. Van den Ende is supported by a grant from the Fund for Scientific Research (FSR, Flanders).