On the propensity of lignin to associate: A size exclusion chromatography study with lignin derivatives isolated from different plant species

Abstract

Despite evidence that lignin associates under both aqueous and organic media, the magnitude and nature of the underlying driving forces are still a matter of discussion. The present paper addresses this issue by examining both solution properties and size exclusion behaviour of lignins isolated from five different species of softwoods, as well as from the angiosperms Eucalyptus globulus and wheat straw. This investigation has used the recently described protocol for isolating enzymatic mild acidolysis lignin (EMAL), which offers lignin samples highly representative of the overall lignin present in the wood cell wall. The molecular weight distributions of these EMALs were found to be dependent upon the wood species from which they were isolated and upon the incubation conditions used prior to size exclusion chromatography. While the chromatograms of EMALs isolated from softwoods displayed a bimodal behaviour, the elution profiles of EMAL from E. globulus and straw were nearly unimodal. A marked tendency to dissociate prevailed under incubation at room temperature for all examined species with the exception of the straw lignin preparation; furthermore, lignin solutions incubated at 4 °C showed an associative behaviour manifested by an increase in the weight and number average molecular weights for some species. The extent of such association/dissociation, as well as the time needed for the process to reach completion, was also found to depend upon the wood species, i.e. lignins from softwoods were found to associate/dissociate to a greater extent than lignins from E. globulus and straw. The origin of such effects within the lignin structure is also discussed.

Introduction

Lignin is a complex natural polymer built up of different interunit linkages such β-O-4′, β-β′, β-5′, β-1′, 5-5′, 4-O-5′, etc. (Fengel and Wegener, 1989). Furthermore, lignin is covalently linked to carbohydrates forming a lignin–carbohydrate network (Yaku et al., 1976, Lawoko et al., 2006). Most softwood lignins consist predominantly of guaiacyl (G) units, whereas the structure of hardwood lignins is more complex due to the presence of both guaiacyl (G) and syringyl (S) units (Fengel and Wegener, 1989).

While the lignin interunit linkage pattern is relatively well known, the three-dimensional structure of lignin and its ultrastructural assembly with in the carbohydrate matrix of plant cell walls remain poorly understood (Besombes and Mazeau, 2005a). Recent experimental observations have suggested the existence of a certain level of coherence in the ultrastructure of lignin in native woody tissues, as a probable consequence of an ordered and controlled process of assembly with the other polysaccharide components during lignin deposition (Terashima and Seguchi, 1988, Besombes and Mazeau, 2005a). Such observations are in agreements with the work of Agarwal and Atalla, 1986, Atalla and Agarwal, 1985, which indicated that the aromatic rings of lignins are oriented preferentially parallel to the surface of the cell wall in spruce. Based on these studies and on the general agreement that the cell wall is formed via successive deposition of cellulose, hemicelluloses and lignin, Atalla (1998) suggested the existence of a strong associative interaction between precursors and the polysaccharide matrix as the dominant organizing influence upon lignin ultrastructure. Nevertheless, the present knowledge of the existence and significance of these associative forces during lignin deposition on the cell wall has been limited mainly to computational studies (Houtman and Atalla, 1995, Besombes and Mazeau, 2005a, Besombes and Mazeau, 2005b). Progress toward understanding such associative interactions have been hindered by the formidable difficulties of the field.

Most experimental evidence that lignin components tend to associate with one another have been obtained by evaluating the behavior of kraft lignin derivatives under alkaline conditions (Lindström, 1979, Sarkanen et al., 1982, Sarkanen et al., 1984, Norgren et al., 2002, Bikova et al., 2004, Gidh et al., 2006). Despite the fact that kraft lignins have undergone significant structural modifications compared with the native polymer (Sarkanen et al., 1982), these efforts have been used to reveal the intricacy of lignin association, which can be further complicated by aggregation between the resulting complexes (Lindström, 1979, Norgren et al., 2002). For example, Lindström (1979) evaluated the colloidal behavior of kraft lignins and emphasized the importance of the hydrogen bonding between the carboxylic groups and ether oxygens in the association process. It is likely though that these changes could have been induced as a result of the kraft delignification processess. He also concluded that the association is thermally irreversible and that prolonged storage of the samples results in the formation of a three-dimensional network. In contrast to his findings, Sarkanen et al. (1982) have reported that the associative/dissociative process is reversible and apparently governed by nonbonded orbital interactions. Furthermore, Sarkanen et al. (1984) have also hypothesized that the associative process, occurring within kraft lignin, is stoichiometrically constrained, i.e. each associated complex possesses a locus that is respectively complementary to only one type of component.

Additional efforts to evaluate association in lignin samples which are less severely modified than kraft lignins include those of Connors et al., 1980, Sarkanen et al., 1981, Cathala et al., 2003, which used Braun’s native lignin; organosolv lignins; milled wood lignin (MWL) and synthetic lignin (DHP), respectively. By using organosolv lignins isolated under relatively mild conditions from different angiosperms, Sarkanen et al. (1981) have reported varying degrees of association, whose extent was dominated by preferential interactions between their lower and higher molecular weight components. They have also confirmed the reversibility of the association/dissociation process described for kraft lignins and have concluded that it involves at least two kinetically distinct steps. On the other hand, Cathala et al. (2003) assessed the association behavior of MWL and lignin model compounds in organic media and concluded that the association of the starting material was not the result of an equilibrium between associated and molecularly dispersed species.

Despite the various studies that point to a prevailing consensus that lignin associates in both aqueous and organic media, the magnitude and the underlying driving forces behind these processes are still a matter of discussion. Accordingly, experiments aimed at supplementing our knowledge of the lignin association process with samples highly representative of native lignin would offer new insights into these processes. Such understanding is of presumed significance as far as the process of lignin deposition in the plant cell wall is concerned. This paper addresses this topic and focuses on the associative behaviour of lignins isolated from different wood species. To overcome the limitations of structural modification and inherent low molecular weights associated with the use of kraft, organosolv and MWL, the recently developed protocol for isolating enzymatic mild acidolysis lignin (EMAL) (Wu and Argyropoulos, 2003, Guerra et al., 2006a) in high yield and purity was used for the first time to address this issue. The combination of derivatization followed by reductive cleavage (DFRC) with quantitative ³¹P NMR (DFRC/³¹P NMR) was also applied in an attempt to better understand the lignin association process.