Mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) re-models hemicelluloses in Equisetum shoots but not in barley shoots or Equisetum callus


  • Kyle E. Mohler,

    1. The Edinburgh Cell Wall Group, Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Edinburgh, UK
    Search for more papers by this author
  • Thomas J. Simmons,

    1. The Edinburgh Cell Wall Group, Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Edinburgh, UK
    Search for more papers by this author
  • Stephen C. Fry

    Corresponding author
    • The Edinburgh Cell Wall Group, Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Edinburgh, UK
    Search for more papers by this author

Author for correspondence:

Stephen C. Fry

Tel: +44 131 650 5320

Email: s.fry@ed.ac.uk


  • Among land-plant hemicelluloses, xyloglucan is ubiquitous, whereas mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) is confined to the Poales (e.g. cereals) and Equisetales (horsetails). The enzyme MLG:xyloglucan endotransglucosylase (MXE) grafts MLG to xyloglucan. In Equisetum, MXE often exceeds extractable xyloglucan endotransglucosylase (XET) activity; curiously, cereals lack extractable MXE. We investigated whether barley possesses inextractable MXE.
  • Grafting of endogenous MLG or xyloglucan onto exogenous [3H]xyloglucan oligosaccharides in vivo indicated MXE and XET action, respectively. Extractable MXE and XET activities were assayed in vitro.
  • MXE and XET actions were both detectable in living Equisetum fluviatile shoots, the MXE : XET ratio increasing with age. However, only XET action was observed in barley coleoptiles, leaves and roots (which all contained MLG) and in E. fluviatile intercalary meristems and callus (which lacked MLG). In E. fluviatile, extractable MXE activity was high in mature shoots, but extremely low in callus and young shoots; in E. arvense strobili, it was undetectable.
  • Barley possesses neither extractable nor inextractable MXE, despite containing both of its substrates and high XET activity. As the Poales are xyloglucan-poor, the role of their abundant endotransglucosylases remains enigmatic. The distribution of MXE action and activity within Equisetum suggests a strengthening role in ageing tissues.


The plant cell wall is a hydrated extraprotoplasmic complex of polysaccharides, proteins and, in some cases, polyphenols (e.g. lignin), silica and other materials. It plays crucial roles in pest and pathogen defence, cell signalling, cell-to-cell adhesion, growth regulation and the determination of cell morphology (Albersheim et al., 2011). Three types of polysaccharide constitute the majority of the dry weight of the cell wall and, in a widely accepted model, each has a distinct function: cellulose microfibrils are the major load-bearing structures of the wall; they are tethered and/or held apart by hemicelluloses, and the resulting cellulose–hemicellulose network is immersed in a pectic gel (Fry, 2011). There is also evidence for covalent bonds between ‘distinct’ polysaccharides (Thompson & Fry, 2000; Popper & Fry, 2008).

Many plant cells synthesize two chemically and functionally distinct cell walls: the primary and secondary. The primary cell wall is a dynamic plastic structure that surrounds virtually all plant cells. By controlled relaxation, it is able to expand to many times its original area, allowing cell expansion whilst maintaining the wall's structural integrity. The secondary cell wall, in comparison, is often rigid and is deposited internal to the primary cell wall of some cells during/following the cessation of growth (Taylor, 2008). This is thought to provide strength to mature tissues.

The presence and fine structure of hemicelluloses vary across the plant taxa in a phylogenetically dependent manner (Popper, 2008; Popper & Tuohy, 2010; Scheller & Ulvskov, 2010; Sørensen et al., 2010). A distinguishing feature of the cell walls of all land plants tested to date is the presence of some form of xyloglucan, the major primary cell wall hemicellulose of dicots and noncommelinid monocots (Popper & Fry, 2003, 2004). Xyloglucan has a β-(1→4)-glucan backbone which has a regular pattern of xylosylation; for many dicots and other plant groups, including Equisetum spp., the first three of every four glucose residues is 6-O-α-xylosylated, forming a heptasaccharide motif termed XXXG (Vincken et al., 1997; Peña et al., 2008; for a description of xyloglucan nomenclature, see Fry et al., 1993). From XXXG ‘base units’, increasingly complex structures can be created by further decoration with a variety of sugars; an updated list of their abbreviations is given in table 1 of Franková & Fry (2012b). The xyloglucan in the Poaceae and other taxa with ‘type II’ walls is markedly different from that of the dicots and noncommelinid monocots (which have ‘type I’ walls), often including XXGGG and other base units and varying in the other constituent sugars which it contains (Hsieh & Harris, 2009). The study of these repeating structures of xyloglucan has been facilitated by the use of xyloglucan endoglucanase (XEG), a xyloglucan-specific endohydrolase that cleaves the β-(1→4) bond in the sequence …GX… (Pauly et al., 1999; Fig. 1).

Figure 1.

Schematic view of the strategy used for the simultaneous detection of the action of mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) in vivo. Freshly sliced tissue was fed a radioactive xyloglucan oligosaccharide (XXXGol or XXLGol), which acts as an exogenous acceptor substrate when endogenous MXE and XET act on their endogenous donor substrates mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) and xyloglucan, respectively. After 16 h of incubation, alcohol-insoluble residue (AIR) was prepared, total hemicellulose was extracted with NaOH and aliquots were digested with lichenase (○) or xyloglucan endoglucanase (XEG) (●), which generate radioactive oligosaccharides diagnostic of prior MXE or XET action, respectively. Key: ···G4G˜G4G4G···, a representative length of Equisetum MLG (where G = β-glucose residue, 4 = (1→4)-linkage, ˜ = (1→3)-linkage); ···LDGXLEGXX···, a representative length of Equisetum xyloglucan (after Peña et al., 2008); XXXGol ★, radiolabelled xyloglucan oligosaccharide. The bond between MLG and the radiolabelled oligosaccharide is shown as ··· because the precise siting of this bond is still undefined.

Mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) is a homopolymer composed of glucopyranose residues connected by a nonrandom order of (1→3) and (1→4) bonds. Typically, two or three (1→4) bonds are followed by a single (1→3) bond, thus:


where G represents β-d-glucopyranose, and 3 and 4 represent (1→3) and (1→4) bonds, respectively. Successive rigid (1→4)-linked segments (underlined) are cellulose-like regions in the polymer, but (1→3) bonds are more flexible and confer water solubility (Burton & Fincher, 2009).

In contrast with the wide phylogenetic distribution of xyloglucan, MLG is sparingly distributed across the plant kingdom. Indeed, the presence of MLG in Equisetum, the only case of MLG in vascular plants other than the evolutionarily distant commelinid monocots, has only been discovered recently (Fry et al., 2008a; Sørensen et al., 2008). The confinement of MLG, within the vascular plants, to two such unrelated taxa indicates convergent evolution. The clearest distinction between poalean MLG and Equisetum MLG concerns the ratio of trisaccharide to tetrasaccharide units; in poalean MLG, the trisaccharide prevails, whereas, in Equisetum, the tetrasaccharide is more abundant (Fry et al., 2008a; Sørensen et al., 2008; Xue & Fry, 2012). Lichenase, an MLG-specific endoglucanase, has frequently been employed to study the fine structure of MLG. Lichenase cleaves the β-(1→4) bond directly following a β-(1→3) bond (in the nonreducing- to reducing-end direction) (Planas, 2000; Fig. 1).

Xyloglucan endotransglucosylase/hydrolases (XTHs) are apoplastic enzymes whose transglucosylase activity (xyloglucan endotransglucosylase (XET)) is found to varying degrees in cell extracts of all land plants. XET activity catalyses the endo-cleavage of xyloglucan (the donor substrate) and the creation of a glycosidic link between the newly formed potentially reducing terminus and the nonreducing terminus of another xyloglucan (or xyloglucan oligosaccharide) molecule (the acceptor substrate) (Fry et al., 1992; Nishitani & Tominaga, 1992).

It has been shown by in-vivo 3H/13C dual-labelling experiments that XET action contributes to wall assembly by integrating newly secreted xyloglucan chains into the wall's architecture (Thompson et al., 1997), as well as to the re-structuring of existing wall material by cutting and re-forming intermicrofibrillar xyloglucan tethers (Thompson & Fry, 2001). This may allow the gradual growth of the cell without severely compromising the wall strength.

Various lines of recent evidence support a role for XTHs in wall expansion (Van Sandt et al., 2007; Lee et al., 2010; Sasidharan et al., 2010; Harada et al., 2011; Miedes et al., 2011). However, the addition of AtXTH14 or AtXTH26 to onion epidermal peels during constant-load extensiometry decreased wall extensibility (Maris et al., 2009). Furthermore, XTH proteins produced during wood secondary wall development are suggested to be involved in the strengthening of xylem tissue (Mellerowicz et al., 2008). It has been proposed that the ratio of newly synthesized xyloglucan to XTH within the Golgi vesicles can dictate whether XET action strengthens or loosens the cell wall (Nishikubo et al., 2011).

An additional enzyme activity, MLG:xyloglucan endotransglucosylase (MXE), recently discovered in extracts of all tested species of Equisetum and several charophytic algae, catalyses a reaction similar to XET activity, but using MLG as a donor substrate as opposed to xyloglucan (Fry et al., 2008b). The product of MXE activity is thus an MLG–xyloglucan hybrid polymer, so that MXE is termed a hetero-endotransglucosylase, making it particularly novel. A positive correlation between tissue age and extractable MXE activity led to the proposal of a wall strengthening role for MXE (Fry et al., 2008b). The finding that Equisetum MLG is immunologically undetectable in immature cortical parenchyma cells, but abundant in structural tissues, such as the epidermis, sclerenchyma, pith parenchyma and some mature cortical parenchyma (Sørensen et al., 2008), is consistent with this hypothesis, although pectins could have masked the detection of MLG and other polysaccharides in that study (Marcus et al., 2008, 2010).

Some additional recently described examples of plant endotransglycosylase activities include trans-β-mannanase (Schröder et al., 2004) and trans-β-xylanase (Franková & Fry, 2011). In addition, ‘exo-transglycosylase’ activities have been detected, including trans-β-xylosidase acting on xylans (Franková & Fry, 2011), and trans-α-xylosidase and trans-β-galactosidase acting on xyloglucans (Sampedro et al., 2010; Franková & Fry, 2012a,b). However, given the diversity of hemicelluloses and the crucial role thought to be played by XET activity across the plant kingdom, it is perhaps surprising that so few other endotransglycosylase activities have been described and characterized. Preliminary evidence exists for activities operating on various donor/acceptor combinations, but these activities have not been characterized in detail (Kosík et al., 2010), or may even be artefacts of the fluorescent label used for detection (Kosík et al., 2011). Interestingly, there have been various reports of MXE and MXE-like activities as ‘side reactions’ catalysed by predominantly XET-active enzymes. For example, Stratilová et al. (2010) isolated an XTH from nasturtium which could catalyse a transglucosylation reaction between a xyloglucan or hydroxyethylcellulose (HEC) donor substrate and laminari-, MLG-, xylo- and pustulan-oligosaccharide acceptor substrates, albeit at much lower rates than in the conventional xyloglucan-to-xyloglucan XET reaction. Hrmova et al. (2007) found that a heterologously produced barley XTH (HvXET5) exhibits MXE activity at c. 0.2% of its XET rate. This enzyme was also able to use soluble cellulose derivatives other than xyloglucan as donor substrates (at 44% of the XET rate in the case of HEC) with xyloglucan oligosaccharides or cello-oligosaccharides as acceptors. Total extracts from the grass Holcus lanatus L. also had an MXE : XET activity ratio of c. 0.002 : 1, whereas total Equisetum enzyme extracts exhibited ratios > 1 : 1 (Fry et al., 2008b).

The possibility of discovering XTHs with > 0.2% MXE activity would be particularly interesting in the Poales (including grasses and cereals), because these plants have a surprisingly large number of putative XTH genes despite having xyloglucan-poor cell walls. For example, rice and maize have 30 and 32 putative XTH genes, respectively (Eklöf & Brumer, 2010). As the Poales include major world crops, with numerous XTHs and a high MLG content, poalean MXE activity seemed to be a reasonable target of research. Indeed, having discovered the lack of appreciable extractable MXE activity, we speculated that poalean cell walls might contain inextractable MXE-active enzymes (Fry et al., 2008b). Alternatively, the low in-vitro MXE activity exhibited by HvXET5 might possibly play a physiologically significant role in vivo (Hrmova et al., 2007). However, at this point, no direct analysis of in-vivo MXE action in poalean cell walls had been undertaken, and therefore evidence for both of these hypotheses was lacking.

The aim of the present work was to devise a novel method by which to test for MXE action in situ within Equisetum tissue (which is known to contain extractable MXE activity) and, if successful, to apply the method to a member of the Poales to test for the presence of inextractable MXE activity. In addition to inextractability, there are several other potential reasons why an enzyme capable of ‘action’ in vivo might not exhibit in-vitro ‘activity’ in extracts (Fry, 2004). We also further characterized the role of MXE in Equisetum by comparing MXE action and activity in several tissues of different ages.

Materials and Methods


Xyloglucan endoglucanase (XEG) was a generous gift from Novo Nordisk A/S, Bagsværd, Denmark (Pauly et al., 1999). Lichenase (from Bacillus subtilis; 330 U mg−1) and β-d-glucosidase (from Aspergillus niger; 52 U mg−1) were from Megazyme, Inc. (Bray, Ireland). Merck silica-gel 60 TLC plates were from VWR (Lutterworth, UK). Dialysis tubing (12–14-kDa cut-off) was purchased from Medicell International, Ltd. (London, UK). Reductively tritiated oligosaccharides ([1-3H]XXXGol and [1-3H]XXLGol) were synthesized in-house, essentially as described by Hetherington & Fry (1993), with oligosaccharides from Megazyme, Inc. Their specific radioactivities were c. 100 and 66 MBq μmol−1, respectively. Miracloth was from Calbiochem (http://www.emdbiosciences.com). Solvents and scintillant were from Fisher Scientific (Loughborough, UK). Other general chemicals came from Sigma-Aldrich (UK).


Equisetum fluviatile L. was grown outdoors in a pond near the laboratory. Equisetum arvense L. and Holcus lanatus L. were grown in a private garden in Edinburgh. Winter barley (Hordeum vulgare L.) cultivar Pearl was grown from seed in a glasshouse under natural light for up to 12 wk.

Preparation of alcohol-insoluble residue (AIR)

Plant material (0.5 g) was ground and washed in 75% (v/v) ethanol. After 2 h, insoluble material was collected, re-washed in 75% (v/v) ethanol at least three times, rinsed in acetone and dried.

Enzyme extraction

Fresh plant tissue (100 g) was homogenized with a food blender in 500 ml of 0.3 M succinate (Na+, pH 5.5) containing 10 mM CaCl2, 20 mM ascorbate and 3% w/v ‘PolyclarAT’ (polyvinylpolypyrrolidone). The homogenate was gently stirred on ice for 2.5 h, then filtered through Miracloth, and the filtrate was centrifuged at 4°C for 5 min at 25 000 g. The supernatant was aliquotted and stored at −80°C; MXE and XET activity were retained for > 2 yr.

Thin layer chromatography (TLC)

TLC was performed in freshly made butan-1-ol : acetic acid : water (2 : 1 : 1; BAW), with two to four ascents. Sugars were stained with thymol (Jork et al., 1994).

Analytical digestion of polysaccharides

Xyloglucan (< 50 μg) was digested for 90 min at 20°C in 20 μl of 0.05% XEG in PyAW (pyridine : acetic acid : water, 1 : 1 : 98 by vol., containing 0.5% chlorobutanol). MLG (< 50 μg) was digested for 90 min at 20°C with 0.1 U of lichenase in 20 μl PyAW. Putative [3H]GnXXXGol (0.5 kBq) was incubated in 20 μl PyAW with 0.125 U ml−1 β-glucosidase. In each case, digestion was stopped by the addition of 0.2–0.5 volumes of formic acid.

Assay of extracted MXE and XET activity in vitro

Endotransglucosylase activities were assayed by radiochemical methods (Fry et al., 1992, 2008b). In brief, extracted enzymes were incubated in 30 μl containing [3H]XXXGol (1 kBq) and 0.32% (w/v) tamarind xyloglucan or barley MLG. Reactions were stopped by the addition of 0.33 volumes of formic acid, and the products were dried onto Whatman 3MM paper (purchased from VWR). The papers were washed overnight in running tap water and bound 3H was assayed.

Preparation of authentic endotransglucosylase products in vitro

For the preparation of authentic radiolabelled MXE products, Equisetum extract (100 μl) was mixed with 100 μl of water containing 100 kBq [3H]XXXGol and 0.5 mg barley MLG. For XET products, Holcus lanatus extract was treated likewise, but with tamarind xyloglucan instead of MLG. After 16 h at 20°C, both reactions were stopped by the addition of 100 μl of formic acid. High-Mr products were precipitated with 14 ml of 75% ethanol containing 1.4% ammonium formate and collected by centrifugation (1700 g for 15 min). The pellet was washed free of 3H-oligosaccharides with 75% ethanol, dried and redissolved in 0.5 ml of water. The MXE and XET products were digested with lichenase and XEG, respectively. Digestion products were subjected to preparative TLC and localized by fluorography (Fry, 2000). The TLC plate was washed free of 2,5-diphenyloxazole (PPO) with toluene and dried, and the major 3H-labelled oligosaccharide was eluted with water.

Initiation of Equisetum callus cultures

Equisetum fluviatile shoots were dipped in 75% ethanol for 5 min, and surface dried. Longitudinal sections from the shoots were placed on solid medium containing Murashige and Skoog basal salts (Sigma M5519), 2% glucose, 2 mg l−1 2,4-dichlorophenoxyacetic acid, pH 5.7, and 1% agar in a Petri dish sealed with Parafilm. Incubation was routinely at 20°C in the dark. Initiated callus was subcultured to fresh medium every month. In some experiments, the medium was supplemented with 1 mM sodium silicate.

Detection of MXE and XET action in vivo

Cross-sections (c. 0.5 mm thick) of Equisetum stems and various barley organs (Supporting Information Fig. S1) were cut by hand with a razor blade. Slices (total 50 mg) were immediately suspended in 250 μl of water containing 0.1 MBq ml−1 [3H]XXXGol or [3H]XXLGol (c. 1 μM) in a 1.5-ml tube, and incubated for 16 h at 20°C. AIR was then prepared, incubated in 1.35 ml of 6 M NaOH at 37°C for 16 h, and centrifuged. The extracted hemicellulose was pooled with two 1-ml water washes of the pellet, slightly acidified with acetic acid and dialysed against water. Aliquots of the dialysate were subjected to digestion with either XEG or lichenase, or left undigested. Reaction products were analysed by TLC either immediately (for Fig. 3) or after the removal of undigested polysaccharides by precipitation with 72% ethanol (for Figs 4, 5).

Detection of radioactivity

Radioactive spots on TLC plates were localized by fluorography (Fry, 2000) on pre-flashed Kodak BioMax MR film. For the quantification of radioactive spots, lanes were cut into segments (5 mm long in regions of interest; up to 50 mm over the remainder of the TLC plate), which were assayed for 3H by scintillation counting in ScintiSafe3 (Fisher Scientific). Background-corrected counts are reported (in some cases, as an accumulation of all readings that constituted a peak). Radioactive products of in-vitro endotransglucosylase assays, on paper squares, were scintillation counted in ScintiSafe3.


Characterization of a diagnostic MXE product

Our method for the detection of MXE action in freshly excised tissue was to supply tracer levels of a 3H-labelled xyloglucan oligosaccharide, such that it would permeate the apoplast and serve as an acceptor substrate wherever endogenous MXE enzyme was acting on endogenous MLG polysaccharide. However, the exogenous oligosaccharides could theoretically be used indiscriminately by any endotransglycosylases able to use them as acceptor (two known examples are MXE and XET). To develop a reliable means of detecting MXE products formed in vivo, we devised a method of distinguishing MXE and XET products formed in model experiments in vitro. The radiolabelled polymeric products of MXE and XET action would be MLG–[3H]XXXGol and xyloglucan–[3H]XXXGol, respectively (where ‘–’ is a glycosidic bond), which we predict would differ in being digestible to low-Mr radioactive products by lichenase and XEG, respectively (Fig. 1).

To validate this strategy, we tested the ability of lichenase and XEG to selectively hydrolyse in-vitro-generated MXE and XET reaction products. Radioactive products of MXE or XET were first made in vitro by the action of an Equisetum extract on [3H]XXXGol plus barley MLG or a Holcus lanatus extract on [3H]XXXGol plus tamarind xyloglucan, respectively. Aliquots of the high-Mr products were then digested separately with lichenase and XEG (Fig. 2a). The MXE-generated polymeric product was completely resistant to XEG, but was digested by lichenase to a radioactive oligosaccharide larger than the acceptor substrate alone. The product was presumed to be an oligoglucosyl chain (a fragment of MLG) attached to [3H]XXXGol, hence Glcn•XXXGol (Hrmova et al., 2007). As predicted, the product of in-vitro XET action was susceptible to XEG (which generated products similar or identical to the starting [3H]XXXGol), but not to lichenase digestion (Fig. 2a).

Figure 2.

Characterization of products formed in vitro by Equisetum mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) activities. (a) [3H]XXXGol was incubated either with Equisetum fluviatile extract plus barley (Hordeum vulgare) mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) or with Holcus lanatus (Yorkshire fog grass) extract plus tamarind xyloglucan, generating radioactive MXE and XET products, respectively. The polymeric products were then analytically digested with lichenase (LIC) or xyloglucan endoglucanase (XEG) and the new products were separated by TLC and fluorographed. The ‘marker’ lane shows the [3H]XXXGol preparation used in this experiment (containing trace impurities of closely related oligosaccharides, probably including [3H]XXX-mannitol). Lichenase digestion of the MXE product results in a larger oligosaccharide, Glcn•XXXGol (arrow). (b) To determine the size of the oligoglucosyl tail of Glcn•XXXGol, we further digested it with β-glucosidase for 0–256 min. Products were analysed as in (a).

The structure of the lichenase-generated Glcn•XXXGol was further investigated by graded digestion with β-d-glucosidase (Fig. 2b). A clear progression from Glcn•XXXGol through a single intermediate to a product which migrates with [3H]XXXGol was observed with increasing incubation times. This indicates the presence of two β-d-glucose residues attached to [3H]XXXGol (Glc2•XXXGol), a result consistent with the source being ‘MLG–[3H]XXXGol’ created by MXE.

Detection of the diagnostic MXE product in vivo

Freshly cut pieces of E. fluviatile mature stem, barley seedling stem (plus sheath) and barley seedling leaf lamina were incubated in aqueous [3H]XXXGol; the endogenous hemicelluloses were then extracted and digested with lichenase or XEG (or left undigested), and any radioactive digestion products were analysed by TLC (Fig. 3). By this method, both MXE and XET actions were concurrently assayed under natural conditions, without potential shortfalls of an in-vitro assay, such as glycosylation differences (for proteins produced in heterologous systems) or artificially buffered pH. In addition, MXE and XET are directly comparable within an individual sample, eliminating biological variability, because both actions are assayed simultaneously.

Figure 3.

Products of mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) action formed by Equisetum fluviatile and barley (Hordeum vulgare) in vivo. Freshly cut explants (three replicate samples of each organ) were incubated with [3H]XXXGol for 16 h. Hemicelluloses were then extracted and digested with lichenase (LIC) or xyloglucan endoglucanase (XEG), or left undigested (None). Products (plus undigested hemicelluloses) were resolved by TLC and detected by fluorography. Markers: lanes 1–3 are 1000, 100 and 10 Bq of [3H]XXXGol; lane 4 is Glc2•[3H]XXXGol produced as in Fig. 2.

In Equisetum stems (Fig. 3a), the amount of MXE diagnostic product (mobile after lichenase digestion; Glc2•[3H]XXXGol) exceeded the amount of XET product (mobile after XEG digestion; [3H]XXXGol) in all three replicate samples.

In barley stems (Fig. 3b), no radioactive polysaccharide was detectably digested by lichenase, indicating that MLG was not appreciably acting as a donor substrate. A portion of the in-vivo product from barley stems was hydrolysed by XEG, indicating XET action, but the majority remained at the origin. This immobile material was tested by digestion with various glycanases (lichenase, endo-β-mannanase, β-xylanase, endopolygalacturonase); however, none gave TLC-mobile, 3H-labelled products (data not shown), and therefore we were unable to provide evidence for novel endotransglycosylases that use XXXGol as acceptor and a non-xyloglucan polysaccharide (e.g. mannan) as donor.

Little of the radioactive substrate was incorporated into barley leaf lamina hemicellulose (Fig. 3c). No [3H]XXXGol was detectable directly after digestion with XEG, but XEG did diminish the radioactivity that remained at the origin of the TLC plate, which implies that the high-Mr product originates from XET action, as in the case of barley stem + sheath. Again, the lack of effect of lichenase shows that no product of barley MXE action was detectable.

To examine whether endotransglucosylase action in barley depends on the endogenous xyloglucan and MLG content of different barley organs, we repeated the above experiment by incubating 13 different organ samples with [3H]XXLGol. The endogenous hemicelluloses were then digested with XEG or lichenase and low-Mr products (only) were analysed on duplicate TLCs: one was cut and assayed for radioactivity (Fig. 4); the other was stained with thymol to show the total oligosaccharides produced (Fig. 5).

Figure 4.

A search for products indicating mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) action in living barley (Hordeum vulgare) tissues. Tissue slices were incubated with [3H]XXLGol. The 3H-labelled hemicellulose formed was then digested with xyloglucan endoglucanase (XEG) or lichenase and the low-Mr products (only) were analysed by TLC, revealing the 3H-oligosaccharide products expected from any XET or MXE action, respectively. Four representative profiles are shown (a–d), as is the profile of the authentic [3H]XXLGol (e). (f), Marker mixture of malto-oligosaccharides. The oligosaccharide zone, defined as in (a), was quantified for each tissue tested (g). All data are corrected for background and are given ± counting error (95% confidence limits).

Figure 5.

Presence of endogenous mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) and xyloglucan in barley (Hordeum vulgare) tissues. Duplicate TLCs of the samples illustrated in Fig. 4, digested with xyloglucan endoglucanase (‘X’) or lichenase (‘L’), were stained with thymol, revealing oligosaccharides originating from endogenous xyloglucan and MLG. M, Marker mixture of malto-oligosaccharides.

Lichenase released stainable concentrations of oligosaccharides from all barley organs tested, indicating the presence of MLG (Fig. 5). The yield was highest in leaves and roots of seedlings, but was also moderately abundant in seedling coleoptiles and in the young leaves of 12-wk-old plants. It was lowest, although still detectable, in old leaves of 12-wk-old plants. The major MLG oligosaccharide detected in all barley organs tested was the trisaccharide, with a smaller proportion of tetrasaccharide, which contrasts with Equisetum MLG, where the tetrasaccharide predominates (Fry et al., 2008a; Xue & Fry, 2012). XEG released a range of oligosaccharides from all barley organs tested, indicating the ubiquitous presence of xyloglucan. The pattern of oligosaccharides produced differed strongly from that obtained by a similar methodology with nonpoalean plants (Xue & Fry, 2012), confirming that poalean xyloglucan has a rather unusual composition (Sims et al., 2000).

XEG released 3H-labelled oligosaccharides from the hemicelluloses of all barley organs tested, but significantly more from younger organs, indicating that, although XET action is ubiquitous throughout the plant, it is more prevalent in younger tissues (Fig. 4). Lichenase, however, was unable to release appreciable amounts of 3H-labelled oligosaccharides, showing that the corresponding MXE action was negligible (Fig. 4).

Testing for MXE and its substrates in Equisetum callus

We were interested in the possibility that Equisetum callus might be a convenient model system in which to study the in-vivo action of MXE and other endotransglucosylases. To our knowledge, no callus cultures of any Equisetum species have been initiated or maintained. Callus cultures can be a very convenient system for the study of in-vivo metabolism. In addition, cells with purely primary walls can be difficult to isolate from whole plants, but abundantly collected from callus. Callus cells are commonly used in the study of cell wall structure and metabolism. For these reasons, we initiated several calli from E. fluviatile stems and screened them for hemicellulose content and extractable endotransglucosylase activity. In view of suggestions that MLG plays an important role in silica metabolism (Fry et al., 2008a; Law & Exley, 2011), we grew the calli in media with and without silicate.

The MLG and xyloglucan contents of AIR from four independently initiated calli were determined by lichenase and XEG digestion (Fig. 6). For comparison, AIR from E. fluviatile mature stems was also digested. Lichenase digestion of stems produced diagnostic MLG oligosaccharides which co-migrated with barley MLG oligosaccharides; however, no such oligosaccharides were produced from Equisetum callus AIR. In contrast, cell cultures of Zea mays (Poales) show clear evidence for MLG by the same methods (Fry et al., 2008a). XEG digestion produced many xyloglucan oligosaccharides from all samples, although these were more prevalent in callus than stem samples.

Figure 6.

Hemicellulose composition of Equisetum calli and stems. Four independently initiated callus lines were maintained on medium either lacking silicate (B, E, F) or containing 1 mM silicate (Bs, Cs, Es, Fs). Alcohol-insoluble residue (AIR) samples from the calli were digested with lichenase (a) or xyloglucan endoglucanase (XEG) (b), and the oligosaccharides produced were resolved by TLC. For comparison, AIR from intact Equisetum stem was analysed in the same way. Markers are oligosaccharides from lichenase-digested barley mixed-linkage (1→3),(1→4)-β-d-glucan (MLG) (bMLGO) and XEG-digested tamarind xyloglucan (tXGO). Individual bMLGOs are labelled: MLG2, laminaribiose; MLG3, 3-O-cellobiosyl-glucose (G4G3G); MLG4, 3-O-cellotriosyl-glucose (G4G4G3G); etc.

Enzymes were extracted from seven calli and assayed radiochemically for endotransglucosylase activities. All calli contained extractable XET activity, but little or no MXE activity (Fig. 7). Only culture ‘Bs’ had a trace of extractable MXE activity significantly above the donor-free control assay.

Figure 7.

Extractable endotransglucosylase activities from Equisetum calli. Callus cultures were grown as in Fig. 6, and extracted enzymes were assayed radiochemically in vitro. The assay mixture (final volume, 30 μl) contained 0.33 μM [3H]XXXGol (1 kBq), 0 or 0.32% donor substrate, and the protein extractable from 6 mg fresh weight of tissue. Reaction mixtures were stopped at seven time points (0–2 h), which gave a linear time course. Results are displayed as radioactive product (cpm) formed per minute of incubation (± SE of the best-fit curve, = 7). A plotted rate of 1 cpm min−1 corresponds to the incorporation of 0.8% (≈ 120 cpm) of the [3H]XXXGol into the polymeric product in 2 h. MXE, mixed-linkage glucan:xyloglucan endotransglucosylase; XET, xyloglucan endotransglucosylase; ‘Bkgd’ is the rate with no deliberately added donor substrate. *, rate is statistically above the background rate (< 0.05).

Quantification of Equisetum in-vivo products

We quantified the in-vivo procedure to examine the profile of MXE and XET action in tissue types that represent the range of developmental stages of Equisetum cells. The product diagnostic of XET action was found in meristematic tissue and callus cultures, but no MXE product was formed in these samples (Table 1). In tissues in which MXE action was present, XET products were also found. However, in more mature tissues, MXE action exceeded XET action. The greatest ratio of MXE : XET action was in the oldest tissue studied – brown, over-wintered leaves. As a general trend, the ratio of MXE : XET action increased from the youngest to the oldest tissues.

Table 1. Mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) action in various Equisetum fluviatile tissues in vivoa
StageTissueYield (% of total 3H-labelled hemicellulose) of product diagnostic ofMXE : XET action ratio
XET actionMXE action
  1. a

    Tissue slices (or pieces of whole callus) were incubated with [3H]XXXGol. The 3H-labelled hemicellulose formed was then digested and analysed as in Fig. 3 to yield products diagnostic of prior MXE or XET action. The diagnostic products were: for XET, [3H]XXXGol after xyloglucan endoglucanase (XEG) digestion; for MXE, Glc2•[3H]XXXGol after lichenase digestion. After a mock digestion in buffer without enzyme, essentially all the radioactivity remained immobile on TLC (data not shown). Data are given ± SE (= 3).

Callus91 ± 22 ± 10.02
YoungestIntercalary meristem89 ± 22 ± 10.02
Immature leaf (green)62 ± 641 ± 50.66
Mature stem (green)29 ± 464 ± 22.2
Over-wintered stem (brown)26 ± 360 ± 32.3
Most matureOver-wintered leaf (brown)15 ± 279 ± 35.3

MXE and XET activity in extracts from Equisetum reproductive tissue

Other tissues not previously studied for MXE activity include the strobilus (spore-bearing ‘cone’) and its specialized stalk, which, in E. arvense (unlike E. fluviatile), is achlorophyllous and very short lived. Total extracts of these E. arvense organs exhibited XET activity, but no appreciable MXE activity (Fig. 8). In contrast, vegetative lateral shoot extracts of E. arvense contained high MXE activity and somewhat less XET activity.

Figure 8.

Extractable endotransglucosylase activities from Equisetum arvense vegetative and reproductive tissues. Total extracts from mature vegetative shoot, strobilus and strobilus stalk were assayed for mixed-linkage glucan:xyloglucan endotransglucosylase (MXE) and xyloglucan endotransglucosylase (XET) activities essentially as in Fig. 7, but with four time points (0–4 h).


The significance of assaying endotransglycosylase action

Enzyme ‘activity’ (as measured in vitro, under optimized conditions, after extraction of the enzyme from the organism) should be clearly distinguished from enzyme ‘action’ (Fry, 2004). An active enzyme may fail to act (either in vivo or in vitro) for any of numerous reasons. For example, if the enzyme never comes into contact with its potential substrates or is in some way inhibited, action will not occur. Conversely, a lack of observed enzyme activity in a cell extract does not prove a lack of action in vivo; enzymes may be inextractable or inactivated during extraction.

Because of this, assaying enzyme action in vivo rather than enzyme activity in vitro provides a way of defining the physiological roles of enzymes, as well as a way of assessing the presence of enzymes that may not otherwise be detectable. The in-vivo method described here is a strategy for the assessment of endogenous endotransglycosylase action specifically on endogenous hemicellulose donor substrates; only the acceptor substrate (3H-labelled oligosaccharide) was exogenous. Thus, in the case of MXE, the data show that the endogenous enzyme can act on its endogenous donor substrate (MLG), which would result in cleavage of the MLG chain. Whether the transglycosylation reaction can then be consummated using an endogenous acceptor substrate remains to be proven, although in dual-labelling experiments on XET action, all three participants (enzyme, donor and acceptor) were endogenous, demonstrating the complete XET transglycosylation action in vivo (Thompson et al., 1997; Thompson & Fry, 2001).

In-vivo assays show MXE action in Equisetum stems, but not in barley

The in-vivo method was able to detect the in-situ action of both previously described extractable activities in Equisetum – MXE and XET. The amounts of in-vivo product indicative of the action of these two activities correlate, relative to each other, with the extractable activities reported in previous studies (Fry et al., 2008b). However, no evidence of appreciable MXE action in any barley organ studied was found. Mature barley leaf laminae and sheaths, as well as seedling organs, were tested as potential sources of MXE-expressing tissues.

Thus, MXE action is readily detectable in Equisetum vegetative shoots, but is undetectable in barley tissues. As no enzyme extraction process had been undertaken and the cells were largely intact, we were effectively testing for the presence of inextractable poalean MXE. A comparison of MXE action in barley and Equisetum provides evidence against the hypothesis of Hrmova et al. (2007) that the very low MXE activities of barley XTHs (e.g. HvXET6), detectable in vitro, could exert significant MXE action in vivo. We cannot discount the possibility that extremely low levels of MXE action occur, similar to the rates of in-vitro activity observed with HvXET6 (Hrmova et al., 2007) or with total extracts of the grass Holcus lanatus (Fry et al., 2008b), but the physiological significance of this is put into doubt by the low rates observed. A major advantage of the in-vivo approach is that a hypothesized action (in our case, poalean MXE) can be tested in a single experiment, circumventing the need to assay individually the proteins encoded by each of the 30 barley XTH genes (Eklöf & Brumer, 2010), as well as candidates from other CAZy families (Cantarel et al., 2009).

Any in-vivo approach (or in-vitro approach using unpurified enzymes) potentially risks being compromised by hydrolysis of the added substrate. For example, many plant tissues are rich in apoplastic α-xylosidase and β-glucosidase activities (Koyama et al., 1983; Franková & Fry, 2011) which, together, might partially hydrolyse our exogenous substrates, [3H]XXXGol or [3H]XXLGol, to products too small to serve as acceptor substrates of endotransglucosylases. However, any such hydrolysis was clearly not responsible for the lack of detectable in-situ MXE action in barley tissues and in the meristems and callus of Equisetum, or inextractable MXE activity in Equisetum callus and strobilus, because XET action and/or activity was readily detected in the same experiments by the use of the same acceptor substrate.

The majority of barley stem endotransglucosylase product was indigestible with XEG in the time allotted. This could be explained in two ways. First, the 3H-labelled hemicellulose formed may have been a standard XET product which, however, became incompletely soluble and thus partly indigestible after the drying step. Poalean xyloglucan has fewer xylosyl substitutions than tamarind or Equisetum xyloglucan, some of the xylose being replaced by O-acetyl groups (Gibeaut et al., 2005), which will have been lost during the NaOH extraction procedure, leaving an under-substituted and less water-soluble xyloglucan in the case of barley.

Alternatively, the TLC-immobile material could be the product of a novel endotransglycosylase which does not use xyloglucan as a donor and whose product would therefore not be susceptible to XEG digestion. As we were unable to produce TLC-mobile oligosaccharides by digestion with various hydrolases (lichenase, endo-β-mannanase, β-xylanase and endopolygalacturonase), we are unable to provide positive evidence for hypothetical endotransglycosylases that use xyloglucan as an acceptor and different polysaccharides as donors.

The lack of MXE action in barley tissues is not a result of the absence of MLG

It could be suggested that the lack of observable MXE action was, in some barley tissues, caused by the absence of sufficient endogenous MLG to serve as donor substrate. However, we found that MLG was detectable in all barley organ samples tested (Fig. 5). In most cases, lichenase-releasable MLG-oligosaccharides exceeded XEG-released xyloglucan-oligosaccharides, indicating that the barley cell walls contained more MLG than xyloglucan. Nevertheless, in-situ XET action was readily detectable in all barley organs tested. Therefore, the absence of appreciable MXE action in barley was not caused by a lack of the appropriate donor substrate.

Equisetum tissue maturity correlates with MXE activity, MXE action and MLG content

Fry et al. (2008b) reported that extracts of mature vegetative Equisetum stems contain a higher MXE : XET activity ratio than extracts of young tissue. This led to the proposal that MXE plays a role in the cessation of cell expansion or in stem strengthening. The unique presence of MXE in Equisetum may compensate for its low lignin content (Siegel, 1962). Our results on MXE action in Equisetum mirror those on extractable MXE activity. The data shown here support the hypothesis that MXE plays a role in strengthening the walls of mature cells, as both MXE action and activity were largely absent in meristematic tissue, and MXE action appeared in maturing tissues and increased in relation to XET action in the oldest tissues studied. Equisetum callus, another ‘immature’ tissue with no obvious wall strengthening, also had a very low (or zero) MXE : XET ratio for both activity and action. Additional Equisetum material tested for the first time in the present work was the E. arvense strobilus and its stalk, both of which, like callus, showed very low or zero extractable MXE : XET activity ratios. The strobilus itself consists of tender tissue packed with spores. The strobilus stalk of E. arvense grows very rapidly in April, before the vegetative shoots appear in May, and is a short-lived, weak, sappy structure. Thus, throughout our work, we support the notion that XET activity predominates in young tender growing tissues, whereas MXE activity predominates in ageing, tough, vegetative stems.

We also show that this correlation of MXE action and activity with age occurs concomitantly with a change in the polysaccharide content of Equisetum cell walls. Specifically, and significantly in the context of this work, we show that both of the MXE substrates (MLG and xyloglucan) differ dramatically in their presence at different developmental stages, MLG being absent in the most ‘immature’ cells tested (callus culture) and highly prevalent in older tissues, whereas xyloglucan content is higher in young tissues.

If MLG is abundant in secondary walls (Sørensen et al., 2008), whereas xyloglucan is mainly located in primary walls, it is possible that MXE provides a mechanism for stitching the secondary wall to the primary wall and thus strengthening the construction of the stem. There is a suggestion (Mellerowicz et al., 2008) that XET action serves a similar role in the wood of seed plants, as some xyloglucan occurs in both the primary and secondary wall layers of xylem.

Implications for the roles of poalean XTHs, xyloglucan and MLG

The lack of detectable MXE action in barley suggests that the role played by MLG in the cell wall of Equisetum and cereals is, at least in some crucial respects, different, a conclusion initially indicated by the different spatiotemporal localization of the MLGs in the cell walls of these two plant groups. Poalean MLG is often synthesized at the onset of cellular elongation, and partially hydrolysed at the cessation of growth (Buckeridge et al., 2004), although small amounts of MLG persist in nonlignified and lignified cell walls of mature barley leaves (Trethewey et al., 2005; Fig. 5 this study). In contrast, Equisetum MLG is abundant in mature tissues, including their secondary cell walls, and, unlike in barley, increases slightly with tissue age (Fry et al., 2008a; Sørensen et al., 2008). Further, the polysaccharide environment in which MLG is found in the two taxa differs significantly. Poalean monocots synthesize proportionally larger quantities of (glucuronoarabino-)xylan (Carpita & Gibeaut, 1993), whereas Equisetum cell walls are richer in mannose-containing polysaccharides (Popper & Fry, 2004; Nothnagel & Nothnagel, 2007; Sørensen et al., 2008).

In addition, the observed lack of MXE action in barley calls for continued investigation into the role of poalean XTHs. The cell walls of the Poales are different from those of other angiosperms in that they contain significantly less (and qualitatively different) xyloglucan. However, the evolution of this unique wall architecture appears to have proceeded without a corresponding decrease in the number of poalean XTH genes. The requirement to preserve a multiplicity of XTH genes in the commelinids may be related to the likelihood that each gene has a specific time and place of expression (Nishitani, 2005). In addition, enzyme extracts from young grass and cereal tissues often have higher total XET activity than those from young dicot tissues (Fry et al., 1992). The high XET activity in poalean walls has led to the suggestion that they mainly act on the other, more prevalent, polysaccharides than xyloglucan (Yokoyama et al., 2004; Fincher, 2009). The discovery of MXE in Equisetum made such an activity in the Poales a more likely hypothesis, as some poalean XTHs might have turned out to possess high MXE activity. However, our discovery that no MXE action occurs in barley is consistent with previous reports of negligible extractable MXE activity from poalean sources (Fry et al., 2008b), and refutes this hypothesis. Nevertheless, our work does not rule out the possibility of other novel endotransglycosylase activities/actions in barley.

In conclusion, this work adds further support to the hypothesized strengthening role of MXE in Equisetum, but counters recent hypotheses about the relevance of MXE in the Poales. Specifically, the hypothesized roles of poalean XTHs as MXE-active enzymes, and of poalean MLG as a substrate for endogenous MXE action, are negated by the present work.


K.E.M. thanks the Darwin Trust of Edinburgh for a stipend; T.J.S. and S.C.F. thank the UK Biotechnology and Biological Sciences Research Council for a studentship and research funding, respectively.