Cross-linking of extracellular feruloyl-polysaccharides in maize by oxidative coupling and the formation of dehydrodiferulate-sugar ether bonds

Test Hypothesis 1: Cross-linking of extracellular polysaccharides is initiated by peroxidase plus hydrogen peroxide, causing oxidative coupling of feruloyl-polysaccharides by forming dehydrodiferulates.
Soluble extracellular polysaccharides (SEPs) were successfully radiolabelled in cultured maize cells with 3H specifically in their pentose residues and/or with 14C specifically in their feruloyl residues. Cross-linking was assayed on a relatively high-throughput basis by size-fractionation on mini-columns of Sepharose CL-2B. Freshly secreted SEPs were eluted at Kav ~0.5 ( 1 MDa), cross-linked SEPs at Kav 0.0 (>20 MDa).
The cross-linking of [3H]- and [14C]SEPs in vivo was shown to be dependent on endogenous H2O2. For example, the process was blocked by the peroxide-scavenger potassium iodide (Fig. 1). The dependency on endogenous H2O2 indicates that the apoplastic enzyme responsible (if any) must be a peroxidase. That an enzyme was in fact necessary was shown in vitro: maize culture filtrates successfully continued to cross-link SEPs only if not heat-denatured. [3H]- and [14C]SEPs present in boiled culture filtrates resumed cross-linking, however, if native (= not boiled) non-radioactive culture filtrate was added. Thus SEP cross-linking was indeed initiated by peroxidase plus hydrogen peroxide, as postulated in Hypothesis 1.
Interestingly, the proposed role of ferulate dimerisation in the cross-linking process could not be supported: very little increase in [14C]diferulate was observed when [14C]SEPs were cross-linked either in vivo or in vitro. On the contrary, the existing (low) [14C]diferulate levels usually decreased further concomitantly with the oxidation of [14C]ferulate residues. Instead of dimers, abundant 14C-labelled saponifiable (= ester-linked) products larger than dimers were formed (trimers, tetramers and probably also larger products).
Thus, the initial hypothesis was supported and modified.

Test Hypothesis 2: Initial cross-linking of esters is quickly reinforced by formation of p-hydroxybenzyl-ether bonds to neighbouring polysaccharides, which may occur via quinone-methide intermediates and which may require a novel lyase or ‘dirigent’.
We consistently found that the initial cross-linking of SEPs is quickly followed by the development of alkali-resistant cross-links both in vivo and in vitro, interpreted as ether bonds. Both cross-linking processes required H2O2. The stronger bonds, formed slightly later than the initial saponifiable cross-links, were resistant to 0.1 M NaOH at 20°C for 24 h, and were thus more alkali-resistant than any known carboxy-esters. They were cleaved by progressively more severe alkaline conditions (concentration, temperature), and were almost completely cleaved by 6 M NaOH at 37°C. Although this is a severe alkali treatment, it did not cause any detectable backbone cleavage of arabinoxylan or xyloglucan chains, and thus its action was deduced to be on the cross-links, not the polysaccharides’ backbones.
Cross-linking of SEPs could not be achieved by use of horseradish peroxidase (HRP) + H2O2; in fact, HRP strongly inhibited the in-vitro cross-linking of SEPs that was catalysed by extracellular maize peroxidases (Fig. 3b–d). We therefore suggest that the peroxidase activities secreted by maize cells possess unique catalytic features not shared by HRP. We have conducted a preliminary characterisation of this peroxidase activity, showing for example that it is more heat-labile than other maize peroxidases. We were not able to separate the peroxidases responsible for initial (alkali-labile) cross-linking from any postulated protein (lyase or dirigent) that might be responsible for ether bond formation. We therefore cannot support the hypothesis that a novel lyase or ‘dirigent’ activity is present in maize cell cultures. More probably, the ether bond formation is non-enzymic.

Test Hypothesis 3: The initiation of cross-linking can be controlled by any or all of (i) peroxidase activity, (ii) hydrogen peroxide supply, and (iii) the concentration and degree of feruloylation of feruloyl-polysaccharides.
We confirmed all three of these possibilities, and added a fourth — the concentration of an endogenous, low-Mr, extracellular antioxidant which inhibits peroxidase action and in particular SEPs cross-linking. Cultured maize cells secreted newly synthesised (= 3H- or 14C-labelled) SEPs at all physiologically active stages of the culture cycle (0–13 d after sub-culture), but did not cross-link them until a particular time-point (often between 8 and 10 d after sub-culture, with slight variation between experiments), at which point the cross-linking occurred very abruptly for ~1 day. After this cross-linking burst, the rate of ‘current’ cross-linking diminished; the ‘cumulative’ cross-linking plateaued, indicating that the cross-links once formed were permanent (Fig. 2).

Adequate concentrations of peroxidase were shown to be essential for SEPs cross-linking, which could be blocked by heat denaturation. However, upon sub-culture, the maize cells very quickly (within ~1–2 h) secreted sufficient peroxidase for rapid SEPs cross-linking. We therefore do not consider peroxidase levels likely to be a ‘limiting factor’ in controlling the cross-linking process in vivo.
There was a brief peak of cross-linking activity late in the culture cycle, which was inhibited by the H2O2 scavenger KI (Fig. 1) and was associated with a transient period of elevated [H2O2] in the medium, as estimated by the xylenol orange assay. We deduce that H2O2 supply is a main factor controlling cross-linking. However, exogenous H2O2, at concentrations sufficient for enzymic cross-linking in vitro, did not enable ‘premature’ cross-linking in vivo (Fig. 1). As a side-line to the main project, we also conducted a detailed study of the extracellular generation of H2O2 in cultured Picea cells and again concluded that H2O2 supply is a tightly controlled and easily measured process (Kärkönen & Fry, 2006).
The presence of feruloyl groups on SEP chains was shown to be essential for cross-linking. SEPs de-feruloylated by NaOH was no longer susceptible to by cross-linking in maize cultures in vivo or in the presence of maize peroxidase + H2O2 in vitro.
The rate of cross-linking was closely correlated with the concentration in the medium of an endogenous, low-Mr, hydrophilic, heat-labile, extracellular compound, as previously reported (Encina & Fry, 2005). We tested the possibility that this compound was ascorbate, but discounted this hypothesis because the substance was not destroyed by ascorbate oxidase; it therefore remains unidentified and further studies on this natural regulator of apoplastic oxidative coupling of feruloyl-polysaccharides will be very valuable.

The results will provide important new information on the repertoire and control of potential mechanisms by which structural polymers may become cross-linked in the intact wall. Soluble extracellular polysaccharides have long been proposed as useful models of the chemical structures of wall polymers, but little use has been made of their ability to undergo metabolic reactions that model those normally occurring in the wall.
The major experiments were: (1) Feeding of [3H]arabinose or [14C]cinnamic acid to maize cell cultures; (2) incubation of [3H]- and [14C]SEPs with maize cell cultures; (3) use of maize culture filtrates, [3H]- and [14C]SEPs, and H2O2 to explore cross-linking in a cell-free system in vitro.

Each of the three hypotheses will be investigated independently of the other two, so the project as a whole would not be fatally flawed by unforeseen difficulties in any one of them. We will explore in detail the mechanisms and control of soluble extracellular polysaccharide cross-linking in non-lignified maize cell cultures. To probe the validity of the above hypotheses, and if necessary to suggest alternative hypotheses, we will seek answers to the following questions, in order of priority:
(1) What range of reactions do the feruloyl groups undergo when a feruloyl-polysaccharide becomes cross-linked?
(a) Esterified products still releasable by mild alkali.
The great majority (>98%) of the [14C]feruloyl groups disappeared during the SEPs cross-linking process, both in vivo and in vitro. About 40% of the [14C]feruloyl groups underwent oxidative coupling to form triferulates and higher oxidatively coupled oligomers. Like feruloyl residues themselves, these products remain releasable from the polysaccharides by mild alkali, and (after acidification) they can still be partitioned into ethyl acetate. This indicates that they had not become bonded to hydrophilic moieties such as carbohydrates.

(b) Ether-like products resisting mild alkali.
The remaining ~60% of the [14C]feruloyl groups that disappeared during the SEPs cross-linking process, both in vivo and in vitro, were recovered in the hydrophilic fraction (partitioning into water, not ethyl acetate) after saponification and acidification. The majority of this material (47% of the total 14C-labelled material) was found to remain of extremely high Mr even after alkali treatment, i.e. were covalently bonded to polysaccharides via ether-like bonds. A small proportion (~12% of the total 14C) was recovered in the form of low-Mr hydrophilic compounds that were mobile on paper chromatography, likely to be ferulate derivatives that had been ester-bonded to polysaccharides and also ether-bonded to small hydrophilic compounds such as monosaccharides and oligosaccharides.

(2) What enzyme activities are involved in the cross-linking?
(a) Initial oxidative coupling of feruloyl-polysaccharides.
The reaction is H2O2-dependent and the enzyme involved is a peroxidase secreted into the cell culture medium, probably requiring certain specific peroxidases with properties not mimicked by horseradish peroxidases (we tested Sigma’s HRPs II and VI). This has been a popular hypothesis, but little previous work had addressed the possibility that laccase activity was responsible (i.e., using O2 instead of H2O2); or the possibility that the cross-linking was non-enzymic, possibly catalysed by apoplastic Cu2+ or dependent on extracellular reactive oxygen species other than peroxide. These questions were resolved in the present work because of our choice to work in vivo, monitoring enzyme action — as opposed to in vitro, monitoring enzyme activity.

(b) Formation of ether-like cross-links.
The formation of ether-like cross-links is also dependent on a prior supply of H2O2, and requires the above-mentioned enzyme for its initiation; but the process is slower than the initial cross-linking reaction and may not itself require a protein catalyst.

(3) Is the presence of a phenolic group essential for a polysaccharide to form ‘ether-like’ cross-links extracellularly?
Phenolic groups do appear to be essential for extracellular ether-like cross-links to form. This observation necessitates a reappraisal of the idea that quinone-methide-type feruloyl dimers can form p-hydroxybenzyl-ether bonds to neighbouring non-feruloylated polysaccharides. The simplest iteration of the hypothesis is that quinone-methide-type feruloyl dimers can form intra-molecular p-hydroxybenzyl-ether bonds to neighbouring sugar residues within the same polysaccharide chain.

(4) What controls the formation of cross-links in SEPs? — feruloyl content, peroxidase activity or H2O2 concentration?
Cross-linking of SEPs in vitro requires non-denatured culture filtrate and H2O2, which evokes gradual (time-dependent) cross-linking (Fig. 3a). Peroxidases are present in the medium throughout the culture cycle. The cross-linking activity appears to be due to an H2O2 burst, supported by the absence of low-Mr anti-oxidants. A controlling role for feruloyl content cannot be ruled out, because maize cultures appear to vary in the degree of feruloyl-esterification of their extracellular polysaccharides; and deliberate removal of feruloyl groups very effectively blocked further
cross-linking of SEPs.

We will study certain characteristic structural elements of plant tissue that govern its texture — namely the cell walls of agriculturally and nutritionally valuable cereals and grasses.

Grasses and cereals have cell walls that differ chemically in several fundamental ways from those of dicotyledons (‘broad-leaved plants’) such as legumes, rape and fruits. However, the majority of biochemical studies on plant cell walls have dealt with dicotyledons. It is therefore timely, and commercially appropriate, to focus some basic work on the unique features of grasses and cereals: major world crops.

Although the project deals with fundamental questions and is not targeted at any one specific crop, the potential outcome of the work is an ability to improve the quality of food/fodder of cereal/grass origin by modification of structural (cell wall) components.

The project deals with a biochemical process — the cross-linking of certain structural polysaccharide molecules of the cell wall via ferulate, a non-carbohydrate side-chain. Such cross-linking is thought to be centrally involved in several biologically/commercially important features of grasses and cereals, e.g. controlling cell expansion and thus plant growth; restricting susceptibility to penetration by pathogens and thus improving disease-resistance; and promoting cell–cell adhesion and thus consolidating the cereal plant’s tissues.

The work aims to describe the chemical nature of these cross-linking processes, especially the formation of novel but poorly characterised cross-links that are believed to be ether bonds between ferulate derivatives and the sugar groups of the polysaccharides. A major question is whether the formation of this type of ether bond is initiated, or at least steered, by novel proteins (enzymes of the lyase class, or dirigent proteins). An ability to assay these proteins would open the way to the identification of their genes, and thus the future ability to modify their action genetically or biotechnologically.

We studied certain characteristic structural elements of plant tissue that govern its texture — namely the cell walls of the agriculturally and nutritionally valuable plant, maize. To simplify the study of basic underlying biochemical processes, we mainly used tissue cultures rather than intact maize plants.

* Grasses and cereals in general have cell walls that differ chemically in several fundamental ways from those of dicotyledons (‘broad-leaved plants’) such as legumes, rape and fruits. However, the majority of biochemical studies on plant cell walls had dealt with dicotyledons. It was therefore timely, and commercially appropriate, to focus some basic work on the unique features of maize: a major world cereal.

* Although the project deals with fundamental questions and is not targeted at any one specific crop, the potential outcome of the work is an ability to improve the quality of food/fodder of cereal/grass origin by modification of structural (cell wall) components.

* The project dealt with a biochemical process — the cross-linking of certain structural polysaccharide molecules of the maize cell wall via ferulate, a non-carbohydrate side-chain. Such cross-linking was thought to be centrally involved in several biologically/commercially important features of grasses and cereals, e.g. controlling cell expansion and thus plant growth; restricting susceptibility to penetration by pathogens and thus improving disease-resistance; and promoting cell–cell adhesion and thus consolidating the cereal plant's tissues.

* The work aimed to describe the chemical nature of these cross-linking processes, especially the formation of novel but poorly characterised cross-links that were believed to be ether bonds between ferulate derivatives and the sugar groups of the polysaccharides. We showed that the cross-linking process as it occurs in living plant cells [many previous studies had been limited to in-vitro studies using purified polysaccharides and enzymes] requires an enzyme (peroxidase, not laccase as had previously been possible) and a reactive oxygen species (hydrogen peroxide, which we confirmed to be formed naturally in maize cell walls).

* The ferulate groups attached to the cell-wall polysaccharides were indeed responsible for the ability of the polysaccharides to cross-link to each other, in a wall-strengthening process; however, we found that the several ferulate groups became coupled to each other, forming oligomers rather than dimers (which had previously attracted most attention). Furthermore, we found that 'coupled' ferulate did not stop at that stage: it went on to form very stable ether bonds to polysaccharide chains, which could account for the tremendous environmental resilience of cereal cell walls. The ether bonds were not formed to neighbouring ferulate-deficient polysaccharide chains, but only to chains which themselves possessed ferulate groups; this contradicted one of our initial assumptions and will be an interesting topic for future investigations.

* We had thought that the formation of these stable ether bonds might depend on the presence of a new cell-wall-located enzyme; however, the results argued in favour of ether bond formation being non-enzymic -- the inexorable result of the chemistry of the initially formed ferulate coupling-products.

* We developed several highly simplified procedures for studying these cell-wall cross-linking reactions in living cells, enabling us and future workers on cereal biochemistry to detect and measure these reactions in living cereal tissues.

* Knowledge of the basic biochemistry of cell-wall cross-linking reactions will enable these processes to be understood and in the future manipulated by chemical and/or genetic means, potentially improving the performance and nutritional value of cereal crops.

* In the course of the work, the laboratory also collaborated on several related projects, which have resulted in 'added value' publications in related fields of plant science.