SEARCH

SEARCH BY CITATION

Keywords:

  • glycosidases;
  • glycanases;
  • trans-β-xylanase;
  • trans-β-xylosidase;
  • trans-α-xylosidase;
  • primary cell walls

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Wall polysaccharide chemistry varies phylogenetically, suggesting a need for variation in wall enzymes. Although plants possess the genes for numerous putative enzymes acting on wall carbohydrates, the activities of the encoded proteins often remain conjectural. To explore phylogenetic differences in demonstrable enzyme activities, we extracted proteins from 57 rapidly growing plant organs with three extractants, and assayed their ability to act on six oligosaccharides ‘modelling’ selected cell-wall polysaccharides. Based on reaction products, we successfully distinguished exo- and endo-hydrolases and found high taxonomic variation in all hydrolases screened: β-d-xylosidase, endo-(1[RIGHTWARDS ARROW]4)-β-d-xylanase, β-d-mannosidase, endo-(1[RIGHTWARDS ARROW]4)-β-d-mannanase, α-d-xylosidase, β-d-galactosidase, α-l-arabinosidase and α-l-fucosidase. The results, as GHATAbase, a searchable compendium in Excel format, also provide a compilation for selecting rich sources of enzymes acting on wall carbohydrates. Four of the hydrolases were accompanied, sometimes exceeded, by transglycosylase activities, generating products larger than the substrate. For example, during β-xylosidase assays on (1[RIGHTWARDS ARROW]4)-β-d-xylohexaose (Xyl6), Marchantia, Selaginella and Equisetum extracts gave negligible free xylose but approximately equimolar Xyl5 and Xyl7, indicating trans-β-xylosidase activity, also found in onion, cereals, legumes and rape. The yield of Xyl9 often exceeded that of Xyl7–8, indicating that β-xylanase was accompanied by an endotransglycosylase activity, here called trans-β-xylanase, catalysing the reaction 2Xyl6[RIGHTWARDS ARROW] Xyl3 + Xyl9. Similar evidence also revealed trans-α-xylosidase, trans-α-arabinosidase and trans-α-arabinanase activities acting on xyloglucan oligosaccharides and (1[RIGHTWARDS ARROW]5)-α-l-arabino-oligosaccharides. In conclusion, diverse plants differ dramatically in extractable enzymes acting on wall carbohydrate, reflecting differences in wall polysaccharide composition. Besides glycosidase and glycanase activities, five new transglycosylase activities were detected. We propose that such activities function in the assembly and re-structuring of the wall matrix.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Growing plant tissues have primary cell walls composed of three major classes of polysaccharide: pectins, hemicelluloses and cellulose (Scheller and Ulvskov, 2010; Albersheim et al., 2011; Fry, 2011). This type of wall is unique to plants, and indispensable for their growth and development – especially for the mechanism of cell expansion, a dramatic feature of plant physiology not seen in animal cells or micro-organisms, during which the wall area increases enormously.

All stages of plant growth and development involve enzyme action in the cell wall. Important reactions considered likely to occur in the primary wall in vivo include the partial hydrolysis of polysaccharides by glycosidases and glycanases* (Labavitch, 1981; Fry, 1995, 2004; de la Torre et al., 2002) and the endotransglycosylation of polysaccharides, e.g. xyloglucan (Thompson and Fry, 2001; Nishitani, 1998; Mellerowicz et al., 2008; Eklof and Brumer, 2010), mixed-linkage (1[RIGHTWARDS ARROW]3,1[RIGHTWARDS ARROW]4)-β-glucan (MLG) (Fry et al., 2008a) and mannans (Schröder et al., 2009). Plant cell walls contain over 15 different glycosidases (β-glucosidase, β-galactosidase, β-xylosidase, α-xylosidase etc.) and over 10 glycanases (β-mannanase, β-xylanase etc.) (reviewed by Fry, 2004). The deconstruction/reconstruction of polysaccharides by these and other wall-localised enzymes probably governs many physiologically important processes, including: tightening and loosening the wall and thus controlling cell expansion by making and breaking bonds in the hemicellulose chains that tether microfibrils (Obel et al., 2007); resisting pathogens by hydrolysing fungal polysaccharides (Gonzalez-Teuber et al., 2010); limiting wall digestibility and increasing wall strength by making bonds between structural polysaccharides (Brett et al., 1999; Fry et al., 2008a; Mellerowicz et al., 2008); modifying tissue cohesion by making and breaking bonds in the pectic polysaccharides of the middle lamella (Lecam et al., 1994; González-Carranza et al., 2007; Zhang et al., 2007); modifying wall porosity by affecting the molecular-weight cut-off of the matrix (Carpita et al., 1979; Baron-Epel et al., 1988); generating intercellular signals (oligosaccharins) by endo-hydrolysis of polysaccharides (McDougall and Fry, 1991; Darvill et al., 1992; Aldington and Fry, 1993; Beňová-Kákošováet al., 2006); and destroying redundant oligosaccharins, when their biological message is no longer relevant (Baydoun and Fry, 1989; Darvill et al., 1992; García-Romera and Fry, 1995).

Plant genomes contain numerous putative ‘cell wall genes’ (Mao et al., 2009). For example, Arabidopsis thaliana has 730 open reading frames encoding putative glycosyltransferases and glycosylhydrolases (Henrissat et al., 2001), including 33 putative xyloglucan endotransglucosylase/hydrolases (XTHs) (Nishitani, 2005) and 69 putative homogalacturonanases (González-Carranza et al., 2007). Unfortunately, few of the putative homogalacturonanases, for example, have been tested for enzymic activity. Also, the strong possibility exists that additional hydrolase and transglycosylase activities are being overlooked because their genes are unidentified or incorrectly annotated. The recent discovery of a unique endotransglucosylase (MXE) activity in Equisetum (Fry et al., 2008a) emphasises the need to demonstrate experimentally the reactions catalysed by wall enzymes. A survey of the activity of enzymes acting on wall carbohydrates would thus complement the on-going exploration of the corresponding genes.

The restriction of appreciable MXE activity, among land plants, to the single genus Equisetum also emphasises the phylogenetic differences in wall metabolism between plant taxa. This view is reinforced by the emerging realisation that land plants exhibit major taxonomic differences in their wall polysaccharide chemistry: many of the landmark steps in plant evolution were accompanied by remarkable changes in the chemistry of the primary cell wall (Popper, 2008; Sørensen et al., 2010; Fry, 2011), which must have necessitated changes in the wall enzyme profile. For example: only the Poales and Equisetales possess MLG (Trethewey et al., 2005; Fry et al., 2008b); only land plants (not their sister group the Charophyta) have xyloglucan as a major wall component in the vegetative tissues (Popper and Fry, 2003; Domozych et al., 2009); the Poales and the Solanales have low-fucose xyloglucan (Carpita and Gibeaut, 1993; McDougall and Fry, 1994; Hoffman et al., 2005); mosses and liverworts uniquely have xyloglucans rich in β-d-galacturonic acid (Peña et al., 2008); Equisetum and Selaginella have xyloglucan containing α-l-arabinopyranose in place of many of the β-d-galactopyranose residues (Peña et al., 2008); Anthoceros, a hornwort, has polysaccharides uniquely rich in α-d-glucuronosyl-(1[RIGHTWARDS ARROW]3)-l-galactose units (Popper et al., 2003); lycopodiophytes (the earliest-diverging vascular plants) share an unusual abundance of 3-O-methyl-d-galactose residues (Popper et al., 2001); walls of charophytes, bryophytes and homosporous lycopodiophytes are rich in 3-O-methylrhamnose, which is undetectable in other land plants (Popper et al., 2004); and the eusporangiate[RIGHTWARDS ARROW]leptosporangiate transition among fern-allies was accompanied by a marked decrease in mannan content (Popper and Fry, 2004). These observations indicate a special role for cell wall composition in adapting plants to major new environments (e.g. the land) and lifestyles (e.g. vascularisation), and suggest a requirement for the evolution of corresponding differences in wall-located enzymes.

Here we report an extensive study of phylogenetic variation in wall polysaccharide-modifying enzymes. The results are available in the readily searchable GHATAbase (glycosylhydrolase and transglycosylase activity database; Data S2; Figures S1–S153 in Supporting Information), which will be a valuable resource for selecting plant organs from which to extract and study enzymes of interest. We found great variation in the occurrence of known enzyme activities between different plants. In addition, our results reveal the existence of several interesting transglycosylase activities, whose presence could not have been predicted by genomic approaches (Mao et al., 2009) and whose biological roles now invite exploration.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Glycosylhydrolase activity profile of 16 dialysed plant extracts

Extracts from 16 plant sources were freed of endogenous sugars by dialysis, as confirmed by thin-layer chromatography (TLC; Figure 2a), then tested for their ability to hydrolyse five oligosaccharide substrates to yield faster-migrating, i.e. smaller, products (e.g. Figure 2b–f; full results in files in Supporting Information). Each substrate tested models a key cell-wall polysaccharide, and each was partially hydrolysed by most or all of the extracts, indicating at least five distinct glycosylhydrolase activities.

image

Figure 2.  Activity of dialysed extracts from 16 plant species on five oligosaccharides. Plant extracts (made in buffer A, see Experimental Procedures) were dialysed, then incubated at 22°C for 24 h with the oligosaccharide indicated (b–f) or none (a). The TLCs of the products are shown [butan-1-ol/acetic acid/water (BAW); one ascent]. The right-hand lanes contain markers (M) and the relevant unhydrolysed substrate (T0). Extracts were from (1) broad bean root, (2) mung bean hypocotyl, (3) broad bean etiolated leaf, (4) cauliflower leaf, (5) parsley shoot, (6) asparagus sprout, (7) chicory leaf, (8) snowdrop leaf, (9) Equisetum September shoot, (10) Marchantia plant, (11) Selaginella plant, (12) onion etiolated leaf, (13) cress etiolated seedling, (14) lettuce cotyledon, (15) alfalfa etiolated seedling, (16) maize coleoptile. Arrows indicate starting material; markers related to the substrate are labelled in the same colour; bands marked ‘DPn’ are reaction products with the indicated degree of polymerisation.

Download figure to PowerPoint

β-Mannosidase and β-mannanase activity.  (1[RIGHTWARDS ARROW]4)-β-d-Mannohexaose (Man6) models the backbone of the mannan-based hemicelluloses (Schröder et al., 2009). The extracts varied markedly in Man6 hydrolysis (Figure 2b), activity being lowest in chicory, Marchantia and Equisetum and highest in alfalfa.

Most extracts produced progressively smaller oligosaccharides plus free mannose, as expected of β-mannosidase action. The time-course for alfalfa extracts (Figure 3a) illustrates this: sequential hydrolysis of the hexasaccharide gave Man5, Man4, Man3 and Man2, peaking at approximately 1, 2, 8 and >36 h respectively, each step being accompanied by increasing free mannose.

image

Figure 3.  Time-courses for enzyme action on oligosaccharides. (a) A dialysed extract from etiolated alfalfa seedlings acting on Man6. (b) Non-dialysed Marchantia extract acting on Xyl6. (c)–(e) Dialysed broad bean, Selaginella and oat extracts acting on Xyl6. (f) As (b) but with different sized xylo-oligosaccharides. Reactions were terminated at the indicated times by the addition of formic acid. Other details as in Figure 2.

Download figure to PowerPoint

With one exception, each 24-h profile (Figure 2b) closely resembles an alfalfa time-point (Figure 3a). For example, broad bean and cauliflower 24-h digests resemble 8-min and 8-h alfalfa digests respectively, indicating that cauliflower extracts have approximately 60 times higher β-mannosidase activity than broad bean.

Mung bean was the exception, giving a profile (unlike any in Figure 3a) indicative of mannanase (Figure 2b): despite the abundant undigested Man6 remaining after 24 h, the moderate yields of Man5, Man4, Man3 and Man2 approximately equalled each other, and little free mannose appeared. A β-mannosidase effecting only approximately 10% Man6[RIGHTWARDS ARROW] Man5 conversion would be likely, during the same interval, to achieve an approximately 10% Man5[RIGHTWARDS ARROW] Man4 conversion and thus a < 1% Man6[RIGHTWARDS ARROW] Man4 conversion with essentially no Man2–3. The relatively high yield of smaller oligosaccharides therefore suggests mannanase activity

  • image

and

  • image

(in addition to limited mannosidase activity producing Man5 + mannose).

α-Xylosidase activity on xyloglucan oligosaccharides.  Xyloglucan-derived heptasaccharide (XXXG; for abbreviated nomenclature, see Fry et al., 1993) was degraded stepwise (Figure 2c), each major step increasing the chromatographic mobility enough to suggest pairwise removal of sugar residues (heptasaccharide [RIGHTWARDS ARROW] pentasaccharide [RIGHTWARDS ARROW] trisaccharide). This will be because the terminal xylose residue of XXXG is first slowly removed by α-xylosidase to form GXXG (hexasaccharide), which never accumulates because its glucose residue is rapidly removed by β-glucosidase to yield XXG (pentasaccharide) (Koyama et al., 1983). Thus the XXXG [RIGHTWARDS ARROW] XXG conversion rate principally reports the occurrence of α-xylosidase.

Relatively high α-xylosidase was found in etiolated onion leaves, cauliflower leaves, parsley shoots, etiolated cress seedlings and maize coleoptiles, with low activity in the non-angiosperms.

β-Galactosidase activity on xyloglucan oligosaccharides.  Comparing the results for XXXG (Figure 2c) and the galactosylated xyloglucan oligosaccharide (XGO) mixture (Figure 2d) reveals β-galactosidase activity. Several extracts converted a xyloglucan nonasaccharide (XLLG) to an octasaccharide (XXLG; see later), but none converted this efficiently to heptasaccharide, indicating a strong preference for one of the two galactose residues in XLLG. Extracts with high XLLG-active β-galactosidase included broad bean, cauliflower and asparagus; non-seed plants had low activity.

β-Xylosidase and β-xylanase activity.  (1[RIGHTWARDS ARROW]4)-β-d-Xylohexaose (Xyl6), representing the β-xylan backbone, was digested to a homologous series (penta- to monosaccharide; Figure 2e). In many extracts (e.g. asparagus and onion), Xyl5 was by far the major product, indicating predominantly β-xylosidase activity. In others (Selaginella, Marchantia; possibly Equisetum and parsley), Xyl4, Xyl3 and Xyl2 exceeded Xyl5; this observation, coupled with abundant undigested Xyl6 and little free xylose, indicates xylanase, catalysing

  • image

and

  • image

α-Arabinosidase activity on (1[RIGHTWARDS ARROW]5)-oligoarabinans.  (1[RIGHTWARDS ARROW]5)-α-l-Arabinohexaose (Ara6) represents a structure from the neutral side-chains of pectic rhamnogalacturonan-I (Albersheim et al., 2011). About half the extracts digested this substrate; the major hydrolysis product was Ara5, indicating α-arabinosidase (Figure 2f). High activities were found in mung bean, asparagus and cress. Similar conclusions were reached with Ara8 as substrate (Figure S123).

Effect of inclusion of proteinase inhibitors

The plant extracts tested above were crude dialysates, and it might be suggested that some enzyme activities were lost by proteinase action. To test for this, we re-assayed selected extracts in the presence of a broad-spectrum proteinase inhibitor cocktail (PIC). The PIC sometimes slightly promoted the release of free mannose or xylose from Man6 or Xyl6, e.g. by broad bean leaves and asparagus sprouts, but not sufficiently to modify the observed pattern of intermediary oligosaccharides or disappearance of substrate (Figure S154). The PIC also promoted the formation of free fructose in chicory extracts, presumably by preventing proteolysis of a β-fructofuranosidase that acts on co-extracted fructan. However, the most prominent effect of PIC was to inhibit the hydrolysis of manno-oligosaccharides by some extracts (e.g. snowdrop, onion and cress; Figure S154a), and for this reason a proteinase inhibitor was not routinely used in our experiments.

Effect of extraction buffer and incubation temperature

Figure 2 shows digestions by low-salt-extracted enzyme, assayed at 22°C. To test the influence of these conditions, we repeated incubations with the same extracts at 40°C and also with high-salt extracts at 22 and 40°C. Patterns similar to those in Figure 2 were obtained with each permutation. For example, with Xyl6 the following trends were reproducible (Figure 4): most extracts (mung bean, cauliflower, asparagus, snowdrop, onion, cress, lettuce, alfalfa and maize) catalysed partial Xyl6[RIGHTWARDS ARROW] Xyl5 hydrolysis, with only traces of Xyl2–4, indicating moderate β-xylosidase activity; broad bean and parsley generated smaller oligosaccharides plus free xylose, indicating high β-xylosidase; and xylanase activity was found in Marchantia and Selaginella. Minor differences were, however, noted: for example, parsley exhibited clearer evidence for xylanase at 22°C than 40°C.

image

Figure 4.  Effect of extraction buffer and incubation temperature on the digestion of xylohexaose. Plant extracts (made in buffer A or B, see Experimental Procedures) were dialysed, then incubated with Xyl6 at 22°C (24 h) or 40°C (3.5 h). (a) High-salt extractant, 22°C. (b) Low-salt extractant, 40°C. (c) High-salt extractant, 40°C; compare low-salt extractant, 22°C (Figure 2e). Other details, including identity of plants, as in Figure 2.

Download figure to PowerPoint

Screening for enzyme distribution using undialysed extracts

Results exemplified by Figure 2 demonstrate extreme phylogenetic variability in wall-hydrolysing enzymes of dialysed plant extracts. To extend this survey, we streamlined the assays by using undialysed extracts. Omitting dialysis saves time; dialysis may also cause selective loss of some enzyme activities (Takeda and Fry, 2004). To validate this approach, we prepared 143 extracts from 51 diverse plant specimens, using three extractants, and assayed activities without further purification (Figure 5b–f and Supporting Information).

image

Figure 5.  Streamlined enzyme screening using non-dialysed extracts. Details as in Figure 2, but extracts were not dialysed. ‘M’ lanes show markers. Extracts were from (1) Equisetum September shoot, (2) Marchantia plant, (3) Selaginella plant, (4) onion etiolated leaf, (5) onion root, (6) cress etiolated seedling, (7) radish hypocotyl, (8) radish etiolated leaf, (9) radish root, (10) rape hypocotyl, (11) rape cotyledon, (12) rape root.

Download figure to PowerPoint

As expected, undialysed extracts contained endogenous sugars (Figure 5a), but these did not usually prevent monitoring of substrate consumption or product formation. Most extracts contained glucose and/or fructose (Figure 5a), which did not conceal other monosaccharide products, e.g. xylose (Figure 5c,d). Selaginella extracts contained an oligosaccharide, possibly raffinose (Figure 5a), that co-migrated with Ara7 but did not prevent us observing the disappearance of Ara8 (Figure 5b). Other extracts (especially snowdrop, chicory and onion) contained fructo-oligosaccharides. Correct interpretation of profiles from undialysed extracts (Figure 5b–f) is possible by comparison with the appropriate substrate-free control (Figure 5a).

The results with undialysed extracts (Figures 5 and 6; Supporting Information), summarised below, strengthen and extend the major conclusion: plants vary greatly in enzyme profile, but extractant composition and assay temperature have only subtle effects. This validates the use of undialysed extracts for extensive screening of enzymes.

image

Figure 6.  Radiochemical detection of α-fucosidase in plant extracts. Non-dialysed extracts (in extractant A, see Experimental Procedures) were incubated with [Fuc-3H]XXFG at 22°C for 24 h, then analysed by TLC in ethyl acetate/pyridine/propan-1-ol/acetic acid/water (EPPAW). (a) Fluorograph. (b) Stained markers. The short vertical lines help to locate lane centres. Extracts were from (1) cucumber root, (2) stonecrop shoot, (3) crocus leaf, (4) crocus stalk, (5) crocus daughter corm, (6) snowdrop leaf, (7) snowdrop stalk, (8) Equisetum September shoot, (9) Marchantia plant, (10) Selaginella plant, (11) onion etiolated leaf, (12) onion root, (13) cress etiolated seedling, (14) radish hypocotyl, (15) radish etiolated leaf. ‘M’ lanes show markers.

Download figure to PowerPoint

(1[RIGHTWARDS ARROW]5)-α-l-Arabinosidase.  Digestion of Ara8 is clearly seen in undialysed extracts, as monitored by the formation of Ara7, Ara6 and sometimes smaller products (Figure 5b). The trends with Ara6 (Figure 5f) broadly support those with Ara8, though the latter loses its first Ara residue much more readily than Ara6. This indicates that the α-arabinosidases preferentially act on larger oligosaccharides and presumably also on arabinans (polysaccharides).

(1[RIGHTWARDS ARROW]2)-β-d-Galactosidase and (1[RIGHTWARDS ARROW]6)-α-d-xylosidase.  Some extracts contained high XGO-active β-galactosidase (e.g. onion shoots, radish, radish, cress and rape) and others very little (e.g. onion roots, Equisetum, Selaginella and Marchantia) (Figure 5c and Data S2).

To determine which of the two galactose residues of XLLG was cleaved, we incubated a crude extract of cress seedlings with the XGO mixture for various times, then analysed the products by TLC and high-pressure liquid chromatography (HPLC) (Figure 7). The TLC system used does not resolve the two octasaccharides XXLG and XLXG, but HPLC does resolve them. The HPLC peaks were distinguished on the basis that XXLG > XLXG in tamarind xyloglucan, whereas XLXG > XXLG in nasturtium (Fanutti et al., 1996). However, TLC showed that the nonasaccharide XLLG had almost disappeared by 16 h and been largely replaced by a relatively stable octasaccharide spot (Figure 7a). HPLC showed that the octasaccharide formed was XXLG rather than XLXG, and that the small amount of XLXG initially present in the tamarind XGO mixture was itself efficiently hydrolysed, presumably to XXXG (Figure 7b–g). Thus cress β-galactosidases hydrolysed the galactose residue attached to the second isoprimeverose unit (counting from the non-reducing end), but not the one on the third, in this respect resembling a β-galactosidase from Copaifera cotyledons (de Alcântara et al., 1999).

image

Figure 7.  Identification of the galactose residue removed from xyloglucan oligosaccharides (XGOs) by cress seedling β-galactosidase. (a) Non-dialysed cress seedling extract (extractant A, see Experimental Procedures) (16 μl) was incubated with tamarind XGOs (final concentration 4.5 mg ml−1; final volume 64 μl); the reaction was stopped at intervals with formic acid and 3.75 μg products analysed by TLC (M = markers). The right-hand panels show HPLC analyses of (b) the tamarind XGOs used as substrate (XXLG > XLXG), (c) nasturtium XGOs (XXLG < XLXG) and (d)–(g) 0.54 μg of the products formed from tamarind XGOs at some of the time-points shown on the TLC. Free galactose co-elutes with glucose (approximately 2.95 min) in this HPLC system.

Download figure to PowerPoint

(1[RIGHTWARDS ARROW]4)-β-d-Xylosidase and (1[RIGHTWARDS ARROW]4)-β-d-xylanase.  Xyl6 was usually hydrolysed to Xyl5 and sometimes Xyl4, accompanied by free xylose, indicating β-xylosidase (Figure 5d). Etiolated leaves of radish and onion and rape cotyledons were particularly active in this respect. Other extracts, especially from the non-angiosperms Marchantia, Selaginella and Equisetum, gave a wide range of digestion products (Xyl2–5, predominantly Xyl3) but negligible free xylose, indicating xylanase exceeding β-xylosidase.

(1[RIGHTWARDS ARROW]4)-β-d-Mannosidase and (1[RIGHTWARDS ARROW]4)-β-d-mannanase.  Almost all undialysed extracts hydrolysed Man6 to Man5 and Man4 (Figure 5e). The most active extracts (e.g. etiolated onion leaves) also produced Man2–3. As with dialysed mung bean extracts (Figure 2b), undialysed extracts of mung bean (Figure S70) and Marchantia (Figure S72) gave a profile, diagnostic of mannanase activity (Xyl6 > Xyl3 ≥ Xyl5), which would not arise by mannosidase activity alone (compare Figure 3a).

Radiochemical screening for α-fucosidase activity

Numerous enzymes besides α-fucosidase potentially hydrolyse XXFG, including α-xylosidase, β-glucosidase and β-galactosidase, leading to complex digestion patterns. To screen specifically for XGO-acting α-fucosidase, we employed [Fuc-3H]XXFG as substrate (Figure 6 and Supporting Information). Extracts of crocus, snowdrop, onion, cress and radish produced a single radioactive product, free [3H]fucose, showing that α-fucosidase strongly exceeded any α-xylosidase and β-glucosidase (Figure 6a). Other extracts (e.g. Equisetum and Selaginella) yielded not only free [3H]fucose, indicating α-fucosidase, but also 3H-oligosaccharides, probably [Fuc-3H]XFG and [Fuc-3H]FG.

Released [3H]fucose was quantified by paper chromatography (Table S1). Almost all dicot extracts (low- or high-salt; dialysed or not) showed high α-fucosidase activity (mean ≈ 57% release of free fucose in 24 h), including lettuce, pea, parsley and nasturtium. The lowest activities among dicots were found in potato and Sedum.

Non-poalean monocots also had consistently high activity (mean ≈ 62% fucose release). Surprisingly, this also applied to the Poales (mean ≈ 50% fucose release), whose xyloglucans contain very little fucose (McDougall and Fry, 1994). The non-flowering plants tested had low α-fucosidase activity (mean ≈ 15% fucose release).

Screening for hydrolase activity on polymeric substrates

The conclusions drawn with the ‘model’ substrate, Xyl6, were confirmed with the polysaccharide, arabinoxylan (Figure 8 and Supporting Information). Maize and oat yielded an oligosaccharide series (indicating xylanase) plus xylose (indicating β-xylosidase) and arabinose (indicating arabinoxylan-active α-arabinosidase). Broad bean and a few other dicots also generated oligosaccharides and some xylose, but not free arabinose, indicating that arabinoxylan-active α-arabinosidase may be limited to the Poales. Marchantia and Selaginella extracts produced oligosaccharides, confirming high xylanase activity (Figures S133, S138 and S139). Thus, conclusions drawn from low-Mr model substrates are relevant to wall polysaccharides.

image

Figure 8.  Xylanase, β-xylosidase and xylan-active α-arabinosidase in plant extracts. Plant extracts ((a,b) in buffer B, non-dialysed, and (c) in buffer C, dialysed; see Experimental Procedures for buffers] were incubated with arabinoxylan at 22°C for 24 h. Products (from 5 μg polysaccharide) were analysed by TLC in butan-1-ol/acetic acid/water (BAW). Extracts were from (1) parsley shoot, (2) asparagus sprout, (3) chicory leaf, (4) cucumber hypocotyl, (5) cucumber cotyledon, (6) cucumber root, (7) stonecrop shoot, (8) crocus leaf, (9) crocus stalk, (10) snowdrop leaf, (11) snowdrop stalk, (12) Equisetum September shoot, (13) Marchantia plant, (14) Selaginella plant, (15) onion etiolated leaf, (16) onion root, (17) cress etiolated seedling, (18) maize root, (19) radish etiolated leaf, (20) radish root, (21) nasturtium hypocotyl, (22) nasturtium etiolated leaf, (23) nasturtium root, (24) pea epicotyl, (25) pea leaf, (26) pea root, (27) maize coleoptile, (28) maize leaf, (29) maize root, (30) oat coleoptile, (31) oat leaf, (32) oat root, (33) broad bean hypocotyl, (34) broad bean leaf, (35) broad bean root. *Endogenous fructose and/or glucose.

Download figure to PowerPoint

None of the extracts released mono- or oligosaccharides from tamarind xyloglucan (data not shown). Therefore, any glycanase (XEH) activity was insufficient to release appreciable amounts of oligosaccharides, and the glycosidases that attack XGOs failed to release monosaccharides from most of the xyloglucan chain.

Exo- and endotransglycosylation of oligoxylans

Many extracts acted on Xyl6 to generate larger as well as smaller products (Figures 2d, 3b, 4, 5d and S154c,d), indicating transglycosylation, not (solely) hydrolysis. For example, Marchantia extracts gave approximately equal yields of Xyl7 and Xyl5 but no free xylose (Figure 3b), indicating trans-β-xylosidase activity

  • image

but no detectable β-xylosidase. Likewise, Selaginella and Equisetum gave Xyl7 and Xyl5 and almost no free xylose (e.g. Figure 3d). In addition, Xyl2–9 were all produced. Xyl8 appeared after Xyl7 (Figure 3b), indicating that Xyl8 arises by continued trans-β-xylosidase action on Xyl7,

  • image

Interestingly, Xyl9 often exceeded Xyl7 and Xyl8 [e.g. in Equisetum (Figure 5d), Marchantia, Selaginella, broad bean and oat (Figures 3b–e and 4)]. This indicates trans-β-xylanase (β-xylan endotransglycosylase) activity,

  • image

exceeding trans-β-xylosidase. The presence of two distinct enzymes (trans-β-xylanase and trans-β-xylosidase) is supported by the observation that Xyl9 production peaked earlier than Xyl8 (exemplified by Marchantia, Selaginella, broad bean and oat; Figure 3b–e).

Products larger than Xyl6 were also formed by several angiosperm (especially root) extracts (Figures 3c,e and 5d). Again, there was often more Xyl9 than Xyl7–8, indicating trans-β-xylanase exceeding trans-β-xylosidase. This was explored in more depth with various growing organs of numerous additional species (Figure 9 and Supporting Information). Many angiosperms (including Poales) produced Xyl7–10, predominantly Xyl9, especially in roots and hypocotyls.

image

Figure 9.  Trans-β-xylosidase and trans-β-xylanase activities in legume and cereal extracts. Plant extracts in buffer C (see Experimental Procedures), dialysed, were incubated with Xyl6 at 40°C for 3.5 h. Products were analysed by TLC in butan-1-ol/acetic acid/water (BAW; one ascent). M = markers lanes. Extracts were from (1) pea epicotyl, (2) pea leaf, (3) pea root, (4) maize coleoptile, (5) maize leaf, (6) maize root, (7) oat coleoptile, (8) oat leaf, (9) oat root, (10) broad bean hypocotyl, (11) broad bean leaf, (12) broad bean root.

Download figure to PowerPoint

To explore the substrate’s size requirements for endo- and exo-transglycosylation of β-xylosyl linkages, we compared the action of Marchantia extracts on Xyl3 to Xyl6 inclusive (Figure 3f). The trisaccharide was not a substrate, as expected since most possible modes of cleaving Xyl3 (exo- or endohydrolysis or endotransglycosylation) would have given free xylose, which had not been observed in earlier experiments. Xyl4 quickly gave Xyl6 + Xyl2, indicating trans-β-xylanase activity, and smaller amounts of Xyl5 + Xyl3, indicating trans-β-xylosidase activity; later, Xyl7 and larger products were also produced, as the early products themselves acted as acceptor substrates. Xyl5 quickly gave approximately equal amounts of Xyl7 and Xyl8 (plus Xyl3 and Xyl2), indicating transfer of a two- or three-sugar fragment with little preference as to whether the second or third glycosyl linkage of the donor substrate was cleaved; and smaller amounts of Xyl6 + Xyl4 indicating trans-β-xylosidase. Xyl6 gave products similar to those reported before (Figure 3b), with Xyl9 quickly predominating. We conclude that oligo-β-xylans as small as the tetrasaccharide are effective model substrates for the Marchantia enzymes that transglycosylate wall-related (1[RIGHTWARDS ARROW]4)-β-d-xylans in endo- and exo-fashion, but that the enzymes do not act on very small oligosaccharides.

Marchantia extracts can catalyse exo-transglycosylation without releasing detectable free xylose, demonstrating a trans-β-xylosidase with no detectable exo-hydrolytic side-reaction when acting on 1.85 mm Xyl6; however, judged by staining intensity, products smaller than Xyl6 slightly exceeded those larger than it (e.g. Figure 3b,f). This suggests that the trans-β-xylanase was accompanied by a competing endo-hydrolase activity. It was thus interesting to determine the effect of oligosaccharide concentration (relative to an almost constant water concentration) on the transglycosylation : hydrolysis ratio. To quantify the products on a molar basis, we used [1-3H]Xyl6-ol (xylohexaitol, radiolabelled in the xylitol moiety) as a substrate (Figure 10). At very low Xyl6-ol concentrations, irreversible endo-hydrolysis had greatly exceeded reversible endotransglycosylation by 24 h. For example, the major radioactive end-products of 8 μm Xyl6-ol were Xyl3-ol and Xyl2-ol, and only 1.5% of the radioactivity was finally present in products larger than the substrate. As the substrate concentration was raised from 64 to 2048 μm, the yield of larger products progressively increased from 1.9 to 28%, indicating that both transglycosylation and hydrolysis were occurring. At 4096 μm substrate, transglycosylation greatly exceeded hydrolysis, with products larger and smaller than Xyl6-ol accounting for 37 and 40%, respectively, of the total 3H (Figure 10b,c), which is close to the 1:1 ratio expected for pure transglycosylation.

image

Figure 10.  Effect of substrate concentration on hydrolysis accompanying trans-β-xylosidase and trans-β-xylanase activities. Dried [1-3H]Xyl6-ol was re-dissolved at various concentrations (8–4096 μm) in undialysed, undiluted Marchantia extract (extractant A, see Experimental Procedures), and incubated at 22°C for 24 h. In each case, 2.5 kBq of products was then run by TLC and detected by fluorography (a). Each lane was quantitatively profiled, as exemplified by the 4096-μm sample (c). The distribution of radioactivity between differently sized products is reported for each concentration tested (b).

Download figure to PowerPoint

In conclusion, diverse land plants possess enzymes catalysing both exo- and endotransglycosylation of Xyl6, a substrate modelling the xylan-based hemicelluloses of the plant cell wall.

Exo-transglycosylation of xyloglucan oligosaccharides

Xyloglucan-derived heptasaccharide (XXXG) also gave both larger and smaller products when incubated with extracts of dicots, poalean and non-poalean monocots and Selaginella (Figure 2c and Supporting Information), strongly suggesting trans-α-xylosidase activity.

Transglycosylation of oligo-arabinans

Broad bean extracts (especially from hypocotyls and roots) converted Ara6 to two slower-migrating products (RAra6 = 0.92 and 0.82; Figure 11b), indicating trans-α-arabinosidase activity, in addition to Ara5 (RAra6 = 1.15). Substrate-free controls demonstrate the absence of endogenous sugars co-migrating with oligo-arabinans (Figure 11a). Since the two slower-migrating products were not in a chromatographic pattern typical of a homologous series, they may be Ara7 isomers differing in the linkage of the additional residue [e.g. (1[RIGHTWARDS ARROW]5) versus (1[RIGHTWARDS ARROW]3), or furanose versus pyranose]. One transglycosylation product (RAra6 0.92) was also formed by cauliflower extracts (Figure 11b). Assays with Ara8 as substrate (Figure 11c) supported the existence of trans-α-arabinosidase activity, giving products slightly larger than Ara8. In addition, broad bean, pea and cauliflower extracts yielded some much larger products (approximately Ara11–14), suggesting endotransglycosylation catalysed by trans-arabinanase, although this activity was largely lost on dialysis (Figures S1–S153, Data S2).

image

Figure 11.  Trans-α-arabinosidase and trans-α-arabinanase activities in diverse plant extracts. Plant extracts in buffer A (see Experimental Procedures), non-dialysed, were incubated with (a) no added oligosaccharide, (b) Ara6 or (c) Ara8 at 22°C for 24 h. Products were analysed by TLC in butan-1-ol/acetic acid/water (BAW; one ascent). M = markers, T0 = substrate. Extracts were from (1) pea epicotyl, (2) pea etiolated leaf, (3) pea root, (4) broad bean hypocotyl, (5) broad bean etiolated leaf, (6) broad bean root, (7) mung bean hypocotyl, (8) oat etiolated leaf, (9) broad bean leaf, (10) Equisetum May shoot, (11) cauliflower leaf, (12) cauliflower floret; *Trans-α-arabinosidase products; †, trans-α-arabinanase products.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

We report here an extensive survey of enzyme activities that may deconstruct or reconstruct the primary cell wall. We found wide phylogenetic variation in activity profiles, suggesting major differences in wall metabolism. The work also revealed several interesting enzyme activities that had not, to our knowledge, previously been reported.

Advantages of assaying total extractable proteins

All the organs surveyed were rapidly growing, and thus rich in primary walls. The enzymes were not purified; instead the activities were screened in total extractable protein preparations. Whilst this approach does not allow us to identify the specific genes encoding those enzymes, it does nevertheless allow the total activity of a given Enzyme Commission-defined entry to be assayed, and admits of the possibility of novel activities; importantly, it does not limit our attention to known proteins. The usefulness of the strategy of screening total extracts is illustrated by its success in revealing the existence in Equisetum of MXE (Fry et al., 2008a), an activity not predictable from sequence data and found only at very low levels in studies of individual XTHs from other plants (Hrmováet al., 2007; Maris et al., 2011). A similar ‘total-extract’ strategy has also suggested the existence of other heterotransglycanase activities in plants (Kosík et al., 2010). Such a strategy is also more efficient (than testing proteins one by one) at indicating the absence of a postulated activity, e.g. the absence of MXE in the Poales (Fry et al., 2008a), the absence of ‘homogalacturonan endotransglycosylase’ (García-Romera and Fry, 1994), and the absence of ‘xyloglucan:pectin endotransglucosylase’ (Popper and Fry, 2008).

Plants exhibit surprisingly large differences in glycosidase activities, sometimes correlating with wall composition

Glycosidases were assayed on selected cell-wall-related hexa- to nonasaccharide substrates, which model polysaccharides more realistically than do disaccharides, nitrophenyl glycosides etc. (Frandsen and Svensson, 1998). We found enormous phylogenetic variation in glycosidases that potentially deconstruct polysaccharides of the primary wall matrix.

Sometimes this variation mirrored known taxonomic variations in wall chemistry. For example, XXFG-active α-fucosidase had low activity in potato and Marchantia, correlating with their low-fucose xyloglucan (Hoffman et al., 2005; Peña et al., 2008). Conflicting with this correlation, however, the Poaceae had high α-fucosidase but low and fucose-poor xyloglucan (Carpita and Gibeaut, 1993; McDougall and Fry, 1994). It is possible that the α-fucosidase activity detected in the Poaceae was a side-reaction catalysed by a specific endo-β-mannosidase (Ishimizu et al., 2007) and of no physiological significance. However, the Poaceae do also have high activities of other xyloglucan-attacking activities, e.g. transglycanase (XET) (Fry et al., 1992; Smith et al., 1996; Genovesi et al., 2008) and α-xylosidase (present paper); therefore, either poacean xyloglucan has a functional significance belying its small quantity, or the enzymes are an evolutionary legacy from high-xyloglucan ancestors (approximately 65 Ma; Kellogg, 2001) that has not been selected against. Also conflicting with the correlation, Equisetum and Selaginella have fucose-rich xyloglucans (Peña et al., 2008) but exhibited little α-fucosidase (Table S1).

Some crude extracts might retain little hydrolase activity because of wholesale proteolysis, denaturation, tannin precipitation, or loss during dialysis (Takeda and Fry, 2004). Selaginella extracts, for example, had little activity in xyloglucan-related assays (α-xylosidase, α-fucosidase); however, they showed high β-xylan- and β-mannan-related activities, demonstrating that these Selaginella extracts indeed retained active enzymes. A proteinase inhibitor cocktail had little effect on the activities detected, and all extracts studied exhibited high activities in at least some assays.

Most angiosperm extracts rapidly hydrolysed xyloglucan oligosaccharides (Figure 2). Marchantia and Selaginella (non-angiosperms) were much less active in this respect, but had unusually high β-xylosidase. Bryophyte vegetative cell walls contain little β-xylan (Kremer et al., 2004; Carafa et al., 2005), whereas Selaginella walls are very rich in xylan (L. Franková and S. C. Fry, unpublished). Thus, curiously, high β-xylosidase activity correlates with low and high xylan content in Marchantia and Selaginella, respectively.

There was particularly high variation in β-mannosidase (Figure 2b). In etiolated alfalfa and cress seedlings, high activity correlates with a mannan-rich endosperm (Dirk et al., 1995), but the plants with least β-mannosidase included the non-angiosperms, despite their mannan-rich primary walls (Popper and Fry, 2003, 2004). It seems likely that in legume seeds the mannans are reserves, hydrolysed after germination to feed the etiolated seedling, whereas in non-angiosperms the mannans mainly serve architectural roles in the cell wall and are therefore not normally subject to extensive breakdown.

In conclusion, the question of whether a high or low content of a given polysaccharide predicts a high or low activity of the glycosidase that attacks it will depend on the physiological role of the specific polysaccharide: for example, food storage versus wall construction. Plants evidently differ greatly in their wall-deconstructing enzyme activities, and we therefore suggest that different plants specialise in different polysaccharide-modifying mechanisms during cell wall assembly and expansion.

Non-dialysed extracts are valuable for surveying wall-degrading enzymes

In general, the extractant (high or low salt; MES or succinate buffer) and assay temperature had only subtle effects on the enzyme profiles recorded. Furthermore, any low-Mr solutes co-extracted with the enzymes did not appreciably sway the results (non-dialysed versus dialysed extracts) and were thus not substrates for either hydrolysis or transglycosylation. We therefore performed extensive screening using non-dialysed extracts (Figure 5, Table S1 and files in the Supporting Information). The results are available in GHATAbase.

Distinguishing glycanases from glycosidases in crude extracts

Glycosidases (exo-hydrolases) usually attack polysaccharides and oligosaccharides at the non-reducing terminus, releasing free monosaccharides. Acting on a hexasaccharide, a glycosidase progressively generates penta-, then tetrasaccharide, etc. Digestion may proceed completely to the monosaccharide; alternatively, digestion may stall at a ‘limit digestion product’. We documented the kinetics of Man6 exo-hydrolysis by alfalfa β-mannosidase (Figure 3a). Any appreciable deviation from this pattern by other extracts may imply mannanase (endo-hydrolase) as well as or instead of β-mannosidase activity.

Glycanases attack the backbones of polysaccharides or suitably large oligosaccharides at mid-chain, producing smaller chains but usually no monosaccharide. A hexasaccharide would typically yield two trisaccharides or a disaccharide plus tetrasaccharide. Di- to tetrasaccharide production accompanied by little or no monosaccharide clearly indicates glycanase activity. In undialysed extracts, the endogenous plant monosaccharides can make it difficult to distinguish glycanase products (e.g. hexasaccharide [RIGHTWARDS ARROW] two trisaccharides) from products of a glycosidase that stalls at the trisaccharide (hexasaccharide [RIGHTWARDS ARROW] trisaccharide + three monosaccharides). Therefore, when endogenous monosaccharides are present, it is important to record what proportion of the original hexasaccharide remains. For example, if substantial hexasaccharide remains and there is trisaccharide but little pentasaccharide, then this provides strong evidence for glycanase action, as observed in mung bean (Figures 2b and S70) but not alfalfa (Figure 3a).

Our conclusions concerning glycanase activities, where distinguishable from glycosidases, are summarised in GHATAbase. Plant taxa differ markedly in glycanase profiles, as they do in glycosidase profiles.

Hydrolase activity on polysaccharide substrates

Although oligosaccharides are useful substrates, ‘modelling’ polysaccharides, in certain respects they are unsuitable. For example, it cannot be assumed that an α-arabinosidase capable of hydrolysing Ara6 can necessarily remove α-Ara side-chains from arabinoxylans. To test for arabinoxylan-active α-arabinosidase, we therefore used arabinoxylan itself as a substrate. Certain plants, mainly Poaceae, exhibited such activity, releasing free arabinose from arabinoxylan. This supports the hypothesis (Darvill et al., 1978) that stripping the arabinosyl side-chains from arabinoxylan is a normal feature of cereal cell wall turnover in vivo.

Glycosidases may modify xyloglucan’s role as an XET substrate

Plant α-xylosidases act on XG, XXG, XXXG (Koyama et al., 1983; Sanchez et al., 2003), certain larger oligosaccharides (Fanutti et al., 1991), and possibly polysaccharides (Guillén et al., 1995), to remove only the single xylose residue furthest from the reducing terminus, leaving other xylose residues in place. Thus, if α-xylosidase can attack the polysaccharide, it probably does so only at the single residue furthest from the reducing terminus, releasing only one xylose molecule per polysaccharide molecule. As expected, we detected little or no xylosidase activity on high-Mr xyloglucan (data not shown).

Removal of only a single xylose or fucose (Augur et al., 1993; Ishimizu et al., 2007) residue from the chain-end of xyloglucan might seem inconsequential in such a large polysaccharide molecule. However, this particular residue probably has a disproportionately large functional significance because it is the only site in a xyloglucan chain known to serve as acceptor substrate for XET activity. XXG is a weak acceptor substrate for total Pisum XET activities, but GXG is not an acceptor substrate (Lorences and Fry, 1993). Thus, a single hit by α-xylosidase (Figure 1e, step 1) may effectively create an ‘end-cap’ (GXXG), abolishing or diminishing (Fanutti et al., 1996) that whole chain’s XET acceptor capability, with major consequences for wall assembly and restructuring. Sequential β-glucosidase and α-xylosidase action might later bring the ‘capped’ chain back into play by eroding the non-reducing end until the next fully functional acceptor structure (XXFG in Figure 1e) is exposed. Unfortunately, there are no direct experimental data on the ability of plant α-xylosidases to attack non-reducing terminal structures of high-Mr xyloglucan. A qualitatively different end-cap would be provided by GXFG, as this could be by-passed only if α-fucosidase and β-galactosidase, and if necessary an esterase to remove acetyl groups from the galactose, also acted (Figure 1f).

image

Figure 1.  Enzyme nomenclature and proposed reaction schemes. (a–d) Four classes of enzyme activity (given in italics). Examples reported in the present paper are those acting on Xyl6: (a) β-xylosidase, (b) β-xylanase, (c) trans-β-xylosidase, (d) trans-β-xylanase. Each string of circles represents an oligosaccharide (dotted circle = original reducing terminus). (e, f) Requirement for two or four glycosidases, respectively, to restore transglycanase (XET) acceptor substrate function to a xyloglucan chain after ‘end-capping’ by (e) GXXG or (f) GXFG. ✓ and × indicate whether the non-reducing terminus is expected to serve as acceptor substrate for XET activity (Lorences and Fry, 1993).

Download figure to PowerPoint

Thus, under certain physiological conditions, α-xylosidase, β-glucosidase, α-fucosidase and β-galactosidase activity could block and/or restore XET action. Sampedro et al. (2010) discuss additional biological consequences of the effects of such glycosidases on xyloglucan. Our work shows that these enzymes occur to different extents in different plants and organs, as would be expected for a set of enzymes fine-tuning the operation of the plant cell wall in vivo.

Biological significance of glycosidase action on oligosaccharins

Some glycosidases may also modulate cell signalling. Certain XGOs, especially XXFG, have oligosaccharin (‘hormone-like’) effects, inhibiting the auxin-induced growth of pea stem segments at nanomolar concentrations (York et al., 1984; Darvill et al., 1992; Aldington and Fry, 1993). Endogenous XXFG has been detected in living cell cultures (McDougall and Fry, 1991), so it may be a naturally occurring signal. The fucose residue of XXFG plays a particularly important role in conferring ‘anti-auxin’ activity, implying that α-fucosidase is biologically significant in modulating its effect in vivo.

Galactoglucomannan-derived oligosaccharides also have physiological effects at nanomolar concentrations (Auxtováet al., 1995). It remains unknown whether such oligosaccharins, or only the corresponding polysaccharides, occur in vivo (Beňová-Kákošováet al., 2006); however, the β-mannosidase studied in our work may modulate the signalling effects of mannan-derived oligosaccharins.

Presence of transglycosidase activities in plant extracts

Several oligosaccharide substrates yielded products slightly larger, as well as smaller, than the starting material, indicating transglycosidase (‘exo-transglycosylase’) activity. For example, Xyl6 and Ara6 gave clear-cut evidence for, respectively, trans-β-xylosidase in many plants and trans-α-arabinosidase in at least two. The xyloglucan heptasaccharide XXXG, at 1.4 mm, also gave evidence for transglycosidase activity, although the interpretation is more complex because both α-xylosyl and β-glucosyl linkages are present. For example, from XXXG, an octasaccharide could in principle be formed either by trans-α-xylosidase, e.g.

  • image

or by α-xylosidase followed by trans-β-glucosidase:

  • (i)
     XXXG + H2O [RIGHTWARDS ARROW] xylose + GXXG;
  • (ii)
     GXXG + XXXG [RIGHTWARDS ARROW] GXXXG + XXG.

GXXXG, an octasaccharide with a non-reducing terminal glucose residue, would be short-lived because all extracts were rich in β-glucosidase; we therefore suggest that the stable octasaccharide formed from XXXG carries an additional xylose (not glucose) residue, implying trans-α-xylosidase activity. Trans-α-xylosidase activity on 1.5 mm XXXG was also recently reported in extracts of wild-type Arabidopsis, but not α-xylosidase-deficient mutants (Sampedro et al., 2010).

Only a single oligosaccharide substrate was added in each of our assays, so the same compound was acting as both donor and acceptor substrate during transglycosylation. Since we obtained similar results with dialysed and undialysed extracts, we can exclude the possibility that an endogenous mono- or oligosaccharide was serving as donor or acceptor in the undialysed extracts.

The trans-β-xylosidase, trans-α-xylosidase and trans-α-arabinosidase activities reported here are potentially very important biologically. They require critical discussion because at high substrate concentrations many retaining glycosidases catalyse ‘mechanistic’ transglycosylation in addition to hydrolysis. For example, Dey (1979) used 0.5–2.0 m substrates to demonstrate transgalactosidase activity in a Prunusα-galactosidase; Nari et al. (1983) showed that a soya bean β-glucosidase catalyses transglycosylation with gentiobiose or cellobiose (high concentrations not quantified); and Hrmováet al. (1998) showed that a barley β-glucosidase catalyses transglycosylation with 100 mm nitrophenyl β-glucoside. The high substrate concentrations necessary in those studies argue against major biological significance.

In contrast, three lines of evidence suggest that the transglycosidase activities observed in the present work are physiologically significant. For example, in the case of the trans-β-xylosidase activity:

  • (i)
     The transglycosylation:hydrolysis ratio varied between plants. Not all extracts with high β-xylosidase (catalysing hydrolysis) had detectable trans-β-xylosidase activity (e.g. radish leaves and rape cotyledons; Figure 5d). Of four species with high trans-β-xylosidase, the β-xylosidase activity (indicated by free xylose yield) varied widely: Marchantia ≪ Selaginella < oat ≪ broad bean (Figure 3). Therefore, transglycosylation is not an inevitable side-reaction accompanying hydrolysis at the substrate concentrations used.
  • (ii)
     Transglycosylation occurred at the relatively low substrate concentration of 1.8 mm (e.g. Figure 3) and transglycosylation products from 64 μm substrate were still detectable even after 24 h (Figure 10). To undergo transglycosylation at 64 μm, Xyl6 was successfully outcompeting an approximately 106-fold molar excess of H2O (the ‘acceptor substrate’ for hydrolysis).
  • (iii)
     Approximately millimolar substrate concentrations may indeed be physiologically important in muro since the wall matrix contains high polysaccharide concentrations (approximately 30%, w/w; Monro et al., 1976) and relatively low water.

These considerations suggest a biological relevance for the observed trans-β-xylosidase reaction with 64–4096 μm substrate. Similar arguments apply to the observed trans-α-xylosidase and trans-α-arabinosidase activities acting on approximately 1.4 mm oligosaccharides of xyloglucan and arabinan respectively. Transglycosylation was also detected when a plant β-glucosidase acted on 5 mm cellotetraose (Crombie et al., 1998).

Presence of trans-β-xylanase and trans-α-arabinanase activities

Kosík et al. (2010) used microarray screening to obtain evidence for several new heterotransglycanase activities, i.e. transglycanases in which the donor substrate is qualitatively different from the acceptor, in addition to known XET and MXE activities. For example, they reported activities transferring segments of mannan or xylan onto oligosaccharides of xyloglucan, cellulose or MLG, and activities transferring segments of xyloglucan onto oligosaccharides of galactomannan, glucomannan and MLG.

We used single substrates to look for ‘homotransglycanase’ activities (the donor and acceptor being qualitatively similar). Kosík et al. (2010) had found no evidence for xylan or mannan homotransglycanases. We also found no clear-cut evidence for trans-β-mannanase activity acting on Man6 in our survey, but we obtained evidence for two homotransglycanase activities: trans-α-arabinanase activity was present in a few extracts, and many of our extracts showed endotransglycosylase activity with β-(1[RIGHTWARDS ARROW]4)-xylohexaose as both donor and acceptor, interpreted as 2 Xyl6[RIGHTWARDS ARROW] Xyl3 + Xyl9.

Many XTHs have XET but no detectable XEH activity. This represents an extreme example of a carbohydrate successfully outcompeting H2O as acceptor substrate. Other cases are less clear cut. For instance, trans-β-mannanase clearly is a β-mannanase (endo-hydrolase), although its additional ability to catalyse endotransglycosylation may be biologically significant (Schröder et al., 2009).

It is expected that the transglycosylation:hydrolysis ratio will rise with increasing substrate concentration, as the oligosaccharide:H2O ratio increases. In our work, trans-β-xylanase was accompanied by endo-hydrolase activity, but the former was detectable at all substrate concentrations above about 64 μm, and had become strongly predominant by 4 mm, suggesting a biological significance. Pure transglycosylation would give equimolar (and thus equally radioactive) products larger and smaller than the substrate, e.g.

  • image

and this situation was approached at 4 mm Xyl6-ol (Figure 10). In competition with this, especially at lower substrate concentrations, hydrolysis also occurred: an irreversible process inexorably tilting the size balance in favour of smaller products. Transglycosylation:hydrolysis ratios as monitored at a single final time-point (e.g. 24 h in Figure 10) will inevitably underestimate transglycosylation since hydrolysis is irreversible, and molecules may potentially undergo transglycosylation several times before finally being taken ‘out of play’ by a single hydrolysis reaction producing [3H]Xyl2-ol or [3H]Xyl3-ol. Figure 3 shows evidence for transglycosylation at early time points, even in assays where hydrolysis ultimately prevailed.

Trans-β-xylanase appears to be a new enzyme activity, clearly detectable at sub-millimolar hexasaccharide substrate concentrations, and which, when operating on high-Mrβ-xylans, would be of great interest for the integration and/or re-structuring of xylan chains in the growing cell-wall matrix. Trans-β-xylanase activity was particularly high in the liverwort Marchantia and the fern-ally Equisetum, but was also found in angiosperms, so this interesting activity that had not, to our knowledge, previously been reported may be of general significance in the cell-wall biology of land plants.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

This work shows that enzymes acting on cell wall polysaccharides exhibit wide variation across the plant kingdom. Simple methods are described which enable the documentation of total enzyme activity and are not constrained by difficulties in enzyme purification. In particular, the strategy of surveying total protein extracts has led to the discovery of exo- and endotransglycosylase activities, especially trans-β-xylosidase and trans-β-xylanase, whose detailed enzymology and ability to act in vivo will now be further explored.

Experimental Procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Plant material

Parsley (Petroselinum crispum), cucumber (Cucumis sativus), onion (Allium cepa), cress (Lepidium sativum), radish (Raphanus sativus var. sativus), rape (Brassica napus ssp. oleifera), lettuce (Lactuca sativa var. crispa), lucerne (alfalfa; Medicago sativa), maize (Zea mays), potato (Solanum tuberosum), Arabidopsis thaliana, pea (Pisum sativum), broad bean (Vicia faba) and oat (Avena sativa) were grown in a greenhouse in vermiculite (cucumber, cress, radish and rape) or soil under daylight or (if stated ‘etiolated’) in dark conditions. Asparagus officinalis, chicory (Cichorium endivia), cauliflower (Brassica oleracea var. botrytis) and bean sprouts (Vigna radiata) were obtained from a supermarket. Crocus vernus, snowdrop (Galanthus nivalis), winter purslane (Claytonia perfoliata), stonecrop (Sedum rubrotinctum), horsetail (Equisetum arvense), liverwort (Marchantia polymorpha) and Selaginella apoda were grown at the King’s Buildings, Edinburgh. Individual plant organs such as hypocotyls, epicotyls, cotyledons, coleoptiles, first true leaves and whole roots were usually separated from 4- to 6-day-old seedlings, except parsley, onion and potato, where young rapidly growing leaves, 2–10-cm-long etiolated leaves and 2–5-cm-long tuberous sprouts were taken, respectively. In the case of mature plants such as asparagus, horsetail, chicory etc., the growing ‘soft’ parts of sprouts, stems or leaves were used for preparation of enzyme extract.

Polysaccharides and oligosaccharides

Monosaccharides were purchased from Sigma (http://www.sigmaaldrich.com/). Wheat arabinoxylan, Man6, Xyl3–Xyl6, Ara8, Ara6 and XXXG were obtained from Megazyme (http://www.megazyme.com/). A set of manno-, arabino- and xylo-oligosaccharides (used as standards) was generated by partial acid hydrolysis of Man6, Ara6 and Xyl6, respectively. A mixture of non-fucosylated XGOs (principally XLLG > XXLG > XXXG) was kindly given by Dr K. Yamatoya, Dainippon Pharmaceutical Co., Ltd., Osaka, Japan. A similar oligosaccharide mixture was prepared by cellulase digestion of nasturtium (Tropaeolum majus) seed xyloglucan. [Fuc-3H]XXFG was prepared by cellulase digestion of alcohol-insoluble residue from cultured spinach cells that had been fed l-[1-3H]fucose (McDougall and Fry, 1991). [1-3H]Xyl6-ol of various specific radioactivities was prepared by NaB3H4-reduction of Xyl6 and purified by paper chromatography in ethyl acetate/acetic acid/water (10:5:6 by vol.), which resolves Xyl6 from Xyl6-ol.

Preparation of the enzyme extracts

All procedures were carried out at 4°C. Plant material was finely chopped and homogenised in liquid nitrogen with a pestle and mortar. Alternatively, organs available in large quantities were homogenised in chilled extractant with a hand-held blender. The extraction ratio varied from 1:3 to 1:5 (gram fresh weight:millilitre extractant), according to the water content of the plant material. Three different extractants were used: (i) 0.2 m succinate (Na+), pH 5.5, 10 mm CaCl2; (ii) 1 m NaCl dissolved in buffer A; (iii) 1 m NaCl dissolved in 0.2 MES (Na+), pH 5.5. Polyvinylpolypyrrolidone (PVPP; 2% w/v) was suspended in all extractants. The homogenate was stirred slowly with a magnetic stirrer for 3 h at 4°C. After filtration through two layers of Miracloth (Calbiochem, http://www.calbiochem.com/united-kingdom), the extract was centrifuged at 12 000g for 40 min. Half the supernatant (termed ‘crude enzyme extract’) was divided into small aliquots for storage. The second half was dialysed against distilled water (2 × 12 h) followed by 0.04 m succinate (Na+) buffer (1 × 12 h). The supernatant obtained after centrifugation (1 min, 1000g) was termed ‘dialysate’. All enzyme extracts were stored at −80°C.

Enzyme assays

A TLC-based assay of hydrolases and transglycosylases acting on oligosaccharides.  Routinely, each assay mixture contained 30–50 μg of substrate, added in 5 μl water and thoroughly mixed with 15 μl of buffered enzyme extract to give final substrate concentrations of: arabinoxylan, 0.25% (w/v); XGOs, approximately 2 mm; pure XXXG, 1.4 mm; Xyl6, 1.85 mm; Man6, 1.52 mm; Ara6, 1.85 mm; or Ara8, 1.40 mm. After incubation (22°C for 24 h or 40°C for 3.5 h), reactions were stopped on ice and products analysed by TLC. In some experiments, 0.075 volumes of a broad-spectrum proteinase inhibitor cocktail (PIC; Sigma catalogue no. P9599; 10-fold diluted in water), or dimethylsulphoxide (1% v/v final concentration) as a control, was added during the incubation.

α-Fucosidase activity on radiolabelled xyloglucan nonasaccharide.  Each assay mixture contained 0.5 kBq [Fuc-3H]XXFG, added in 2.5 μl water to 7.5 μl of buffered enzyme extract. After incubation at 22°C for 24 h, 0.2 kBq was subjected to TLC and fluorographed. A 0.3-kBq portion was analysed by paper chromatography: the fucose band and the (immobile) oligosaccharide band were separately assayed for radioactivity.

TLC and paper chromatography

Reaction products were routinely analysed on Merck silica-gel ‘60’ TLC plates (Merck, http://www.merck.co.uk) developed in butan-1-ol/acetic acid/water (BAW; 2:1:1). For analysis of XGO digests, ethyl acetate/pyridine/ethanol/water (EPEW; 6:3:1:1) was also used. Fucosidase products were analysed by TLC in ethyl acetate/pyridine/propan-1-ol/acetic acid/water (EPPAW; 5:2:2:1:1; TLC, three ascents) or by paper chromatography in butan-1-ol/acetic acid/water (12:3:5). For staining, TLC plates were dipped in thymol solution [0.5% thymol (w/v) and 5% (v/v) concentrated H2SO4 in 96% (v/v) ethanol], dried and heated at 105°C for 5–10 min. The plates were scanned within 15 min of staining. The intensity of the individual bands was semi-quantified visually.

High-pressure liquid chromatography

In some experiments, XGO digestion products were analysed by HPLC on a CarboPac PA1 column (‘high-performance’ anion-exchange chromatography, HPAEC; Dionex UK, http://www.dionex.com/) with elution at 1 ml per min in: 0–5 min, 100 mm NaOH/50 mm sodium acetate [RIGHTWARDS ARROW] 100 mm NaOH/70 mm sodium acetate (linear gradient); 5–15 min, 100 mm NaOH/70 mm sodium acetate [RIGHTWARDS ARROW] 100 mm NaOH/100 mm sodium acetate (linear gradient); 15–25 min, 100 mm NaOH/100 mm sodium acetate [RIGHTWARDS ARROW] 100 mm NaOH/150 mm sodium acetate (linear gradient); 26–31 min, 800 mm NaOH isocratic; 31–37 min, 100 mm NaOH/50 mm sodium acetate isocratic.

Detection of radioactivity

The TLC plates were fluorographed after dipping through 7% (w/v) 2,5-diphenyloxazole in diethyl ether, and quantitatively profiled by 60-min counting in a LabLogic AR2000 radioactivity scanner (http://www.lablogic.com/). Strips from paper chromatograms were assayed by scintillation counting.

Footnotes
  • *

    In this paper we name exo- and endo-acting glycosylhydrolases as ‘glycosidases’ and ‘glycanases’, respectively (Figure 1a–d). Thus, an enzyme that splits off a single non-reducing terminal β-xylose residue from xylohexaose is a β-xylosidase; an enzyme that cleaves a mid-chain linkage in the same substrate is a β-xylanase. The term ‘enzyme’ follows Enzyme Commission usage: it is defined as an ‘activity’, which may be shared by multiple isozymes encoded by different genes. When the proteins responsible for these enzyme activities are known, they can be found in the CAZy database (Cantarel et al., 2009).

  • (G) β-d-Glc residue of the xyloglucan backbone with no side-chain attached. (X) α-d-Xyl-(1[RIGHTWARDS ARROW]6)-β-d-Glc** (isoprimeverose). (L) β-d-Gal-(1[RIGHTWARDS ARROW]2)-α-d-Xyl-(1[RIGHTWARDS ARROW]6)-β-d-Glc** (trisaccharide). (F) α-l-Fuc-(1[RIGHTWARDS ARROW]2)-β-d-Gal-(1[RIGHTWARDS ARROW]2)-α-d-Xyl-(1[RIGHTWARDS ARROW]6)-β-d-Glc** (tetrasaccharide). **The β-d-Glc in each structure is part of the (1[RIGHTWARDS ARROW]4)-β-glucan backbone of the xyloglucan. For further information see Fry et al., 1993.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank the Biotechnology and Biological Sciences Research Council (BBSRC) (UK) for financial support of this work.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Conclusion
  7. Experimental Procedures
  8. Acknowledgements
  9. References
  10. Supporting Information

Figures S1–S153. Scans of TLC plates documenting the presence of various enzymes according to the reaction products generated by their action. (A Word file to which the Excel file ‘Data S2’ becomes linked if saved in the same folder.)

Figure S154. Effect of proteinase inhibitor cocktail on hydrolytic and transglycosylation activity. Plant extracts (in buffer A) were dialysed and then incubated with (a, b) Man6 or (c, d) Xyl6, in the presence (a, c) or absence (b, d) of PIC. Other details as in Figure 2.

Table S1. α-Fucosidase activity in extracts of 57 plant organs, assayed on the xyloglucan nonasaccharide [Fuc-3H]XXFG. Extracts were prepared in buffer A (low-salt) or buffers B or C (high-salt), then either dialysed or not, and incubated with [fucosyl-3H]XXFG for 24 h as in Figure 6. [3H]Fucose, resolved by paper chromatography, was assayed by scintillation-counting. Remaining [3H]XXFG and any intermediary 3H-oligosaccharides all remain very close to the origin in the chromatography solvent used. Within each taxomonic group, the specimens studied are listed in descending order of fucosidase activity.

Data S1. GHATAbase users’ manual.

Data S2. Database of multiple cell wall enzyme activities detected in crude extracts of various plant species. A searchable Excel file, which becomes linked to the Word document ‘Figures S1–153’ if saved in the same folder.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
TPJ_4625_sm_DataS1.doc933KSupporting info item
TPJ_4625_sm_DataS2.xls3775KSupporting info item
TPJ_4625_sm_FiguresS1-S153.doc13113KSupporting info item
TPJ_4625_sm_FigureS154.doc3800KSupporting info item
TPJ_4625_sm_Legends.doc26KSupporting info item
TPJ_4625_sm_TableS1.doc123KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.