Elucidation of new pathways in normal and perturbed lignification

It seems incomprehensible that new structures in lignins and new pathways in lignification remain to be discovered after all this time. With new technologies and the improved understanding of the process that comes from new ways to perturb lignification, such discoveries continue to be made, however. Some may be of rather academic interest only, while others underlie important new approaches to improved utilization of cell walls. An assault on the existing theory of the process of lignification has lead not to its demise but its strengthening as it continues to account for old and new observations. That said, aspects of lignin structure still hold many biosynthetic mysteries. Although it is time to summarily discard the proposed new hypothesis/theory for the absolute control of lignin primary structure, lignification is complex and an all-encompassing theory remains elusive. Nevertheless, lignins and lignification are becoming increasingly well understood. New details on the coupling processes are being revealed. The β–1-pathway produces novel spirodienone structures that are, now that we know how to look for them, surprisingly easy to detect in lignins. What used to be the well-accepted β–β-coupling pathway to produce resinols is now being shown to proceed by alternate pathways that give other structures in lignins. β–β-Coupling when the γ-alcohol is acylated in natural lignins is revealing a range of fascinating structures and, more importantly, that acylated lignins derive from acylated monolignols, which therefore must be recognized as authentic lignin monomers. To date, the benefit of such acylation to the plant is unknown, and the components and pathways remain unexploited. A small range of monomers are now known to substitute for the conventional monolignols in various natural and transgenic plants. Substitution appears to be most successful when the replacement monomer behaves, in its chemical radical coupling and cross-coupling reactions, like a normal monolignol. The post-coupling reactions that may be altered by the different functionality on the monomer seem to have less impact. Thus massive changes in the lignin structure occur in COMT-deficient plants when 5-hydroxyconiferyl alcohol substitutes for sinapyl alcohol, for example — the coupling reactions are analogous, but post-coupling steps produce novel benzodioxane structures that drastically change the lignin structure. Such biosynthetic malleability functions well for the plant, but also provides significant opportunities for engineering the polymer. Already it has been demonstrated that natural and industrial processes ranging from ruminant digestibility to chemical pulping can be both positively and negatively impacted by alterations to lignin, and we are beginning to get a mechanistic handle to explain these effects. Future work should reveal opportunities beyond the interesting deviations achieved by upand down-regulating genes on the monolignol pathway to date. INTRODUCTION In addition to the numerous reviews on lignins and lignification, we have recently published two reviews, one dealing more with the biosynthetic pathways and transgenics, and one concentrating more on aspects of structure. Consequently, this abstract will NOT be extensively referenced — references can be found in those two reviews, which are available electronically (including from the Dairy Forage Research Center’s web site at http://www/dfrc.ars.usda.gov). The aim here is to provide a commentary on some of the recent revelations regarding lignin structure and the process of lignification, and to note some of the implications for fiber utilization. THE MONOLIGNOL BIOSYNTHETIC PATHWAY Various aspects of the pathway by which the monolignols (the primary lignin monomers, p-coumaryl, coniferyl, and sinapyl alcohol) are biosynthesized have been clarified. Briefly, the pathway through the previously entertained metabolic grid has been somewhat streamlined, particularly with outstanding work from several groups. In particular, 5-hydroxylation and methylation of the 5-OH (to ultimately produce sinapyl alcohol in angiosperms) has now been shown to occur principally at the aldehyde level, Fig. 1. Taken together with discoveries regarding 3-hydroxylation (see below), these data argue that none of the C3 and C5 substitutions of the aromatic ring take place at the cinnamic acid level in monolignol biosynthesis. Evidence suggests that wall ferulate esters may also be produced via the aldehydes rather than directly from their corresponding acids. For a synthetic organic chemist, one of the most pleasing recent discoveries concerns the 3-hydroxylation on the pathway from p-coumaric to caffeic acid. The pathway was always considered to proceed from p-coumaric acid via the CoA-thioester to caffeoyl-CoA (Fig. 1). But every organic chemist knows that you must be foolhardy to go and activate an acid, say as its acid chloride (c.f. the biological analog, the thioester), and then not directly utilize that group but go off and do some other modification elsewhere on the molecule, expecting your activated acid to remain intact. Of course plants are smarter than organic chemists, and perhaps the enzymes for hydroxylation require a rather common SCoA binding site. But it was always agonizing to have to teach this step in the pathway. It has been a delight to recently find, in arabidopsis but likely in all plants producing syringyl/guaiacyl lignins, that the SCoA ester is merely an intermediate and that the substrate for the hydroxylation is in fact a shikimic or quinic acid ester — see the expansion in Fig. 1. An HCT enzyme takes p-coumaroyl-CoA and makes the shikimate ester which then becomes hydroxylated (via the C3H enzyme); the resultant caffeoyl shikimate is then activated back to the SCoA ester ready for the next transformations — methylation and reduction to the aldehyde. Although this process involves more steps, it is very much analogous to the use of protecting groups by the synthetic chemist. Nature makes sense again! Rather intriguingly, shikimate is a crucial intermediate in the pathway from carbohydrates to the aromatic pools (including the amino acid phenylalanine, and the monolignols) in the plant. It is interesting that shikimate is used again in this monolignol pathway, essentially as a temporary protecting group. A practical benefit of the discovery of the new enzyme required on the pathway from p-coumarate to coniferyl and sinapyl alcohols is that there is now an additional target that can be used to down-regulate the biosynthesis of these two primary monolignols. Thus downregulating either C3H or HCT can both produce similar effects. To the extent that inducing plants to produce a greater percentage of their lignins from p-coumaryl alcohol (rather than coniferyl and sinapyl alcohols) may have some advantages, there are now two distinct target genes for achieving this. LIGNINS ARE RACEMIC POLYMERS Although there is nothing new in the above heading, this concept bears bringing up once more as it continues to be overlooked and misunderstood. And yet this facet is crucial for understanding the biosynthesis of the polymer and its consequent chemical and physical properties. In 1991 I tried to make the point that a regular but racemic β-ether polymer (not lignin!) with 110 units had about the same number of isomers as there were atoms in the Galaxy. The molecular size was artificial and the statement somewhat silly, and in fact was a couple of orders of magnitude out. It has been repeatedly ridiculed by one scientist. Unfortunately, it remains true!! As any first year organic chemist knows, when you have a molecule with n optical centers you have 2 possible optical isomers, and can have half that number of physically distinct isomers. Thus, with a β-ether 110-mer, there are actually only 218 optical center (not the 220 I originally assumed), and therefore 2 physically distinct isomers. This is about 4x10 — an astronomically large number in line with the original claim. Since the racemic nature of the polymer continues to be underappreciated, we recently tried to make the point again using models for lignins from normal poplar and a COMT-deficient transgenic. These are only models, but represent the compositional data available in a structure of 20 units in a manner similar to the famous models by Adler and the more recent advancement by Brunow. The two models are shown in Fig. 2; these and the molecular models themselves can be found on our website (http://www.dfrc.ars.usda.gov/LigninModels. html). For the single wild-type lignin structure drawn, without considering isomers based on altering the linkage sequence but merely the stereochemistry of each unit, and taking into account the isomer restriction in units such as the phenylcoumaran and dibenzodioxocin units (where the number of isomers is reduced by a factor of two from those theoretically possible), this simple little model 20-mer still has an astounding 17 billion isomers. The COMT-deficient lignin model has fewer isomers, a mere 134 million since it has more of the restricted phenylcoumaran and dibenzodioxocin structures (see Fig. 2 caption). Any flaw in the above will hopefully be addressed by those choosing to ridicule these notions further! The major point of course is that lignins are produced by a process which creates two new optical centers with each coupling reaction when the coupling is at the favored βposition, and there really appears to be no stereo-control over their creation. Thus, unlike in the case reported for Fig. 1. Simplified monolignol biosynthetic pathway. Note that, at least in Arabidopsis, the p-coumaric acid to caffeic acid conversion involves more steps than previously thought, as shown in the expansion. Traditional names for the enzymes involved are: 4CL = 4-coumarate:CoA ligase; C3H = pcoumarate 3-hydroxylase; HCT = p-hydroxycinnamoyl-CoA: quinate shikimate p-hydroxycinnamoyltransferase; CCoAOMT = caffeoyl-CoA O-methyltransferase; CCR = cinnamoylCoA reductase; F5H = ferulate 5-hydroxylase; CAld5H = coniferaldehyde 5-hydroxylase; COMT = caffeic acid Omethyltransferase; AldOMT = 5-hydroxyconiferaldehyde Omethyltransferase; CAD = cinnamyl alcohol dehydrogenase. Many of these enzymes are no longer thought to function on the substrates for which they were originally named. OH O O HO HO O OH OH O O HO HO OH O OH