Widespread Occurrence of a Covalent Linkage Between Xyloglucan and Acidic Polysaccharides in Suspension-cultured Angiosperm Cells

Abstract

INTRODUCTION

Xyloglucan and pectins are quantitatively and dynamically important primary cell wall (PCW) polymers. Xyloglucan, a neutral hemicellulose, has a widespread occurrence in land plant PCWs (Popper and Fry, 2003), typically contributing 20–25 % of the dry weight of the dicot PCW (Keegstra et al., 1973) and 2–5 % of the dry weight of the monocot PCW (Fry, 1989a; Hayashi, 1989). It is thought to function in cell enlargement by serving as a substrate for xyloglucan endotransglucosylase (XET) activity, which is able to cut and rejoin intermicrofibrillar xyloglucan chains (Fry et al., 1992; Nishitani and Tominaga, 1992; Thompson et al., 1997; Thompson and Fry, 2001). Pectins are acidic polysaccharides, rich in α-d-galacturonate (GalA) residues. There are three major pectic domains: homogalacturonan and the rhamnogalacturonans RG-I and RG-II (O'Neill et al., 1990; Ridley et al., 2001). Pectins constitute about 35 % of the dry weight of dicot PCWs (Fry, 2000) and are major targets for a variety of hydrolytic enzymes both endogenous and pathogen/herbivore-derived [examples include endopolygalacturonase (EPG), pectin methyl esterase, rhamnogalacturonase and arabinanase]. A covalent linkage between xyloglucan and pectin would be likely to make a major contribution to cell wall structure and metabolism. A better understanding of the occurrence of this linkage is therefore necessary for attempts to model the PCW and to explain and manipulate plant growth.

The occurrence of covalent xyloglucan–pectin linkages within the dicot PCW was first proposed in the PCW-model of Albersheim and co-workers (Keegstra et al., 1973). They proposed that the reducing terminus of xyloglucan is glycosidically linked to what would today be termed an RG-I side-chain. Evidence for these linkages was centred on the products released after EPG digestion of Acer PCWs. EPG solubilized approx. 50 % of the total PCW ‘pectin’ (as defined by the presence of uronic acid residues) and approx. 10 % of the ‘xyloglucan’ [as defined by the presence of xylose (Xyl) residues]; a large proportion (approx. 70 %) of this Xyl-containing polysaccharide bound to the anion-exchanger, DEAE-Sephadex, and co-eluted with ‘pectin’, suggesting a bond between xyloglucan (which is neutral) and an acidic pectin (Talmadge et al., 1973). EPG-digested PCWs released a further proportion of the wall (approx. 16 %) on extraction with 8 m urea or with 0·5 m NaOH containing 100 mm NaBH₄, or on treatment with endo-β-glucanase. The solubilized wall material was shown to contain sugar residues thought to be characteristic of xyloglucan and pectins (Bauer et al., 1973). In the 1973 model (Keegstra et al., 1973) Xyl residues were not distinguished from 2-O-methylxylose and it was assumed that all Xyl was in xyloglucan and all uronic acid residues were in pectin. Later research showed that Xyl (and/or 2-O-methylxylose) and uronic acid residues could be derived from a single species of polymer [such as glucuronoarabinoxylan (GAX) or RG-II] without necessarily implying the existence of a xyloglucan–RG-I complex (Spellman et al., 1983; Prade et al., 1999). Many PCW models now place xyloglucan and RG-I in independent networks (McCann and Roberts, 1991).

However, recent evidence supports the existence of xyloglucan–RG-I conjugates. Thompson and Fry (2000) found approx. 30 % of xyloglucan (extracted from non-lignified suspension-cultured rose cells by 6 m NaOH) to be covalently bonded to anionic material, as judged by high-voltage electrophoresis and by anion-exchange chromatography. The anion-associated xyloglucan did not lose its charge after treatment with 8 m urea, 6 m NaOH or proteinase. Femenia et al. (1999) reported the presence of ‘pectic-xylan–xyloglucan’ complexes in alkali extracts from lignified cauliflower stems. This is also compatible with xyloglucan–pectin covalent complexes, although the possibility that lignin was responsible for the cross-linking of these two polysaccharides cannot be excluded. Enzymic digestion of the complex gave evidence consistent with hemicellulose–pectin complexes. Endo-xylanase and EPG treatments each (separately) caused a decrease in the apparent M_r of the pectin, xyloglucan and xylan components. Abdel-Massih et al. (2003) extracted nascent pectin from particulate enzyme preparations (from etiolated pea shoots) incubated with UDP-[U-¹⁴C]galactose. The nascent [¹⁴C]pectin had a strong affinity for paper (a characteristic of xyloglucan; Fry et al., 1992). The paper-binding ability of the [¹⁴C]pectin was greatly reduced by treatment with endo-1,4-β-glucanase, an observation which is consistent with the presence of xyloglucan in the radioactive paper-binding complex. Further support for a linkage between xyloglucan and RG-I has recently been presented by Vidal et al. (2003) who, similarly to Thompson and Fry (2000), observed strong binding of xyloglucans to an anion-exchange resin and subsequent co-elution with highly acidic RGs under high salt conditions.

In the present paper we report evidence for the presence of alkali-stable xyloglucan–pectin linkages in non-lignified suspension-cultured cells of all angiosperm cultures tested, both dicots and monocots

MATERIALS AND METHODS

RESULTS

Evidence for anionic ³H-hemicelluloses

The 50 % ethanol-precipitated material (xyloglucan-enriched hemicellulose) was subjected to anion-exchange chromatography in the presence of 8 m urea. This led to the elution of only 7–16 % of the ³H in the neutral pool (fractions 1 and 2; eluted with 11 mm PyOAc, which is the same concentration as the loading buffer); in all species, the majority of the ³H was eluted as acidic material [fractions 6–13 (175–1400 mm PyOAc), 14 (2 m sodium acetate, pH 7·0) and 15 (1 m NaOH)] (Fig. 1). Less than 1 % of the total radioactivity was eluted in fractions 3–5.

Open in a separate window

Fig. 1.

Behaviour of (pentosyl-³H)-labelled, xyloglucan-enriched hemicellulose, from various cell-suspension cultures, during anion-exchange chromatography. ³H-Hemicellulose was extracted with 6 m NaOH, neutralized, and precipitated with 50 % ethanol; the insoluble material was subjected to anion-exchange chromatography on Q-Sepharose FastFlow with a gradient of acetate buffer.

The anionic ³H-hemicellulose from spinach and tomato cultures was found to be eluted in two peaks (Fig. 1): fraction 11 (1050 mm PyOAc) and fraction 14 (2 m sodium acetate, pH 7·0). However, anionic ³H-hemicellulose extracted from Arabidopsis differed in its elution pattern, with a greater proportion being eluted in the more anionic fraction 14. Anionic ³H-hemicellulose from maize was eluted in two major peaks, both of which were less anionic than in spinach, Arabidopsis and tomato. That from rose was eluted in one anionic peak. The ³H-elution patterns of xyloglucan-enriched hemicellulose were similar to those observed in a replicate experiment in which the cells were fed one-quarter of the amount of radioactivity (data not shown).

Evidence for anionic ³H-xyloglucans

The 50 %-ethanol-precipitable anionic ³H-polysaccharides shown in Fig. 1 could have included GAXs, RGs and arabinogalactan-proteins (AGPs) as well as xyloglucans. To distinguish specific hemicelluloses, we applied Driselase digestion.

A proportion of each appreciably radiolabelled fraction, from 1 (neutral) to 15 (increasingly anionic), was Driselase-digested. Digestion products were separated by paper chromatography in solvent systems 1, 2 and 3. Exact co-chromatography of a peak of ³H with a non-radioactive, stainable internal marker of isoprimeverose (in each of the three solvent systems) confirmed that all cultures tested contained anionic [³H]xyloglucan. This is illustrated for Arabidopsis (Fig. 2); the other five species tested gave similar results (data not shown).

Open in a separate window

Fig. 2.

Paper chromatography of Driselase-digestion products of the Q-Sepharose fractions of the (pentosyl-³H)-labelled, xyloglucan-enriched hemicellulose extracted from a 7-d-old Arabidopsis culture. The fractions (Fr.) were selected from those shown in Fig. 1 (Arabidopsis). The paper chromatogram was developed in solvent system 2 and the non-radioactive internal marker of authentic isoprimeverose (indicated by the dark line) was stained with dilute aniline hydrogen-phthalate. The chromatogram was then cut into strips, which were assayed for radioactivity. The figure shows only the relevant zone (4–9 cm from the origin); the solvent front typically ran approx. 45 cm past the origin in solvent system 2. R_F values of major Driselase digestion-products of this solvent system are: isoprimeverose, 0·195; xylobiose, 0·21; galacturonic acid, 0·23; galactose, 0·27; glucose, 0·285; mannose, 0·325; arabinose, 0·35; xylose, 0·375. Additional chromatograms, in solvent systems 1 and 3, confirmed the identity of isoprimeverose (data not included).

The data in Fig. 2 and many related experiments (not shown) were quantified so that the elution profile of [³H]xyloglucan could be determined (Fig. 3). Between 30 % (in Arabidopsis) and 55 % (in rose and spinach) of the [³H]isoprimeverose-yielding polymer (=[³H]xyloglucan) was eluted in the neutral fraction (Fig. 3). In a subsequent experiment, a value of 70 % was obtained for barley (data not shown).

Open in a separate window

Fig. 3.

Anion-exchange chromatography profiles of polymers containing individual (pentosyl-³H)-labelled residues. The Q-Sepharose fractions reported in Fig. 1 were Driselase-digested and the radioactive products (isoprimeverose, xylobiose, Xyl, Ara and Driselase-indigestible material) were separated by paper chromatography in solvent system 1 and assayed by scintillation-counting. Fractions 2–5 contained little radioactivity and were therefore not analysed.

The remaining 30–70 % of the [³H]xyloglucan was eluted in anionic fractions. The distribution of [³H]xyloglucan between the various anionic fractions differed between species, Arabidopsis [³H]xyloglucan being particularly highly anionic.

DISCUSSION

Xyloglucan is a neutral polysaccharide (Fry, 1989a; Hayashi, 1989), composed of a β-(1→4)-d-glucan backbone substituted with α-d-Xyl, β-d-Gal, and in some cases α-l-Fuc and/or α-l-Ara. Widely accepted PCW models propose at least two independent polymer networks: a cellulose–xyloglucan network held together by hydrogen-bonds, and a pectic network held together by Ca²⁺ bridges. In onion epidermal cell walls, both the pectic and the cellulose–xyloglucan networks have a preferred orientation (Chen et al., 1997), which is suggestive of communication between the two networks. A high-M_r complex of xyloglucan linked to another polysaccharide, in particular another abundant PCW polymer such as pectin, would be expected to play a more effective tethering role in PCW architecture (Fry, 1989b; Kerr and Fry, 2003) than lower-M_r, free xyloglucans. The widespread taxonomic distribution of the xyloglucan–pectin linkage, reported here, suggests its importance to angiosperm PCW structure and function. Elucidation of the chemical structure of the linkage and its mode of synthesis are likely to promote our understanding of, and in future enable manipulation of, important physical properties of PCWs.

The extracted anionic [³H]xyloglucan is unlikely to be ester-bonded to pectin. Very mild conditions of NaOH hydrolysis (1 m NaOH at 20 °C) break all detectable ester bonds within 24 h, whereas we extracted anionic [³H]xyloglucan with 6 m NaOH at 37 °C. This suggests the existence of a highly alkali-stable bond between xyloglucan and an acidic PCW polymer.

Kerr and Fry (2003) found that in cultured maize cells, the average M_r of pulse-labelled intraprotoplasmic [³H]xyloglucans increased up to 30 min after labelling but prior to secretion. The increase in average M_r of the ³H-polymers was too large to be entirely ascribed to on-going (NDP-sugar-dependent) chain elongation within the Golgi cisternae (which could theoretically account for at most a doubling of the average M_r of the [³H]xyloglucans; see fig. 10 of Kerr and Fry, 2003), so it is likely that the increase was due to the post-synthetic bonding of [³H]xyloglucans to each other or to other (non-radioactive) polymers. Their bonding to RGs could account for the widespread formation of ‘anionic xyloglucan’, reported here.

It has also been reported that nascent pectin extracted from pea stems is complexed with xyloglucan (Abdel-Massih et al., 2003), supporting this conclusion. In agreement with the observations of Abdel-Massih et al. (2003), the cellulose- (paper-) binding ability of hot-oxalate-extracted [GalA-¹⁴C]pectin, noted in the present work on Arabidopsis cells, indicates an unexpected ability of some pectins to bind to cellulose. Paper-binding is not a characteristic of pectins (particularly if the sample is not dried on to the paper). It is therefore more likely that the [GalA-¹⁴C]pectin, solubilized by hot oxalate, was attached to, and able to bind to paper via, a hemicellulose (such as xyloglucan) that does have strong paper-binding ability. We probably underestimated the proportion of pectin that was hemicellulose-linked in this way, since the extractant used (hot oxalate) would not efficiently break the hemicellulose–cellulose hydrogen-bonds that may have been present in muro. Some hemicelluloses (especially xylans and glucuronomannans) contain GlcA and/or MeGlcA residues, which would be radioactive in the [¹⁴C]GlcA-fed cells; such hemicelluloses would also show up as paper-binding ¹⁴C-polymers. However, it is unlikely that these hemicelluloses would have been extracted from the cell walls by hot oxalate. We verified that the ¹⁴C-residues present in the paper-binding polymers reported here were indeed GalA (diagnostic of pectins), not GlcA or MeGlcA. Thus, we conclude that some of the pectin in Arabidopsis PCWs was strongly (probably covalently) linked to a cellulose-binding hemicellulose (e.g. xyloglucan).

We investigated the existence of a xyloglucan–pectin linkage in a wide variety of angiosperm cell-suspension cultures chosen to represent the diversity within angiosperm PCW structure. The major hemicellulose in Poales (gramineous monocots) is GAX, not xyloglucan, and the PCWs of Poales (represented here by maize and barley) often contain only about 20 % of the xyloglucan present in non-gramineous monocot and dicot PCWs (Fry, 1989a; Hayashi, 1989). They also contain high concentrations of feruloylated GAXs (present but not feruloylated in dicots) and have lower galacturonan contents than dicot PCWs (Shibuya et al., 1983; Jarvis et al., 1988). Additionally, PCWs of plants within the Poales contain mixed-linkage glucan, a hemicellulose absent from all other angiosperm PCWs (Nevins et al., 1978; Smith and Harris, 1999; Popper and Fry, 2003). PCWs of the Centrospermae (represented by spinach) are rich in ferulic acid, which in these dicots forms esters with Ara and Gal residues of pectic polysaccharides (Fry, 1982; Micard et al., 1997). It is therefore differently linked from the ferulic acid found in gramineous monocots. Xyloglucans so far extracted from different sources show some differences. The majority of xyloglucans extracted from both gymnosperms and angiosperms are reported to have a backbone consisting of repeat units of XXXG, whereas xyloglucans extracted from solanaceous plants (represented by tomato) have two, instead of three, consecutive branched Glc residues, which alternate with two unsubstituted Glc residues (XXGG; Vincken et al., 1997). [For an explanation of the abbreviated nomenclature of xyloglucan oligosaccharides (XXXG etc.), see Fry et al. (1993).] Additionally, xyloglucans isolated from the Solanaceae differ from those of other dicots in the branching of their xylosyl substituents; XXXG and XXFG are absent (Vincken et al., 1997) and Xyl residues are 2-O-substituted predominantly with α-l-Ara and β-d-Gal (York et al., 1996). Xyloglucan extracted from Poales also differs in composition, containing less Xyl, Gal and Fuc than typical dicot xyloglucan (Hayashi, 1989; McDougall and Fry, 1994; Vincken et al., 1997).

It is interesting that despite the wide variability in xyloglucan structure within angiosperms we found that in all angiosperms tested a substantial proportion of the xyloglucan was attached to an acidic polymer. It is possible that the structure of the acidic polymer is highly variable. RG-I is known to vary in methylation and/or acetylation of the GalA residues in its backbone (Perrone et al., 2002) as well as in the degree of substitution with arabinan and galactan side-chains (McNeil et al., 1982; Lerouge et al., 1993). It seems remarkable that a linkage between xyloglucans and an acidic polysaccharide, probably RG-I, appears to be conserved among angiosperms. This suggests that the linkage is important to PCW integrity and function.

Xyloglucan is likely to have a major functional role in gramineous monocot PCWs since the xyloglucan-modifying enzyme XTH extracted from grass tissues has a high XET specific activity (Fry et al., 1992). This is especially significant since gramineous monocot PCWs are widely reported to contain very much less xyloglucan than do dicot PCWs (Bacic et al., 1988; Smith and Harris, 1995). However, maize xyloglucan has a much higher molecular weight than that of dicot xyloglucan (Kerr and Fry, 2003). This may partially account for the survival of gramineous monocots despite their low xyloglucan content. Anionic xyloglucan was readily detected in suspension-cultured cells of gramineous monocots (maize and barley) as well as in those of dicots.

In conclusion, we obtained evidence for anionic xyloglucan in all angiosperm PCWs investigated despite wide variation in their xyloglucan structures as well as differing overall PCW compositions. This indicates that the xyloglucan–pectin linkage is evolutionarily conserved and may be required for effective PCW structure and function.

Supplementary Material

Acknowledgments

LITERATURE CITED