11
Review Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana Thomas Goujon a , Richard Sibout b , Aymerick Eudes a , John MacKay c , Lise Jouanin a, * a Laboratoire de Biologie Cellulaire, Institut National de la RechercheAgronomique (INRA), Route de Saint Cyr, Versailles cedex 78026, France b Centre de Foresterie des Laurentides, 1055 rue du PEPS, P.O. Box 3800, Sainte Foy, Quebec, Canada G1V4C7 c Centre de Recherche en Biologie Forestière, Université Laval, Quebec, Canada G1K 7P4 Received 6 December 2002; accepted 12 February 2003 Abstract Lignin is a complex polymer assembled from monolignol precursors derived from phenylalanine after several hydroxylation and methylation steps of the aromatic ring and reduction of the lateral chain. Three main monolignols, the p-coumaryl, coniferyl and sinapyl alcohols, give rise, respectively, to the hydroxyphenyl (H), guaiacyl (G), and syringyl (S) units of the polymer.A complete inventory of the genes potentially involved in the monolignol pathway in the model plant, Arabidopsis thaliana, is presented in this review. Genes encoding enzymes implicated in constitutive lignin synthesis were identified on the basis of their homology to monolignol biosynthesis genes of other plants and their high expression in lignified tissues (floral stems, roots). This overview shows that most of these genes belong to multigene families and that some (PAL, 4CL, CAD) are duplicated in this model plant. The genes encoding the cytochrome P450 monooxygenases (C4H, C3H, F5H) are unique except for F5H that has at least one homologue gene present in the complete genome. Mutants and transgenic Arabidopsis lines deregulated in the monolignol biosynthesis pathway are listed and the impact of the target gene deregulation on growth, and lignin content and structure are reported. © 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: Arabidopsis; Lignin; Monolignol biosynthesis; Mutant; Secondary cell wall 1. Introduction Lignin is a plant phenolic biopolymer of complex struc- ture made up of three main p-hydroxycinammyl alcohol precursors or monolignols, namely p-coumaryl, coniferyl and sinapyl alcohols. Lignin (from the Latin lignum: wood), a characteristic feature of secondary cell walls, accounts for 20–30% of the dry mass of wood, second only to cellulose. Its appearance in evolution is linked to the development of the upright growth habit of terrestrial plants. Mechanical support and water conductive properties of vascular tissues are augmented by embedding specific cells like vessel ele- ments and fibers with lignin. The lignification of tissues is also part of the defense arsenal of plants to limit pathogen invasion. The monolignols differ structurally from one another by the number of methoxyl groups present on their aromatic ring; they possess zero, one and two methoxyl groups, re- spectively (Fig. 1). Lignification is the process of polymer- ization of the monolignols, p-coumaryl, coniferyl and si- napyl alcohols, each giving rise to the hydroxyphenyl (H), guaiacyl (G) or syringyl (S) lignin units, respectively. Depo- sition of lignin in the cell wall occurs simultaneously as the polymer is formed. Lignin units are linked by different bond- types within the same lignin macromolecule [11,24]. The most frequently encountered linkage is the labile b-O-4 ether bond, which is the target of most degradation techniques used to analyze the chemical structure of lignins and of delignification processes like pulping and bleaching. Lignin Abbreviations: CAD, cinnamyl alcohol dehydrogenase; Cald5H, conife- raldehyde 5-hydroxylase; AldOMT, 5-hydroxyconiferaldehyde O-methyltransferase; CaMV, cauliflower mosaic virus; CCoAMT, caffeoyl coenzyme A O-methyltransferase; CCR, cinnamoyl coenzyme A reductase; COMT, caffeic acid O-methyltransferase; C3H, coumaroyl coenzyme A 3-hydroxylase; C4H, cinnamate-4-hydroxylase; 4CL, 4-coumarate- coenzyme A ligase; EST, expressed sequence tag; F5H, ferulate 5-hydroxylase; G, guaiacyl unit; HCA, hydroxycinnamic acids; PAL, phe- nylalanine ammonia-lyase; S, syringyl unit. * Corresponding author. E-mail address: [email protected] (L. Jouanin). Plant Physiology and Biochemistry 41 (2003) 677–687 www.elsevier.com/locate/plaphy © 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. doi:10.1016/S0981-9428(03)00095-0

Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

Embed Size (px)

Citation preview

Page 1: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

Review

Genes involved in the biosynthesis of lignin precursorsin Arabidopsis thaliana

Thomas Goujon a, Richard Sibout b, Aymerick Eudes a, John MacKay c, Lise Jouanin a,*a Laboratoire de Biologie Cellulaire, Institut National de la Recherche Agronomique (INRA), Route de Saint Cyr, Versailles cedex 78026, France

b Centre de Foresterie des Laurentides, 1055 rue du PEPS, P.O. Box 3800, Sainte Foy, Quebec, Canada G1V 4C7c Centre de Recherche en Biologie Forestière, Université Laval, Quebec, Canada G1K 7P4

Received 6 December 2002; accepted 12 February 2003

Abstract

Lignin is a complex polymer assembled from monolignol precursors derived from phenylalanine after several hydroxylation andmethylation steps of the aromatic ring and reduction of the lateral chain. Three main monolignols, the p-coumaryl, coniferyl and sinapylalcohols, give rise, respectively, to the hydroxyphenyl (H), guaiacyl (G), and syringyl (S) units of the polymer. A complete inventory of thegenes potentially involved in the monolignol pathway in the model plant, Arabidopsis thaliana, is presented in this review. Genes encodingenzymes implicated in constitutive lignin synthesis were identified on the basis of their homology to monolignol biosynthesis genes of otherplants and their high expression in lignified tissues (floral stems, roots). This overview shows that most of these genes belong to multigenefamilies and that some (PAL, 4CL, CAD) are duplicated in this model plant. The genes encoding the cytochrome P450 monooxygenases (C4H,C3H, F5H) are unique except for F5H that has at least one homologue gene present in the complete genome. Mutants and transgenicArabidopsis lines deregulated in the monolignol biosynthesis pathway are listed and the impact of the target gene deregulation on growth, andlignin content and structure are reported.

© 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.

Keywords: Arabidopsis; Lignin; Monolignol biosynthesis; Mutant; Secondary cell wall

1. Introduction

Lignin is a plant phenolic biopolymer of complex struc-ture made up of three main p-hydroxycinammyl alcoholprecursors or monolignols, namely p-coumaryl, coniferyland sinapyl alcohols. Lignin (from the Latin lignum: wood),a characteristic feature of secondary cell walls, accounts for20–30% of the dry mass of wood, second only to cellulose.Its appearance in evolution is linked to the development of

the upright growth habit of terrestrial plants. Mechanicalsupport and water conductive properties of vascular tissuesare augmented by embedding specific cells like vessel ele-ments and fibers with lignin. The lignification of tissues isalso part of the defense arsenal of plants to limit pathogeninvasion.

The monolignols differ structurally from one another bythe number of methoxyl groups present on their aromaticring; they possess zero, one and two methoxyl groups, re-spectively (Fig. 1). Lignification is the process of polymer-ization of the monolignols, p-coumaryl, coniferyl and si-napyl alcohols, each giving rise to the hydroxyphenyl (H),guaiacyl (G) or syringyl (S) lignin units, respectively. Depo-sition of lignin in the cell wall occurs simultaneously as thepolymer is formed. Lignin units are linked by different bond-types within the same lignin macromolecule [11,24]. Themost frequently encountered linkage is the labile b-O-4 etherbond, which is the target of most degradation techniquesused to analyze the chemical structure of lignins and ofdelignification processes like pulping and bleaching. Lignin

Abbreviations: CAD, cinnamyl alcohol dehydrogenase; Cald5H, conife-raldehyde 5-hydroxylase; AldOMT, 5-hydroxyconiferaldehydeO-methyltransferase; CaMV, cauliflower mosaic virus; CCoAMT, caffeoylcoenzyme A O-methyltransferase; CCR, cinnamoyl coenzyme A reductase;COMT, caffeic acid O-methyltransferase; C3H, coumaroyl coenzyme A3-hydroxylase; C4H, cinnamate-4-hydroxylase; 4CL, 4-coumarate-coenzyme A ligase; EST, expressed sequence tag; F5H, ferulate5-hydroxylase; G, guaiacyl unit; HCA, hydroxycinnamic acids; PAL, phe-nylalanine ammonia-lyase; S, syringyl unit.

* Corresponding author.E-mail address: [email protected] (L. Jouanin).

Plant Physiology and Biochemistry 41 (2003) 677–687

www.elsevier.com/locate/plaphy

© 2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved.doi:10.1016/S0981-9428(03)00095-0

Page 2: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

content and lignin subunit composition are known to varyaccording to taxa, tissues, cell types and cell wall layers[9,14]. Angiosperms possess guaiacyl (G) and syringyl (S)lignins and the S/G ratio calculated from the labile ligninfraction is used to make inferences on the overall ligninstructure.

The monolignol biosynthesis pathway can be divided intotwo parts: the general phenylpropanoid pathway (from phe-nylalanine to the hydroxycinnamic acids (HCA) and theirCoA-esters and the monolignol specific pathway (reductionof the HCA-CoA esters into monolignols). The first enzymeof the general phenylpropanoid pathway is phenylalanine

ammonia-lyase (PAL), which deaminates phenylalanine, anaromatic amino acid derived from the shikimate pathway.Further enzymatic reactions include hydroxylation of thearomatic ring, methylation of selected phenolic hydroxylgroups, activation of cinnamic acids to cinnamoyl CoA estersand the reduction of these esters firstly to cinnamaldehydesand then to cinnamyl alcohols (Fig. 1). The lignin precursorsthus synthesized are then linked via oxidative coupling cata-lyzed by both peroxidases and laccases. This pathway wasproposed to function as a “metabolic grid” [25] leading to Gand S units with hydroxylation and methylation reactionspotentially occurring at different steps of side chain oxida-

Fig. 1. Principal monolignol biosynthetic pathway in Angiosperms (adapted from [27,40,55]). PAL, phenylalanine ammonia-lyase; C4H, cinnamate-4-hydroxylase; C3H, coumaroyl-CoA-3-hydroxylase; 4CL, 4-coumarate-CoA ligase; CCoAOMT, caffeoyl CoA O-methyltransferase; CCR, cinnamoyl CoAreductase; F5H, ferulate-5-hydroxylase or coniferylaldehyde-5-hydroxylase; COMT, caffeic acid O-methyltransferase or 5-hydroxyconiferaldehydeO-methyltransferase; CAD, coniferyl alcohol dehydrogenase; SAD, sinapyl alcohol dehydrogenase. The main pathway is in black and alternative ways in blue.

678 T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 3: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

tion. However, recent data suggest a more restricted route forthe synthesis of G units and branching with 5' hydroxylationand methylation from coniferaldehyde to form sinapyl alco-hol precursors [13,28,41].

A large part of our current understanding of the monoli-gnol biosynthesis pathway comes from purification of en-zymes found in the xylem of forest trees like Eucalyptus,poplar and conifers or from annual plants like tobacco, al-falfa and others (reviewed in [4,6,43]). The cDNAs wereisolated based upon amino acid sequences information andon the basis of sequence similarity to genes identified in otherspecies. In recent years, Arabidopsis thaliana has become amodel plant for molecular studies related to development.Until recently, few studies of genes encoding the enzymes oflignin biosynthesis have focused on Arabidopsis, howeverwith the determination of its entire genomic DNA sequence[2] it has become possible to perform a systematic search forthese genes. Arabidopsis has a typical dicot lignificationpattern with G + S lignins [12]. The interfascicular fibers andthe xylem bundles form a ring of lignified tissues in itsmature stems (Fig. 2). Xylem elements and fiber lignins arecomposed of G and G + S units, respectively [44]. Dry floralstems contain 12–18% lignin according to the growth condi-tions with a S/G ratio of 0.4–0.5. However, these values candiffer slightly from one ecotype to another.

This report presents an overview of the genes potentiallyinvolved in each step of the monolignol biosynthesis path-way based upon an inventory performed with the entireArabidopsis genome sequence. In addition, the expressionpattern of the genes believed to be involved in lignin biosyn-thesis is reported, and the identification and characterizationof mutants and transgenic lines under- or over-expressinglignin genes are reviewed. In this review, we focus on thebiosynthesis pathway of monolignols, the precursors of thelignin polymer and will not consider the polymerization step,

where laccases and peroxydases are involved. Laccases andperoxydases belong to large multigene families of 17 and 73genes, respectively. Until now it has been unclear whichmembers of these families are directly involved in ligninpolymerization in plants and no Arabidopsis mutants havebeen characterized. Other genes could be important in thispathway including glycosyl transferases, beta-glucosidases,dirigent proteins and others, but available information are toopartial to be included in a review. An inventory of character-ized mutants which have defects in either the control ofsecondary cell wall deposition or secondary cell wall cellu-lose or lignin biosynthesis is made in a recent review [64].Another review lists the Arabidopsis mutants affected invascular differentiation and pattern formation [68].

2. State of the art

Taking advantage of cDNAs already cloned and character-ized in other plant species, Arabidopsis orthologs weresearched on the entire Arabidopsis genome [2] using thetblastn program [1]. This study (Table 1) shows that, in mostcases, genes involved in the monolignol pathway belong tosmall multigene families except for two enzymes involved inthe hydroxylation of cinnamate and p-coumarate, the P450-dependent monooxygenases (C4H, C3H). In many cases, theinvolvement of one of the gene family members is proposedon the basis of high amino acid sequence similarity withknown proteins of the monolignol pathway in other plants.

In order to estimate the expression level and distributionof these genes, two complementary approaches were taken,EST (Expressed Sequence Tag from random sequencing ofcDNAs) information available in public databases, and theexpression pattern in different plant organs or tissues (seed-lings, roots, leaves, stems). The number of ESTs identified indifferent public Arabidopsis cDNA libraries and their distri-bution in roots, plantlets, green siliques, aboveground or-gans, flower buds and developing seeds are indicated in Table1. These data show that the proposed “lignification” genesare identified in many different libraries. However, it must bepointed out that none of the ESTs sequencing has beencarried out from a floral stem cDNA library, the most ligni-fied part of the Arabidopsis plant. The expression patterns ofsome of the cDNAs and in particular those proposed to beinvolved in constitutive lignification have been determinedand are reported in Table 2. Their expression in most parts ofthe plant correlates with the presence of lignified vasculartissues.

Arabidopsis mutants have been identified for some of themonolignol biosynthesis genes in chemically (EMS) orT-DNA insertion collections (Table 3). The main phenotypiccharacteristics, lignin content and structure of previouslystudied mutants and antisense or sense Arabidopsis lines arereported in Table 4.

Fig. 2. Mäule staining of a cross-section of a floral stem of wild typeArabidopsis thaliana. The lignified tissues are stained in red if they containS units (F, interfascicular fibers) and in brown if only G units are present (X,xylem). Co, cortex; M, medular parenchyma.

679T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 4: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

Table 1Identification of the genes potentially involved in the monolignol biosynthesis pathway

Enzyme First identification of theprotein (origin)[reference]

Gene (accession number) Number of ArabidopsisESTs found in GenBank[1] Score > 200

Identity percentage withthe first identifiedprotein (Blast2) [61]

Reference

PAL (EC 4.3.1.5) AAA33770(P. vulgaris)[16]

AtPAL–1 (At2g37040) 39 83% [65]

AtPAL–2 (At3g53260) 35 83% [65]AtPAL–3 (At5g04230) 6 73% [65]AtPAL–4 (At3g10340) 21 80% This study

4CL (EC 6.2.1.12) CAA31696 (P. crispum)[15]

At4CL–1 (At1g51680) 13 69% [17]

At4CL–2 (At3g21240) 13 69% [17]At4CL–3 (At1g65060) 5 65% [17]At4CL-like1 (At3g21230) 0 60% This studyAt4CL-like2 (At4g05160) 3 41% This study

C4H (CYP73A5) JC1458 (V. radiata) [46] AtC4H (At2g30490) 16 78% [5]

C3H (CYP98A3) O22203 (A. thaliana)[55]

AtC3H (At2g40890) 22 100% [55]

F5H/Cald5H (CYP84A1) Q42600 (A. thaliana)[44]

AtF5H (At4g36220) 2 100% [44]AtF5H–2 (At5g04330) 0 65% This study

CCoAOMT (EC2.2.1.104)

A40975 (P. crispum)[56]

AtCCoAOMT–1(At4g34050)

13 85% This study

AtCCoAOMT–2(At4g26220)

2 55% This study

AtCCoAOMT–3(At1g67980)

0 54% [70]

AtCCoAOMT–4(At1g67990)

0 52% This study

AtCCoAOMT–5(At1g24735)

0 53% This study

AtCCoAOMT–6(At4g39060)

0 No significativealignment

This study

COMT/AldOMT (EC2.1.1.68)

Q00763 (P. tremuloides)[8]

AtOMT–1 (At5g54160) 84 79% [69]AtOMT–2 (At1g21100) 44 47% [23]AtOMT–3 (At1g77520) 7 48% [23]AtOMT–4 (At1g63140) 1 47% [23]AtOMT–5 (At3g53140) 1 38% [23]AtOMT–66 (At5g53810) 0 46% [23]AtOMT–77 (At4g35160) 4 34% [23]

CCR (EC 1.2.1.44) T10733 (E. gunnii) [36] AtCCR–1 (At1g15950) 32 73% [37]AtCCR–2 (At1g80820) 6 67% [37]

CAD (EC 1.1.1.195) CAD, P30359(N. tabucum) [35]

AtCAD-C (Atg19450) 25 76% (CAD) 53% (SAD) [61] and this study

SAD, AAK58693(P. tremuloides) [41]

AtCAD-D (At4g34230) 20 80% (CAD) 53% (SAD) [58,62]AtCAD–1 (At4g39330) 23 53% (CAD) 67% (SAD) [58,62]AtCAD-A (At4g37970) 0 53% (CAD) 68% (SAD) [58,62]AtCAD-B1 or Eli3–1(At4g37980)

18 50% (CAD) 71% (SAD) [58,62]

AtCAD-B2 or Eli3–2(At4g37s990)

15 52% (CAD) 72% (SAD) [58,62]

AtCAD-E (At2g21730) 0 51% (CAD) 64% (SAD) [58,62]AtCAD-F (At2g21890) 0 51% (CAD) 63% (SAD) [58,62]AtCAD-G (At1g72680) 7 45% (CAD) 49% (SAD) [58]

680 T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 5: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

2.1. Phenylalanine ammonia-lyase

PAL (EC 4.3.1.5) is considered to be the first enzyme ofthe general phenylpropanoid pathway and is responsible forthe phenylalanine deamination. Wanner et al. [65] identifiedthree PAL genes (AtPAL–1, –2, –3) which were differentiallyexpressed. AtPAL–1 and –2 are highly expressed in roots andalso in other tissues, but AtPAL–1 is unambiguously pre-dominant.An AtPAL–1 promoter-GUS fusion revealed a highexpression in the entire seedling which became progressivelyrestricted to vascular tissues as plants developed [48] sug-gesting its relationship to constitutive lignification in matureplants. A total of 39 and 35 ESTs for AtPAL–1 and –2,respectively, have been reported in cDNA libraries. AtPAL–3expression is very low and was observed only in young seed-lings [64] and is consequently rarely found in cDNA libraries(6 ESTs). We identified another PAL gene (AtPAL–4), whichis well represented in cDNA libraries (21 ESTs) but has notbeen previously studied.

Until now, PAL mutants have not been reported. Due to theputative functional redundancy of AtPAL–1 and AtPAL–2, a

mutation in one of these genes may not induce an easilydetectable phenotype, whereas a double AtPAL–1/ AtPAL–2null mutant may be lethal.

2.2. 4-Coumarate-coenzyme A ligase

4-Coumarate-coenzymeA ligase (4CL; EC 6.2.1.12) cata-lyzes the formation of CoA thioesters of hydroxycinnamicacids. Lee et al. [38] identified a gene (At4CL–1) highlyexpressed in seedlings and in floral stems. Ehlting et al. [17]reported the cloning of three 4CL genes (At4CL–1, –2 and–3). At4 CL–1 and –2 amino acid sequences are more similarto each other (86% identity) than to the At4 CL–3 sequence(71% and 73% identity, respectively). Phylogenetic analysisof available full length 4CL sequences placed At4CL–1 and–2 in 4CL class I and At4CL–3 in 4CL class II [17]. Theproteins At4CL–1 and –2 do not possess the same substratespecificity, while both enzymes convert 4-coumarate, onlyAt4CL–1 is capable of converting ferulate to the correspond-ing CoA-ester [18]. ESTs of At4CL–1 and –2 appear at equalfrequencies (13) in cDNA libraries whereas ESTs of

Table 2Expression profiles of different genes supposed to be involved in the monolignol biosynthesis pathway. NR, non-reported

Gene Seedlings Leaves Stem Root ReferencePAL–1 ++ + NR +++ [65]PAL–2 + +/– NR ++ [65]PAL–3 +/– – NR +/– [65]C4H ++ + +++ +++ [5,46]4CL–1 ++ + +++ ++ [17]4CL–2 ++ +/– + +++ [17]4CL–3 +/– +/– +/– – [17]C3H NR +/– +++ +/– [55]CCoAOMT–1 +/– +/– +++ NR This workF5H + + ++ a NR [44] and this workCOMT–1 + + +++ ++ [23]CCR–1 NR + +++ NR [37]CCR–2 NR – – NR [37]CAD–1 – +/– +++ ++ Eudes et al., unpublishedCAD-A – – + +/– Eudes et al., unpublishedCAD-B1 – + + + Eudes et al., unpublishedCAD-B2 – + + + Eudes et al., unpublishedCAD-C +/– ++ ++ +++ [58]CAD-D +/– + +++ +++ [58]CAD-E – – + + Eudes et al., unpublishedCAD-F – – – – Eudes et al., unpublishedCAD-G – – + + Eudes et al., unpublished

a The F5H expression is higher in the basal part of the stem than in the apical part.

Table 3Arabidopsis lines mutated in genes involved in the monolignol biosynthesis pathway

Mutant name Mutated gene Type of mutant Ecotype Referencefah1 F5H EMS and T-DNA Columbia [10]irx4 CCR–1 EMS Langsberg erecta [29]ref8 C3H EMS Columbia [20,21]ref3 C4H EMS Columbia [54]Atomt1 OMT–1 T-DNA Wassilevskija [23,30]AtcadC CAD-C T-DNA Wassilevskija [58]AtcadD CAD-D T-DNA Wassilevskija [31,58]AtcadG CAD-G T-DNA Wassilevskija Eudes et al., unpublished

681T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 6: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

At4CL–3 are less abundant [5]. At4CL–1 and –2 are ex-pressed in various tissues but the highest expressions werefound in floral stems and roots with a higher level of expres-sion for At4CL–1. At4CL–3 is expressed at a lower level(only detectable by RT-PCR). Two other 4CL related se-quences (At4CL– like 1 and –2) were identified in the Arabi-dopsis genome sequence (Table 1). The only At4CL– likeESTs identified in public cDNA databases were three se-quences of the At4CL– like–2 gene.

No 4CL Arabidopsis mutant has been reported but asmentioned for PAL genes, only a double At4CL–1 and –2mutant would be expected to present an obvious phenotypeand may be lethal if null. Antisense plants for the At4CL–1gene have been obtained and characterized [39]. Both theAt4CL–1 and the CaMV 35S promoters have been used withthe same efficiency to reduce 4CL activity down to 8% ofwild type. When 4CL activity is highly reduced, the Arabi-dopsis transgenic lines possess a normal development ingreenhouse with a 50% lignin content reduction and a higherS/G ratio due to an increase in S units and a decrease in Gunits (Table 4).

2.3. Hydroxylation steps (cinnamate-4-hydroxylase,coumaroyl CoA 3-hydroxylase and ferulate-5-hydroxylase)

The first hydroxylation in the phenylpropanoid pathwayoccurs at the C4 position of the aromatic ring and is carriedout by the cinnamate-4-hydroxylase (C4H; CYP73A5), aP450-dependant monooxygenase. One gene encoding C4His present in the Arabidopsis genome and 16 ESTs are foundin cDNA libraries. Northern experiments revealed that C4HmRNA is expressed in all tissues with highest levels in stemsand roots [5,47]. C4H promoter-GUS transcriptional fusionshows a diffuse staining in different tissues with a very high

staining in vascular tissues, such as veins of the leaves, xylemand sclerified parenchyma in the stems [5].

A line named ref3 for “Reduced Epidermal Fluorescence”was identified [54]; it contains a mutation in the C4H gene(cited in [20]). This mutant presents a decreased lignin con-tent and a modified composition (reduction in S units); it alsopossesses an altered development (dwarf, increased branch-ing, male sterility) [54]. The residual C4H activity in thismutant is not reported and may not be null.

The gene involved in the hydroxylation at the C3 positionhas been identified very recently [55]. The coumaroyl CoA3-hydroxylase (C3H; CYP98A3) is a cytochrome P450 en-zyme which catalyzes the 3'-hydroxylation of coumaroylquinate/shikimate leading to caffeoyl CoA and then to ligninmonomers. Present as a unique copy in Arabidopsis, it isexpressed in all plant tissues (northern experiments and 16ESTs) but its expression level is by far highest in stems andthen in roots and siliques [20,55].

A C3H mutant (ref8) was obtained and characterized byFranke et al. [20,21]. The mutant is dwarf and possesses alignin formed primarily from p-coumaryl alcohol, a mono-mer that is a minor component in the lignin in wild typeplants [20]. However, the mutant is probably not null sinceanother C3H mutant (T-DNA insertion) was lethal in thehomozygous condition (D. Werck, personal communica-tion).

Ferulate 5-hydroxylase (F5H) or coniferaldehyde5-hydroxylase (Cald5H; CYP84A1) is a cytochrome P450-enzyme which catalyzes the hydroxylation at the C5 posi-tion. This enzyme was first believed to act at the level offerulic acid but recent studies have demonstrated that F5Hpreferentially hydroxylates coniferaldehyde [27,49] and co-niferyl alcohol [27]. The gene was isolated [44] following the

Table 4Phenotype and lignin characteristics in Arabidopsis lines deregulated in the lignin monolignol biosynthesis pathway

Line (gene) Residual activity Development (stemsize)

Lignin content %/WT

Lignin composition Presence of unusualcompounds

References

AS4CL (4CL–1) 8% Normal 50 Increase S [39]Decrease G

ref3 (C4H) ND Dwarf Reduced Decrease S [53]ref8 (C3H) ND Dwarf Reduced H unit [20,21]irx4 (CCR–1) ND Reduced 50 [29]ASCCR (CCR–1) 22% Reduced 50 Coniferyl and

ferulic acids[22]

fah1 (F5H) ND Normal 100 G unit [10]35SF5H (Surex F5H) ND Normal 100 Increase S [45,57]C4HF5H (Surex F5H) ND Normal Reduced Mainly S 5-OH-G [42,45]Atomt–1 (OMT–1) 15% 1 Normal 100 G unit 5-OH-G [23,30]SOMT–22 (SurexOMT) 285%1 Normal 100 WT lignins [23,30]Atcad-C (CAD-C) 77% 2 Normal Normal WT lignin [58]

34% 3

Atcad-D (CAD-D) 20% 2 Normal Reduced Decrease S Sinapaldehyde [31,58]2% 3

Atcad-G (CAD-C) 75% 2 Normal ND ND Eudes et al.,unpublished60% 3

ND, non-determined; 1, using stem crude protein extracts and 5-OH coniferaldehyde as substrate; 2 and 3, using stem crude protein extracts and coniferylalcohol or sinapyl alcohol respectively as substrates surex, lines overexpressing a specific lignin enzyme.

682 T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 7: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

identification of the fah1 mutant which lacks sinapate ester inleaves [10]. F5H is expressed in stems, at a higher level in thebasal part [45]. It is rarely observed in the cDNA libraries (2ESTs) likely due to the lack of stem tissue cDNA librariesand perhaps, to its high G + C nucleotide content. Neverthe-less, Ruegger et al. [53] observed a significant expression insenescent leaves and roots. Expression in these tissues, how-ever, is not consistent with the absence of de novo sinapate-derived metabolite synthesis and extremely low levels ofS-unit lignin in these organs [57]. Moreover F5H is to ourknowledge, the only lignification gene where the role of the3' downstream region is known to be involved in tissuespecificity [53]. Finally, a second gene that we namedAtF5H–2 (Table 1) is present in the Arabidopsis genome, itshares 65% sequence identity with AtF5H [44] but no ESTshave previously been sequenced.

The fah1 (EMS and T-DNA insertion) mutants lackingF5H expression have a normal development in greenhouseconditions, and possess a normal lignin content but synthe-size lignin which is composed mainly of G units [10]. Inaddition, these mutants lack sinapic-acid-derived compo-nents typical of crucifers. The fah1–2 EMS mutant can becomplemented by constructs where the F5H gene is ex-pressed either under the control of the CaMV 35S or theArabidopsis C4H promoters [45,57]. Rescued lines with thehighest F5H mRNA expression possess lignin enriched in Sunits when compared to wild type. Moreover, some linesoverexpressing F5H under the C4H promoter contain ligninalmost entirely composed of S units in all lignified cells[42,45]. The tissue-specific distribution of S-lignin is thusabolished since xylem elements are normally devoid of

S-lignin. These data have provided strong evidence that F5Hacts in controlling lignin composition.

2.4. Methylation steps (CCoAOMT and COMT/AldOMT)

The first O-methylation at the C3 position involves thecaffeoyl Coenzyme A O-methyltransferase (CCoAOMT; EC2.1.1.104). This enzyme, first characterized in a parsley cellsuspension culture [50], was latter hypothesized to be in-volved in monolignol biosynthesis on the basis of its expres-sion profile [66,67]. The characterization of transgenic plantsunderexpressing CCoAOMT has confirmed its role in thispathway (reviewed in [13,43]). Only one ArabidopsisCCoAOMT gene (named AtCCoAOMT–3 in this work) hadbeen described until now [70]. We have identified five otherCCoAOMT genes in the Arabidopsis genome but one ofthem, AtCCoAOMT–6, is highly divergent from the otherprotein sequences (Table 1) and contains none of the aminoacid regions characteristic of CCoAOMT proteins. For thisreason, it was eliminated for construction of the phylogenetictree of AtCCoAOMT based upon a comparison of amino acidsequences (Fig. 3). AtCCoAOMT–1 shares the highest simi-larity with CCoAOMT proteins previously shown as in-volved in constitutive lignification in other species. In con-trast, AtCCoAOMT–3, –4 and –5 belong to another groupwhere only one homolog has been identified in poplar(P93711). The position of AtCCoAOMT–2 remains unclearand no homologs have yet been identified in other species.The high expression level of AtCCoAOMT–1 in floral stemssuggests its involvement in the lignification process

Fig. 3. Phylogenetic relationships among CCoAOMT from Arabidopsis thaliana and other plant species. The radial tree was constructed by neighbor-joiningdistance using a Kimura matrix (Phylip software) after alignment with Bioedit and Clustalw. Line lengths indicate the relative distances between nodes.Bootstrap values > 50% of 100 replications for all branches are shown. Sequence of proteins potentially encoding CCoAOMT used for alignment are:AtCCoAOMT1; AtCCoAOMT2; AtCCoAOMT3; AtCCoAOMT4; AtCCoAOMT5; f, Nicotiana tabacum (AAC49914); g, Nicotiana tabacum (AAC49915); h,Nicotiana tabacum (AAC49916); i, Pinus taeda (AAD02050); j, Populus tomentosa (AAF44689); k, Populus alba × Populus glandulosa (AAK16714); l,Populus balsamifera subsp. trichocarpa (CAA10217); m, Populus kitakamiensis (P93711); n, Populus tremuloides (Q43095); o, Medicago sativa(AAC28973); p, Petroselinum crispum (A40975); q, Solanum tuberosum (BAC23054); r, Oryza sativa (BAA81777).

683T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 8: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

(Table 2). ESTs were found for only 2 of the AtCCoAOMTgenes (13 for AtCCoAOMT–1 and 2 for AtCCoAOMT–2) incDNA libraries.

No Arabidopsis mutant or antisense lines for AtC-CoAOMT–1 have been reported. However, it is expected thatthe phenotype of such mutant or antisense lines would besimilar to those of antisense tobacco and poplar plants whichhave a reduced lignin content and a modified lignin compo-sition with a decrease in G units (reviewed in [13,43]).

Caffeic acid O-methyltransferase (COMT; EC 2.1.1.68)or 5-hydroxyconiferaldehyde O-methyltransferase (Al-dOMT) was thought to methylate caffeic and5-hydroxyferulic acids [8]. However, analysis of lignins fromCOMT antisense plants (reviewed in [4,13,41]) suggestedthat COMT was preferentially involved in the second methy-lation step responsible for the formation of S units. Morerecently, Osakabe et al. [49] and Li et al. [40] demonstratedthat COMT preferentially uses 5-hydroxyconiferaldehydeinstead of caffeic or 5-hydroxyferulic acids. A COMT gene(AtOMT–1) was identified by Zhang et al. [69] and its expres-sion determined in some tissues was consistent with a role inlignification. Analysis of the complete Arabidopsis genomesequence reveals the presence of at least six other relatedsequences [23]. Among these genes, AtOMT–1 is the mostclosely related to those already reported as involved in ligni-fication in other dicot plants. A role in monolignol biosynthe-sis is also consistent with its high expression level in stems(Table 3). Numerous ESTs (84) of this gene are present incDNA databases. This high frequency could also be relatedto the involvement of AtOMT1 in both sinapate ester andmonolignol biosynthesis [23]. One other member of thisclass, the AtOMT–2 gene is also represented by several ESTs[44] but has an unknown role, whereas the other genes arepoorly or not represented in cDNA libraries.

The Atomt-1 null mutant [30] possesses a very lowO-methyltransferase residual activity for caffeic and5-hydroxyferulic acids, coniferaldehyde and5-hydroxyconiferyl alcohol [23]. It has a normal develop-ment in greenhouse conditions and the same lignin contentthan the wild type. However, its lignin structure is highlymodified since no S units are incorporated in the polymer. Sunits are partly replaced by 5-hydroxyguaicyl units (5-OH-G) which forms new structures, the benzodioxanes [52].In addition, the sinapate ester content of this line is reduced[23]. Introduction of a functional poplar OMT cDNA is ableto complement the Atomt1 mutant and restores the S/G ratioto the wild type level. However, overexpression of the generesulting in increased OMT activity had no impact on lignincomposition. In contrast, F5H overexpression altered ligninsubunit composition significantly [45]. It has thus been con-cluded that OMT is not the major limiting enzyme for S-unitbiosynthesis in xylem of Arabidopsis [23,30].

2.5. Cinnamoyl coenzyme A reductase

Cinnamoyl coenzyme A reductase (CCR; EC 1.2.1.44)catalyzes the reduction of the hydroxycinnamoyl CoA esters

and is considered as the first committed enzyme of the mono-lignol specific pathway [36]. Two cDNAs (AtCCR–1 and –2)have been identified and characterized [37]. Expression ofAtCCR–1 was high in stems whereas no expression ofAtCCR–2 was detected in several tissues. In contrast, thelevel of AtCCR–2 mRNA increased strongly and transientlyafter inoculation by a pathogen while the AtCCR–1 transcriptlevel did not change [37]. These expression patterns are incomplete concordance with the number of ESTs found foreach gene in cDNA libraries (32 and 6, respectively). Severalother related CCR genes have been identified by Jones et al.[29], however, their sequences were too divergent to beconsidered in this work.

The irx4 (for irregular xylem) mutant [63] is mutated inthe AtCCR–1 gene [29]. This line has a dwarf phenotype withcollapsed vessels and the lignin content is decreased (50%reduction). An antisense strategy was also used to obtainlines with a 20% residual CCR activity [22]. These plantswere dwarf and had the same lignin content as irx4. Inaddition, lignins of these antisense CCR Arabidopsis aremore condensed (less b-O-4 bonds) and contain sinapic andferulic acids [22] also observed in CCR antisense tobaccoplants [51].

2.6. Cinnamyl alcohol dehydrogenase

Several cinnamyl alcohol dehydrogenase (CAD; EC1.1.1.195) proteins and genes have been isolated and charac-terized in different plants (reviewed in [4,13]) but their roleshave not been clearly identified. Until recently, CAD proteinswere generally believed to catalyze the reduction of the 3cinnamaldehydes to cinnamyl alcohols, the last step ofmonolignol biosynthesis. In each plant species, one or twogenes were identified as involved in lignification. The isola-tion of AtCAD-related genes in Arabidopsis (Eli3 [34],CAD–1 [59], CAD–2 [3]) has been reported with very fewindications on their expression pattern. Tavares et al. [62]performed a bioinformatic search for different gene familiesand identified eight CAD-like genes comprising those al-ready known (Eli3, AtCAD–1 and –2). Analysis of the com-plete Arabidopsis genome sequence identified one additionalgene (AtCAD-G; Table 1). A phylogenetic analysis separatedthe CAD proteins in three main classes [58]. A first clusterincludes the two AtCAD proteins (AtCAD-C and AtCAD-D)related to CADs of other plants well documented to beinvolved in monolignol synthesis. The second cluster is moreclosely related to the new CAD identified in poplar by Li etal. (named SAD for Sinapyl Alcohol Dehydrogenase [41])and in alfalfa by Brill et al. (expressed after wounding [7]).AtCAD-G constitutes a separate third cluster by itself, withno similar genes cloned in other species. Most of theseArabidopsis genes are expressed in stem tissues althoughdifferent expression levels were observed (Table 2). Only theexpression of AtCAD-C, AtCAD-D and AtCAD–1 was re-vealed by northern blot in several tissues, RT-PCR was nec-essary to detect the others. The number of ESTs in cDNAlibraries is in agreement with these results since ESTs of

684 T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 9: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

AtCAD–1, AtCAD-C and AtCAD-D are found most fre-quently (23, 20 and 25, respectively). Also noteworthy, ESTsof AtCAD-B1 and -B2 (also called Eli 3–1 and –2) arerelatively frequently encountered (15 and 18, respectively)and are mainly found in plantlet and root cDNA libraries.Their expression has been previously linked to pathogeninfection [34]. Until mutants are available, the function ofthese genes is difficult to verify. In addition, Somssich et al.[60] conducted substrate affinity studies and suggested thatElI3–2 could be a benzyl alcohol dehydrogenase and not acinnamyl alcohol dehydrogenase. Recently, we have identi-fied null mutants for AtCAD-C, -D and -G genes ([31,58],Eudes et al., unpublished result). Enzymatic and chemicalanalysis carried out on these mutants show that CAD-D isresponsible for some lignin features in stems and roots [31].Further investigations of other single and double mutants ofthe CAD genes are underway to identify the function of eachmember of this family.

3. Conclusion

This review reports our current knowledge of lignin pre-cursor biosynthesis in Arabidopsis. The availability of thecomplete genome sequence enabled a thorough inventory ofputative genes encoding enzymes of this metabolic pathway.Most of the genes belong to small multigene families. Ho-mologies with cDNAs and genes known to be involved in thispathway and published expression studies completed by un-published data helped to identify which genes may be themost important for the precursor biosynthesis of constitutivelignin. Numbers of ESTs in cDNA libraries are generallyconsistent with mRNA abundance observed by northern orRT-PCR analysis (as observed for the AtCAD gene family).The frequency of ESTs thus constitutes a relatively reliableindicator of gene expression and helps to infer potentialinvolvement in a given pathway. In some cases, expressionpatterns identified two genes encoding the same enzyme asbeing involved in this pathway, (AtPAL–1 and –2, At4CL–1and –2, AtCAD-C and -D) and in other cases, only one geneseems to be involved (AtOMT–1, AtCCoAOMT–1, AtCCR–1). Chemical analysis of lignin from mutant rachis confirmedthe role of some of these genes in lignification. For example,the AtOMT–1 and the AtF5H mutants possess lignins devoidof S units and a AtCCR–1 mutant contains less lignin. Thesituation is still unclear for the AtCAD gene family whereseveral genes are candidates for a role in constitutive lignifi-cation. Moreover, AtCAD paralogs may be involved in bioticand abiotic stress, as it was shown for AtCCR–2.

The impact of deregulation (under- and over-expression)of specific monolignol biosynthesis genes in Arabidopsis isof interest since it seems, in some cases, to be similar to thatobserved in tobacco plants, in trees like poplar, and in cropslike maize and alfalfa. This has been shown for F5H[19,45,57], for CCR [22,51,63] for OMT [4,30] and CAD([31], Lapierre et al., personal communication). In contrast,down-regulation of 4CL has been shown to induce different

lignin phenotypes according to the target plants (Arabidopsis[39], tobacco [32,33], and poplar [26]). Reduction in lignincontent was observed in the three species but the impact onlignin structure was different (increase S/G ratio in Arabi-dopsis, decrease S/G in tobacco, no change in poplar). Thiscould be due to the target plant species but also to otherfactors such as the methods used to analyze lignin (only thenon-condensed lignin part is analyzed) and the type of 4CLgenes (4CL is a multigene family). A complete collection ofsingle and double Arabidopsis mutants with characterizedlignin phenotypes could lead to a more comprehensive un-derstanding of the consequences of down-regulation of spe-cific lignin biosynthesis genes and help to develop predictivemodels for crop plants and trees. Modifications of ligninquality and quantity constitute important targets to improveagro-industrial end uses such as pulp and paper making fromforest trees and digestibility forage crop for livestock[6,13,30,43].

Acknowledgements

The authors are grateful to Nicolas Feau for his helpfulassistance with the phylogenetic analysis, to Danièle Werckand Catherine Lapierre for communication of unpublishedresults.

References

[1] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, Basiclocal alignment search tool, J. Mol. Biol. 215 (1997) 403–410.

[2] Arabidopsis genome initiative, Analysis of the genome sequence ofthe flowering plant Arabidopsis thaliana, Nature 208 (2000) 796–815.

[3] M. Baucher, J. Van Doorsselaere, J. Gielen, M. Van Montagu, D. Inzé,W. Boerjan, Genomic nucleotide sequence of an Arabidopsis thalianagene encoding a cinnamyl alcohol dehydrogenase, Plant Physiol. 107(1995) 285–286.

[4] M. Baucher, B. Monties, M. Van Montagu, W. Boerjan, Biosynthesisand genetic engineering of lignin, Critical Rev. Plant Sci. 17 (1998)125–197.

[5] D. Bell-Lelong, J.C. Cusumano, K. Meyer, C. Chapple, Cinnamate-4-hydroxylase in Arabidopsis, Plant Physiol. 113 (1997) 729–738.

[6] A.M. Boudet, Lignins and lignification: selected issues, Plant Physiol.Biochem. 38 (2000) 81–96.

[7] E.M. Brill, S. Abrahams, C.M. Hayes, C.L.D. Jenkin, J.M. Watson,Molecular characterisation and expression of a wound-induciblecDNA encoding a novel cinnamyl-alcohol dehydrogenase enzyme inlucerne (Medicago sativa L.), Plant Mol. Biol. 41 (1999) 279–291.

[8] R.C. Bugos, V.L.C. Chiang, W.H. Campell, cDNA cloning, sequenceanalysis and seasonal expression of lignin-bispecific caffeic acid/5-hydroxyferulic acid O-methyl transferase of aspen, Plant Mol. Biol.17 (1991) 1203–1215.

[9] M.M. Campbell, R.R. Sederoff, Variation in lignin content and com-position. Mechanisms of control and implications for the geneticimprovement of plants, Plant Physiol. 110 (1996) 3–13.

[10] C.C.S. Chapple, T. Vogt, B.E. Ellis, C.S. Sommerville, An Arabidop-sis mutant defective in the general phenylpropanoid pathway, PlantCell 4 (1992) 1413–1424.

685T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 10: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

[11] L.B. Davin, N.G. Lewis, Phenylpropanoid metabolism: biosynthesisof monolignols, lignans, lignins and suberins, in: H.A. Stafford, R.K.Ibrahim (Eds.), Recent Advances in Phytochemistry, vol. 26, PlenumPress, New York, pp. 325–375.

[12] D.P. Dharmawardhana, B.E. Ellis, J.E. Carlson, Characterization ofvascular lignification in Arabidopsis thaliana, Can. J. Bot. 70 (1992)2238–2244.

[13] R.A. Dixon, F. Chen, D. Guo, K. Parvathi, The biosynthesis ofmonolignols: a “metabolic grid”, or independant pathway to guaiacyland syringyl units?, Phytochemistry 57 (2001) 1069–1084.

[14] D.A. Donaldson, Lignification and lignin topochemistry—an ultra-structural view, Phytochemistry 57 (2001) 859–873.

[15] C. Douglas, H. Hoffmann, W. Schulz, K. Hahlbrock, Structure andelicitor or U.V. light-stimulated expression of two 4-coumarate:CoAligases in parsley, EMBO J. 6 (1987) 1189–1195.

[16] K. Edwards, C.L. Cramer, G. Bolwell, R.A. Dixon, W. Schulch,C. Lamb, Rapid transient induction of phenylalanine ammonia-lyasemRNA in elicitor-treated bean cells, Proc. Natl. Acad. Sci. USA 87(1985) 9057–9061.

[17] J. Ehlting, D. Büttner, Q. Wang, C.J. Douglas, I.E. Somssich, E. Kom-brink, Three 4-coumarate:coenzyme A ligases in Arabidopsisthaliana represent two evolutionarily divergent classes inangiosperms, Plant J. 19 (1999) 9–20.

[18] J. Ehlting, J.J. Shin, C.J. Douglas, Identification of 4-coumarate:coenzyme A ligase (4CL) substrate recognition domains, Plant J. 27(2001) 455–465.

[19] R. Franke, C.M. McMichael, K. Meyer, A.M. Shirely, J.C. Cusumano,C. Chapple, Modified lignin in tobacco and poplar plants overexpress-ing the Arabidopsis gene encoding ferulate 5-hydroxylase, Plant J. 22(2000) 223–234.

[20] R. Franke, R.M. Hemm, J.W. Denault, M.O. Ruegger,J.M. Humpheys, C. Chapple, Changes in secondary metabolism anddeposition of an unusual lignin in the ref8 mutant of Arabidopsis,Plant J. 30 (2002) 33–45.

[21] R. Franke, J.M. Humpheys, M.R. Hemm, J.W. Denault, M.O. Rueg-ger, J.C. Cusumano, C.C. Chapple, The Arabidopsis REF8 geneencodes the 3-hydroxylase of phenylpropanoid metabolism, Plant J.30 (2002) 47–59.

[22] T. Goujon, V. Ferret, I. Mila, B. Pollet, K. Ruel, V. Burlat, J.P.Joseleau, Y. Barrière, C. Lapierre, L. Jouanin, Down-regulation ofAtCCR1 in Arabidopsis thaliana: Effects on phenotype, lignins andcell degradability, Planta (2003) in press.

[23] T. Goujon, R. Sibout, B. Pollet, B. Maba, L. Nussaume, N. Bechtold,F. Lu, J. Ralph, I. Mila, Y. Barriere, C. Lapierre, L. Jouanin, A newArabidopsis mutant deficient in the expression ofO-methyltransferase 1: impact on lignins and on sinapoyl esters, PlantMol. Biol. 51 (2003) 973–989.

[24] R. Hatfield, W. Vermerris, Lignin formation in plants. The dilemma oflinkage specificity, Plant Physiol. 126 (2001) 1351–1357.

[25] M. Higuchi, Lignin biochemistry: biosynthesis and degradation,Wood Sci. Technol. 24 (1990) 23–63.

[26] W. Hu, G.J. Harding, J. Lung, J.L. Popko, J. Ralph, D.D. Stokke,C.-J. Tsai, V.L. Chiang, Repression of lignin biosynthesis promotescellulose accumulation and growth of transgenic trees, Nature Bio-technol. 17 (1999) 808–812.

[27] J.M. Humphreys, M.R. Hemm, C. Chapple, New routes for ligninbiosynthesis defined by biochemical characterization of recombinantferulate 5-hydroxylase, a multifunctional cytochrome P450-dependant monooxygenase, Proc. Natl. Acad. Sci. USA 96 (1999)10045–10050.

[28] J.M. Humphreys, C.C. Chapple, Rewriting the lignin roadmap, Cur.Opin. Plant Biol. 5 (2002) 224–229.

[29] L. Jones, A.R. Ennos, S.R. Turner, Cloning and characterization ofirregular xylem 4 (irx4): a severely lignin deficient mutant of Arabi-dopsis, Plant J. 26 (2001) 206–216.

[30] L. Jouanin, T. Goujon, R. Sibout, B. Pollet, I. Mila, B. Maba, J. Ralph,M. Petit-Conil, C. Lapierre, Tuning lignin structure through geneti-cally silencing, restoring or increasing caffeic acidO-methyltransferase activity: evaluation with poplar and Arabidopsisthaliana models, Proceedings of the 11th International Symposium onWood and Pulping Industry. Nice, France, pp. 25–28.

[31] L. Jouanin, T. Goujon, R. Sibout, B. Pollet, I. Mila, J.-C. Leplé,G. Pilat, M. Petit-Conil, J. Ralph, C. Lapierre, Comparison of theconsequences on lignin content and structure of COMT and CADdownregulation in poplar and Arabidopsis thaliana models, in:M. Carson, C. Walter (Eds.), Plantation Forest Biotechnology for the21st Century, Research Signpost, Kerala, India, 2003 (in press).

[32] S. Kajita, Y. Katayama, S. Omori, Alterations in the biosynthesis oflignin in transgenic plants with chimeric genes for 4-coumarate:coenzyme A ligase, Plant Cell Physiol. 37 (1996) S957–S965.

[33] S. Kajita, S. Hishiyama, Y. Tomimura, Y. Katayama, S. Omori, Struc-tural characterization of modified lignin in transgenic tobacco plantsin which the activity of 4-coumarate:coenzyme A ligase is depressed,Plant Physiol. 114 (1997) 871–879.

[34] S. Kiedrowski, P. Kwalleck, K. Hahlbrock, I.E. Somssich, J.L. Dangl,Rapid activation of a novel plant defense gene is strictly dependent onthe Arabidopsis RPM1 disease resistance locus, EMBO J. 11 (1992)4677–4684.

[35] R.E. Knight, C. Halpin, W. Schuch, Identification and characterizationof cDNA clones encoding cinnamyl alcohol dehydrogenase fromtobacco, Plant Mol. Biol. 19 (1992) 793–801.

[36] E. Lacombe, S.W. Hawkins, J. van Doorsselaere, J. Piquemal,D. Goffner, O. Poeydomenge, A.M. Boudet, J. Grima-Pettenati, Cin-namoyl CoA reductase, the first committed enzyme of the ligninbranch biosynthetic pathway: cloning, expression and phylogeneticrelationships, Plant J. 11 (1997) 429–441.

[37] V. Lauvergeat, C. Lacomme, E. Lacombe, E. Lasserre, D. Roby,J. Grima-Pettenati, Two cinnamoyl-CoA reductase (CCR) genes fromArabidopsis thaliana are differentially expressed during developmentand in response to infection with pathogenic bacteria, Phytochemistry57 (2001) 1187–1195.

[38] D. Lee, M. Ellard, L.A. Wanner, K.R. Davis, C.J. Douglas, TheArabidopsis thaliana 4-coumarate:CoA ligase gene: stress and devel-opmentally regulated expression and nucleotide sequence of itscDNA, Plant Mol. Biol. 28 (1995) 871–884.

[39] D. Lee, K. Meyer, C. Chapple, C.J. Douglas, Antisense suppression of4-coumarate:coenzyme A ligase activity in Arabidopsis leads toaltered lignin subunit composition, Plant Cell 9 (1997) 1985–1998.

[40] L. Li, J.L. Popko, T. Umezawa, V.L. Chiang, 5-Hydroxyconiferylaldehyde modulates enzymatic methylation for syringyl monolignolformation, a new view of monolignol biosynthesis in Angiosperms, J.Biol. Chem. 275 (2000) 6537–6545.

[41] L. Li, X.F. Cheng, J. Leshkevich, T. Umezawa, S.A. Hardling,V.L. Chiang, The last step of syringyl monolignol biosynthesis inAngiosperms is regulated by a novel gene encoding sinapyl alcoholdehydrogenase, Plant Cell 13 (2001) 1567–1585.

[42] J.M. Marita, J. Ralph, R.D. Hatfield, C. Chapple, NMR characteriza-tion of lignins in Arabidopsis altered in the activity of ferulate5-hydroxylase, Proc. Natl. Acad. Sci. USA 96 (1999) 12328–12332.

[43] E. Mellerowicz, M. Baucher, B. Sundberg, W. Boerjan, Unravellingcell wall formation in the woody dicot stem, Plant Mol. Biol. 47(2001) 239–274.

[44] K. Meyer, J.C. Cusumano, C. Somerville, C.C.S. Chapple, Ferulate-5-hydroxylase from Arabidopsis thaliana defines a new family ofcytochrome P450-dependant monooxygenases, Proc. Natl. Acad. Sci.USA 93 (1996) 6869–6874.

[45] K. Meyer, A.M. Shirley, J.C. Cusumano, D.A. Bell-Lelong,C. Chapple, Lignin monomer composition is determined by theexpression of a cytochrome P450-dependent monooxygenase in Ara-bidopsis, Proc. Natl. Acad. Sci. USA 95 (1998) 6619–6623.

686 T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687

Page 11: Genes involved in the biosynthesis of lignin precursors in Arabidopsis thaliana

[46] M. Mizutani, D. Ohta, R. Sato, Purification and characterization of acytochrome P450 (trans-cinnamic acid 4-hydroxylase) from etiolatedmung bean seedlings, Plant Cell Physiol. 34 (1993) 481–488.

[47] M. Mizutani, S. Ohta, R. Sato, Isolation of a cDNA and a genomicclone encoding cinnamate 4-hydroxylase from Arabidopsis and itsexpression manner in planta, Plant Physiol. 113 (1997) 755–763.

[48] S. Ohl, S.A. Hebrick, J. Chory, C.J. Lamb, Functional properties of aphenylalanine ammonia-lyase promoter from Arabidopsis, Plant Cell2 (1990) 837–848.

[49] K. Osakabe, C.C. Tsao, L. Li, J.L. Popko, T. Umezawa, D.T. Carr-away, R.H. Smeltzer, C.P. Joshi, V.L. Chiang, Coniferyl aldehyde5-hydroxylation and methylation direct syringyl lignin biosynthesis inangiosperms, Proc. Natl. Acad. Sci. USA 96 (1999) 8955–8960.

[50] A.E. Pakusch, R.E. Kneusel, U. Matern, S-Adenosyl-L-methionine:trans-caffeoyl-coenzymeA 3- O-methyltransferase from elicitor-treated parsley cell suspension cultures, Arch. Biochem. Biophys. 271(1989) 488–494.

[51] J. Piquemal, C. Lapierre, K. Myton, A. O’Connell, W. Schuch,J. Grima-Pettenati, A.M. Boudet, Down-regulation of cinnamoyl-CoA reductase induces significant changes of lignin profiles in trans-genic tobacco plants, Plant J. 13 (1998) 71–83.

[52] J. Ralph, C. Lapierre, F. Lu, J.M. Marita, J. Van Doorseleare, W. Boer-jan, L. Jouanin, NMR evidence for benzodioxane structure resultingfrom incorporation of 5-hydroxyconiferyl alcohol in lignins ofO-methyltransferase-deficient plants, J. Agri. Food Sci. 49 (2001)86–91.

[53] M. Ruegger, K. Meyer, J. Cusumano, C. Chapple, Regulation offerulate-5-hydroxylase expression in Arabidopsis in the context ofsinapate ester biosynthesis, Plant Physiol. 119 (1999) 101–110.

[54] M. Ruegger, C. Chapple, Mutations that reduce sinapoylmalate accu-mulation in Arabidopsis thaliana define loci with diverse roles inphenylpropanoid metabolism, Genetics 159 (2001) 1741–1749.

[55] G. Schoch, S. Goepfert, M. Morant, A. Hehn, D. Meyer, P. Ullmann,D. Werck-Reichhart, CYP98A3 from Arabidopsis thaliana is a3'-hydroxylase of phenolic esters, a missing link in the phenylpro-panoid pathway, J. Biol. Chem. 276 (2001) 36566–36574.

[56] D. Schmitt, A.E. Pakusch, U. Matern, Molecular cloning, induction,and taxonomic distribution of caffeoyl-CoA-3-O-methyltransferase,an enzyme involved in disease resistance, J. Biol. Chem. 266 (1991)17416–17423.

[57] R. Sibout, M. Baucher, M. Gatineau, J. Van Doorsselaere, I. Mila,B. Pollet, B. Maba, G. Pilate, C. Lapierre, W. Boerjan, L. Jouanin,Expression of a poplar cDNA encoding a ferulate-5-hydroxylase/coniferaldehyde-5-hydroxylase increases S lignin depo-sition in Arabidopsis thaliana, Plant Physiol. Biochem. 40 (2002)1087–1096.

[58] R. Sibout, A. Eudes, B. Pollet, T. Goujon, I. Mila, F. Granier,A. Séguin, C. Lapierre, L. Jouanin, Expression pattern of two paralogsencoding cinnamyl alcohol dehydrogenases in Arabidopsis. Isolationand characterization of the corresponding mutants, Plant Physiol. 132(2003) in press.

[59] D.A. Somers, J.P. Nourse, J.M. Manners, S. Abrahms, J.M. Watson, Agene encoding a cinnamyl alcohol dehydrogenase homolog in Arabi-dopsis thaliana, Plant Physiol. 108 (1995) 1309–1310.

[60] I.E. Somssich, P. Wernert, S. Kiedrowski, K. Hahlbrock, Arabidopsisthaliana defense-related protein ELI3 is an aromatic alcohol:NADP+oxidoreductase, Proc. Natl. Acad. Sci. USA 93 (1996) 14199–14203.

[61] T.A. Tatusova, T.L. Madden, BLAST sequences, a new tool for com-paring protein and nucleotide sequences, FEMS Microbiol. Lett. 174(1999) 247–250.

[62] R. Tavares, S. Aubourg, A. Lecharny, M. Kreis, Organization andstructural evolution of four multigene families in Arabidopsisthaliana: AtLCAD, AtLGT, AtMYST and AtHD-GL2, Plant Mol. Biol.42 (2000) 703–717.

[63] S. Turner, C.R. Sommerville, Collapsed xylem phenotype of Arabi-dopsis thaliana identifies mutants deficient in cellulose deposition inthe secondary cell wall, Plant Cell 9 (1997) 689–701.

[64] S.R. Turner, N. Taylor, L. Jones, Mutations of the secondary cell wall,Plant Mol. Biol. 47 (2001) 209–219.

[65] L.A. Wanner, G. Li, D. Ware, I.E. Somssich, K.R. Davis, The pheny-lalanine ammonia-lyase gene family in Arabidopsis thaliana, PlantMol. Biol. 27 (1995) 327–338.

[66] Z.-H. Ye, R.E. Kneusel, U. Matern, J.E. Varner, An alternative methy-lation pathway in lignin biosynthesis in Zinnia, Plant Cell 6 (1994)1427–1439.

[67] Z.-H. Ye, R.E. Kneusel, U. Matern, J.E. Varner, Differential expres-sion of two O-methyltransferases in lignin biosynthesis in Zinniaelegans, Plant Physiol. 108 (1995) 459–467.

[68] Z.-H. Ye, Vascular tissue differentiation and pattern formation inplants, Annu. Rev. Plant Biol. 53 (2002) 183–202.

[69] H. Zhang, J. Wang, H.M. Goodman, An Arabidopsis gene encoding aputative 14-3-3-interacting protein, caffeic acid/5-hydroxyferulic acidO-methyltransferase, Biochem. Biophys. Acta 1353 (1997) 199–202.

[70] J. Zou, D.C. Taylor, Isolation of an Arabidopsis thaliana cDNAhomologous to parsley (Petroselinum crispum) S-adenosyl-L-methionine: trans-caffeyol-Coenzyme A 3- O-methyltransferase, anenzyme involved in disease resistance, Plant Physiol. Biochem. 32(1994) 423–427.

687T. Goujon et al. / Plant Physiology and Biochemistry 41 (2003) 677–687