4-Coumarate 3-hydroxylase in the lignin biosynthesis pathway is a cytosolic ascorbate peroxidase

Abstract

Lignin biosynthesis is evolutionarily conserved among higher plants and features a critical 3-hydroxylation reaction involving phenolic esters. However, increasing evidence questions the involvement of a single pathway to lignin formation in vascular plants. Here we describe an enzyme catalyzing the direct 3-hydroxylation of 4-coumarate to caffeate in lignin biosynthesis as a bifunctional peroxidase that oxidizes both ascorbate and 4-coumarate at comparable rates. A combination of biochemical and genetic evidence in the model plants Brachypodium distachyon and Arabidopsis thaliana supports a role for this coumarate 3-hydroxylase (C3H) in the early steps of lignin biosynthesis. The subsequent efficient O-methylation of caffeate to ferulate in grasses is substantiated by in vivo biochemical assays. Our results identify C3H as the only non-membrane bound hydroxylase in the lignin pathway and revise the currently accepted models of lignin biosynthesis, suggesting new gene targets to improve forage and bioenergy crops.

Fig. 1

The lignin biosynthetic pathway highlighting the 4-coumarate 3-hydroxylase (C3H) reaction characterized in this study. Shadings illustrate key findings in the current understanding of lignin biosynthesis in plants. Model of lignin biosynthesis as understood in the 80 s (in gray), shikimate shunt and substrate preferences of P450 enzymes discovered in the 90 s (in brown), subsequently identified 5H- and C-lignin monomers (in blue), and most recent step discovered (in green). PAL l-phenylalanine ammonia-lyase, PTAL bifunctional l-phenylalanine/l-tyrosine ammonia-lyase, C4H cinnamate 4-hydroxylase, C3H 4-coumarate 3-hydroxylase, COMT caffeate/5-hydroxyferulate 3-O-methyltransferase, F5H ferulate 5-hydroxylase/coniferaldehyde 5-hydroxylase, 4CL 4-hydroxycinnamate:CoA ligase, HCT 4-hydroxycinnamoyl CoA:shikimate/quinate hydroxycinnamoyltransferase, C3’H 4-coumaroyl shikimate/quinate 3’-hydroxylase, CSE caffeoyl shikimate esterase, CCoAOMT caffeoyl CoA 3-O-methyltransferase, CCR cinnamoyl CoA reductase, CAD, cinnamyl alcohol dehydrogenase

Fig. 2

Characterization and identification of maize C3H. a C3H activity in several tissues from different plant species: Brachypodium distachyon 30-day-old stems and leaves, Zea mays 30-day-old in vitro grown roots, Zea mays 40-day-old field grown stems and leaves, Panicum virgatum mature tillers, Medicago truncatula young stems, and Arabidopsis thaliana 30-day-old stems. TAL and CSE denote the presence in the genomes of these plants of bifunctional l-phenylalanine/l-tyrosine ammonia-lyase and orthologs of caffeoyl shikimate esterase genes, respectively. b C3H activity in crude protein extracts from maize roots tested under the conditions indicated. Abbreviations and concentrations: 4CA 4-coumarate (1 mM), BSA bovine serum albumin (75 μg/ml), NaP sodium phosphate buffer pH 6 (75 mM), ASC ascorbate (4 mM), O₂ incubation with oxygen (+) or under N₂ atmosphere (–), DTT dithiothreitol (100 mM), BME 2-mercaptoethanol (250 mM), H₂O₂ hydrogen peroxide (0.03%), CAT catalase (11 KU ml^–1), GSH glutathione (2 mM), NADPH reduced nicotinamide adenine dinucleotide phosphate (2 mM), NADH reduced nicotinamide adenine dinucleotide (2 mM), (NH₄)₂SO₄ ammonium sulfate (1 mM). c C3H (black line) and TAL (gray line) activity in different FPLC fractions prepared from maize root extracts. d SDS–PAGE gel used to separate the proteins of the fractions obtained from gel filtration chromatography. The main band from fraction 14 (red arrow) was excised and subjected to trypsin digestion and peptide mass mapping by ESI-MS/MS. The dashed box outlines the main protein band correlated with the C3H activity (+, –, n.d., high, low, or not detected, respectively). The full uncropped gel is provided as a Source Data file. e MS/MS fragmentation of one peptide matching the C3H protein sequence. The y-, and b- ions are indicated, ++ denotes doubly charged ions and * denotes loss of NH₃ group. f Mascot summary obtained after using the monoisotopic masses of the tryptic digest to match against NCBI plant databases. Error bars indicate mean ± SD, two-sided unpaired t-test, n = 3

Fig. 3

Phenotypic characterization of Brachypodium c3h mutants. a Activation tagging construct pJJ2LBA and diagram of the T-DNA insertion in line JJ25124 (http://jgi.doe.gov). b Relative expression by qPCR, extractable activity and protein level by immunoblotting in c3h mutants compared with wild-type and T-DNA control line JJ22251 (apx3). The antibodies raised against C3H showed no cross-reactivity and detect a band of 29.5 kDa. Line numbers are indicated in each lane. The full uncropped gel and blot shown are provided as a Source Data file. c Growth phenotype and transverse stem sections (UV-autofluorescence and phloroglucinol-HCl staining) of c3h mutants and apx3 and wild-type controls. d Total lignin (upper panel), S/G ratio (middle panel), and relative monolignol composition (lower panel) determined by thioacidolysis in c3h mutants and apx3 and wild-type controls. e Correlation plots for C3H activity with total lignin amount and individual lignin monomers. f Metabolite concentrations in mature stems of c3h mutants compared with apx3 and wild-type controls. Lignans were recognized from their fragmentation patterns. H/G lignan is 14.58 267 297 hydroxyphenyl guaiacyl lignan; G-lignan is 15.88 297 411 323 guaiacyl lignan, and S-lignan is 16.14 327 361 239 syringyl lignan glycoside (the first number is the retention time in min and the others are key mass-to-charge ratios, m/z). CWr cell wall residue. Error bars indicate mean ± SD, two-sided unpaired t-test. The Rsquared value (R²) was calculated from the linear regression model using Excel. Data points for all biological replicates are shown

Fig. 4

Phenotypic characterization of Arabidopsis c3h1 mutants. a Position of the T-DNA insertion, mature Arabidopsis plants and transverse stem sections (UV-autofluorescence and phloroglucinol-HCl staining) of c3h1 mutants and wild-type controls. b Lignin levels (upper panel) and relative monolignol composition (lower panel) determined by thioacidolysis of c3h mutants in both Col-0 and Wassilewskija (Ws) backgrounds. c Screening for Arabidopsis c3h1/cse2 double mutants. T- and G- are T-DNA and gene specific primers used for genotyping. No c3h1/cse2 double mutants were obtained from over 300 F2 plants screened, and so cse2C3H1+/– (i.e. homozygous for cse2 and heterozygous for c3h1; lanes 4, and 11) and c3h1CSE2+/– (i.e., homozygous for c3h1 and heterozygous for cse2; lanes 5 and 14) were subsequently generated and their lignin content and composition estimated (Supplementary Fig. 8). The full uncropped gels are provided as a Source Data file. d Aborted seeds in c3h1CSE2+/– and cse2C3H1+/– mutant lines determined in 8 to 18 individual siliques (segregation ratios estimated in Supplementary Table 2). e Comparison of the self-fertilized F3 seeds of mutants and WT plants by light microscopy in young (middle panel) and mature (right panel) siliques. The left panel shows a magnification of the aborted ovules (circles) in cse2C3H1+/– mutants. f Visual phenotype and transverse stem sections (UV-autofluorescence and Mäule staining) of c3h1/CSE-RNAi lines and c3h1 controls. g Total lignin amount and composition of c3h1/CSE-RNAi lines and c3h1 controls. h Visual phenotype and transverse stem sections (UV-autofluorescence and Mäule staining) of cse2/C3H-RNAi lines and cse2 controls. i Total lignin amount and composition of cse2/C3H-RNAi lines and cse2 controls. Mäule staining protocol was used in f and h to assess lignin distribution and composition. CWr cell wall residue. Error bars indicate mean ± SD, two-sided unpaired t-test. Data points for all biological replicates are shown

Fig. 5

Biochemical assays to study the fate of caffeate in Brachypodium and Arabidopsis. a Specific activities of individual caffeate 3-O-methyltransferase (COMT) and 4-hydroxycinnamate:CoA ligase (4CL) reactions in crude stem protein extracts from 1-month-old Brachypodium and Arabidopsis plants, performed at 10 and 50 μM caffeate concentrations (left panel), and double reactions performed by co-incubating caffeate with both cofactors required to perform the CoA activation (CoA+ATP) and 3-methoxylation (S-adenosyl methionine, SAM) reactions (right panel). The final common product of both parallel activities is feruloyl CoA. b Scheme of the studied reactions including cofactors and showing the proposed most favored pathways in the model monocot Brachypodium (red arrows) and the model dicot Arabidopsis (black arrows). The bar plot displays the activity ratios of competing enzymatic activities for both species and substrate (caffeate) concentrations calculated from the data shown in panel a. c Labeling patterns of lignin monomers in isotopic feeding experiments (m/z, mass-to-charge ratio). d Percentage of ¹³C-labeled ferulate incorporated into different monolignols (H-, G-, and S-units) and total lignin (T) in roots of Brachypodium and Arabidopsis seedlings. C3H 4-coumarate 3-hydroxylase, CSE caffeoyl shikimate esterase, CCoAOMT caffeoyl CoA 3-O-methyltransferase. Error bars indicate mean ± SD, two-sided unpaired t-test. n = 3