Modeling Lignin Polymerization. I. Simulation Model of Dehydrogenation Polymers

Lignin is a heteropolymer that is thought to form in the cell wall by combinatorial radical coupling of monolignols. Here, we present a simulation model of in vitro lignin polymerization, based on the combinatorial coupling theory, which allows us to predict the reaction conditions controlling the primary structure of lignin polymers. Our model predicts two controlling factors for the β-O-4 content of syringyl-guaiacyl lignins: the supply rate of monolignols and the relative amount of supplied sinapyl alcohol monomers. We have analyzed the in silico degradability of the resulting lignin polymers by cutting the resulting lignin polymers at β-O-4 bonds. These are cleaved in analytical methods used to study lignin composition, namely thioacidolysis and derivatization followed by reductive cleavage, under pulping conditions, and in some lignocellulosic biomass pretreatments.Lignins are aromatic polymers that are predominantly present in secondarily thickened cell walls. These polymers make the cell wall rigid and impervious, allowing transport of water and nutrients through the vascular system and protecting plants against microbial invasion. Lignins are heterogeneous polymers derived from phenylpropanoid monomers, mainly the hydroxycinnamyl alcohols coniferyl alcohol (G-monomer) and sinapyl alcohol (S-monomer) and minor amounts of p-coumaryl alcohol (H-monomer). These monolignols differ in their degree of aromatic methoxylation (-OCH₃ group; Fig. 1). The resulting units in the lignin polymer are the guaiacyl (G), syringyl (S), and p-hydroxyphenyl (H) units. They are linked by a variety of chemical bonds (Fig. 2) that have different chemical properties (Boerjan et al., 2003; Ralph et al., 2004; Vanholme et al., 2008).Open in a separate windowFigure 1.Chemical structures of three monolignols. A, H-monomer (p-coumaryl alcohol). B, G-monomer (coniferyl alcohol). C, S-monomer (sinapyl alcohol). G- and S-monomers are considered in our simulations. The G-monomer is methoxylated (-OCH₃ group) on position 3, and the S-monomer is methoxylated on positions 3 and 5.Open in a separate windowFigure 2.Chemical structures resulting from the possible bonding between two monomers (A) or a monomer and the bindable end of an oligomer (B). X and Y in the monomers denote the absence (for a G-unit) or presence (for an S-unit) of a methoxyl group at position 5 (see Fig. 1). The red line indicates the bonds generated by couplings of the B position and B, 4, or 5 position.Lignification is the process by which monomers and/or oligomers are polymerized via radical coupling reactions and typically occurs after the polysaccharides have been laid down in the cell wall. Lignin composition varies among cell types and can even be different in individual cell wall layers (Ruel et al., 2009). Lignin composition is also influenced by environmental conditions; for example, lignin in compression wood is enriched in H-units (Timell, 1986). Hence, both developmental and environmental parameters influence the composition and thus the structure of the lignin polymer (Boerjan et al., 2003; Ralph et al., 2004).Lignin is one of the main negative factors in the conversion of lignocellulosic plant biomass into pulp and bioethanol (Lynd et al., 1991; Hill et al., 2006). In these processes, lignin needs to be degraded by chemical or mechanical processes that are expensive and often environmentally polluting. Hence, major research efforts are devoted toward understanding lignin biosynthesis and structure. It has already been shown that reducing lignin content and modifying its composition in transgenic plants can result in dramatic improvements in pulping efficiency (Pilate et al., 2002; Baucher et al., 2003; Huntley et al., 2003; Leplé et al., 2007) and in the conversion of biomass into bioethanol (Stewart et al., 2006; Chen and Dixon, 2007; Custers, 2009). These altered biomass properties are related to the alterations in lignin composition and structure in terms of the frequencies of the lignin units and the bond types connecting them and possibly also their interaction with hemicelluloses (Ralph et al., 2004; Ralph, 2006).To study the parameters that influence lignin structure, lignin polymerization has been mimicked in vitro by experiments with dehydrogenation polymers (DHPs; Terashima et al., 1995). Indeed, lignification can be mimicked by oxidizing monolignols using a peroxidase, such as horseradish peroxidase (HRP), and supplying its cofactor hydrogen peroxide, producing synthetic DHP lignins. Monolignol oxidation can also be achieved without enzymes (e.g. by using transition metal one-electron oxidants, such as copper acetate). Some of these biomimetic DHPs have been suggested to be better models for wood lignins than HRP-generated DHPs (Landucci, 2000).In DHP experiments, the monolignols are either added in bulk (Zulauf experiment) or dropwise (Zutropf experiment) to the reaction mixture, yielding lignin polymers with very different bond frequencies (Freudenberg, 1956). Zutropf experiments approach the in vivo formation of lignin, which depends on the slow introduction of monolignols into the wall matrix via diffusion to the site of incorporation (Hatfield and Vermerris, 2001). Because the exact reaction conditions are known, such in vitro experiments have provided insight into the lignification process in planta. In this way, numerous factors were shown to influence lignin structure, including the relative supply of the monolignols, the pH, the presence of polysaccharides, hydrogen peroxide concentrations, and cell wall matrix elements in general (Grabber et al., 2003; Vanholme et al., 2008).Computer simulations of lignin polymerization can help explain and predict lignin structure from low-level chemical kinetic factors, including subunit-coupling probabilities and monolignol synthesis rates. Such models are helpful in explaining the mechanism behind a range of controlling factors identified in the experimental work, including (1) the ratio of coniferyl versus sinapyl alcohol monolignols, (2) the monolignol supply rate, and (3) the abundance of alternative monomers present during lignin biosynthesis in mutants and transgenics. Thus, computer models will also help in suggesting new targets for controlled lignin biosynthesis.Here, we propose a simulation model of synthetic lignin polymerization that is based upon an emerging consensus from a variety of observations and derives from a series of previous models of lignin polymerization (Glasser and Glasser, 1974; Glasser et al., 1976; Jurasek, 1995; Roussel and Lim, 1995). Our model uses a symbolic grammar to describe a constructive dynamical system (Fontana, 1992) or a rule-based system (Feret et al., 2009) in which it is not necessary to define all possible products in advance. We assume that G- and S-monomers and newly formed oligomers couple in a well-mixed medium, depending on coupling rules and experimentally measured coupling probabilities. To develop the model, we have used information from DHP experiments rather than natural lignins, as they are formed in a well-mixed medium and their reaction conditions are well known (e.g. the influx rate of monomers). Using information from natural lignin would have further complicated our model, as the structures of natural lignin polymers are influenced by many factors, including the possible involvement of dirigent proteins (Davin and Lewis, 2005), steric hindrance by polysaccharides, spatiotemporal regulation, and modifications during isolation procedures (Boerjan et al., 2003; Ralph et al., 2004).Using our simulation models, we study how putative controlling factors of lignin primary structure, including the influx rate of monomers and the relative amount of S-monomers, affect in silico lignin synthesis, and we compare our predictions with in vitro experiments. To predict the degradability of lignins formed in our simulations, we apply an in silico thioacidolysis, which cleaves the polymers at their β-O-4 positions. This simulates the molecular action of two of the most used methods to analyze lignin composition, thioacidolysis (Lapierre, 1993; Baucher et al., 2003) and derivatization followed by reductive cleavage (Lu and Ralph, 1997). The G+S-monomer yield is often taken as a reflection of the fraction of units bound by β-O-4 bonds. Cleavage of β-O-4 bonds is also the most important reaction in kraft pulping of wood (Baucher et al., 2003). The model predicts from first principles (1) that DHP lignins formed under Zutropf conditions have a higher β-O-4 content than those formed under Zulauf conditions, (2) that DHP lignins formed with high S content have a higher β-O-4 content than those formed with high G content, and (3) that a higher β-O-4 content does not necessarily reduce the average length of lignin fragments generated during in silico thioacidolysis.