Chara — a living sister to the land plants with pivotal enzymic toolkit for mannan and xylan remodelling

Abstract Land‐plant transglycosylases ‘cut‐and‐paste’ cell‐wall polysaccharides by endo‐transglycosylation (transglycanases) and exo‐transglycosylation (transglycosidases). Such enzymes may remodel the wall, adjusting extensibility and adhesion. Charophytes have cell‐wall polysaccharides that broadly resemble, but appreciably differ from land‐plants'. We investigated whether Chara vulgaris has wall‐restructuring enzymes mirroring those of land‐plants. Wall enzymes extracted from Chara were assayed in vitro for transglycosylase activities on various donor substrates — β‐(1→4)‐glucan‐based [xyloglucan and mixed‐linkage glucans (MLGs)], β‐(1→4)‐xylans and β‐(1→4)‐mannans — plus related acceptor substrates (tritium‐labelled oligosaccharides, XXXGol, Xyl6‐ol and Man6‐ol), thus 12 donor:acceptor permutations. Also, fluorescent oligosaccharides were incubated in situ with Chara, revealing endogenous enzyme action on endogenous (potentially novel) polysaccharides. Chara enzymes acted on the glucan‐based polysaccharides with [3H]XXXGol as acceptor substrate, demonstrating ‘glucan:glucan‐type’ transglucanases. Such activities were unexpected because Chara lacks biochemically detectable xyloglucan and MLG. With xylans as donor and [3H]Xyl6‐ol (but not [3H]Man6‐ol) as acceptor, high trans‐β‐xylanase activity was detected. With mannans as donor and either [3H]Man6‐ol or [3H]Xyl6‐ol as acceptor, we detected high levels of both mannan:mannan homo‐trans‐β‐mannanase and mannan:xylan hetero‐trans‐β‐mannanase activity, showing that Chara can not only ‘cut/paste’ these hemicelluloses by homo‐transglycosylation but also hetero‐transglycosylate them, forming mannan→xylan (but not xylan→mannan) hybrid hemicelluloses. In in‐situ assays, Chara walls attached endogenous polysaccharides to exogenous sulphorhodamine‐labelled Man6‐ol, indicating transglycanase (possibly trans‐mannanase) action on endogenous polysaccharides. In conclusion, cell‐wall transglycosylases, comparable to but different from those of land‐plants, pre‐dated the divergence of the Charophyceae from its sister clade (Coleochaetophyceae/Zygnematophyceae/land‐plants). Thus, the ability to ‘cut/paste’ wall polysaccharides is an evolutionarily ancient streptophytic trait.


| INTRODUCTION
Numerous papers mention charophycean algae as attractive models for plant physiologists, biochemists and molecular biologists (Gylete et al., 2021, Domozych et al., 2016, Domozych and Bagdan, 2022), especially Chara spp.owing to their flexible anatomical structure, large size and easy manipulation.Charophytes are a group of green algae considered the closest living relatives of land plants.Indeed, charophytes plus land plants together constitute a broad taxon, the Streptophyta, which excludes all other algae.Chara and Nitella (stoneworts) are typical representatives of the class Charophyceae which among other classes (Coleochaetophyceae, Zygnematophyceae, Klebsormidiophyceae Chlorokybophyceae and Mesostigmatophyceae) constitute the charophytes (Domozych et al., 2016;McCourt et al., 2004, Nishiyama et al., 2018).
The Charophyceae were formerly considered as the closest living relatives to the land plants (Karol et al., 2001); however, more recent studies favour the Coleochaetophyceae (Finet et al., 2010;Leliaert et al., 2012) or Zygnematophyceae (Leebens-Mack et al., 2019;Puttick et al., 2018;Timme et al., 2012;Turmel et al., 2006;Nishiyama et al., 2018;Domozych & LoRicco, 2023), especially the latter.Nevertheless, the Charophyceae (including Chara and Nitella) exhibit many features found in land plants, such as plasmodesmata, phragmoplast, branching, cellulose synthase rosettes, auxin signalling, apical cell growth etc. (Nishiyama et al., 2018;Foissner &Wastenys, 2014).On the other hand, the biochemistry and architecture of algal cell walls differ substantially from those of land plants.Briefly, as shown in Table 1, the major polysaccharide classes reported for many charophytes are cellulose, mannan (a hemicellulose) and homogalacturonan (HG; a pectic domain), but the results of conventional biochemical analysis and immuno-labelling do not always match.It seems that typical land-plant polysaccharides such as xyloglucan, xylans and rhamnogalacturonan-I (RG-I) are not equally distributed (or present in detectable quantities) in charophytic algae.This implies that there must be a broad diversity of poorly understood wall polymers in charophytes.

| Land-plant polysaccharides
The three major polysaccharide classes in land plants are pectin, hemicelluloses and cellulose.It cannot be expected that the structures of these will be precisely conserved between charophytes and land-plants; nevertheless, all three classes are present in Chara and Arabidopsis, and a brief description of these polysaccharide classes in land-plants will set the scene for Chara.
Land-plant pectin is a complex, multi-domain polysaccharide that is, by definition, rich in α-D-galacturonic acid (GalA) residues.
The three major domains are HG, RG-I and RG-II, which can readily be distinguished after digestion of whole cell walls with endo-polygalacturonase (EPG).This enzyme cleaves the HG domains (after pre-treatment with mild alkali to remove the methyl and acetyl ester groups) to yield GalA 1 , 2 and 3 , and in the process also releases intact RG-I (M r $ 50-500 kDa; Kaczmarska et al., 2022) and RG-II (M r $ 5 kDa; Kobayashi et al., 1999;Begum & Fry, 2023).It is suspected that land-plant pectin comprises a large polysaccharide in which HG, RG-I and RG-II are glycosidically linked like beads on a string.Thus, whole pectin molecules can be cross-linked via any of their domainsfor example, Ca 2+ bridges (between HG domains) and borate bridges (between RG-II domains; Sanhueza et al., 2022), and also potentially via xyloglucan-RG-I glycosidic bonds (Thompson & Fry, 2000), RG-Icellulose bonds (Lin et al., 2016) and GalA-polyamine amide bonds (Perrone et al., 1998).
Xylans are the second most abundant plant polymers on Earth (Scheller and Ulvskov, 2010).The xylan backbone is formed by β-1-4-linked xylose (Xyl) residues.The main classes of xylan are glucuronoxylans (decorated with glucuronate residues), arabinoxylans (arabinofuranosyl-substituted) and glucuronoarabinoxylans (branched with α-L-arabinofuranose (Ara) and glucuronate (GlcA) side chains; Curry et al., 2023;Smith et al., 2017).Glucuronoarabinoxylans are typical of grasses, cereals and palms, and are thought to be the main cellulose-tethering hemicelluloses in both primary and secondary cell walls.A common feature of grass xylans is further substitution of Ara with feruloyl, coumaroyl and β-xylosyl residues, which make a significant contribution to the stability of the plant cell wall and its resistance to digestion by microbes and herbivores (Buanafina, 2009).
Xyloglucan is a typical land-plant hemicellulose which comprises a linear backbone of (1!4)-linked β-Glc residues, many of which carry an α-Xyl residue on position 6 (Fry et al., 2011).Typically, every fourth Glc residue is unsubstituted, resulting in structures based on XXXG [for nomenclature, see Fry et al. (1993)].Some of the Xyl residues (especially the third from the non-reducing end) carry an additional β-Gal residue on position 2, and the β-Gal itself often carries α-Fuc on its 2-position (forming XLXG and XLFG, respectively).The detailed structure and sequence of diverse side chains along the backbone of xyloglucan has been reviewed (Fry, 2011;Franková & Fry, 2012b).

| Enzymic transglycosylation reactions
This paper focuses on wall-localised enzymes that catalyse transglycosylation.Transglycosylation is a reaction, often enzyme-catalysed, that cleaves a glycosidic bond (in a 'donor substrate') and conserves the energy of that bond in making a new glycosidic bond (to an 'acceptor substrate').Transglycosylation may be described as 'cutting and pasting' of sugars (Franková & Fry, 2013).In some cases, the energy of the cut bond greatly exceeds that of the new bond that is made, such that the reaction is essentially irreversible.Classic examples of this situation are in polysaccharide synthases, where the donor is a nucleoside diphosphate sugar (e.g.UDP-glucose; thus, the linkage that is broken is a sugar-phosphate bond), and the product is a new sugar-sugar linkage (e.g.within a polysaccharide).In the present paper, however, we focus on cases where the energy of the broken bond approximately equals that of the newly produced bond, such that the reaction is freely reversible.Such reactions can be summarised as  (Franková & Fry, 2013).
Often, the acceptor substrate is similar or identical to the donor; enzymes catalysing such reactions are thus called homo-transglycosylases, examples (expressed as 'donor: acceptor') being xyloglucan:xyloglucan (Fry et al., 1992), mannan:mannan (Schröder et al., 2006) and xylan:xylan (Franková & Fry, 2011;Johnston et al., 2013)  It may be envisaged that a transglycanase (endo-activity) could, by a single catalytic event, exert a profound effect on the mechanics of the cell wall, greatly enhancing wall extensibility, for example.In contrast, a transglycosidase (exo-activity) would have only a small direct mechanical effect, though potentially a major knock-on effect by altering the ability of a whole polysaccharide molecule to serve as an acceptor substrate for a transglycanase (Franková & Fry, 2011).
Heterotransglycanases can be of great interest in that they interlink dissimilar polysaccharides within the cell wallfor instance bonding cellulose to xyloglucan and thus potentially anchoring xyloglucan to the cellulosic microfibrils (Herburger et al., 2021).

| Transglycosylation reactions in charophyte cell walls
The great majority of information on wall-localised transglycosylases (catalysing reversible or dedicated transglycosylation reactions) comes from land-plants, especially angiosperms, with few studies having focused on algae (Fry et al., 2008b;van Sandt et al., 2007;Herburger et al., 2018;Franková & Fry, 2021).In the present paper, we explore to what extent comparable activities may occur in a charophytic cousin of the land plants -Chara vulgaris, a representative of the Charophyceae.The aim is to explore the ancient evolutionary origin of wall-remodelling activities among the streptophytes.
In this study, we focus on profiling and characterising transglycanase and transglycosidase activities that modify mannans and xylanstwo main hemicelluloses abundant in charophycean algae.We aim to find out whether both TBM (trans-β-mannanase) and TBX (trans-

| Enzyme preparation
Enzyme extracts of axenic Chara were prepared as described by Franková & Fry (2021); all steps were performed at 0-4 C. Algal cultures were thoroughly rinsed with de-ionised water and then vigorously homogenised by mortar and pestle in extraction buffer (0.3 M succinate [Na + , pH 5.5] supplemented with 15% v/v glycerol and 20 mM ascorbate) containing 3% (w/v, suspension) polyvinyl polypyrrolidone at an extraction ratio of 1:4 (g fresh weight:ml extractant).After 3 h of gentle stirring, the homogenate was filtered through three layers of nylon gauze and then centrifuged at 10,500 g for 45 min.For the isoelectric-focusing experiment, a different extract was prepared: axenic algal cultures were freeze-dried and then extracted with 0.2 M succinate (Na + , pH 5.5) containing 1 M NaCl, 0.02% Triton and 5% glycerol.The extraction ratio varied from 1:10 to 1:20 (g freeze-dried culture: ml extractant) according to the density of the final homogenate.After homogenisation with a pestle and mortar followed by 3 h stirring, the homogenate was centrifuged (10,500 g, 45 min) and dialysed twice against 2.5 mM succinate [Na + , pH 5.5].

| Isoelectric focusing
Extract containing 5% glycerol and 2% ampholites (Bio-Lite 3-10, 163-1112, BioRad Inc.) was isoelectric-focused in a Rotofor Cell IEF (isoelectric focusing) apparatus (BioRad) according to the manufacturer's instructions.Electrophoresis was performed at 12 W until the voltage and current stabilized.The pH of the fractions was then measured, the fractions were de-salted on a centrifugal concentrator (Vivaspin, 5 kDa molecular weight cut off) and 5 μL was assayed for transglycanase activities.

| Transglycanase (endo-) activity assays
The following transglycanase substrates were employed for assaying For testing optimal substrate concentrations, the reaction mixtures (final volume 10 μL) contained the following combination of substrates: After incubation for 2 h, the reactions were stopped by addition of 6 μL of 15 M ammonia, which does not affect mannan solubility.The reactions for time-courses (used in Figure 8) were stopped after 0-32 h.
The 3 H-labelled transglycosylase products were separated from unreacted oligosaccharides by a glass-fibre blotting method (Franková & Fry, 2020) except for the products generated from Man 6 or Man 6 -ol, which were separated by the TLC method followed by 3 H-scanning (Franková & Fry, 2011; used in Figure 4C, Figure 6).The error bars represent three to four technical replicates (n = 3-4) ± SE.

| Choosing Chara for enzyme screening
We selected healthy Chara vulgaris specimens grown under laboratory axenic and natural conditions.Light microscopy (Figure 1) showed that the selected specimens of axenic Chara exhibited typical features of wild-grown specimens such as a well-developed whorl of branchlets and stipulodes, cortex cells, crystals of calcium carbonate, characteristic red antheridia and oogonia, and numerous meristematic nodal zones.
This means that our axenic cultures of Chara vulgaris were healthy and continued their life cycle as if grown in their natural freshwater habitat.
In our earlier screens of numerous algal and fern extracts (Franková & Fry, 2021), we found that the Charophyceae, among other classes, contain extractable glycanase activities that are capable Figure 2 summarises the main activities tested for, including those that were and those that were not detected.

| XET-like and MXE-like activities detected in Chara extracts
It is still not certain whether Chara cell walls contain any typical xyloglucan or MLG (well-known land-plant polysaccharides, the former ubiquitous and the latter found only in Equisetum and the Poales; Table 1).Nevertheless, in vitro, endo-transglycosylation could be found between xyloglucan and MLG (as donors) and tritiated oligoxyloglucan (as acceptor).If the transglycosylation occurs between xyloglucan and oligoxyloglucan, or between MLG and oligoxyloglucan, it indicates the presence of enzyme activities known as XET and MXE, respectively.Since the presence of conventional xyloglucan and MLG with typical linkages and structure as in land plants was not chemically proven in Chara, we will specify these activities as "XETand MXElike" (Figure 3A).[It should be noted that we quote all the transglycosylation reactions as being 'from' donor 'to' acceptor.] All the values were corrected to a control containing heat-or formic acid-inactivated enzyme.3A).

| Survey of extractable homo-and heterotransglycanase activities acting on hemicellulosic donor substrates
In in-vitro assays, Chara extracts were tested for transglycanase activities with diverse polysaccharides as potential donors and either [ 3 H] Man 6 -ol or [ 3 H]Xyl 6 -ol as acceptors (Figure 3B,C).
For both enzyme activities (TBM and TBX), the yield of product increased with acceptor substrate concentration, as expected (Figure 5).Surprisingly, the yield plateaued above $300-400 nM concentration of acceptor substrate.This suggests that the K M of these enzymes for their acceptor substrates is sub-μM.The diminished activities seen at 700 nM acceptor concentration (Figure 5A,B) are likely to be caused by detection limit and sensitivity of scintillation counter used in detection of 3 H-labelled polymers.
The transglycosylation products of both TBM and TBX were relatively stable over time (up to at least 24 h; Figure 6), so it can be concluded that TBM and TBX catalyse non-mechanistic transglycosylation that may have physiological relevance in vivo.The attachment of mannans to [ 3 H]Man 6 -ol was remarkable, especially that of glucomannan, where $55% of initial acceptor provided became incorporated into polymers. The

| Transglycosidase (exo-acting) activities on oligosaccharides
Using specific oligosaccharides to assay transglycosylase activities has the advantage over using polysaccharides in that the structures of the substrates are precisely defined, whereas polysaccharides are of imprecisely known composition and molecular masses.
Oligosaccharide-to-oligosaccharide transglycosylation with mannan hexasaccharides is shown in Figure 4C, Figure 9A and For better interpretation of activities detected in Chara crude extracts, consider first the radioactive products formed from a trace of [ 3 H]Man 6 -ol ($0.1 μM) in the presence of a higher concentration ($3 mM) of a qualitatively different non-radioactive oligosaccharide, Xyl 6 (Figure 6A, lanes 1 and 3).

| Mannan hydrolases
The [ 3 H]Man 6 -ol had been largely hydrolysed within 24 h, the major products being The transglycanase products were separated from unreacted products by the glass-fibre blotting method except for those from water-soluble cellulose acetate (A).Three controls were prepared -(i) with formic acidinactivated enzyme, (ii) donor free (to correct for endogenous substrate which might be possibly co-extracted with proteins), and (iii) an enzymefree control (to correct for nonenzymic trapping of acceptor in the donor).All values were corrected to control (i).Thus each reaction mixture had its own control for each donor).
Both controls (ii) and (iii) are corrected to background and are thus higher than baseline; the value in control (iii) was subtracted from the value obtained with enzyme.Since control (ii) was minimal in A, B and C we conclude that our extracts had negligible co-extracted polysaccharides with donor capability, and thus values shown in the histograms are reliable.Negative values in some samples are given by the slight hydrolytic activity of Chara extract hydrolysing the acceptor.So if control (i) was higher than the undenatured enzyme sample it was attributable to the trapping of 3 H in the polysaccharide during the glassfibre blotting step.Error bars represent four technical replicates both catalysed by (endo) β-mannanase acting near the middle of the hexasaccharide, but with negligible catalysed by β-mannanase acting near the reducing terminus, nor Man 5 -ol catalysed by (exo) β-mannosidase.
Judged by the 3 H-labelled hydrolysis products observed, concentrated Xyl 6 did not compete with the hydrolysis of [ 3 H]Man 6 -ol, suggesting the presence of a mannan-specific hydrolase.

| Lack of xylan:mannan heterotransglycosylases
No radioactive products larger than [ 3 H]Man 6 -ol were detected with Xyl 6 as the potential donor (Figure 6A, lanes 1 and 3), indicating negligible X:M transglycosylase actions such as catalysed by a hypothetical heterotransxylanase, or catalysed by a hypothetical heterotransxylosidase.

| Xylan hydrolases
In the case of tracer [ 3 H]Xyl 6 -ol plus (non-interfering) non-radioactive Man 6 , after 24 h some [ 3 H]Xyl 6 -ol remained intact but the majority There was no evidence for a hypothetical β-xylanase endo-activity, which would have given [ 3 H]Xyl 3 -ol as the major product.Judged by the 3 H-labelled hydrolysis products observed, concentrated Man 6 did not compete with the hydrolysis of [ 3 H]Xyl 6 -ol, suggesting the presence of a xylan-specific hydrolase.
3.6.5 | Homo-trans-β-mannosidase and homo-trans-β-mannanase Next consider the thymol-stained (bulk) oligosaccharides in the reaction mixtures described above (Figure 6D).Even at 0.25 h, the Man 6 contained traces of Man 3-5 and bigger products (possibly Man $9 ) (lane 5); however, compared with these, large amounts of Man 2-5 and Man 7-10 and a chromatographically immobile polysaccharide had formed by 24 h (Figure 6D, The formation of larger products e.g.[ 3 H]Man 8-10 -ol and even [ 3 H]polysaccharides (designated as indicates homo-trans-mannanase activity, catalysing, for example, The possible production of Xyl 9 by a hypothetical progressive trans-β-xylosidase (exo) activity, with the product of each step being the substrate for the next: 2 Xyl 6 !Xyl 7 þ Xyl 5 followed by : Xyl and finally some : would have produced more Xyl 8 than Xyl 9 .Higher-homologue products ([ 3 H]Man 8-10 -ol and [ 3 H]polysaccharides), indicative of trans-β-mannanase activity, were also clearly visible when the donor was ≥505 μM, though not at 101 μM (Figure 7B and D).These findings suggest that trans-β-mannosidase has a lower K M for Man 6 (donor) and is a more 'dedicated' transglycosylase activity than trans-β-mannanase.
Keeping the donor concentration constant at 1000 μM, we tracked the 0-24-h time-course of formation of transglycosylation products (Figure 7A and C).The [ 3 H]Man 9 -ol and [ 3 H]polysaccharide products (indicative of endo-transglycosylation) appeared first, becoming detectable within 10-30 min.The yield of [ 3 H]Man 9 -ol plateaued at $1 h, and thereafter remained steady for at least 24 h, whereas the yield of [ 3 H] polysaccharide continued to rise for at least 12 h.This suggests that [ 3 H] Man 9 -ol was an intermediate, en route to the production of (relatively insoluble and thus stable) [ 3 H]polysaccharides: The [ 3 H]Man 7 -ol product (indicative of exo-transglycosylation) appeared more gradually, continuing to rise up to at least 24 h (Figure 7A,C).The results thus indicate that the endo-enzyme transβ-mannanase, despite having a lower affinity (higher K M ) for Man 6 , was ultimately more active than endo-enzyme trans-β-mannosidase.

| Setting the scene
Several land-plant transglycanase and transglycosidase activities have been detected, which certainly in vitro and in some cases demonstrated in vivo, are capable of re-structuring the major cell-wall polysaccharides of land-plants.Many of those activities are found throughout the (land-) plant kingdom, though some are confined to (or only prominent in) certain taxa such as trans-β-mannanase in some ferns, hornworts and lycophytes (Franková &Fry, 2021), andMXE andCXE in Equisetum (Simmons et al., 2015;Herburger et al., 2021).In this work, we extended the search for such enzymes to a charophytic alga, Chara vulgaris.Chara is a latediverging (formerly termed 'higher') charophyte, i.e., relatively closely related to land plants.Indeed, Chara shares several cell-wall polysaccharides with land-plants, but those from Chara are not fully characterised, especially their side chains and linkages; therefore, we used land-plant polysaccharides as substrates in our assays.charophytes -Nitella, Klebsormidium, Spirogyra, Zygnema and Coleochaete (Franková & Fry, 2021;Herburger et al., 2018).However, it was mainly Charophyceae (Chara and Nitella sp.) exhibiting high XETand MXE-like activities (Franková & Fry, 2021).

| XET-like and MXE-like activities detected
Chara does possess cellulose, the donor substrate for Equisetum CXE activity.However, there was no evidence of a 'CXE' activity capable of acting on the model donor substrate, water-soluble cellulose acetate.But this may not be surprising since Equisetum HTG has low 'CXE' activity on water-soluble cellulose acetate (Simmons et al., 2015) but prefers unmodified cellulose type II and I (Herburger et al., 2020b) .
Chara extracts did not exhibit either of the transglycosidase  Indeed, both these activities were readily detectable in vitro in Chara extracts (summarised in Figure 2).Routinely, we used [ 3 H]Man 6 -ol as the acceptor substrate, representing a fragment of the backbone of mannan.Trans-β-mannanase activity was detected with any of a range of donor substrates: pure (1!4)-β-D-mannan, glucomannan, galactomannan and mannohexaose.We suggest that this enzyme activity serves the role of 'cutting and pasting' chains of the major hemicellulose, mannan, in the Chara cell wall, enabling cell expansion in the manner previously proposed for XET activity (on xyloglucan) in land-plants (Thompson & Fry, 2001;Hayashi & Kaida, 2011;Franková & Fry, 2013).
The exo-enzyme activity, homo-trans-β-mannosidase, was also detected (Figure 6, Figure 7, Figure 2).This was demonstrated by the use of the hexasaccharide fragment, Man 6 , as donor substrate and either a second Man 6 molecule or [ 3 H]Man 6 -ol as acceptor.Since this activity attacks the donor substrate specifically at the non-reducing terminus of the chain, it is more easily detected acting on a hexasaccharide than on a polysaccharide (in which very few termini are present relative to the total mass).It is unclear what the role of this exo-activity might beslightly 'nibbling' the substrate from the end rather than 'biting' it in the middle.But, in general terms, it is possible that tweaking the non-reducing end could render that terminus a better or worse acceptor substrate for mannan-to-mannan endotransglycosylation.For example, by removing the terminal β-Man residue it could leave the polysaccharide with a substituted Man residue (e.g. one with a galactosyl or acetyl side-chain) at the terminus, which might be a poor acceptor substrate for trans-β-mannanase; this would take that polysaccharide molecule 'out of the game' (at least as an acceptor substrate) for wall re-structuring.
Trans-glycanase action was also detected on the intrinsic cell wall components of Chara, with exogenous fluorescent Man 6 -ol-SR as the acceptor substrate and endogenous polysaccharide(s) as donor (Figure 10).Potentially, the observed activity was TBM, though the possibility cannot be excluded that the in-situ donor substrate was not a mannanfor example, the activity could be xylan:oligomannan hetero-transxylanase, though this is unlikely because no such activity was detectable in enzyme extracts in vitro (Figure 2B).
Man 6 -ol-SR as acceptor substrate was also used in the dot-blot method (Figure S3), which demonstrated Chara TBM's in-vitro activity on glucomannan.Despite both Equisetum and Chara cell walls being rich in Man residues (Nothnagel & Nothnagel, 2007;Silva et al., 2011), only Chara extracts formed bright fluorescent spots on the test papers.This indicates mannan-remodelling is more favoured in charophytes than in the Equisetales.
No hetero-trans-β-xylanase activity was detected with xylans as donor and [ 3 H]Man 6 -ol as acceptor.Thus, we did not detect an enzyme capable of creating xylan-mannan hybrid polysaccharides.
We conclude that Chara can potentially make mannan!xylanhybrid hemicelluloses (where '!' is a glycosidic bond) but not xylan!mannanhybrids.The significance of this distinction concerning charophyte wall architecture remains to be elucidated.
Thus Chara appears unable to 'nibble' xylans in the manner proposed above for mannans.

| Dedicated versus mechanistic transglycosylases: crucial importance of acceptor substrate concentration
There is little fundamental difference between a transglycosylase activity and a hydrolase activity (Franková & Fry, 2013).The latter can be regarded as a special case of the former in which the acceptor substrate is H 2 O rather than an organic molecule such as a carbohydrate.
Indeed, many hydrolases can catalyse transglycosylation reactions in vitro if a high enough concentration of carbohydrate is present.On the other hand, some enzymes are 'dedicated' transglycosylases, such as most XET-active enzymes (Shi et al., 2015), unable to catalyse appreciable rates of hydrolysis.On the spectrum between 'dedicated' transglycosylases and predominant hydrolases, the type of reaction catalysed in vivo will depend on the concentration of potential acceptor substrate molecules (relative to H 2 O, which is always $55 M).The concentration of the donor (or hydrolysable) substrate will have little influence on the transglycosylation versus hydrolysis balance (except that the intended 'donor' substance may also act as an acceptor if concentrated enough).
In many of our experiments, the acceptor leading to detectable transglycosylation was a radiolabelled oligosaccharide present in the order of 0.1-2.0μM concentrations.Since we detected transglycosylation under those conditions, we conclude that the observed transglycanase or transglycosylase was capable of catalysing transglycosylation at vanishing low acceptor concentrations and was thus not purely a hydrolase.Therefore, the enzyme is likely to be capable of catalysing transglycosylation in vivo at substrate concentrations naturally occurring in the cell wall.
Indeed, the reaction rate in trans-β-mannanase assays appeared to plateau above about 0.3-0.4μM [ 3 H]Man 6 -ol (acceptor substrate) concentration, suggesting that the K M of the enzymes was at or below this range.This observation indicates a very high affinity (low K M ) of the enzyme for non-reducing terminal Man residues, again strongly supporting the view that the enzymic reaction (trans-β-mannanase) can operate in the cellular environment.
Since the acceptor substrate site is assumed to be the nonreducing terminus of the hemicellulose backbone (i.e., one site per polysaccharide molecule), it is relevant to consider the molar concentration of the acceptor substrate rather than its w/v concentration.
Calculating the in-vivo molarity of a hemicellulose depends on the w/v concentration of the polysaccharide in the cell wall and its relative molecular mass (M r ).As an order of magnitude, the value of 1 μM mentioned above seems reasonable.Therefore, we conclude that the enzymes detected in this work by use of radiolabelled acceptor substrates would be capable of operating in vivo.
In other experiments, we used non-radioactive oligosaccharides as both donor and acceptor.In these cases, the oligosaccharide (thus acceptor substrate but also, irrelevantly, the donor) concentration was 0-3 mM.The lowest concentration showing clear-cut transglycosylation under these conditions was $100 μM, which we suggest is also about within the range of polysaccharide concentrations occurring in the cell-wall matrix, depending on M r .Thus, again, the data support the proposed ability of the detected enzymes to re-structure Chara hemicelluloses in vivo.

| Dependence of trans-glycanase and transglycosidase action on donor substrate concentrations
The concentration-dependence of the donor substrate is less relevant to understanding the main role (hydrolysis versus transglycosylation) of these enzymes.However, we can record that when mannan, glucomannan, Man 6 and xylan were used as donor, the transglycosylation rate reached a plateau at roughly 1 mg/mL, and above that did not increase.Since the donor substrate can be attacked by the transglycanases at essentially any appropriate glycosidic linkage along the backbone chain, the more relevant units to quote are w/v concentration rather than molarity.The donors tested had a wide range of M r values, thus the common figure of $1 mg ml À1 (w/v; Figure 4) corresponds to a wide range of molarities ($2 μM glucomannan, $400 μM for pure β-1,4-mannan, 1000 μM Man 6 , and $ 10 μM xylan.It is clear from this that, as expected, the donor molarity is largely irrelevant, and that the 1 mg ml À1 concentration found to give maximal rates is what matters.A polysaccharide concentration of $1 mg ml À1 in the cell-wall matrix in vivo is low compared with realistic estimates of typical cell-wall compositions (hemicellulose:H 2 O ratio; Monro et al., 1976; 20% hemicellulose of total mass present in the primary cell wall, Varner & Lin, 1989).
We cannot rule out the possibility that the two pI-4.3activities are catalysed by a single protein with a relatively lax donor substrate specificity.Future work will be required to test whether these enzymes are themselves multiple and whether they have homology with related land-plant enzymes.In conclusion, we present evidence that Chara has a sophisticated arsenal of enzymic 'tools' capable of re-structuring its cell-wall matrix, thereby potentially contributing to the mechanism and regulation of cell expansion, which may represent the evolutionary origins of the wall hemicellulose re-structuring mechanisms much studied in landplants.

AUTHOR CONTRIBUTIONS
Both SCF and LF planned and designed the study, LF performed the experiments, prepared the figures and processed the data.Both LF and SCF drafted and edited the manuscript.

Figure
Figure 5A) 0.3% 1,4-β-xylan + 0.005-0.6μM [ 3 H]Xyl 6 -ol (used in Figure 5B) Chara specimens and in situ detection of transglycanase activitiesNon-fixed, unlabelled Chara specimens were photographed with a Leica camera using Leica LAS AF software.In-situ detection of transglycanase action on native algal walls was performed by use of fluorescently labelled acceptor substrate (Man 6 -ol-SR).Water-washed algal specimens were incubated in fresh algal culture medium containing 5 μM fluorescently labelled acceptor substrates for 6 h, then rinsed with ethanol/formic acid/water (5:0.5:4.5; 2 Â for 20 min).This procedure removes all unreacted fluorescent acceptor substrates.The specimens were then rinsed with 5% formic acid (gentle rotation overnight) then briefly rinsed in water and mounted onto microscopy slides containing a drop of 5% glycerol.The fluorescence was visualised with a fluorescence microscope (Leica DM2000 LED) using a standard setting for all specimens.Controls contained boiled Chara samples (105 C for 1 h), which were incubated in fluorescent substrates and subjected to the same procedure as above.
of in-vitro remodelling of their endogenous cell-wall polysaccharides as well as commercial extrinsic land-plant-derived mannan, xylan and (surprisingly) xyloglucan polysaccharides.As transglycanase and transglycosidase activities from Chara and Nitella were often higher than those in land plants, we decided to explore transglycanases in Chara in greater detail and to assess the potential of cell-wall remodelling in charophytes and compare it with XET-based xyloglucan remodelling in early-and late-diverging land-plants.Therefore, we prepared crude extracts from Chara using a 'universal' extractant (see material and methods) and screened a wide range of extrinsic land plant and algal polysaccharides, which could act as donor substrates for transglycanase activities in Chara.To help the reader navigate this paper, Microscopy of Chara vulgaris showing characteristic features of wild Chara vulgaris and laboratory-maintained cultures.Specimens (A-J) grown under laboratory axenic and (K-R) natural conditions.Light microscopy (A-O and R) and fluorescence microscopy (P and Q).(A) Chara branchlets (B; leaf-like structures) growing from the node (N) of the first whorl; (B) antheridium with red antheridial filaments (AF) and transparent shield cells (SC); BO, bracteole; BC, bract cell; (C) detail of spermatogenous filaments (F) leaving the mature antheridium (A); (D) upper and lower row of stipulodes (white and red arrows respectively) at the multicellular node; AI, axial internode of giant cells (stem-like structure); (E) oogonium (O) with nocule (NC) and corona (C); (F) zygote (Z); (G) apex (AP) of a bract cell (BC) with a gelatinous layer (GL); (H) axis cortex with spine cells (SP) and cortex cells (CC); (I) detail of a spine cell (SP) with thin cell wall (CW) and discoid chloroplasts (DC) located at the periphery of the cytoplasm; (J) detail of the bract cell showing central vacuole (CV), peripheral chloroplasts (PC), and the cell wall (CW) surrounded by a gelatinous mucilage layer (GL); (K) a branchlet whorl with antheridia (A); both branchlets (B) and main axial internodes (AI) are covered by crystals of calcium carbonate; (L) lateral branchlet apex (AP; free of calcium crystals) and calcified branchlet internode cells (BI); (M) a detail of a branchlet internode with crystals of calcium carbonate (white arrows); (N) bract cell (BC) with a 'talon'-like ending coated with a mucilage layer (GL); (O) multicellular meristematic apex of 'talon'-like branchlets with corticate (4-7) and ecorticate (1-3) segments; (P) autofluorescence of antheridium (A) and oogonium (O); BO, bractiole; (Q) fluorescence micrographs of axial internode cells (AI) with a nodal zone (N); (R) a detail of branchlet apex with short branchlet internodes (BI) and numerous meristematic nodal zones (N).Scale bar = 200 μm.F I G U R E 2 Legend on next page.FRANKOV Á and FRY 9 of 23 Physiologia Plantarum The activity was detected both with linear mannans (including Man 6 ) and with two heteromannans, optimally with glucomannan.With [ 3 H]Xyl 6 -ol as acceptor, the extracts exhibited TBX (homotrans-β-xylanase activity) with both unsubstituted xylan and heteroxylans (Figure 3C).TBX activity preferred pure unsubstituted xylan, but also acted on arabinoxylan ≥ glucuronoxylan and > > fluorescent (sulphorhodamine-labelled) xylan.A short oligomannan (Man 6 ) was a poor substrate for TBX owing the methodology used [glass-fibre blotting method, detecting mostly polymeric 3 H-labelled products (Franková & Fry, 2015, 2020)].Interestingly, substantial hetero-trans-mannanase activity (MXT) was detected with galactomannan as donor [(galacto) mannan:xylan transglycanase].Negligible activity was found withother polysaccharides as potential donor (Figure3C).

3. 4 |
Non-mechanistic trans-β-mannanase and trans-β-xylanase activities start at physiologically relevant donor substrate concentrations Figure 8).IEF revealed that Chara contains two isoforms of transglycanase acting on xyloglucanone possessing XET-like activity [the neutral isoform, pI (isoelectric point) = 7.4, in fraction 13 of Figure 8A] and one exhibiting both XET-and MXE-like (the acidic isoform, pI = 4.3, in fraction 4) and thus capable of remodelling both xyloglucan and MLG.Likewise, TBM activity from Chara was present in at least two isoformsneutral and acidic (pI = 7.0 and 4.3, in fractions 11 and 4, respectively, Figure 8B).

Figure
Figure 6B,D.To further explore how Chara enzymes can catalyse homo and hetero-transglycosylation on oligosaccharides, we prepared diverse reaction mixtures with non-radioactive Man 6 , Xyl 6 or xyloglucan oligosaccharides and radioactive Man 6 -ol, Xyl 6 -ol and XXXGol (Figure 6).Products longer than the starting material, as judged by lower mobility on TLC, must indicate the occurrence of transglycosylation; smaller products are the result of hydrolysis and/or transglycosylation reactions.These experiments potentially reveal homo-and hetero-activities of transglycosidase and transglycanase (exo-and endo-enzymes, respectively)which are dealt with individually below.
Dependence of the reaction rates on donor substrate concentrations.Trans-β-mannanase with donor substrates (A) glucomannan, (B) β-1,4-mannan, (C) mannohexaose; (D) trans-β-xylanase with donor substrate birchwood β-1,4-xylan The acceptor substrate was constant 100-200 nM [ 3 H]Man 6 -ol (A-C) or 100 nM [ 3 H]Xyl 6 -ol (D).All the activities were assayed by glass-fibre blotting method except for those on mannohexaose where TLC approach followed by quantitative 3 H-scanning was employed.Error bars represent three technical replicates (n = 3) ± SE. had been hydrolysed to [ 3 H]Xyl 5 -ol, with smaller yields of [ 3 H]Xyl 4 -ol and a trace of [ 3 H]Xyl 3 -ol (Figure 6A, lanes 2 and 4), suggesting a progressive exo-hydrolysis by β-xylosidase activity:3 Figure 6B shows experiments looking for homo-transglycosylase activities as a comparison with Figure 6D but allowing us to trace Testing hetero-and homo-transglycosylase activities on model oligomeric substrates.Different reaction mixtures containing radioactive and non-radioactive oligomers of mannan, xylan and xyloglucan were incubated with an enzyme preparation for 0.25 or 24 h, then analysed by TLC.(A,B) Fluorographic visualisation of 3 H-labelled oligosaccharides; (C,D) thymol-stained non-radioactive sugars on the same TLCs (compounds that are only radiolabelled are not visible in C and D).(A,C) Hetero-transglycosylase activities; (B,D) homo-transglycosylase activities.One of the two replicates of each TLC is shown.Repeated several times, this would generate a radioactive polymer, a process which might be rendered irreversible by the polysaccharide coming out of solution.Considering the bulk (thymolstained) sugars, we found that, as expected, Man 6 in the presence of a trace of [ 3 H]Man 6 -ol (Figure 6D track 1) behaved exactly as Man 6 in the presence of a trace of [ 3 H]Xyl 6 -ol (Figure 6C track 2; discussed above).3.6.8 | Alternative method for seeking homo-transβ-xylosidase (not found) and homo-trans-β-xylanase (found)Hydrolysis of [ 3 H]Xyl 6 -ol was almost undetectable when nonradioactive Xyl 6 was also present (Figure6B, lanes 2 and 6), indicating that the β-xylosidase action on [ 3 H]Xyl 6 -ol (detected in Figure6A, lane 2) was out-competed by the $3 mM Xyl 6 , which must therefore have been a concentration well above the K M .Nevertheless, a detectable proportion of the non-radioactive Xyl 6 was converted to Xyl 8-10 (especially Xyl 9 ), accompanied by Xyl 4 and Xyl 3 (Figure6D, lane 2 compared with lane 6).It is difficult to judge the possible production of Xyl 7 (which would suggest trans-β-xylosidase activity) since the hexasaccharide substrate was already slightly contaminated by the heptasaccharide.However, the yield of Xyl 9 exceeded that of Xyl 8 , so it appears that a trans-β-xylanase predominated, catalysing the endotransglycosylation reaction 2 Xyl 6 !Xyl 9 þ Xyl 3

3. 6
.9 | Lack of xyloglucan-acting α-xylosidase, trans-α-xylosidase, β-galactosidase and transβ-galactosidase To look for (exo) enzyme activities that might act on xyloglucan oligosaccharides, we monitored the action of Chara enzymes on a trace ($0.1 μM) of [ 3 H]XXXGol in the presence higher concentrations ($3 mM) of non-radioactive xyloglucan oligosaccharides.The nonradioactive oligomers tested were a relatively pure heptasaccharide (XXXG) and a mixture of nona-to heptasaccharides (mainly XLLG > XXLG > XXXG).Figure 6B (lanes 3,4,7 and 8) shows no reaction of the radioactive substrate (which could have been a donor or acceptor).Furthermore, Figure 6D (same lanes) shows no hydrolysis or transglycosylation of the non-radioactive substrates.Thus, the Chara extract contained no detectable α-xylosidase, transα-xylosidase, β-galactosidase or trans-β-galactosidase.It was not expected that endo-enzyme activities capable of acting within xyloglucan oligosaccharides would be present as these are not known from land plants and because Chara lacks biochemically detectable xyloglucan.3.7 | Kinetics of trans-β-mannosidase and transβ-mannanase activities from CharaAs trans-β-mannosidase and trans-β-mannanase were major wallpolymer remodelling activities detected in Chara extracts (Figure6B,C,D), we examined the kinetics of these enzymes.We tested 0-3027 μM Man 6 as the donor and 0.1 μM [ 3 H]Man 6 -ol as acceptor, which allowed us to track the formation of the transglycosylation products (FigureS4).The radioactive products were separated by TLC and quantitatively scanned.After 24 h incubation, the exotransglycosylation product [ 3 H]Man 7 -ol was detectable at the relatively low donor substrate concentration of 101 μM (Figure7Band 8D).With higher donor concentrations (≥ 505 μM), the 24-h yield of [ 3 H]Man 7 -ol had peaked, and there was little difference between 505 and 3027 μM donor.

3. 8 |
Transglycanase activity utilising oligomannan as acceptor substrate detected in situ acting on endogenous wall polysaccharides An in-situ approach performed by fluorescence microscopy (Figure 10) revealed that Man 6 -ol-SR can serve as acceptor substrate for endogenous transglycanase(s) acting on intrinsic donor polysaccharides in Chara cell walls.Fluorescent signals were detected in the multicellular meristematic apex of axillary whorls, in the side-and cross-wall between two adjacent branchlets and especially in the branchlet cells.Judged by the fluorescent signal from wall-incorporated Man 6 -ol-SR, the detected transglycanase(s) appear to 'cut and paste' wall polysaccharide(s), with oligomannan as acceptor substrate, in most Chara cells.
Kinetic properties of oligomannan-remodelling transglycosylases extracted from Chara vulgaris.(A) Time-dependent action of Chara transglycosylases and hydrolases on 0.1 μM [ 3 H]Man 6 -ol (acceptor) mixed with 3 μM Man 6 (donor).Trans-β-mannanase products formed in the first 4 h are indicated with red arrows; trans-β-mannosidase products formed later have pink arrows.(B) Effect of substrate concentration on products formed by Chara transglycosylases and hydrolases after 24 h.Dried [ 3 H]Man 6 -ol was re-dissolved at various concentrations (0-3027 μM) of Man 6 , mixed with Chara extract and incubated at 22 C for 24 h.(C), (D) The distribution of radioactivity between differently sized products on the TLCs shown in (A) and (B).The histogram bars represent the average of two replicates.The values were corrected to heatinactivated controls shown in Figure S4.(E) Controls with heat-inactivated enzyme for (A); (F) controls with heat-inactivated enzyme for (B).
(i.e., exo-) activities assayedtrans-α-xylosidase and transβ-galactosidasewhen xyloglucan-derived oligosaccharides were tested as potential substrates.This tallies with the consensus that Chara lacks land-plant-like xyloglucan and the enzymes specifically adapted to restructure it.
In-situ action of trans-glycanase(s) on native endogenous polysaccharides of Chara vulgaris.Fluorescence and corresponding bright-field micrographs of specimens fed with 5 μM fluorescently tagged Man 6 -ol-SR, to which Chara cell-wall polysaccharides became grafted.(A) Transglycanase action in the basal node (BN) and multicellular meristematic apex (MA) of axillary whorl of 'talon'-like branchlets (WB).(C) The third whorl of branchlets of the main axis.(E) Fluorescent side walls (SW) and cross-wall (CrW) between two adjacent branchlet cells (BC).(G) A detail of Man 6 -ol-SR-tagged side cell wall (SW) of elongated branchlet cell.(I) A detail of a fluorescent branchlet apex (AP) and branchlet cells (BC).Controls (B,D, F,H,J) show pre-boiled Chara (thus containing heat-denatured transglycanases) incubated with 5 μM Man 6 -ol-SR.Scale bar = 250 μm.4)-β-D-Mannans are major hemicelluloses of the Chara cell wall, so it seemed more likely that trans-β-mannanases and transβ-mannosidases might be present to allow their restructuring in vivo.
This is the first study to explore in depth the transglycanase and transglycosidase (endo-and exo-, respectively) activities likely to act on the cell walls of Chara, a 'late-diverging' charophytic alga relatively closely related to the land-plants.We show that Chara possesses activities capable of re-structuring two of its major hemicelluloses: mannan-based (by endo-and exo-transglycosylation) and xylan-based (by endo-transglycosylation).There was also evidence for mannan-to-xylan heterotransglycosylation (but not xylan-to-mannan transglycosylation), thus providing Chara with the capacity to produce mannan!xylan 'hybrid' polysaccharides (Figure2).More surprisingly, we found enzyme activities that can catalyse xyloglucan-to-xyloglucan endo-transglycosylation and MLG-to-xyloglucan endo-transglycosylation, suggesting that Chara possesses hemicelluloses functionally resembling xyloglucan and/or MLG, but differing in sufficient structural features that these polymers cannot be detected by chemical analysis using techniques which very sensitively reveal xyloglucan and MLG in land-plant cell walls.Evidence for traces of epitopes detected by xyloglucan-and MLG-'specific' antibodies(Mikkelsen et al., 2021) is not conclusive because the precise range of epitope(s) recognised by these antibodies can never be comprehensively defined.Results entirely based on antibodies are not always reliable, as reported byPfeifer et al. (2023), who found that antibodies raised against arabinogalactan-protein (AGP) epitopes cross-reacted with unconventional methylated galactan in four Chara species that lack AGPs.Future work will be required to demonstrate to what extent Chara possesses xyloglucan-and MLG-like hemicelluloses.Chara certainly did not exhibit any trans-glycosidase activities capable of acting on fragments of landplant xyloglucan.Based on the predominant detected homo-and hetero-TBM > TBX and XET/MXE-like activities, we assume that mannan-, xylan-and xyloglucan-like polysaccharide-remodelling is well represented in Chara and thus may have originated prior to the divergence of the Charophyceae from their sister clade (Li et al., 2020) which contains the Zygnematophyceae, Coleochaetophyceae and land-plants.
in Equisetum that can use MLG or cellulose as the donor substrate and xyloglucan as the acceptor, creating a 'hybrid product' that is MLG-xyloglucan or cellulose-xyloglucan.Such enzyme activities can be named, for exam- Further work on Chara hemicelluloses is required to identify the native substrate of the 'XET-like' and 'MXE-like' activities detected here.Nevertheless, our conclusion is that Chara shares with the land-plants certain fundamental enzymic tools involved in cell-wall restructuring.These tools were also found in other representatives of