Lignin, a major component of the cell wall of vascular plants, has long been recognized for its negative impact on forage quality, paper manufacturing, and, more recently, cellulosic biofuel production. Over the last two decades, genetic and biochemical analyses of brown midrib mutants of maize, sorghum and related grasses have advanced our understanding of the relationship between lignification and forage digestibility. This work has also inspired genetic engineering efforts aimed at generating crops with altered lignin, with the expectation that these strategies would enhance forage digestibility and/or pulping efficiency. The knowledge gained from these bioengineering efforts has greatly improved our understanding of the optimal lignin characteristics required for various applications of lignocellulosic materials while also contributing to our understanding of the lignin biosynthetic pathway. The recent upswing of interest in cellulosic biofuel production has become the new focus of lignin engineering. Populus trichocarpa and Brachypodium distachyon are emerging as model systems for energy crops. Lignin research on these systems, as well as on a variety of proposed energy crop species, is expected to shed new light on lignin biosynthesis and its regulation in energy crops, and lead to rational genetic engineering approaches to modify lignin for improved biofuel production.
The exploitation of fossil fuels during the global industrialization of the past century has propelled development of the world economy. As the demand for energy continues to surge in the 21st century, societal concerns about soaring oil prices, global warming due to greenhouse gas emissions, and energy security are prompting policymakers and scientists to explore the feasibility of renewable, sustainable biofuels as alternatives and/or supplements to petroleum-based fuels (Kerr and Service, 2005; Koonin, 2006; Schubert, 2006). The biofuels currently available on market, mainly ethanol and biodiesel, are predominantly produced from corn grain, sugar cane and soybean oil. According to recent statistics, approximately 5 billion gallons of corn ethanol were produced in the USA in 2006, which is equivalent to only 3.6% of the total volume of gasoline consumed in that year. Similarly, the 100 million gallons of biodiesel produced in the USA in 2006 accounted for <0.2% of the total diesel used domestically (Yacobucci and Schnepf, 2007). To achieve a substantial substitution of gasoline by biofuels, it has been proposed by many that lignocellulosic biomass from a variety of sources, including agricultural residues such as corn stover and sugar cane bagasse, trees and grasses, is a potential source for ‘cellulosic ethanol’ (Demirbas, 2005; Lynd et al., 1991; Somerville, 2006, 2007; Wang et al., 1983). Cellulose from plant materials has already played many important roles in human history: cotton fiber is used for cloth making, paper is made from cellulose fibers extracted mostly from wood, and cellulose is also the major nutritional component in forage crops for ruminant livestock. Indeed, a large proportion of the population of the developing world depends not upon oil but upon wood for their energy needs (Madubansi and Shackleton, 2007). If we can find a way of efficiently converting cellulose from plant biomass into liquid fuels, it will provide humans with a renewable and carbon-neutral energy source for future sustainable development.
Despite the potential promise of cellulosic ethanol and other cellulose-derived biofuels, major technical obstacles need to be addressed to make the process feasible and economically viable for large-scale adoption. One of the major problems lies in the fact that cellulose in the plant cell wall is crystalline and thus recalcitrant to hydrolysis. This problem is exacerbated by the fact that cellulose is embedded in a complex matrix that includes the phenolic polymer lignin, the presence of which interferes with access of hydrolytic enzymes to the cellulose polymer. Lignin can also adsorb hydrolytic enzymes that are used to generate monosaccharides from lignocellulose, and some lignin degradation products inhibit subsequent fermentation steps (Keating et al., 2006).
To ensure successful biological conversion, the interactions between lignin and the polysaccharide components of the cell wall must be reduced through pre-treatment, a process that is considered to be one of the most costly steps in the whole process (Wyman et al., 2005). For this reason, it has been suggested that engineering feedstock crops with cell-wall structures that are more susceptible to pre-treatment and thus more amenable to hydrolysis, or are sufficiently altered that they require no pre-treatment at all, will make biofuels cost-competitive with fossil fuels, and thus contribute to making the entire process economically viable.
Cellulose and lignin are the two most abundant biopolymers on earth (Boerjan et al., 2003). Lignin is deposited within the secondary cell walls of all vascular plants, and accounts for approximately 30% of the terrestrial organic carbon fixed in the biosphere annually (Battle et al., 2000). As a major component of plant cell walls, lignin has a far-reaching impact on agriculture, industry and the environment. For example, in the paper industry, lignin has to be removed from wood chips during the pulping process to make high-quality paper. Lignin also reduces the quality and digestibility of forage crops such as alfalfa, which in turn has an impact on the livestock industry. Driven by its significance in the economics of these industries, lignin has been one of the most intensively studied subjects in plant biochemistry for more than a century. In the last decade, our growing knowledge of lignin biosynthesis, coupled with advances in plant transformation technologies, has allowed researchers to manipulate lignin content and composition in a variety of plant species using genetics and genetic engineering (Boudet et al., 2003; Humphreys and Chapple, 2002). These efforts have resulted in transgenic plants with improved pulping efficiency or digestibility, some of which, such as brown midrib (bm or bmr) sorghum × sudangrass hybrids, are already in use in agriculture.
The aromatic lignin polymers commonly found in angiosperms are primarily composed of three monolignols, namely p-coumaryl, coniferyl and sinapyl alcohols, which, when polymerized, form p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) lignin. These monolignols are synthesized via the phenylpropanoid pathway, which begins with deamination of phenylalanine to form cinnamic acid, followed by a series of ring hydroxylations, O-methylations and side-chain modifications (Figure 1). Lignin found in gymnosperms and ferns generally lacks S units (Boerjan et al., 2003), suggesting that the branch leading to sinapyl alcohol biosynthesis may be a relatively recent addition to plants’ biochemical repertoire. Our understanding of the monolignol biosynthetic pathway has undergone multiple major revisions over the past decade, due to the data generated from both in vitro kinetic studies on lignin biosynthetic enzymes and genetic studies on mutants and transgenic plants with altered expression levels of phenylpropanoid pathway genes (Boerjan et al., 2003; Humphreys and Chapple, 2002). Current thinking suggests that 10 enzymes are required for monolignol biosynthesis: phenylalanine ammonia lyase (PAL), cinnamic acid 4-hydroxylase (C4H), 4-(hydroxy)cinnamoyl CoA ligase (4CL), hydroxycinnamoyl CoA:shikimate hydroxycinnamoyl transferase (HCT), p-coumaroylshikimate 3′-hydroxylase (C3′H), caffeoyl CoA O-methyltransferase (CCoAOMT), (hydroxy)cinnamoyl CoA reductase (CCR), ferulic acid 5-hydroxylase (F5H), caffeic acid/5-hydroxyferulic acid O-methyltransferase (COMT) and (hydroxy)cinnamyl alcohol dehydrogenase (CAD). In the cell wall, dehydrogenative polymerization of the monolignols is thought to be catalyzed by peroxidases (PER) and laccases (LAC) (Boerjan et al., 2003); however, the exact roles of these enzymes in the regulation of lignin biosynthesis in planta have not been defined.
In this review, we focus on studies that have demonstrated altered and often improved cell-wall traits brought about by altering lignin biosynthesis in mutants and transgenic plants. Although most of this research has been conducted in the context of forage crops and the pulp and paper industry, the knowledge generated from this research is already providing the basis for studies aimed at the rational genetic engineering of lignin deposition in bioenergy crops.
Analysis of mutants with improved forage digestibility
There are many reports indicating that forage digestibility is negatively correlated with lignin concentration. It is believed that lignin physically hinders the accessibility of cell-wall polysaccharides to the hydrolytic enzymes during digestion, thereby limiting the nutritional value of forage crops (Moore and Jung, 2001). Thus, many breeding efforts for improving forage quality have focused on identifying forage crops with altered lignification. In 1924, a naturally occurring maize mutant with a bm phenotype was isolated at the University of Minnesota (Jorgensen, 1931). After the mutant plant has produced 4–6 leaves, the midribs of subsequent leaves are reddish brown in comparison to the grey-green midribs of wild-type plants. This pigmentation is associated with xylem and other sclerified tissue, and remains in the cell-wall residues after soluble metabolites and cell-wall polysaccharides are removed, indicating that it is intrinsic to the lignin polymer itself. This mutant was named as bm1 when three other naturally occurring bm maize mutants (bm2, bm3 and bm4) were isolated subsequently (Kuc and Nelson, 1964). All of the four bm mutants are recessive and follow simple Mendelian inheritance. More importantly, they have altered lignin composition and/or reduced lignin content (Gee et al., 1968; Kuc and Nelson, 1964; Kuc et al., 1968; Muller et al., 1971). Similar brown midrib mutants in sorghum and pearl millet (often referred as bmr in these species) have been generated by chemical mutagenesis (Figure 2; Cherney et al., 1991). Detailed reviews have been published on the characteristics of bm/bmr mutants in the context of cell-wall composition, forage digestibility and animal performance (Barriere et al., 2004; Cherney et al., 1991). Among these studies, the work on the bm3 and bm1 mutants can be best compared with the results obtained from genetic engineering aimed at modifying lignin in dicotyledonous species used for forage and pulp production because in these mutants the affected genes have been identified.
The Klason lignin content of the bm3 mutant is about 20% less than that of wild-type maize (Barriere et al., 2004; Gaudillere and Monties, 1989; Marita et al., 2003), a phenotype that is thought to explain the greater in vitro cell-wall digestibility of the mutant (Keith et al., 1979). The in vivo digestibility in ruminant animals was also reported to be higher for the bm3 mutant than for wild-type plants (Keith et al., 1979; Muller et al., 1972; Tjardes et al., 2000; Weller and Phipps, 1986), presumably for similar reasons. In an attempt to reveal the nature of the bm3 mutation, Grand et al. (1985) measured enzyme activities of PAL, 4CL, CCR, CAD, COMT and PER in shoots at three different ages (10, 20 and 30 days old) for both bm3 mutant and wild-type maize plants. Of the six enzymes tested, only COMT activity was affected in bm3 mutants, and was reduced to only about 10% of that in wild-type plants. Because COMT is responsible for converting 5-hydroxyguaiacyl (5-OH-G) to S lignin units, it was expected that the strongly decreased COMT activity in the bm3 mutant would result in a reduction of S units and concomitant accumulation of 5-OH-G units, a prediction verified by thioacidolysis (Gaudillere and Monties, 1989; Lapierre et al., 1988). These early findings were confirmed by recent studies (Marita et al., 2003) using more advanced analytical methods such as NMR and derivatization followed by reductive cleavage (Lu and Ralph, 1998). Cloning of the maize COMT gene (Collazo et al., 1992) made it possible to characterize the bm3 mutation at the molecular level. The COMT gene was found to contain either an insertion or a deletion in three independent bm3 mutants (Morrow et al., 1997; Vignols et al., 1995). Moreover, down-regulation of COMT in maize by a transgenic approach resulted in biochemical alterations of lignin content and composition similar to that observed in bm3 mutants, as well as increased cell-wall digestibility (Piquemal et al., 2002).
There are conflicting reports regarding lignin content alterations in the maize bm1 mutant. Halpin et al. (1998) reported a 20% reduction of Klason lignin in the bm1 mutant, but Marita et al. (2003) reported that there was no reduction at all. This discrepancy may be due to the different genetic background and/or different alleles of the bm1 mutants used in the two studies. Related phenotypic differences attributed to background effects have also been observed in studies of the bm3 mutant (Weller et al., 1985). The bm1 mutant has an S/G ratio that is comparable to that of wild-type; however, its lignin includes elevated levels of coniferaldehyde and sinapaldehyde (Barriere et al., 2004). Like the bm3 mutant, in vivo cell-wall digestibility is enhanced in the bm1 mutant, but to a lesser extent (Cymbaluk et al., 1973). Interestingly, a study using synthetic lignin and a cell-wall model system showed that coniferaldehyde-containing lignin has more of an inhibitory effect on cell-wall degradability than coniferyl alcohol-containing lignin when measured using an in vitro enzyme digestion assay (Grabber et al., 1998b). This suggests that the apparent increased digestibility of bm1 may be primarily associated with the reduction of lignin content rather than the observed compositional changes, although the differences between the experimental sources of these data may confound the clarity of this conclusion.
The first clue to the nature of the bm1 mutation came from studies on the sorghum bmr-6 mutant, which, like the bm1 mutant, has increased accumulation of aldehyde groups in its lignin. Bucholtz et al. (1980) reported that the bmr-6 mutant has reduced CAD activity, suggesting that its phenotype is probably due to a deficiency in this enzyme. Given the similarity of the lignin phenotypes observed in bmr-6 sorghum and bm1 maize, it seemed likely that CAD deficiency of activity was also the cause of the bm1 phenotype. However, it was later discovered that, as in bm3 maize, COMT activity was also significantly decreased in bmr-6 plants (Pillonel et al., 1991), making it unclear whether a deficiency in CAD activity alone could cause the brown midrib phenotype. Purification of a CAD from tobacco stems (Halpin et al., 1992) and the cloning of its corresponding cDNA (Knight et al., 1992) made it possible to test the above hypothesis by specifically reducing CAD activity using antisense technology. Tobacco with severely reduced CAD activity showed a red-brown color in xylem tissues (Halpin et al., 1994), reminiscent of the bmr-6 phenotype. Similar phenotypes were also observed in CAD-down-regulated transgenic poplar trees (Baucher et al., 1996). These results indicated that a reduction of CAD activity is sufficient to cause the brown midrib phenotype. Halpin et al. (1998) reported that CAD activity in bm1 maize mutant was reduced by 60–70%, and the expression of a CAD gene was greatly reduced at both the mRNA level and the protein level. This CAD gene was mapped to the same region in which bm1 is located, but it was not determined whether it contained a mutation that could lead to the bm1 phenotype. A recent investigation on the expression of phenylpropanoid pathway genes in each of the four bm maize mutants revealed that, in addition to the CAD gene mentioned above, expression of many phenylpropanoid pathway genes, including four putative CAD genes, was down-regulated in bm1 (Guillaumie et al., 2007). This suggests that either the bm1 mutation occurs in the aforementioned CAD gene, the down-regulation of which causes a more globalized change in phenylpropanoid gene expression, or, more likely, the bm1 mutation occurs in a regulatory gene that affects expression of the CAD gene family.
Despite the fact that little is known about the nature of the majority of bm/bmr mutations, brown midrib corn, sorghum and sorghum–sudangrass hybrids have been commercialized and are routinely grown for forage production. It has been observed that forage with higher digestibility is associated with higher intake by livestock, which has been attributed to the increased rate of forage digestion and its passage through the digestive tract (Cherney et al., 1991; Jung and Allen, 1995). Interestingly, some bm/bmr mutants also show improved palatability (Figure 3). Even in short-term grazing experiments during which digestibility is not a factor, sheep choose a bmr sorghum × sudangrass hybrid over the wild-type. The reason for this grazing preference is unclear. It is possible that some organoleptic factor in bmr material makes it more palatable, but, whatever the reason, this increased acceptability may further enhance the value of bmr mutants.
Metabolic engineering for improved forage digestibility
Perhaps inspired by the utility of naturally occurring lignin mutants, many genetic engineering efforts aimed at improving forage digestibility have focused on down-regulation of lignin biosynthetic pathway genes. The targets of these experiments and their outcomes are summarized in Table 1. It is notable that, in some cases, different studies have obtained different results on the impact of down-regulation of a certain gene on lignin content, quality or digestibility. This may be explained by differences in the promoters used to drive transgene expression, the extent of enzyme activity reduction, or the environment growth conditions between studies. It is also possible that plants with different genetic backgrounds have different responses to the same manipulation. Despite this complication, some general conclusions can still be drawn. Generally, suppression of genes early in the monolignol biosynthetic pathway, such as PAL, C4H, HCT and C3′H, is the most effective for reducing lignin content. In contrast, down-regulation of F5H or COMT, which resides on a branch pathway converting G to S, greatly reduces the lignin S/G ratio but has little effect on lignin content. In this context, it is noteworthy that, with the exception of CAD down-regulation, the transgenic alfalfa plants listed in Table 1 were generated in the same genetic background and the same promoter was used to drive the corresponding transgenes. Comparison of the digestibility of this isogenic collection of plants revealed a strong negative linear relationship between the forage digestibility and lignin content, but not lignin composition (Reddy et al., 2005). This is consistent with the finding that the Arabidopsis fah1 mutant, which lacks S lignin but has an unaltered lignin content, does not show altered cell-wall digestibility (Jung et al., 1999). These genetic data are also consistent with studies showing that the S/G ratio has no impact on the digestibility of artificially lignified maize cell walls derived from a tissue culture system (Grabber et al., 1997). It should be noted, however, that only small S/G ratio changes occurred in these transgenic alfalfa plants, and S subunits constitute only about one quarter of the total lignin monomers in wild-type Arabidopsis. Moreover, in neither case was the S/G ratio increased to such an extent that S subunits became the dominant monomers. This leaves open the possibility that greater changes in lignin quality, particularly high S/G ratios, as observed in F5H-over-expressing Arabidopsis, tobacco and poplar (Franke et al., 2000; Meyer et al., 1998), may still hold promise for improved forage digestibility.
Table 1. Genetic engineering of lignin biosynthetic genes and its effect on lignin and forage digestibility
Arrows indicate up- or down-regulation of genes, and increases or decreases in lignin composition.
Unfortunately, significant improvements in forage digestibility resulting from reduced lignin content are often accompanied by perturbations in plant growth and development. For example, some of the maize bm mutants have been reported to have reduced biomass yield (Lee and Brewbaker, 1984; Miller et al., 1983), reduced stalk strength (Zuber et al., 1977) and reduced resistance to pathogens (Nicholson et al., 1976), at least under certain growing conditions or environments. Similarly, strong reduction of the activities of C4H, C3′H and HCT in alfalfa resulted in dwarf plants with reduced biomass (Reddy et al., 2005; Shadle et al., 2007). This phenotype may be the result of the collapsed xylem phenotype that has been observed in plants with strongly decreased lignin content, such as the Arabidopsis cad-c cad-d double mutant and the CCR-deficient irregular xylem 4 (irx4) mutant (Jones et al., 2001; Sibout et al., 2005). Lignin-deficient tracheary elements probably collapse when their cell walls cannot stand the negative pressure generated during transpiration, and as a result water transport is impeded, with deleterious effects on plant growth. On the other hand, despite the apparent correlation between low lignin and reduced plant growth, it is still unclear whether the abnormal growth phenotype is caused by the low lignin content per se. An alternative explanation is that perturbations in lignin biosynthesis affect the accumulation of some other derivatives of the phenylpropanoid pathway that are required for, or have direct or indirect deleterious effects on, normal plant growth and development. For example, in tobacco, dehydrodiconiferyl alcohol glucosides, lignan glycosides derived from coniferyl alcohol, were found to have a cytokinin-like ability to stimulate cell division (Teutonico et al., 1991). It is possible that, as a result of genetic engineering of the lignin biosynthetic pathway, reductions in coniferyl alcohol pools lead to deficiency of dehydrodiconiferyl alcohol glucosides and altered cell division. Alternatively or additionally, perturbations in other plant growth regulators may arise when lignin biosynthesis is down-regulated. Besseau et al. (2007) reported that the abnormal growth phenotype of HCT-down-regulated Arabidopsis is associated with reduced auxin transport due to increased flavonoid accumulation when lignin biosynthesis is blocked. Both the growth and auxin phenotypes can be rescued by down-regulation of chalcone synthase (CHS), the first committed enzyme for flavonoid biosynthesis. Thus it is possible that simultaneous down-regulation of CHS and HCT may provide a viable genetic engineering strategy for the improvement of forage crops. Finally, a recent molecular phenotyping study that combined transcript and metabolite profiling of CCR-, CAD- and double down-regulated tobacco plants revealed that perturbation of the lignin biosynthesis pathway has global biochemical impacts on other metabolic pathways, such as starch metabolism and photorespiration (Dauwe et al., 2007). It is expected that similar studies in the future will greatly improve our understanding of the interaction networks between lignin biosynthesis and other metabolic pathways, and enhance our ability to design rational genetic engineering approaches to modify lignin with minimal effects on plant growth.
Modification of lignification to improve pulping efficiency
Wood is the major raw material for paper making. To make high-quality papers, lignin must be removed, usually by chemical (Kraft) pulping and bleaching processes, which are energy-intensive and environmentally unfriendly. For this reason, the development of trees with a lower amount of and/or more extractable lignin is an attractive approach to improve pulping efficiency and potentially alleviate some of the negative environmental impacts of the paper-making industry. Despite the extensive literature on genetic modification of lignin biosynthesis in a variety of plants, only a few studies have reported the impact of modified lignin content and composition on pulping and bleaching processes. Nevertheless, significant progress has been made in this field, as summarized in Table 2.
Table 2. Genetic engineering of lignin biosynthetic genes and its effect on lignin and pulping efficiency
Arrows indicate up- or down-regulation of genes, and increases or decreases in lignin composition.
It has been observed that wood from angiosperm trees (hardwoods) has a better Kraft pulping efficiency than the wood from conifers (softwoods) (Chiang et al., 1988). This has been attributed to the differences of lignin composition between these species of plants, although this conclusion is obviously confounded by the myriad differences in anatomy and cell-wall chemistry that exist between the relatively distantly related lineages to which these trees belong. Angiosperms contain both G and S lignin, whereas, as mentioned previously, most gymnosperms lack S units in their lignin. A major difference between these two lignin monomers is that the aromatic C5 position of a G unit is free to form carbon–carbon linkages during lignin polymerization, but this position is occupied by a methoxy group in S units. Thus, angiosperm lignin contains fewer carbon–carbon linkages than gymnosperm lignin, possibly explaining the observation that the former is more easily removed than the latter during the pulping process (Baucher et al., 2003; Chiang and Funaoka, 1990). Taken together, these observations suggested that better pulping efficiency might be achieved by increasing the S monomer content of plant lignins.
In Arabidopsis, the F5H-deficient fah1 mutant lacks S units in its lignin, and over-expression of the F5H gene under the control of the C4H promoter results in almost pure S lignin (Chapple et al., 1992; Marita et al., 1999; Meyer et al., 1996, 1998). These data clearly demonstrate the essential role of F5H expression in determining the relative abundance of G and S units in the lignin of Arabidopsis. The success of these experiments led to similar efforts to over-express the F5H gene in woody plants to determine whether this strategy was generally applicable, and, if so, whether it would lead to lignin with a high proportion of S units and improved chemical pulping performance. When the Arabidopsis F5H gene was over-expressed in tobacco and poplar using the C4H promoter, the S/G ratio was greatly increased but lignin content did not change (Franke et al., 2000). Interestingly, wood from the transgenic poplars showed a striking improvement in pulping and bleaching efficiency (Figure 4; Huntley et al., 2003). Under the same conditions, pulp made from F5H-over-expressing plants contained much less residual lignin, as indicated by a lower kappa number (a measurement used in the pulping industry to quantify residual lignin content) than pulp made from control plants when processed for the same length of time. This increase in the efficiency of delignification during the pulping process was correlated with the increase in the S/G ratio. Stated differently, to reach the same kappa value when processed under the same pulping conditions, wood from the best F5H-over-expressing trees required less than half of the time needed for wild-type samples (Huntley et al., 2003). Just as importantly, and unlike what has been observed when lignin content is decreased, F5H over-expression has not been reported to have detrimental effects on plant growth and development, although Li et al. (2003) suggested that accelerated maturation occurs in the secondary xylem of F5H-over-expressing poplar.
As mentioned previously in the context of the maize bm1 mutant, down-regulation of CAD in tobacco and poplar results in increased incorporation of sinapaldehyde and coniferaldehyde monomers into their lignin (Baucher et al., 1996; Halpin et al., 1994; Kim et al., 2003; Ralph et al., 2001b). More lignin can be released by mild alkaline treatment from these CAD-down-regulated plants than from the corresponding controls. In-depth structure analysis of the lignin in CAD-down-regulated poplar trees revealed that it contains more free phenolic groups, which is correlated with the increased alkaline solubility of lignin and increased pulping efficiency (Lapierre et al., 1999, 2004). Like the maize bm3 mutant, COMT-down-regulated poplar has a lower S/G ratio and incorporates 5-OH-G units in its lignin (Jouanin et al., 2000; Marita et al., 2001; Ralph et al., 2001a,b; VanDoorsselaere et al., 1995). In contrast to the CAD-down-regulated plants, COMT-down-regulated poplars were more resistant to delignification during the Kraft pulping process (Jouanin et al., 2000; Lapierre et al., 1999; Pilate et al., 2002). This may be because the benzodioxane structures formed by 5-OH-G units are more resistant to alkaline treatment in addition to concomitant decreases in S unit content.
Down-regulation of CCoAOMT or CCR in tobacco and poplar results in reduced lignin content and improved pulping efficiency; however, transgenic plants in which the expression of these genes is down-regulated sufficiently to improve pulping efficiency exhibit reduced growth (Leple et al., 2007; O’Connell et al., 2002; Petit-Conil et al., 1999; Pincon et al., 2001; Piquemal et al., 1998). These results highlight the challenges and limitations of lignin down-regulation approaches: it is essential but difficult to find a level of lignin reduction that is sufficient to be advantageous but not so severe as to affect normal plant growth and development.
Biofuel-oriented lignin research
Despite the fact that a great deal of work in the lignin field has been directed at improving forage crops and the pulping process, only a few examples can be found in which research has been aimed specifically at improving biofuel production. A notable exception to this is the recent work of Chen et al. who analyzed the biomass digestibility after both acid pre-treatment and enzymatic hydrolysis in six transgenic alfalfa lines (down-regulated for C4H, HCT, C3′H, CCoAOMT, F5H and COMT) that were the subject of a previous study on forage digestibility (Chen and Dixon, 2007; Chen et al., 2006). The authors reported dramatic differences in the cellulase/cellobiase saccharification efficiencies of acid-pre-treated cell walls of the various lines. Among these lines, the C3′H- and HCT-down-regulated lines showed enzymatic saccharification efficiencies that were almost double that of the controls. These C3′H- and HCT-down-regulated plants have previously been reported to have increased accumulation of H lignin and decreased total lignin (Chen et al., 2006). Indeed, even without the acid pre-treatment step, the C3′H- and HCT-down-regulated lines showed higher enzymatic saccharification efficiencies than acid-pre-treated controls. These results are similar to those of a previous study in which the cell walls of the Arabidopsis C3′H-deficient ref8 mutant were found to be much more susceptible to enzymatic hydrolysis than wild-type cell walls (Franke et al., 2002). These results suggest that genetic modification of lignin biosynthesis can facilitate the processing of lignocellulosic materials for biofuel production.
In a related study that focused on existing natural variation in lignification, Davison et al. (2006) studied biomass saccharification efficiency in five genetic variants of poplar generated from a single segregating F2 hybrid family (Davison et al., 2006). They reported that one variant with low lignin content and low S/G ratio released the highest amount of xylose after acid hydrolysis, whereas another variant with high lignin content and high S/G ratio released the lowest amount of xylose. It is worth noting that the results indicated correlations but not causal relationships between lignin differences and acid hydrolysis efficiency, and other genetic factors segregating within this F2 population may have also contributed to the observed effects.
Our current understanding of lignin biosynthesis is derived primarily from studies in dicotyledonous herbaceous plants, such as Arabidopsis and alfalfa; however, cellulosic biofuel crops may ultimately include grasses and dicotyledonous plants ranging from woody plants (e.g. hybrid poplar and eucalyptus) to legumes (Department of Energy, 2005). Although lignin is common to all vascular plants, this does not necessarily mean that all the mechanisms underlying lignin biosynthesis and polymerization are conserved. Arabidopsis can be induced to make secondary xylem from vascular cambium in its hypocotyl, but the fact that perennial woody plants devote enormous resources to lignification during secondary growth is only one example of how lignin deposition may be regulated differently between woody plants and herbaceous plants. While established models such as Arabidopsis and alfalfa will continue to serve as platforms for studying lignification, a number of other species have recently been selected as potential model systems for functional genomics studies, including the temperate grass Brachypodium distachyon (Draper et al., 2001) and the black cottonwood tree Populus trichocarpa (Bradshaw et al., 2000; Wullschleger et al., 2002). Both of these species have features that make them suitable as experimental subjects for investigating lignin biosynthesis in a context more closely related to bona fide biomass crops. These attributes include a relatively small genome size, a short lifecycle in the case of Brachypodium or a rapid growth rate in the case of poplar, and the ease with which transformation can be accomplished. A recent B. distachyon EST project revealed that its genome contains homologs of all 10 monolignol biosynthetic genes known in Arabidopsis (Vogel et al., 2006), most of which have more isoforms in B. distachyon than in Arabidopsis. Analyzing the P. trichocarpa whole genome sequence identified 37 potential lignin biosynthetic genes (Tuskan et al., 2006), and, as in B. distachyon, the families to which these genes belong are generally larger than those in Arabidopsis. A tissue-specific transcript profiling study in Populus has shown that the two Populus C4H genes are co-expressed in tissues related to wood formation, whereas the three Populus C3 ′H genes are expressed exclusively in different tissue types. Taken together, these data indicate a level of potential functional redundancy and divergence among the C4H and C3′H enzyme isoforms, respectively (Schrader et al., 2004). The rapidly growing genomic resources for these model systems provide huge opportunities for functional genomics research. The knowledge generated from studying the biosynthesis of lignocellulosic cell walls in these model systems will be extremely important for optimal genetic engineering of relevant bioenergy crops.
To date, most genetic engineering of lignin biosynthesis has relied on up- or down-regulating genes in the monolignol pathway. However, some of these strategies, especially those aimed at reducing total lignin content, are often achieved at the expense of plant fitness and viability, limiting their applications in genetic engineering of biofuel crops. If rational manipulation of lignin biosynthesis in plants for biofuel production is to be truly successful, more opportunities should be explored beyond the scope of mis-regulation of monolignol biosynthetic genes. One such opportunity lies in the flexibility of the lignin polymerization process. Studies have revealed that phenolic components other than the three conventional monolignols can be incorporated into the lignin polymer (Boerjan et al., 2003). This suggests that it may be possible to manipulate plants to synthesize novel lignins that contain linkages that are easily broken by enzymes or by alkaline hydrolysis. In nature, many plant species have been reported to accumulate phenolic compounds with these properties, such as hydroxycinnamoyl esters and amides, usually in response to pathogen attack (Franke et al., 1998; Jang et al., 2004; Von Roepenack-Lahaye et al., 2003). Some of these compounds can be transported to the wall at the site of pathogen attack (McLusky et al., 1999), and some hydroxycinnamoyl esters have been found to be incorporated into the lignin polymer (Lu and Ralph, 1999, 2002; Lu et al., 2004; Morreel et al., 2004; Ralph et al., 2004). Elucidation of how these compounds are synthesized in the cell, targeted to the wall, and incorporated into the lignin polymer will potentially expand our toolbox for engineering lignin biosynthesis in planta. Ideally, modified lignins with hydrolysable inter-unit linkages would still fulfil the structural roles played by lignin, but would be easily degraded and removed during post-harvest processing, thereby greatly increasing the efficiency of cellulosic biofuel production.
Grass cell walls contain a substantial fraction of ferulate, which is cross-linked to both xylan and lignin (Grabber et al., 2004). Like lignin, cell-wall-bound ferulates have a negative impact on cell-wall degradability (Casler and Jung, 1999; Grabber et al., 1998a). NMR spectroscopy analysis of bonding patterns of ferulate–cell wall linkages further suggests that ferulates may serve as nucleation sites for lignin polymerization (Ralph et al., 1995). For these reasons, genetically modified plants with lowered ferulate content or impaired ferulate–cell wall cross-linking capability could also increase cell-wall degradability. Although previous models for lignin suggested that ferulate was an intermediate in lignin biosynthesis, analysis of the Arabidopsis ref1 mutant revealed that, at least in Arabidopsis, ferulic acid is synthesized via oxidation of coniferaldehyde by the REF1 aldehyde dehydrogenase (ALDH; Nair et al., 2004). Thus, ferulic acid appears to be an end-product of phenylpropanoid metabolism rather than a lignin biosynthetic intermediate, therefore its synthesis can be manipulated independently from lignin deposition. Phylogenetic analysis of the ALDH gene family has led to the identification of REF1 homologs in grass genomes (Kirch et al., 2004), making these genes very interesting targets for genetic engineering efforts aimed at reducing ferulate content and reducing cell-wall cross-linking.
Many studies have shown that phenylpropanoid pathway genes are developmentally regulated by various classes of trans-acting factors (Raes et al., 2003). Transgenic plants or mutants with ectopic expression levels of some of these trans-acting factors have been shown to have increased or decreased levels of phenylpropanoid end products, including lignin (Jin et al., 2000; Preston et al., 2004; Rogers and Campbell, 2004; Tamagnone et al., 1998). Although almost all the genes encoding lignin biosynthetic enzymes have been previously targeted for modification of lignin biosynthesis, few studies have taken advantage of the possibility of mis-regulating the trans-acting factors involved in phenylpropanoid metabolism. Presumably, altering the expression level of appropriate transcription factors would coordinately affect a group of genes in the pathway. By doing so, it might be possible to avoid the deleterious phenotypes associated with the accumulation of pathway intermediates that are the result of single gene/enzyme down-regulation. Furthermore, promoters with different tissue specificities could also be explored in the future to control the expression of sense/antisense transgenes. For example, as mentioned previously, it has been suggested that many of the deleterious phenotypes associated with lignin-down-regulated plants are due to collapse of the tracheary elements in the xylem that are responsible for water transport (Jones et al., 2001). By using a fiber- or sclerenchyma-specific promoter to modify lignin-associated gene expression in a tissue-specific fashion, lignin properties could be specifically changed only in those cells where lignin is dispensable, but left intact in those xylem elements where lignin is required for water transport, as previously proposed (Jung and Casler, 2006; Jung and Engels, 2002).
Another possibility would be to up-regulate lignin biosynthesis to increase biomass energy density rather than down-regulating lignin biosynthesis to engineer biomass to facilitate enzymatic hydrolysis. As the lignin polymer is more reduced than polysaccharides, biomass with a higher lignin content would be a better raw material for gasification and Fischer–Tropsch-type processes such as those that have been used to convert coal to liquid fuel (Agrawal et al., 2007; Schubert, 2006).
In conclusion, thanks to decades of research into lignin biosynthesis, much of which has been motivated by the desire to improve forage crops and the pulping process, lignin is now seen as the ‘low hanging fruit’ for genetic engineering of bioenergy crops. Although genetic engineering of lignin promises to increase the accessibility of cell-wall polymers to hydrolytic enzymes during the biofuel production process, the potential problems associated with this approach must be addressed before large-scale application. For example, it should be expected that plants with altered lignin chemistry may become more susceptible to pathogens or may be attacked by insects against which they are normally resistant. To address these and other potential problems, it will be necessary to advance our understanding of plant biology in a much broader sense, including aspects of biochemistry, microbial biology, entomology and ecology that may be affected by alterations in the plant cell wall. Combined efforts from scientists in these fields should identify the optimal approaches to engineering of bioenergy crops and thereby ensure the sustainability of this new agricultural paradigm.
Preparation of this manuscript was supported by the Office of Science (BER), US Department of Energy, grant number DE-FG02-06ER64301, and the National Science Foundation, grant number IOB-0450289. This is journal paper number 2008-18300 of the Purdue University Agricultural Experiment Station.