Introduction

Lignin is one of the major components of the secondary cell wall of vascular plants, and fills the spaces between cell wall polysaccharides. It confers mechanical strength and imperviousness to the cell wall (Boerjan et al. 2003). Therefore, lignin biosynthesis is closely related to the evolution of land plants. Lignin has several properties that present obstacles to chemical pulping, forage digestion, and conversion of plant cell wall polysaccharides into biofuels. For these processes, it would be beneficial for plant materials to either have less lignin, or to have lignin that is easier to remove. For these reasons, lignin biosynthesis is an area of great interest (Dixon and Reddy 2003; Chiang 2006; Vanholme et al. 2008; Weng et al. 2008a).

Lignin is biosynthesized via oxidative coupling of p-hydroxycinnamyl alcohols (monolignols) and related compounds that are formed in the cinnamate/monolignol pathway. Recently, the pathway has been extensively revised, and the pathway’s control mechanisms are being studied intensively (Chiang 2006; Vanholme et al. 2008; Weng et al. 2008a). In this review, the major routes and the transcriptional control of the cinnamate/monolignol pathway are outlined briefly in relation to lignification.

Earlier views of the cinnamate/monolignol pathway

Monolignols are biosynthesized from cinnamic acids in the cinnamate/monolignol pathway (Fig. 1). This pathway supplies precursors for various phenylpropanoid compounds such as lignins, lignans, neolignans, norlignans, p-hydroxycinnamate esters, coumarins, suberins, flavonoids, and stilbenes (Ishii 1997; Umezawa 2001, 2003, 2005; Bernards 2002; Boerjan et al. 2003; Davin and Lewis 2003; Dixon and Reddy 2003; Ralph et al. 2004a, b, Bunzel et al. 2005; Chiang 2006; Bourgaud et al. 2006; Suzuki and Umezawa 2007; Ferrer et al. 2008). As illustrated in Fig. 1, many parallel routes can be envisaged for this pathway. The pathway towards sinapyl alcohol via ferulic acid, sinapic acid, and sinapaldehyde had long been proposed for angiosperm syringyl lignin biosynthesis (Sarkanen 1971; Higuchi 1985, 1997). This was based on tracer experiments with isotope-labelled phenylpropanoid monomers and associated enzymatic experiments (Sarkanen 1971; Higuchi 1985, 1997). For example, administration of 14C-labelled p-hydroxycinnamic acids such as ferulic, 5-hydroxyferulic, and sinapic acids to various plant species, including wheat (Triticum vulgare var. Thatcher) resulted in the formation of 14C-labelled syringyl lignin (Brown and Neish 1959; Higuchi and Brown 1963).

Fig. 1
figure 1

The cinnamate/monolignol pathway. Thick open arrow represents the major routes for lignin biosynthesis. C4H, cinnamate 4-hydroxylase; C3H, p-coumarate 3-hydroxylase; CAldOMT, 5-hydroxyconiferaldehyde O-methyltransferase; 4CL, 4-hydroxycinnamate CoA ligase; HCT, hydroxycinnamoyl CoA: shikimate/quinate hydroxycinnamoyl transferase; CCoAOMT, caffeoyl CoA O-methyltransferase; CCR, cinnamoyl CoA reductase; CAld5H, coniferaldehyde 5-hydroxylase; CAD, cinnamyl alcohol dehydrogenase; SAD, sinapyl alcohol dehydrogenase; F5H, ferulate 5-hydroxylase; CAOMT, caffeic acid O-methyltransferase; AEOMT, hydroxycinnamic acid/hydroxycinnamoyl CoA esters O-methyltransferase; C3'H, p-coumaroyl shikimate/quinate 3-hydroxylase. Asterisks: Only the structure of shikimate ester is shown

Full size image

The new route for sinapyl alcohol biosynthesis

The first revision of the cinnamate/monolignol pathway resulted from the discovery of a new O-methyltransferase (OMT), caffeoyl CoA O-methyltransferase (CCoAOMT). This OMT methylates caffeoyl CoA to form feruloyl CoA, and 5-hydroxyferuloyl CoA to form sinapoyl CoA. At first, this new OMT was thought to play a role in defense. Later, the involvement of CCoAOMT in lignin biosynthesis was demonstrated (Ye et al. 1994, 2001).

Around that time, lignins in transgenic Nicotiana tabacum (tobacco) plants with altered expressions of phenylalanine ammonia-lyase (PAL) and cinnamate 4-hydroxylase (C4H) were analyzed. The results of those studies suggested that biosynthesis of monolignols from phenylalanine might occur by more than one route (Howles et al. 1996; Sewalt et al. 1997a). In addition, characteristics of lignin synthesis in a number of transgenic plants with downregulated expressions of caffeic acid O-methyltransferase (CAOMT or COMT) and/or CCoAOMT were reported. The downregulation of CAOMT in N. tabacum cv. Havana (Dwivedi et al. 1994), Populus tremula × Populus alba (Van Doorsselaere et al. 1995; Marita et al. 2001), Populus tremuloides (Tsai et al. 1998), N. tabacum cv. Samsun NN (Atanassova et al. 1995), and N. tabacum cv. Xanthi (Zhong et al. 1998) caused significant decreases in the S/G ratio, no large change in lignin content (0–11% decrease), and the appearance of a 5-hydroxyguaiacyl unit. On the other hand, suppression of CCoAOMT in N. tabacum cv. Xanthi (Zhong et al. 1998) and P. tremula × P. alba (Zhong et al. 2000; Meyermans et al. 2000) resulted in decreased lignin content and lower levels of guaiacyl and syringyl lignins.

These results suggested that CCoAOMT was involved in the biosynthesis of guaiacyl and syringyl lignins at least in these plant species, while CAOMT catalyzed the second methylation leading to the syringyl unit, although there were reports of different results that in CAOMT-silenced transgenic N. tabacum cv. Xanti and P. tremula × P. alba, substantial lignin reduction occurred (17–35% decrease) (Ni et al. 1994; Sewalt et al. 1997b; Jouanin et al. 2000). At this point, the role-sharing of CCoAOMT and CAOMT in the lignin biosynthetic pathway remained unclear.

The roles of CCoAOMT and CAOMT were clarified by the discovery of a novel pathway for syringyl lignin biosynthesis in angiosperm trees. In this pathway, syringyl lignin is synthesized via coniferaldehyde, 5-hydroxyconiferaldehyde, sinapaldehyde and sinapyl alcohol (Osakabe et al. 1999; Li et al. 2000, 2001; Chiang 2006). Similar pathways were also proposed for the herbaceous model plant Arabidopsis thaliana (Humphreys et al. 1999).

The discovery of the novel pathway via 5-hydroxyconiferaldehyde was based on kinetic analyses of recombinant and plant CAOMTs and ferulic acid 5-hydroxylases (F5Hs) of various angiosperm trees including Liquidambar styraciflua (sweetgum) and P. tremuloides (aspen) (Osakabe et al. 1999; Li et al. 2000). Individually incubating caffeic acid, 5-hydroxyferulic acid, and 5-hydroxyconiferaldehyde with CAOMT in the presence of S-adenosyl-l-methionine gave rise to the corresponding methylation products; ferulic acid, sinapic acid, and sinapaldehyde, respectively. When incubated individually with F5H, ferulic acid and coniferaldehyde were efficiently hydroxylated yielding 5-hydroxyferulic acid and 5-hydroxyconiferaldehyde, respectively. However, 5-hydroxyconiferaldehyde inhibited the CAOMT-catalyzed methylation of caffeic acid and 5-hydroxyferulic acid, while coniferaldehyde inhibited the F5H-catalyzed 5-hydroxylation of ferulic acid. This indicated that 5-hydroxyconiferaldehyde is the preferred substrate of CAOMT, and that coniferaldehyde is the preferred substrate of F5H (Osakabe et al. 1999; Li et al. 2000). Therefore CAOMT was renamed 5-hydroxyconiferaldehyde OMT (CAldOMT), and F5H was renamed coniferaldehyde 5-hydroxylase (CAld5H) (Osakabe et al. 1999; Li et al. 2000, 2001).

On the other hand, a different lignin biosynthetic pathway was reported for Medicago sativa (alfalfa). Significant lignin reduction (24–26% decrease) was observed in CAldOMT-silenced transgenic M. sativa (Guo et al. 2001; Chen et al. 2006). In contrast, downregulation of CCoAOMT led to reduced lignin levels (14–21% decrease) and a reduction in guaiacyl units without decreasing the amount of syringyl units (Guo et al. 2001; Chen et al. 2006). On the basis of these results, Guo et al. proposed a new biosynthetic pathway for syringyl lignin formation that was independent of guaiacyl lignin biosynthesis. In this new pathway, syringyl lignin is formed via caffealdehyde, coniferaldehyde, 5-hydroxyconiferaldehyde, sinapaldehyde and sinapyl alcohol. The new pathway differs from those proposed previously for angiosperm trees (Osakabe et al. 1999; Li et al. 2000) and A. thaliana (Humphreys et al. 1999) in two ways: caffealdehyde is involved as an intermediate and CCoAOMT does not participate in syringyl lignin formation (Guo et al. 2001; Parvathi et al. 2001; Dixon et al. 2001). Later, Chen et al. proposed that 3-methylation of the 3,4-dihydroxyphenyl moiety can be catalyzed by either CAOMT or another OMT in M. sativa (Chen et al. 2006). By contrast, Marita et al. (2003) reported that downregulation of CCoAOMT in M. sativa resulted in the reduction of both guaiacyl and syringyl units and thereby proposed a syringyl lignin formation pathway where CCoAOMT operated in addition to guaiacyl lignin formation. The biosynthetic pathway in M. sativa proposed by Marita et al. (2003) is essentially the same as those proposed for angiosperm trees (Osakabe et al. 1999; Li et al. 2000) and A. thaliana (Humphreys et al. 1999).

As well as methylation of 5-hydroxyconiferaldehyde, CAldOMTs can also efficiently transform 5-hydroxyconiferyl alcohol into the corresponding methylated alcohol, sinapyl alcohol (Humphreys et al. 1999; Parvathi et al. 2001; Zubieta et al. 2002; Nakatsubo et al. 2007a; Nakatsubo et al. 2008). In addition, using a radio-tracer method, Matsui et al. (2000) suggested the conversion of coniferyl alcohol to sinapyl alcohol in dicotyledonous angiosperms. However, identification of a sinapyl alcohol dehydrogenase (SAD) in P. tremuloides, a cinnamyl alcohol dehydrogenase (CAD) catalyzing the selective reduction of sinapaldehyde to sinapyl alcohol, reinforced the pathway from coniferaldehyde to sinapyl alcohol via 5-hydroxyconiferaldehyde and sinapaldehyde at least in tree species (Li et al. 2001; Chiang 2006). Although sinapaldehyde-specific CAD activity (i.e. true SAD) has not yet been found from other plant species, CAD with catalytic activity towards both coniferaldehyde and sinapaldehyde are known to exist, for example, in A. thaliana (Kim et al. 2004) and Oryza sativa (Zhang et al. 2006b).

Kinetic analyses of a recombinant A. thaliana CAldOMT encoded by At5g54160 indicated that 5-hydroxyconiferaldehyde inhibited the AtCAldOMT-catalyzed methylation of 5-hydroxyconiferyl alcohol with a K i of 0.5 μM, and that the K i value of the inhibition of 5-hydroxyconiferaldehyde methylation by 5-hydroxyconiferyl alcohol was 73.4 μM. Thus, the O-methylation of 5-hydroxyconiferyl alcohol by AtCAldOMT was significantly inhibited in the presence of 5-hydroxyconiferaldehyde. That is, when both substrates were present, AtCAldOMT preferentially catalyzed the O-methylation of 5-hydroxyconiferaldehyde (Nakatsubo et al. 2008). Similar results were obtained for Carthamus tinctorius (safflower) CAldOMT-catalyzed methylation of the two substrates (Nakatsubo et al. 2007a). These results further support the physiological role of the 5-hydroxyconiferaldehyde pathway for syringyl lignin biosynthesis, provided that 5-hydroxyconiferaldehyde and 5-hydroxyconiferyl alcohol exist together at the reaction site. The 5-hydroxyconiferaldehyde pathway for syringyl lignin biosynthesis is now considered to be a major in vivo pathway (Osakabe et al. 1999; Humphreys et al. 1999; Li et al. 2000, 2001; Ralph et al. 2001; Marita et al. 2003; Boerjan et al. 2003; Umezawa 2005; Chiang 2006).

It is often thought that syringyl lignin is found only in angiosperms. However, it has also been detected in lycophytes. Recently, a distinct CAld5H (F5H) was reported in the lycophyte Selaginella moellendorffii, which is a basal vascular plant with syringyl units in its lignin (Weng et al. 2008b). Phylogenetic analysis suggested that SmCAld5H evolved independently of angiosperm CAld5Hs (Weng et al. 2008b).

The new route for 3-hydroxylation of the p-hydroxyphenyl moiety

The 3-hydroxylation had been an enigmatic step in the cinnamate/monolignol pathway (Ehlting et al. 2006). Soon after the establishment of the pathway from coniferaldehyde to sinapyl alcohol, a new route for 3-hydroxylation of the p-hydroxyphenyl moiety was proposed (Schoch et al. 2001; Franke et al. 2002b). The A. thaliana CYP98A3 was found to hydroxylate p-coumaric acid esters, especially the coumaroyl esters of shikimic and quinic acids (Schoch et al. 2001; Franke et al. 2002a). Similar substrate specificity was also reported for Ocimum basilicum CYP98A13 (Gang et al. 2002). In the wild-type, AtCYP98A3 was expressed in inflorescence stems and wounded tissues (Schoch et al. 2001) and lignified interfascicular fibers and xylem vessels in the stem (Nair et al. 2002). A mutant defective in the AtCYP98A3 gene deposited smaller amounts of lignin compared with the wild-type, and the mutant lignin was formed primarily from p-coumaryl alcohol (Franke et al. 2002b). Similarly, down-regulation of the homologous genes in M. sativa cv. Regan SY (Ralph et al. 2006) and Populus grandidentata × alba (hybrid poplar) (Coleman et al. 2008) increased the proportion of p-hydroxyphenyl units relative to the normally dominant guaiacyl and syringyl units.

Later, a cDNA encoding hydroxycinnamoyl CoA:shikimate/quinate hydroxycinnamoyl transferase (HCT) was isolated from N. tabacum (Hoffmann et al. 2003). Downregulation of the HCT gene in Nicotiana benthamiana resulted in lower lignin content but a higher proportion of p-hydroxyphenyl lignin. In addition, an A. thaliana mutant in which the corresponding A. thaliana HCT was downregulated showed lower syringyl lignin content (Hoffmann et al. 2004), indicating that HCT was involved in lignification. The new pathway consisted of the ester exchange catalyzed by HCT and the subsequent 3-hydroxylation of the formed p-coumaroyl ester of shikimic acid or quinic acid by CYP98A3. This gives rise to the corresponding caffeoyl ester, followed by its conversion to caffeoyl CoA by HCT (Fig. 1) (Schoch et al. 2001; Franke et al. 2002b; Hoffmann et al. 2004). Thus, CYP98A3 was designated as p-coumaroyl shikimate/quinate 3-hydroxylase (C3′H) (Schoch et al. 2001). This enzyme is also often known as p-coumarate 3-hydroxylase (C3H) (Franke et al. 2002a, b).

Recently, the Medicago truncatula HCT gene was used to silence the endogenous HCT gene in M. sativa cv. Regen SY plants, and the Pinus radiata HCT gene was silenced in tracheary element-forming P. radiata callus culture. In both cases, silencing of these HCT genes resulted in an increased proportion of p-hydroxyphenyl lignin units but lower lignin contents (Chen et al. 2006; Wagner et al. 2007). These results suggested that the new pathway is conserved among diverse plant groups.

The initially proposed pathway included both shikimic and quinic acid esters as possible intermediates (Schoch et al. 2001; Franke et al. 2002b; Hoffmann et al. 2004). Later, cDNAs encoding hydroxycinnamoyl CoA:quinate hydroxycinnamoyl transferase (HQT) were isolated from N. tabacum and Lycopersicon esculentum (tomato) (Niggeweg et al. 2004). Transgenic L. esculentum plants were generated in which the LeHQT gene was silenced or overexpressed. Both types of transformants had similar lignin contents to that of the wild-type. The amount of chlorogenic acid (caffeoyl quinate) was decreased in the silenced plants, whereas it was increased in overexpressing plants (Niggeweg et al. 2004). Together, those results suggested that HQT was involved in chlorogenic acid formation, but did not supplement the activity of HCT in lignin biosynthesis (Niggeweg et al. 2004). Niggeweg et al. (2004) suggested that the p-coumaroyl ester of shikimic acid was the intermediate for lignin formation in the figure in their report, although it was not mentioned in the text.

Additional routes and enzymes in the cinnamate/monolignol pathway

Figure 1 shows revised routes for lignin biosynthesis in the cinnamate/monolignol pathway. These revisions were based on the cumulative results from a number of plant species including A. thaliana, Populus spp., and N. tabacum. It is plausible that these same routes operate in other plant species, but routes of lignin biosynthesis should still be validated in other plants. It is also possible that physiologically important routes in the pathway and genes for lignin synthesis in secondary xylem formation might differ from those for biosynthesis of other phenylpropanoid compounds, such as lignans, norlignans, and neolignans. As well, biosynthetic routes and genes may differ in lignin formation after fungal attack or physical wounding.

For example, Kawasaki et al. (2006) reported that O. sativa CCR1 is involved in lignin synthesis associated with the defense response but not in developmental lignin synthesis. In addition, metabolic analyses of the cinnamate/monolignol and lignan pathways were carried out to characterize biosynthetic processes in maturing seeds of C. tinctorius. These analyses, which included a stable-isotope-dilution method combined with the administration of stable-isotope-labelled precursors, strongly suggested that ferulic acid was the intermediate in lignan biosynthesis in C. tinctorius (Sakakibara et al. 2007).

Further research showed that AtCAldOMT can compensate for the function of AtCCoAOMT, which methylates the 3-hydroxyl group of the 3,4-dihydroxyphenyl moiety. This suggests the possible involvement of another shunt for sinapyl alcohol formation (Do et al. 2007; Nakatsubo et al. 2008). The At5g54160 gene had been annotated as a CAOMT (CAldOMT) gene based on its high sequence homology to the P. tremuloides CAOMT (CAldOMT) (Zhang et al. 1997). Later, it was reported that a recombinant protein of the gene methylated flavonoids (Muzac et al. 2000). Recently, the function of this gene was re-evaluated. Its recombinant enzyme catalyzed the O-methylation of phenylpropanoid monomers and flavonoid substrates with high efficiencies. The specificity constants (k cat/K m) for 5-hydroxyconiferaldehyde and quercetin were both 0.11 μM−1 min−1. Lignins of At5g54160-knockout A. thaliana mutants lacked syringyl units. Taken together, these results strongly suggested that the gene was involved in both syringyl lignin and flavonoid biosynthesis, which was consistent with the results of gene co-expression network analysis (Tohge et al. 2007). Interestingly, silencing of the At5g54160 gene also resulted in significant accumulation of caffeyl alcohol, strongly suggesting that AtCAldOMT encoded by At5g54160 can methylate both 5-hydroxyconiferaldehyde and 3,4-dihydroxyphenyl compound(s) in syringyl lignin synthesis (Nakatsubo et al. 2008). These findings agreed with a recent report that an At4g34050 (the gene encoding A. thaliana CCoAOMT)-knockout mutant produced both guaiacyl and syringyl lignins, but had slightly lower total lignin content, while the At5g54160 (AtCAldOMT)-knockout mutant showed drastically decreased syringyl lignin content (Do et al. 2007). In addition, the At4g34050 and At5g54160 double-knockout mutant had very low lignin content, and its development was arrested at the plantlet stage. These results indicated that other members of AtCCoAOMT and AtCAldOMT gene families can not compensate for the functions of AtCCoAOMT (At4g34050) and AtCAldOMT (At5g54160). Importantly, the AtCCoAOMT-catalyzed methylation of 3,4-dihydroxyphenyl units in the wild type was replaced by AtCAldOMT in the AtCCoAOMT-knockout mutant (Do et al. 2007). Together, these results indicated that AtCAldOMT can catalyze methylation of 3,4-dihydroxyphenyl in vivo. A similar methylation of 3-hydroxyl group of the 3,4-dihydroxyphenyl moiety, that is, caffealdehyde methylation to coniferaldehyde catalyzed by CAldOMT in M. sativa was proposed (Guo et al. 2001; Parvathi et al. 2001), and should be tested in other plant species.

There may be additional enzymes that catalyze reactions in the cinnamate/monolignol pathway. For example, an OMT potentially involved in lignification has been identified in Pinus taeda (Li et al. 1997). This OMT methylated caffeic and 5-hydroxyferulic acids and caffeoyl and 5-hydroxyferuloyl CoA esters with similar specific activities; therefore, it was named hydroxycinnamic acids/hydroxycinnamoyl CoA esters OMT (AEOMT). The deduced protein sequence of AEOMT is partially similar to, but clearly distinguishable from, that of CAldOMTs. It does not exhibit any significant similarity to CCoAOMT protein sequences (Li et al. 1997). A putative promoter fragment of AEOMT was isolated from the P. radiata gene (Moyle et al. 2002). Histochemical analysis in transgenic N. tabacum plants revealed that the AEOMT promoter induced GUS expression in cell types associated with lignification, such as developing vessels, phloem and wood fibers, xylem parenchyma and non-lignifying phloem parenchyma, suggesting the possible role of the OMT in biosynthesis of phenylpropanoid compounds including monolignols (Moyle et al. 2002).

In addition, an OMT with higher preference for caffeic acid than for 5-hydroxyferulic acid was isolated from M. sativa cv. Apollo (alfalfa) (Inoue et al. 2000). This enzyme (COMT II) differed from CAOMT (CAldOMT) in its very low K m for caffeic acid and its native molecular weight. MsCOMT II may be a true caffeic acid OMT (Inoue et al. 2000).

A cDNA encoding another OMT that is possibly involved in monolignol biosynthesis was cloned from C. tinctorius. Phylogenetic analysis showed that the novel OMT did not belong to the typical CAldOMT cluster. The recombinant protein of the OMT catalyzed 3- (or 5-) O-methylation of hydroxycinnamaldehydes and hydroxycinnamyl alcohols, while it showed only weak or moderate activity towards hydroxycinnamates and hydroxycinnamoyl CoA esters (Nakatsubo et al. 2007b).

Similarly, an additional type of CAD was reported (Goffner et al. 1992, 1998; Damiani et al. 2005). The CAD, named CAD1, was first purified from Eucalyptus gunnii xylem. It reduced coniferaldehyde to coniferyl alcohol, but showed no activity towards sinapaldehyde (Goffner et al. 1992). Interestingly, a cDNA encoding EgCAD1 showed highest similarity to CCRs, and not the conventional CADs (Goffner et al. 1998). The corresponding CAD1 cDNAs were also isolated from N. tabacum, and the recombinant NtCAD1 showed activity towards coniferaldehyde and coniferyl alcohol, but not sinapaldehyde (Damiani et al. 2005).

As mentioned above, we can envision the outline of the cinnamate/monolignol pathway, especially for lignin biosynthesis. Our knowledge of the biosynthetic routes is largely based on the results of studies on enzyme kinetics and transgenic plants. The completion of entire genome sequences of some plant species, including A. thaliana, has enabled us to take advantage of the cDNAs already cloned and characterized in other species. These studies have provided lists of genes for which the functions have been annotated (Dixon and Reddy 2003; Goujon et al. 2003; Raes et al. 2003). For example, Raes et al. (2003) listed up to 61 gene family members that possibly encode A. thaliana cinnamate/monolignol pathway enzymes. Functions of a number of the annotated A. thaliana genes in the pathway have been characterized experimentally using their recombinant proteins, for instance, for C3H (Schoch et al. 2001; Franke et al. 2002a), CAldOMT (Nakatsubo et al. 2008), CAld5H (Humphreys et al. 1999), 4CL (Ehlting et al. 1999; Schneider et al. 2003; Hamberger and Hahlbrock 2004), CCR (Lauvergeat et al. 2001), and CAD (Kim et al. 2004). Kim et al. (2004) prepared recombinant proteins of nine A. thaliana CAD homologs, six of which were biochemically competent to reduce p-hydroxycinnamaldehyde substrates. Of those six, two had the highest activity and homology relative to CADs from other plant species (Kim et al. 2004). However, functions of many of the annotated genes in the pathway are yet to be determined experimentally, including biochemical characterization of their recombinant proteins. Hence, the identification of the bona fide functions of annotated genes has become one of the key subjects in plant biosciences. Systems-biological strategies are being used to unravel the complex and comprehensive control mechanisms that underlie metabolic processes of plants.

MYB transcription factors in the control of the cinnamate/monolignol pathway and lignin biosynthesis

Transcription factors (Kranz et al. 1998; Rogers and Campbell 2004; Olsen et al. 2005; Zhong and Ye 2007) and microRNAs (Lu et al. 2005; Zhang et al. 2006a; Groover and Robischon 2006) play important roles in the comprehensive control of plant metabolism. Lu et al. (2005) reported microRNA-mediated control of cell wall formation, and there are many recent reports on metabolic control of cell wall formation by transcription factors (Rogers and Campbell 2004; Zhong and Ye 2007). The MYB family of transcription factors is one of the most abundant classes of transcription factors in plants. The largest subfamily is R2R3-MYB, which contains the two-repeat R2R3 DNA-binding domain (Stracke et al. 2001). A number of R2R3-MYB genes that regulate phenylpropanoid biosynthesis in plants were outlined by Rogers and Campbell (2004).

Tamagnone et al. (1998) demonstrated that overexpression of an Antirrhinum majus R2R3-MYB gene (AmMYB308) repressed lignin biosynthesis in transgenic N. tabacum plants. The expression of the PAL gene was not affected by AmMYB308-overexpression in tobacco plants, whereas the expressions of C4H, 4CL, and CAD genes were significantly reduced in plants overexpressing AmMYB308 (Tamagnone et al. 1998). Similarly, 4CL gene expression was reduced in N. tabacum plants overexpressing AmMYB330 (Tamagnone et al. 1998). Experiments with yeast indicated that AmMYB308 can act as a very weak transcriptional activator. Thus, it was hypothesized that AmMYB308 out-competes endogenous strong activators and thereby competitively inhibits gene activation (Tamagnone et al. 1998).

While AmMYB308 was overexpressed in N. tabacum (Tamagnone et al. 1998), Jin et al. (2000) analyzed an A. thaliana mutant defective in the A. thaliana ortholog of AmMYB308, AtMYB4. The mutant had enhanced levels of sinapate esters in its leaves. In addition, AtMYB4 negatively regulated the C4H gene, but positively regulated the CCoAOMT gene (Jin et al. 2000). On the other hand, the transcript level of CCoAOMT in the T-DNA insertion mutant of AtMYB32 (a gene with high similarity to AtMYB4) was unchanged, while the transcript level of CAOMT was increased (Preston et al. 2004). Recently, increased expression of AtMYB4 was found in A. thaliana in response to cadmium exposure (van de Mortel et al. 2008).

Another A. thaliana R2R3-MYB, AtPAP1 (=AtMYB75), has a marked effect on activating flavonoid biosynthesis (Borevitz et al. 2000; Tohge et al. 2005). In the A. thaliana mutant pap1-D, the AtPAP1 gene shows massive and widespread activation throughout plant development. This mutant was characterized by strongly increased concentrations of glycosylated anthocyanins and flavonols compared with the wild-type, and also by increased lignin content (Borevitz et al. 2000). Although both accumulation of PAL transcripts and PAL activity in pap1-D were increased compared with the wild-type (Borevitz et al. 2000), the genes involved in the monolignol biosynthetic pathway were not upregulated by AtPAP1, except 4CL (Tohge et al. 2005). This result suggests that upregulation of PAL enhances lignin accumulation.

Patzlaff et al. (2003b) characterized a member of the R2R3-MYB family, PtMYB4, from P. taeda. PtMYB4 may play a direct role in regulating lignin biosynthesis during lignification. This MYB was expressed in lignifying P. taeda cells. In transgenic N. tabacum plants overexpressing PtMYB4, the expression of C3H, CCoAOMT, CAOMT, CCR, and CAD was increased, while PAL expression was decreased. In addition, the transgenic N. tabacum plants accumulated greater quantities of lignin and showed ectopic lignification. Together, these results strongly suggested that PtMYB4 enables lignification (Patzlaff et al. 2003b).

Overexpression of PtMYB4 in A. thaliana caused ectopic lignin deposition, and the PtMYB4 overexpressor shared some phenotypic traits, including ectopic lignification, with the A. thaliana de-etiolated 3 (det3) mutant (Newman et al. 2004). Expression analyses of A. thaliana R2R3-MYB genes that showed high similarity to PtMYB4 suggested that AtMYB61 was misexpressed in the det3 mutant. Overexpression of AtMYB61 produced a phenocopy of the det3 mutant, including the ectopic accumulation of lignin. Taken together, these results suggested that AtMYB61 has a role in regulation of lignin biosynthesis in A. thaliana (Newman et al. 2004).

Several other R2R3-MYBs have been characterized from conifers. PtMYB1 was suggested to regulate lignin biosynthesis (Patzlaff et al. 2003a). Later, the roles of PtMYB1 and PtMYB8 were further studied using transformants of another gymnosperm Picea glauca (spruce) (Bomal et al. 2008). Overexpression of PtMYB1 and PtMYB8 in P. glauca caused ectopic secondary wall deposition and increased lignin content. In addition, overexpression of a number of genes including PAL, 4CL, and CAOMT was observed in the PtMYB1- and PtMYB8-overexpressors, suggesting that PtMYB1 and PtMYB8 may be involved in secondary wall deposition (Bomal et al. 2008). P. glauca MYBs, PgMYB1, PgMYB2, PgMYB4 and PgMYB8, which are the closest homologs to PtMYB1, PtMYB2, PtMYB4 and PtMYB8, respectively, were expressed specifically in xylem tissues (Bedon et al. 2007).

In the hybrid aspen P. tremula × P. tremuloides, an R2R3-MYB designated as PttMYB21a was expressed strongly in xylem tissues. Transgenic aspen plants expressing the MYB gene in an antisense orientation showed higher transcript levels of CCo-AOMT in the phloem, suggesting that PttMYB21a might work as a repressor (Karpinska et al. 2004).

Goicoechea et al. (2005) and Legay et al. (2007) reported two R2R3-MYB family members, EgMYB1 and EgMYB2, from E. gunnii. EgMYB2 was expressed preferentially in the differentiating secondary xylem in stems and roots. The recombinant EgMYB2 protein specifically bound to the cis-regulatory regions of the promoters of EgCCR and EgCAD2 (Goicoechea et al. 2005). Overexpression of EgMYB2 in N. tabacum plants caused ectopic lignification and drastic upregulation of genes encoding enzymes committed to monolignol synthesis; C3H, CCoAOMT, F5H, CAOMT, HCT, CCR, CAD, and 4CL. Taken together, the results strongly suggested that EgMYB2 is involved in transcriptional regulation of lignin biosynthesis as a transcriptional activator (Goicoechea et al. 2005). On the other hand, EgMYB1 may be a negative regulator of lignin biosynthesis (Legay et al. 2007). EgMYB1 was preferentially expressed in differentiating xylem tissues in stems and roots of Eucalyptus. EgMYB1 bound to cis-regulatory elements in the promoters of EgCCR and EgCAD2 and repressed the transcription of these genes (Legay et al. 2007).

An R2R3-MYB family member from Vitis vinifera cv. Cabernet Sauvignon (grape), VvMYB5a, was expressed in skin, flesh, and seeds, mainly during the early stages of berry development (Deluc et al. 2006). Overexpression of the MYB gene in N. tabacum resulted in decreased expression of CCoAOMT in stamens (Deluc et al. 2006).

Fornalé et al. (2006) characterized two R2R3-MYB transcription factors ZmMYB31 and ZmMYB42 from Zea mays (maize). Expression of CAOMT genes was downregulated in A. thaliana plants overexpressing ZmMYB31 and ZmMYB42 genes. In addition, overexpression of the MYB genes also affected expression of other genes in the cinnamate/monolignol pathway, and resulted in decreased lignin content in transgenic plants (Fornalé et al. 2006).

LIM transcription factors in the control of the cinnamate/monolignol pathway and lignin biosynthesis

LIM proteins derive their name from the initials of the homeodomain proteins Lin11, Isl-1 and Mec-3 (Bach 2000). A member of the LIM proteins, NtLIM1, was isolated from N. tabacum on the basis of its capacity to bind to an AC element, which is an important cis-acting element for gene expression in lignification (Kawaoka et al. 2000; Kawaoka and Ebinuma 2001). Transgenic N. tabacum plants with antisense NtLIM1 had low transcript levels of some lignin biosynthetic pathway genes, such as PAL, 4CL, and CAD. Furthermore, these transformants showed a 27% reduction in lignin content (Kawaoka et al. 2000).

A homologous gene, EcLIM1, was isolated from Eucalyptus camaldulensis (Kawaoka et al. 2006). Antisense NtLIM1 was introduced into E. camaldulensis, and transgenic Eucalyptus plants showed decreased expression levels of EcLIM1 as well as several lignin biosynthesis genes, including PAL, C4H, and 4CL, and a 29% reduction in lignin content. These results show that EcLIM1 is one of the key transcription factors involved in lignin biosynthesis in E. camaldulensis (Kawaoka et al. 2006).

Transcriptional network regulating secondary wall biosynthesis including lignification

Recently, transcriptional profiles were determined for genes that are differentially expressed during secondary wall formation. These profiles led to the discovery of a hierarchy of transcription factors involved in regulating secondary wall biosynthesis, including lignification (Fig. 2) (Zhong and Ye 2007; Demura and Fukuda 2007; Yang et al. 2007; Zhong et al. 2008, 2009). Functional redundancy of transcription factors often complicates characterization of their functions. This can be overcome by using chimeric repressor silencing technology (Hiratsu et al. 2004).

Fig. 2
figure 2

Hierarchy of transcriptional activation in biosynthesis of lignin, cellulose and xylan in Arabidopsis thaliana. This regulation diagram is a summary of overexpression and direct promoter activation experiments of the transcription factors in addition to their chimeric repressor silencing, double RNAi inhibition, and double T-DNA knockout experiments (Mitsuda et al. 2005; Kubo et al. 2005; Zhong et al. 2006, 2007a, b, 2008; Mitsuda et al. 2007; Yang et al. 2007; Zhou et al. 2009). CesA7, Cellulose synthase catalytic subunit 7; CesA8, Cellulose synthase 8; Fra8, Fragile fiber 8 (a member of glycosyltransferase family 47); IRX9, putative family 43 glycosyl transferase; the abbreviations of enzymes involved in lignin synthesis are described in Fig. 1 legend

Full size image

Mitsuda et al. (2005) used the silencing technology to characterize two NAC transcription factors from A. thaliana, AtNST1 (NAC SECONDARY WALL THICKENING PROMOTING FACTOR 1) and NST2. The silencing results, and results from other techniques including phenotyping of T-DNA double knockout mutants, indicated that AtNST1 and AtNST2 are redundant in regulating secondary wall thickening in anther endothecium cell walls. Overexpression of AtNST1 induced ectopic lignified secondary wall thickening in various tissues of A. thaliana, including stems. Overexpressors also showed upregulated expression of several genes involved in secondary wall biosynthesis, including CCR (Mitsuda et al. 2005). Subsequently, it was shown that AtMYB26 regulates expressions of AtNST1 and AtNST2 (Yang et al. 2007).

On the other hand, Kubo et al. (2005) found that two A. thaliana NAC transcription factors, AtVND6 (VASCULAR-RELATED NAC-DOMAIN 6) and AtVND7 induced transdifferentiation of various types of cells into metaxylem- and protoxylem-like vessel elements, respectively, in A. thaliana and P. tremula × P. tremuloides (hybrid aspen). These results suggest that AtVND6 and AtVND7 are transcriptional switches for plant metaxylem and protoxylem vessel formation, respectively (Kubo et al. 2005).

Zhong et al. (2006) showed that expression of a NAC transcription factor gene of A. thaliana, AtSND1 (secondary wall-associated NAC domain protein 1), is associated with secondary wall thickening in A. thaliana fibers. Dominant repression of AtSND1 in A. thaliana led to a severe decrease in the secondary wall thickness of fibers. Overexpression of AtSND1 in A. thaliana resulted in activation of genes including CCoAOMT and 4CL, which are involved in lignin biosynthesis, leading to ectopic deposition of secondary walls in cells that are normally nonsclerenchymatous. As well, AtSND1 overexpression in A. thaliana resulted in upregulation of a number of transcription factors, such as NACs (AtSND2, AtSND3), MYBs (AtMYB20, AtMYB46, AtMYB85, AtMYB103), and a homeodomain transcription factor (AtKNAT7). Zhong et al. concluded that AtSND1 is a transcriptional activator involved in secondary wall thickening in fibers (Zhong et al. 2006, 2007a).

On the other hand, overexpression of AtSND1 inhibited normal secondary wall thickening in A. thaliana fibers (Zhong et al. 2006). Ko et al. (2007) reported a similar inhibition by AtSND1 (also known as AtANAC012), and concluded that AtANAC012 may act as a negative regulator of secondary wall thickening in A. thaliana xylary fibers (Ko et al. 2007), in contrast to the above-mentioned conclusion of Zhong et al. (2006).

As well as its role in regulating secondary wall thickening in anther endothecium cell walls, AtNST1 has a role in regulating formation of A. thaliana fibers. AtNST1 and AtSND1 (also known as AtNST3 and AtANAC012) redundantly regulate secondary wall thickening in fibers (Mitsuda et al. 2007; Zhong et al. 2007b).

Secondary wall biosynthesis research has subsequently focused on identifying factors downstream in the hierarchy of transcriptional regulation (Fig. 2) (Zhong and Ye 2007; Demura and Fukuda 2007; Zhong et al. 2008; Zhou et al. 2009). First, AtMYB46 was found to be a direct target of AtSND1. In addition, AtSND1 and its close homologs AtNST1, AtNST2, AtVND6, and AtVND7, were able to activate the AtMYB46 promoter. Dominant repression of AtMYB46 caused a drastic reduction in the secondary wall thickening of both fibers and vessels in A. thaliana, while overexpression of AtMYB46 activated the biosynthetic pathways of cellulose, xylan, and lignin. Thus, the expression of cellulose synthase genes CesA7 and CesA8, the xylan biosynthetic gene FRA8, and lignin biosynthetic genes 4CL1 and CCoAOMT were drastically induced (Fig. 2). Concomitantly, overexpression of AtMYB46 in A. thaliana led to ectopic deposition of secondary walls in cells that are normally nonsclerenchymatous (Zhong et al. 2007a).

In addition to AtMYB46, three transcription factors, AtSND3, AtMYB103, and AtKNAT7, were found to be direct targets of AtSND1, AtNST1, AtNST2, AtVND6, and AtVND7, except that the activation of AtSND3 by AtNST1 was indirect (Zhong et al. 2008). In AtSND1 and AtNST1 double RNAi plants, 11 transcription factors were downregulated as well as AtMYB46. These included two NACs (AtSND2 and AtSND3), eight MYBs (AtMYB103, AtMYB85, AtMYB52, AtMYB54, AtMYB69, AtMYB42, AtMYB43, and AtMYB20), and a homeodomain transcription factor (AtKNAT7) (Zhong et al. 2008). Dominant repression of AtSND2, AtSND3, AtMYB103, AtMYB85, AtMYB52, AtMYB54, AtMYB69, and AtKNAT7 caused a severe reduction in secondary wall thickening in both interfascicular fibers and xylary fibers in inflorescence stems, whereas secondary wall thickening in vessels was not significantly affected. Overexpression of AtSND2, AtSND3, and AtMYB103 resulted in increased secondary wall thickening in fiber cells. In addition, overexpression of AtMYB85 brought about ectopic deposition of lignin in epidermal and cortical cells in stems. AtSND2, AtSND3, and AtMYB103 induced the GUS reporter gene expression driven by the promoter of a cellulose synthase gene, CesA8, whereas AtMYB85 specifically induced GUS expression driven by the promoter of the lignin biosynthetic gene, 4CL. AtMYB52 and AtMYB54 induced GUS expression driven by promoters of CesA8, 4CL, and a xylan biosynthetic gene, IRX9. These results indicated that AtSND2, AtSND3, and AtMYB103 are involved in cellulose synthesis, while AtMYB85 specifically induce lignin biosynthesis. On the other hand, AtMYB52 and AtMYB54 are involved in the biosynthesis of the three major cell wall components (Fig. 2) (Zhong et al. 2008).

Thus, AtSND1/NST3/ANAC012 and AtNST1 function redundantly in activating secondary wall biosynthesis in fibers (Zhong et al. 2006, 2007b; Mitsuda et al. 2007). AtVND6 and AtVND7 regulate the differentiation of the metaxylem and protoxylem, respectively, in primary roots (Kubo et al. 2005), while AtNST1 and AtNST2 function redundantly in regulating secondary wall thickening in endothecium cells of the anther (Mitsuda et al. 2005).

Recently, two AtSND1-regulated A. thaliana MYB transcription factors, AtMYB58 and AtMYB63, were found to specifically activate lignin biosynthetic genes during secondary wall formation in A. thaliana (Zhou et al. 2009). Expression studies showed that AtMYB58 and AtMYB63 were specifically expressed in fibers and vessels undergoing secondary wall thickening. Simultaneous RNAi inhibition of both AtSND1 and AtNST1 led to drastic reduction in the expressions of AtMYB58 and AtMYB63. AtMYB58 expression was significantly upregulated by overexpression of AtSND1 and AtNST1. On the other hand, AtMYB63 expression was induced by AtNST1 overexpression but not by AtSND1. Dominant repression of AtMYB58 and AtMYB63 reduced secondary wall thickening of fibers and vessels. As well, the amount of guaiacyl lignin and syringyl lignin was significantly reduced in plants with dominant repression and simultaneous RNAi inhibition of AtMYB58 and AtMYB63. Overexpression of AtMYB58 and AtMYB63 specifically activated lignin biosynthetic genes (PAL1, C4H, 4CL, HCT, C3H, CCoAOMT, CCR, CAOMT, and CAD), alongside a concomitant ectopic deposition of lignin in cells that are normally unlignified. However, there was no activation of cellulose synthase genes CesA4, CesA7, and CesA8, or xylan biosynthetic genes IRX8 and IRX9. In addition, AtMYB58 directly activated the expression of lignin biosynthetic genes. These results indicated that AtMYB58 and AtMYB63 are specific transcriptional activators of lignin biosynthesis in the transcription network that regulates secondary wall biosynthesis (Fig. 2) (Zhou et al. 2009).

Summary

The cinnamate/monolignol pathway has been studied intensively, especially in relation to lignification, and routes within the pathway have been revised several times over the last decade. In the post-genomic era, lists of genes for which functions have been annotated are available for some plant species, including A. thaliana. The identification of the bona fide functions of annotated genes of the cinnamate/monolignol pathway is being studied intensively in relation to the elucidation of comprehensive control mechanisms of plant metabolism, in which transcription factors play important roles. Recently, a number of transcription factors that function as the master switches for xylem formation have been discovered. At present, research is focused on the transcription factors downstream in the hierarchy of transcriptional controls regulating the biosynthesis of secondary wall components, including hydroxycinnamates and lignins.