three 4-coumarate:coenzyme a ligases in arabidopsis thaliana represent two evolutionarily divergent...

Three 4-coumarate:coenzyme A ligases in Arabidopsisthaliana represent two evolutionarily divergent classes inangiosperms

JuÈ rgen Ehlting1, Daniela BuÈ ttner1, Qing Wang2,

Carl J. Douglas2, Imre E. Somssich1 and

Erich Kombrink1,*

1Max-Planck-Institut fuÈr ZuÈchtungsforschung, Abteilung

Biochemie, Carl-von-LinneÂ Weg 10, D-50829 KoÈln,

Germany, and2Department of Botany, University of British Columbia,

Vancouver, British Columbia V6T 1Z4, Canada

Summary

The enzyme 4-coumarate:CoA ligase (4CL) plays a key

role in channelling carbon ¯ow into diverse branch

pathways of phenylpropanoid metabolism which serve

important functions in plant growth and adaptation to

environmental perturbations. Here we report on the

cloning of the 4CL gene family from Arabidopsis thaliana

and demonstrate that its three members, At4CL1, At4CL2

and At4CL3, encode isozymes with distinct substrate

preference and speci®cities. Expression studies revealed

a differential behaviour of the three genes in various

plant organs and upon external stimuli such as wounding

and UV irradiation or upon challenge with the fungus,

Peronospora parasitica. Phylogenetic comparisons

indicate that, in angiosperms, 4CL can be classi®ed into

two major clusters, class I and class II, with the At4CL1

and At4CL2 isoforms belonging to class I and At4CL3 to

class II. Based on their enzymatic properties, expression

characteristics and evolutionary relationships, At4CL3 is

likely to participate in the biosynthetic pathway leading

to ¯avonoids whereas At4CL1 and At4CL2 are probably

involved in lignin formation and in the production of

additional phenolic compounds other than ¯avonoids.

Introduction

The general phenylpropanoid pathway channels carbon

¯ow from primary metabolism to different branch path-

ways of secondary phenolic metabolism via the sequential

action of the enzymes phenylalanine ammonia-lyase (PAL),

cinnamate 4-hydroxylase (C4H) and 4-coumarate:CoA

ligase (4CL). The latter enzyme, 4CL (EC 6.2.1.12), converts

4-coumaric acid and other substituted cinnamic acids such

as caffeic acid and ferulic acid into the corresponding CoA

thiol esters which are used for the biosynthesis of

numerous phenylpropanoid-derived compounds including

¯avonoids, iso¯avonoids, lignin, suberins, coumarins and

wall-bound phenolics (Dixon and Paiva, 1995; Douglas,

1996). It has been proposed that 4CL isoforms differing in

their substrate speci®cities may direct the ¯ux from general

phenylpropanoid metabolismo into the different end

product-speci®c pathways by supplying appropriate mix-

tures of substrates for subsequent reactions and/or speci®c

metabolic needs. In support of such functions, native 4CL

isoforms with distinct capacities to convert different

substituted 4-hydroxycinnamates have been reported in

soybean, petunia and pea (Allina et al., 1998, and references

therein; Hahlbrock and Grisebach, 1979).

4CL genes have been cloned from numerous plant

species, where they exist in small gene families. In some

plants, e.g. parsley, potato and loblolly pine, the cloned

genes encode identical or nearly identical proteins (Becker-

AndreÂ et al., 1991; Douglas et al., 1987; Zhang and Chiang,

1997), whereas in other plants, e.g. tobacco, soybean,

hybrid poplar (Populus trichocarpa 3 P. deloides) and

aspen (Populus tremulopides), structurally divergent forms

have been isolated (Allina et al., 1998; Hu et al., 1998; Lee

and Douglas, 1996; Uhlmann and Ebel, 1993). However,

only in few cases has the existence of functionally different

4CL proteins been correlated with the differential expres-

sion of divergent 4CL gene family members (Allina et al.,

1998; Hu et al., 1998). In aspen, Pt4CL1 has been associated

with lignin biosynthesis since it has activity towards

substituted hydroxycinnamic acids including 5-hydroxy-

cinnamate, and the corresponding gene is expressed in

lignifying xylem (Hu et al., 1998). Aspen Pt4CL2 is consid-

ered to be involved in the biosynthesis of other phenolics

such as ¯avonoids, since it has highest activity towards

coumarate, lacks activity towards 5-hydroxyferulate, and

the corresponding gene is preferentially expressed in the

epidermis of leaves and stems (Hu et al., 1998). Likewise,

the two soybean 4CL genes are differentially regulated.

Based on its expression pattern, the pathogen-inducible

GM4CL16 is speculated to encode a ¯avonoid-speci®c

isoform involved in the biosynthesis of phytoalexins,

whereas the biochemical role of the second, non-inducible

isoform is unclear (Uhlmann and Ebel, 1993). Thus, despite

its central position in phenylpropanoid metabolism, the

precise function of 4CL isoforms in directing metabolic ¯ux

to speci®c phenolic compounds is largely unknown.

Received 18 January 1999; revised 26 April 1999; accepted 3 May 1999.*For correspondence (fax +49 2215062 313;e-mail [email protected]).

The Plant Journal (1999) 19(1), 9±20

ã 1999 Blackwell Science Ltd 9

Con¯icting evidence exists concerning the ability of 4CL

to activate sinapate in preparation for reduction to sinapyl

alcohol, a lignin monomer incorporated into the syringyl

lignin of angiosperms. Isoforms from soybean, petunia

and pea have been reported to convert sinapate to sinapyl-

CoA, whereas native or recombinant enzymes from most

other plants do not (Allina et al., 1998, and references

therein). Antisense suppression of 4CL activity in Arabi-

dopsis and tobacco has also led to con¯icting results

(Kajita et al., 1997; Lee et al., 1997).

Phenylpropanoid metabolism is regulated primarily via

transcriptional control of the corresponding genes (Dixon

and Paiva, 1995). The genes PAL, C4H and 4CL have been

demonstrated to be coordinately expressed in an organ-,

cell-type-, and stimulus-speci®c manner (Koopmann et al.,

1999; Logemann et al., 1995; Reinold and Hahlbrock, 1997).

4CL expression is induced upon environmental stresses

such as wounding or UV irradiation, conditions under

which phenylpropanoid derivatives play an important

protective role (Dixon and Paiva, 1995; Lee et al., 1995).

Transcriptional activation of 4CL has also been shown after

elicitor treatment of cultured cells from different plants

including soybean, parsley and Arabidopsis (Douglas

et al., 1987; Trezzini et al., 1993; Uhlmann and Ebel, 1993),

and it has been demonstrated in soybean and parsley

following inoculation with Phytophthora sojae and in

potato infected with Phytophthora infestans (Becker-AndreÂ

et al., 1991; Schmelzer et al., 1989; Uhlmann and Ebel,

1993). In Arabidopsis, transient accumulation of 4CL

mRNA was detected upon in®ltration of leaves with an

avirulent strain of the bacterial pathogen, Pseudomonas

syringae pv maculicola (Lee et al., 1995).

Previously, 4CL was reported to be encoded by a single

gene in Arabidopsis and the corresponding cDNA was

characterized (Lee et al., 1995). In this study, we report on

the isolation of two additional, divergent Arabidopsis

cDNAs as well as the three genes encoding 4CL. Differ-

ences in enzymatic properties, endogenous expression

patterns and transcript accumulation following infection

with the fungus, P. parasitica, wounding or UV irradiation

strongly suggest different physiological functions of the

three isozymes in Arabidopsis. Based on phylogenetic

reconstruction, two classes of 4CL isoforms were identi®ed

among angiosperms and the functional differences of

these evolutionary divergent classes are discussed.

Results

Isolation and characterization of the 4CL genes and

cDNAs

A genomic 4CL clone was isolated from an Arabidopsis

Col-0 genomic library, using the published Arabidopsis

4CL cDNA as a probe (Lee et al., 1995). This gene contained

1.6 kb of putative promoter sequence, an open reading

frame interrupted by three introns, and 3.2 kb of 3¢-non-

coding sequence. Although 4CL was presumed to be a

single-copy gene in Arabidopsis (Lee et al., 1995), the open

reading frame of the genomic clone showed only 80%

identity to the cDNA. We designated the cDNA At4CL1 and

the isolated gene At4CL2 (Figure 1). To isolate the corre-

sponding At4CL2 cDNA, an Arabidopsis cDNA library was

screened with a subclone of At4CL2 containing the ®rst

exon. The resulting cDNA contained an open reading

frame of 1671 bp, 51 bp 5¢-untranslated region, 211 bp 3¢-non-coding region, and a poly(A) tail. The sequence of the

open reading frame was identical to that of At4CL2.

At4CL1 was PCR-ampli®ed from genomic DNA using

gene-speci®c primers derived from the 5¢ and 3¢ untrans-

lated regions of the cDNA. The product obtained contained

an open reading frame identical to that of the At4CL1

cDNA, interrupted by three introns (Figure 1). The promo-

ter sequence of At4CL1 was PCR-ampli®ed from genomic

DNA using the RAGE procedure (Cormack and Somssich,

1997). The PCR product contained 474 bp of the coding

region and 1 kb of 5¢ upstream promoter sequence. The

sequence of the coding region was in complete agreement

with that of the At4CL1 cDNA and the authenticity of the

At4CL1 promoter region was veri®ed by sequence com-

parison to an independent PCR ampli®cation product.

A third divergent gene, designated At4CL3, was identi-

®ed upon sequence analysis of a genomic clone previously

isolated by cross-hybridization to a parsley 4CL cDNA

(Trezzini et al., 1993). This truncated clone contained the

major part of the 4CL coding region interrupted by six

introns of different sizes and 0.6 kb of 3¢-non-coding

sequence (Figure 1). The coding region of At4CL3 showed

only 65% and 64% identity to At4CL1 and At4CL2,

respectively. The missing 5¢ region of At4CL3 and 645 bp

of putative promoter sequence was PCR-ampli®ed using

the RAGE procedure as described above. The correspond-

ing At4CL3 cDNA was isolated by screening a cDNA library

with the genomic At4CL3 insert. Sequence analysis of the

At4CL3 cDNA showed that it contains the complete coding

region of 1686 bp identical to that of At4CL3, ¯anked by

46 bp of 5¢ and 147 bp of 3¢ untranslated sequences.

Sequence comparisons of At4CL1, At4CL2 and At4CL3

indicated that the positions of three introns found in these

genes are conserved but that the introns differ both in

length and sequence. At4CL3 contains three additional

introns, one interrupting the region corresponding to the

®rst exon of At4CL1 and At4CL2, and two in the region of

their second exons (Figure 1). The 5¢ untranslated and

putative promoter regions show no overall sequence

similarities. Putative TATA boxes are located 124, 103

and 116 bp upstream of the ATG translation initiation sites

of At4CL1, At4CL2 and At4CL3, respectively (Figure 1).

Three sequence motifs, box P, box A and box L, which are

10 JuÈrgen Ehlting et al.

ã Blackwell Science Ltd, The Plant Journal, (1999), 19, 9±20

conserved among several plant PAL, C4H and 4CL genes,

including those in Arabidopsis (Logemann et al., 1995), are

also present in the At4CL3 gene promoter, whereas the

promoters of At4CL1 and At4CL2 contain only boxes P and

L (Figure 1).

Genomic DNA blot analyses

The limited sequence identity amongst the three genes

may explain why 4CL was previously assumed to be a

single-copy gene in Arabidopsis (Lee et al., 1995; Trezzini

et al., 1993). Indeed, unique sets of restriction fragments

were observed on genomic DNA blots using distinct

probes for each gene and high-stringency ®lter washing

conditions (Figure 2). When restriction enzymes were used

that do not cut within the probes, single hybridization

signals were observed in each case; with other enzymes,

all hybridization signals could be explained on the basis of

the available genomic sequences. Even under low-strin-

gency conditions, no clear cross-hybridization was

observed to additional fragments or between the Arabi-

dopsis 4CL genes (data not shown).

Enzymatic activities of the recombinant At4CL proteins

To examine the biochemical properties of the Arabidopsis

4CLs, the three cDNAs were expressed in E. coli in two

forms, as fusion proteins both containing and lacking N-

terminal His6-tags. Figure 3 shows that the recombinant

proteins had apparent molecular masses of approximately

66 kDa and reacted with an antibody raised against parsley

4CL. At4CL2 migrated to a similar position in the denatur-

ing polyacrylamide gel as did the 4CL present in protein

extracts of bolting stems. The slower migration of the

recombinant At4CL1 and At4CL3 is due to both having

longer N-terminal extensions (see Experimental proce-

dures).

Extracts from bacteria expressing the three recombinant

proteins lacking the His6-tag were tested for their abilities

to utilize different substituted cinnamic acids as substrates.

Figure 1. Structural features of the three 4CLgenes from Arabidopsis.(a) Schematic representation of the genestructures of At4CL1, At4CL2 and At4CL3.Exons are represented as dark grey boxes,open reading frames corresponding to eachother are indicated by the light greyshading.(b) Relative positions of the putative CAAT,TATA, P-, A- and L-boxes within the ®rst250 bp of the three At4CL promoters.(c) Comparison of the consensus sequencesof P-, A- and L-boxes (Logemann et al.,1995) with the corresponding motifs foundwithin the At4CL promoters.

Arabidopsis 4-coumarate:CoA ligase genes 11


All recombinant proteins converted 4-coumarate to the

CoA-ester and hence are bona ®de 4CLs. The Km values of

At4CL1 and At4CL3 for coumarate are similar to those

reported for many puri®ed plant 4CLs, whereas that of

At4CL2 is considerably higher, resulting in a 10-fold lower

speci®city (Vmax/Km) for this substrate (Table 1). The

apparent Km, relative Vmax and relative Vmax/Km values

determined for the different substrates tested are listed in

Table 1. It should be noted that cinnamate was also

converted with low ef®ciency by all three enzymes,

whereas sinapate was not converted at all.

No differences in 4CL activities were detected when the

proteins were expressed with or without an N-terminal

His6-tag (data not shown). In extracts of bacterial strains

harbouring the empty expression vector (pQE), no 4CL

protein (Figure 3) or 4CL activity (not shown) was detect-

able.

Quanti®cation of low-abundance mRNAs by RT±PCR

Since no cross-hybridization between the three 4CL genes

was observed by DNA blot analyses, we used the same

probes and hybridization conditions to study expression of

At4CL1, At4CL2 and At4CL3 by RNA blot analysis. How-

ever, At4CL3 mRNA was not detectable, presumably due

to its low abundance. To determine the relative mRNA

amounts of At4CL3 and compare its pattern of expression

to that of At4CL1 and At4CL2, we developed a rapid and

simple RT±PCR method for quanti®cation of low-abun-

dance mRNA. Using At4CL3-speci®c primers (one span-

ning an exon±intron junction to avoid ampli®cation of

genomic DNA), a PCR product of 340 bp was generated

and quanti®ed by staining with the ¯uorescent dye Sybr-

Green I after separation on polyacrylamide gels. Within the

range of 10±180 ng of total cotyledon RNA used as

template, a linear increase in the amount of PCR product

and ¯uorescence signal output was observed. A similar

relationship was obtained using At4CL3 cDNA as template

for PCR, whereas no ampli®cation was detectable from

At4CL1 or At4CL2 cDNA templates under the same

conditions. Up to a relative ¯uorescence of 25 000 units,

no saturation of the signal intensity was observed. Thus,

for further analyses, 100 ng of total RNA was used for RT±

PCR with relative ¯uorescence values not exceeding

Figure 2. Genomic DNA blot analysis of At4CL genes.Arabidopsis DNA (10 mg lane±1) was digested to completion with theindicated restriction enzymes, separated on agarose gels and blottedonto a nylon ®lter. The ®lter was ®rst hybridized with the At4CL1 cDNA,and after washing off the probe, subsequently rehybridized with agenomic 1.5 kb HindIII fragment of At4CL2 containing the ®rst exon, and®nally with a genomic 4.5 kb EcoRI/XbaI fragment of At4CL3.

Figure 3. Immunoblot analysis of therecombinant Arabidopsis 4CLs.Three 4CLs from Arabidopsis wereexpressed as fusion proteins with N-terminal His6-tag in E. coli and puri®edfrom bacterial extracts using Ni-NTAagarose columns. The puri®ed, recombin-ant proteins (0.5 mg lane±1), total proteinextracts from Arabidopsis bolting stemsand from elicitor-treated parsley cells(5 mg lane±1) were separated by SDS±PAGE,transferred to nitrocellulose membranesand probed with antiserum raised againstpuri®ed 4CL from parsley at a dilution of1:5000 for 1 h. Protein extracts (0.5 mglane±1) from bacteria containing the emptyexpression vector (pQE) served as controllane.



25 000 units. Using gene-speci®c primers, the relative

abundance of At4CL1 and At4CL2 mRNAs was also

determined by this method and resulted in expression

patterns virtually identical to the RNA blots shown below.

Expression patterns of the At4CL genes

The differential expression of the Arabidopsis 4CL gene

family was assessed by RNA blot analysis or quantitative

RT±PCR, as outlined above. Since the general phenylpro-

panoid pathway is pathogen-activated in all plants ana-

lysed to date, we inoculated Arabidopsis with the fungus

Peronospora parasitica, the causal agent of the downy

mildew disease, and studied expression of the At4CL

genes in comparison to other inducible Arabidopsis

phenylpropanoid and defence-related genes (Trezzini

et al., 1993). When challenged with P. parasitica pathovar

Noks1, an incompatible interaction is established on

Arabidopsis ecotype Ler-1 cotyledons which is character-

ized by formation of necrotic cavities without sporulation

of the fungus (Holub et al., 1994). Figure 4 shows that

At4CL1 and At4CL2 mRNA accumulation was strongly

induced, starting 12 h and 24 h after inoculation with

fungal spores, respectively. In contrast, the amount of

At4CL3 mRNA was not affected by infection (Figure 4b). As

expected, strong accumulation of AtPAL and AtC4H

mRNAs was also induced by infection, with patterns

similar to those of At4CL1 and At4CL2, respectively.

Expression of the peroxidase gene, AtPOX (Trezzini et al.,

1993), reached a maximum 4 days or later after inoculation

with fungal spores (Figure 4a).

Abiotic stresses such as wounding or irradiation with UV

light also activate phenylpropanoid gene expression

(Dixon and Paiva, 1995). At4CL1 mRNA accumulated

rapidly and transiently in wounded Arabidopsis leaves

(Figure 5a), consistent with previous reports (Lee et al.,

1995). The wound-induced expression pattern of At4CL2

was similar to that of At4CL1; however, the maximum

amount of At4CL2 mRNA was considerably lower than that

of At4CL1 or the wound-inducible allene oxide synthase

gene, AtAOS (Laudert et al., 1996), which served as a

positive control for wounding ef®ciency. At4CL3 mRNA

levels were not affected by wounding (Figure 5a).

When dark-adapted Arabidopsis plants were illuminated

with UV-containing white light, both At4CL1 and At4CL2

mRNAs accumulated rapidly and transiently, reaching

maximum levels 6 h after the onset of irradiation (Figur-

e 5b). Quantitative RT±PCR revealed that At4CL3 mRNA

accumulated more strongly after UV irradiation, but in

contrast to At4CL1 and At4CL2 remained at this elevated

level from 6 h to at least 24 h (Figure 5b). However, even

these elevated At4CL3 mRNA levels remained undetect-

able on RNA blots containing 20 mg of total RNA (data not

shown).

Analysis of the organ-speci®c expression pattern re-

vealed that mRNA levels of At4CL1 were highest in

seedling roots and in bolting stems of mature plants,

siliques contained intermediate levels, whereas mature

leaves and ¯owers contained only low levels (Figure 5c).

At4CL2 mRNA amounts were also high in seedling roots,

whereas low amounts were present in all other organs.

Both At4CL1 and At4CL2 were also expressed in 2- and 3-

Table 1. Kinetic properties of recombinant Arabidopsis 4CLs

Km Relative Vmax Relative Vmax /Km

(mM) (% of coumarate) (1/mM)

At4CL1 Cinnamate 6320 103 0.024-coumarate 38 100 2.6Caffeate 11 27 2.5Ferulate 199 53 0.26Sinapate n.c. ± ±

At4CL2 Cinnamate 6630 21 0.0034-coumarate 252 100 0.39Caffeate 20 74 3.7Ferulate n.c. ± ±Sinapate n.c. ± ±

At4CL3 Cinnamate 2070 164 0.084-coumarate 23 100 4.4Caffeate 374 129 0.35Ferulate 166 86 0.52Sinapate n.c. ± ±

Km and Vmax were determined by linear regression of v against v/s (Eadie±Hofstee plot) with at least six data points. Absolute Vmax

values with 4-coumarate as substrate were 1.13, 0.28 and 0.19 pkat ml±1 protein extract for At4CL1, At4CL2 and At4CL3, respectively.Relative Vmax values were obtained by setting Vmax for 4-coumarate for each enzyme to 100%. The data shown are representative forthree independent measurements. n.c., no conversion.



day-old seedlings, with higher levels of both transcripts in

3-day-old seedlings as previously reported (Lee et al., 1995)

(data not shown). A different expression pattern was

observed for At4CL3. The highest relative mRNA amount

was detected in ¯owers, intermediate mRNA amounts

were present in mature leaves and siliques, and low mRNA

amounts were observed in seedling roots and bolting

stems of mature plants (Figure 5c).

Phylogenetic comparison of 4CL genes

Although clearly divergent, the deduced amino acid

sequences of the isolated At4CL genes show signi®cant

similarity throughout their complete lengths, with the

predicted At4CL1 and At4CL2 amino acid sequences being

more similar to each other (86% identity) than to the

predicted At4CL3 sequence (71% and 73% identity, respec-

tively). Amino acid sequence similarity is most pro-

nounced in the central and C-terminal parts of the

proteins. Beginning at position G407 of At4CL1, the amino

acid motif GEICIRG is conserved at the same relative

positions in all plant 4CLs analysed to date, including

At4CL2 and At4CL3. This motif has been proposed to be

associated with stability and catalytic activity of 4CL and

related enzymes (Becker-AndreÂ et al., 1991). In addition, all

At4CLs contain a motif within the putative AMP binding

domain that is highly conserved in all known 4CLs as well

as in other enzymes that share a similar reaction mechan-

ism (Uhlmann and Ebel, 1993).

A phylogenetic reconstruction based on the different

plant 4CL amino acid sequences available from databases

is shown in Figure 6, and revealed that within the

angiosperms, two major 4CL groups have evolved, which

we designated class I and class II. At4CL1 and At4CL2 are

most closely related to each other within the class I cluster,

whereas At4CL3 is in the divergent class II cluster.

Discussion

Although 4CL was previously assumed to be encoded by a

single-copy gene in Arabidopsis, we have shown in this

paper that it actually comprises at least three genes. Based

Figure 4. Induction of 4CL and other defence-related mRNAs inArabidopsis upon infection with P. parasitica.(a) Arabidopsis Ler-1 seedlings were spray-inoculated with P. parasiticapv Noks1 and harvested at the times indicated. Control plants weresprayed with water. Total extractable RNA (20 mg lane±1) was separatedon agarose gels, blotted and hybridized with DNA probes speci®c forAt4CL1, At4CL2, C4H, PAL, peroxidase (POX). (b) At4CL3 mRNA wasampli®ed from 100 ng total RNA by quantitative RT±PCR as described inExperimental procedures. The products were separated on poly-acrylamide gels, stained with SybrGreen I and visualized with a¯uorescence imager. A false colour image is shown.

Figure 5. Wound-, UV-light- and organ-speci®c accumulation of At4CL mRNAs.(a) For wounding, Arabidopsis leaves werecut into 3±5 mm strips and incubated athigh humidity for the time periodsindicated. Time point 0 h represents leavesof untreated plants. Total extractable RNA(20 mg lane±1) was separated on agarosegels, blotted and hybridized with DNAprobes speci®c for At4CL1, At4CL2 and awound-inducible control gene, AtAOS,encoding allene oxide synthase. At4CL3mRNA quanti®cation by RT±PCR wasperformed as described in Figure 4.(b) For UV irradiation, dark-adaptedArabidopsis seedlings were exposed to UV-containing white light for the timesindicated and used for RNA extraction andanalysis as above.(c) Organs were separated and used forRNA extraction and analysis as above.



on the enzymatic activities of the recombinant proteins, all

three genes encode bona ®de 4CLs. However, the three

isoforms show clear differences in their substrate speci®-

cities. Detailed expression studies also demonstrated

differential regulation of the three At4CL genes, indicative

of different physiological functions. Because we failed to

isolate clones containing additional 4CL genes during

extensive screenings of both genomic and cDNA libraries,

we believe that these may represent the full complement

of 4CL genes in Arabidopsis. Search for `4CL-like'

sequences in the Arabidopsis databases and analysis of

two such sequences (gene F13C5.180 and EST 144C1t7)

suggested that they are unlikely to encode authentic 4CLs

since the degree of amino acid identity relative to bona

®de 4CLs is low (38±44%) and the amino acid motif

GEICIRG within the putative catalytic domain is poorly

conserved (T. Otke and C. Douglas, unpublished data).

Phylogenetic reconstruction with available amino acid

sequences revealed that 4CL forms two major clusters

among the angiosperms (Figure 6). Within the class I

cluster, the Arabidopsis 4CL1 and 4CL2 isoforms are most

closely related to each other, as are the isoforms from

hybrid poplar, tobacco, parsley and potato. This indicates

that class I isoforms have evolved relatively recently after

independent gene duplications within the respective plant

lineages. In contrast, the At4CL3 isoform groups within the

class II cluster of angiosperm 4CL sequences. The greater

evolutionary distance between At4CL3 and the other two

Arabidopsis genes is also obvious at the structural level,

with At4CL3 containing three additional introns that are

not present in At4CL1 and At4CL2. Genes encoding

divergent 4CL isoforms represented in both clusters have

also been cloned from soybean, aspen and Lithospermum,

indicating an early gene duplication event during angio-

sperm evolution. Clear differences in expression patterns

and/or enzymatic properties have been described that

distinguish class I and class II 4CL isoforms in Arabidopsis

(this study), soybean (Uhlmann and Ebel, 1993) and aspen

(Hu et al., 1998). In contrast, the analysis of class I isoforms

has revealed few if any enzymatic differences in several

plants (Allina et al., 1998, and references therein). How-

ever, in Arabidopsis, the class I isoforms do show

divergence in expression patterns and enzymatic proper-

ties. Therefore, an early gene duplication most likely has

lead to the evolution of functionally divergent 4CL iso-

forms of classes I and II, whereas the relatively recent

independent gene duplication events within the class I

cluster have not yet resulted in functionally distinct 4CLs in

all cases.

Although still based on limited data, we hypothesize that

a primary function of class II 4CL isoforms is to channel

activated 4-coumarate to chalcone synthase and subse-

quently to different branch pathways of ¯avonoid second-

ary metabolism leading to ¯ower pigments and UV

protective ¯avonols and anthocyanins (Figure 7). For

Arabidopsis, this proposal is based on our observations

that: (a) At4CL3 is expressed to relatively high levels in

¯owers, but not in ligni®ed organs such as bolting stems

and roots; (b) At4CL3 expression is strongly induced upon

UV irradiation whereas fungal infection and wounding

have no effects, and (c) At4CL3 has a high speci®city for 4-

coumarate, relative to caffeate and ferulate (the more

immediate precursors for lignin biosynthesis). A similar

function in the biosynthesis of non-lignin-related phenyl-

propanoids has been postulated for the related class II 4CL

from aspen (Pt4CL2), which is preferentially expressed in

immature organs but not in lignifying secondary xylem,

and whose enzymatic properties are consistent with it

being involved in ¯avonoid biosynthesis (Hu et al., 1998).

In addition, the class II 4CL isoform from soybean (4CL16)

is speci®cally induced in infected plants in coordination

with other enzymes of phenylpropanoid metabolism, such

as PAL and chalcone synthase, and is thus assumed to

participate in the biosynthesis of anti-microbial phytoalex-

ins derived from the ¯avonoid pathway (Uhlmann and

Ebel, 1993). Unlike soybean and other legumes, the

phytoalexins in Arabidopsis (Tsuji et al., 1992) are not

¯avonoids, perhaps explaining the lack of At4CL3 induc-

tion upon infection with P. parasitica.

The apparent low overall expression levels of At4CL3 in

all organs and upon all treatments tested may not

Figure 6. Phylogenetic reconstruction based on 4CL amino acidsequences.The most parsimonious tree was found using the TBR heuristic branchswapping algorithm with the PAUP 3.1.1 program. The number of stepsbetween nodes is shown above the branch length, and bootstrap values(500 replications) are shown below the branch length. The tree has aconsistency index of 0.793. The allelic pine 4CL1 and 4CL2 sequenceswere used as the outgroup to root the tree.



necessarily re¯ect the true cellular situation. At present it

cannot be excluded that At4CL3 mRNA accumulation is

highly restricted to a few speci®c cell types, e.g. epidermal

cells, as suggested for Pt4CL2 in aspen (Hu et al., 1998) Use

of At4CL3 promoter±reporter gene constructs for trans-

genic studies, and in situ localization of At4CL3 mRNA, will

certainly help to clarify this point.

All 4CL isoforms known so far from parsley, tobacco,

potato and hybrid poplar group within the class I cluster.

One possibility is that divergent class II isoforms exist in

these plants, but remain to be isolated. Indeed, a PCR-

based approach has revealed the existence in hybrid

poplar of an orthologue of the aspen gene Pt4CL2 (D.

Cukovic and C. Douglas, unpublished data). It is also

conceivable that class II isoforms have been lost during

evolution and have been functionally replaced by class I

isoforms in some plant species.

The two Arabidopsis class I 4CLs described here display

signi®cant differences in their enzymatic properties and in

their organ-speci®c expression patterns. At4CL1 shows a

high speci®city for 4-coumarate and caffeate in compar-

ison to ferulate. At4CL1 is also the only gene family

member strongly expressed in the ligni®ed bolting stems

of adult plants (Figure 5c), where in situ hybridization

showed that its expression is speci®c to the xylem and

sclerenchyma (Q. Wang and C. Douglas, unpublished

data). It is also expressed at the onset of lignin deposition

in seedling roots and cotyledons (data not shown) (Lee

et al., 1995). Together, these data strongly support a

primary role for At4CL1 in the biosynthesis of p-hydro-

xyphenyl and guaiacyl lignin subunits, the latter requiring

methylation of caffeoyl-CoA to feruloyl-CoA (Figure 7). A

similar role in lignin biosynthesis has also been postulated

for the class I Pt4CL1 isoform from aspen (Hu et al., 1998),

for genes encoding other class I 4CL isoforms such as

tobacco 4CL1 and 4CL2, which are highly expressed in

lignifying stems (Lee and Douglas, 1996), and for the

parsley 4CL1 gene, whose promoter directs strong expres-

sion to lignifying tissue (Hauffe et al., 1991).

The enzymatic properties of At4CL2 are unusual in so far

as caffeate is converted with higher ef®ciency than 4-

coumarate, and ferulate is not converted at all. But the

existence of an O-methylation pathway that operates on

caffeoyl-CoA to produce feruloyl-CoA (Ye et al., 1994)

suggests that At4CL2 could participate in the biosynthesis

of lignin and/or other caffeoyl-CoA derived phenolic

compounds, especially in seedling roots where At4CL2 is

expressed along with At4CL1. Since the existence of

caffeoyl-derived phenolics such as chlorogenic acid is

unknown in Arabidopsis, this possibility remains to be

experimentally addressed.

The lack of any detectable conversion of sinapate by any

of the Arabidopsis 4CL isoforms con®rms and extends

similar observations made with a number of other plants

and supports the hypothesis that the biosynthesis of

sinapyl alcohol and syringyl lignin may occur via a 4CL-

independent, as yet uncharacterized, pathway (Allina et al.,

1998; Lee et al., 1997). Independent gene duplications

within the class I cluster of different angiosperm plant

lineages may re¯ect the on-going evolution of divergent

and functionally specialized 4CL isoforms, such as At4CL1

and At4CL2 in Arabidopsis, as opposed to functionally

redundant 4CL isoforms, present in plants such as parsley,

potato and tobacco.

The fact that not only At4CL3, but also At4CL1 and

At4CL2, are induced after irradiation with UV-containing

white light, and that all three isoforms support the

biosynthesis of 4-coumaroyl-CoA, may argue against a

strict functional separation between class I and class II

isoforms. However, the light stress response of class I

isoforms may be required to support the activity of the

class II isoforms by increasing the supply of 4-coumaroyl-

CoA esters for the ¯avonoid biosynthetic pathway. Alter-

natively, other phenylpropanoid compounds in addition to

¯avonoids may be required to protect the plant from

damage by UV irradiation, and the differences in the extent

and timing of induction of At4CL3, as opposed to that of

At4CL1 and At4CL2, may re¯ect their distinct roles in

protection. Furthermore, it is likely that treatment with UV

light initiates developmental processes requiring the

synthesis of other phenylpropanoid compounds derived

from pathways controlled by At4CL1 and At4CL2.

The expression patterns of At4CL1 and At4CL2 in

response to UV irradiation, wounding or infection with

Figure 7. Proposed roles of Arabidopsis 4CL isoforms in biosyntheticpathways of phenylpropanoids.For 4CL, the arrow size re¯ects relative substrate speci®cities andconversion rates (Vmax/Km) by the indicated isoforms, as determined inthis paper. PAL, phenylalanine ammonia-lyase; C4H, cinnamate-4-hydroxylase; C3H, coumarate-3-hydroxylase; COMT, caffeic acid O-methyltransferase; F5H, ferulate-5-hydroxylase; CHS, chalcone synthase;CCo3H, coumaroyl-CoA-3-hydroxylase; CCoOMT, caffeoyl-CoA 3-O-methyltransferase; n.d., not determined; n.c., no conversion.



the fungus P. parasitica, showed no signi®cant differences,

except that the amounts of At4CL1 mRNA were, in general,

5±10 times higher than those of At4CL2 (data not shown).

These and earlier results, demonstrating rapid accumula-

tion of At4CL1 mRNA in leaves in®ltrated with an avirulent

strain of the bacterial pathogen P. syringae (Lee et al., 1995),

indicate that both genes may have important functions in

pathogen defence. These may include the biosynthesis of

lignin and other cell-wall-bound phenolic compounds at

the site of pathogen ingress, aimed at restricting the growth

of the micro-organisms. Cell-wall thickening and the

appearance of auto¯orescent, phenolic material at infection

sites are consistent with such a function and have been

observed in Arabidopsis (data not shown) and numerous

other plants (Kombrink and Somssich, 1995). In addition,

cell wall reinforcement and/or sealing of wounded tissue

with phenolics may help to limit subsequent, secondary

infections caused by opportunistic micro-organisms. A

defensive role for 4CL in Arabidopsis is further supported

by results obtained using Arabidopsis plants transgenic for

a parsley class I 4CL promoter±GUS fusion. In these plants,

a strong and rapid activation of the GUS reporter gene is

induced by Peronospora infection and locally restricted to

the vicinity of fungal infection sites (data not shown).

The interaction between Arabidopsis and P. parasitica is

an intensively studied plant pathosystem (Kunkel, 1996)

and a large number of genetically well-de®ned interaction

types have been described (Holub et al., 1994). Several

signal transduction mutants in defence response pathways

showing altered susceptibility to P. parasitica have been

isolated (Dong, 1998), along with a major resistance gene,

RPP5, conferring resistance against P. parasitica pv Noko1

(Parker et al., 1997). Little is known about the molecular and

biochemical mechanisms leading to the resistant pheno-

type and to the transcriptional activation of defence-related

genes at sites of fungal ingress. In this respect, further

studies of At4CL1 and At4CL2 should prove extremely

valuable, particularly for mutagenesis screens aimed at

identifying regulators of this locally activated plant defence

reaction. Finally, the isolation and analysis of knock-out

mutants in the three divergent 4CL genes should help to

elucidate the speci®c metabolic functions of the corre-

sponding enzymes.

Experimental procedures

Plant material and stress treatments

For organ isolation, Arabidopsis thaliana Col-0 plants were grownunder short-day conditions (8 h light, 16 h dark, 20°C) in phyto-chambers or under long-day conditions (16 h light, 8 h dark) in thegreenhouse. Healthy organs were isolated and immediatelyfrozen in liquid nitrogen.

Peronospora parasitica pathovar Noks1 was maintained byweekly inoculation of susceptible Arabidopsis ecotype Col-5 as

previously described (Dangl et al., 1992). For infection studies, 10-day-old Arabidopsis (Ler-1) seedlings grown in soil or on AMplates (0.5 3 MS salts supplemented with vitamins (Sigma,Deisenhofen), 1% sucrose, 0.6% agar) were spray-inoculated withconidia suspensions (50 per ml), maintained in humidity chambersunder short-day conditions (except that the temperature in thedark was 16°C). For each sample, approximately 5000 seedlingswere infected and harvested by freezing the whole pot or agarplate in liquid nitrogen. Aerial parts of the seedlings (cotyledonsand hypocotyl) were scratched off the frozen surface and stored at±80°C. Control plants were sprayed with water.

For wounding experiments, Arabidopsis Col-0 plants weregrown under short-day conditions for 3 weeks in phytochambers.Healthy looking, fully expanded leaves were detached, cut into 3±5 mm strips, placed on MS-soaked Whatman paper in sealed Petridishes, and incubated in the light. Samples were frozen in liquidnitrogen at the time points indicated. Control leaves weredetached and incubated without further wounding.

For UV-light treatment, Arabidopsis Col-0 plants were grownunder short-day conditions for 12 weeks. Plants were dark-adapted for 36 h and subsequently irradiated with UV-containingwhite light (Hartmann et al., 1998). Plants were harvested atdifferent times after onset of irradiation and stored at ±80°C.

Isolation of At4CL genomic and cDNA clones

Genomic regions of At4CL1 were PCR-ampli®ed from 100 nggenomic Col-0 DNA with gene-speci®c primers deduced from theArabidopsis 4CL1 cDNA (Lee et al., 1995) (forward primer B6F: 5¢-AATGGCGCCACAAGAACAAGCAGTTTCTCA-3¢, position 5±34 inthe cDNA; reverse primer B8R: 5¢-CACATTGAAATCATCTTTGC-TTTTTGGTTG-3¢, position 1731±1702). PCR was performed withpwo polymerase (Boehringer, Mannheim) with an annealingtemperature of 60°C for 30 cycles. PCR products of twoindependent ampli®cations were blunt-end-cloned into EcoRV-digested pBluescriptSK vector. Promoter regions of At4CL1 wereisolated using the RAGE procedure (Cormack and Somssich,1997). Brie¯y, genomic Col-0 DNA was digested with HindIII andsubsequently polyadenylated with terminal deoxynucleotidetransferase. The ®rst PCR ampli®cation was done with a gene-speci®c upstream directed primer, B4R2 (5¢-AGTCGATGACTTCT-GATGCCTCGGTGGTC-3¢, position 591±562 in the Arabidopsis4CL1 cDNA) and a universal-T17 primer using Taq+ polymerase(30 cycles at 94°C for 15 sec, 60°C for 30 sec, and 72°C for 3 min).Re-ampli®cation was performed under the same conditions with anested gene-speci®c primer, B7R (5¢-ATGACTACTCCGTCGTCG-TTTTGAAGTGGT-3¢, position 490±461) and the universal primer.The resulting 1686 bp PCR product was blunt-end-cloned into thepBluescriptSK vector. To verify the DNA sequence of the RAGEproduct, the isolated promoter region was ampli®ed fromgenomic DNA using the pwo polymerase with two speci®cprimers deduced from the RAGE product, B11F (5¢-ATTAACTG-CAGACACTTTTAGCCCATAACTTTC-3¢) and B11R (5¢-CTTGTGGC-GCCATGGTAAATAGTAAATATTGTG-3¢), and the resulting PCRproduct was cloned and sequenced.

At4CL2 was isolated by screening a lEMBL4 genomic libraryusing the Arabidopsis 4CL1 cDNA insert (Lee et al., 1995) as aprobe under high-stringency hybridization and washing (0.23

SSC, 0.1% SDS, 65°C) conditions. Overlapping restriction frag-ments were subcloned into the pBluescriptSK vector.

Sequence analysis of a previously isolated genomic 4CL clone(Trezzini et al., 1993) revealed that it contained the major part ofthe At4CL3 coding region. Clones containing the 5¢ end and thepromoter region of At4CL3 were isolated by the RAGE procedure



as described above using XhoI-digested genomic DNA. The ®rstPCR was performed with the universal-T17 primer and the genespeci®c primer R3 (5¢-GATTATGAGTTTGGCTCCGGAAGATT-TAAGC-3¢), the second PCR with the universal primer and anested gene-speci®c primer R2 (5¢-CGGTAGTGATCAGCGT-GAGGTTTTCTCC-3¢). The isolated region was subsequentlyampli®ed from genomic DNA using speci®c primers deducedfrom the RAGE product and cloned into the pBluescriptSK vector.

cDNA clones for At4CL2 and At4CL3 were isolated from a lZAPcDNA library (Korfhage et al., 1994). Approximately 160 000plaques were screened with a 1.6 kb HindIII genomic subclonespanning the ®rst exon of At4CL2 or with the 4.5 kb EcoRI/XhoIgenomic subclone of At4CL3 (Trezzini et al., 1993). For At4CL2 andAt4CL3, 16 and two positive clones, respectively, were identi®edin the initial screening round. Two full-length cDNA clonescorresponding to both genes were puri®ed and subsequentlysubcloned into the pBluescriptSK vector.

DNA sequences of all isolated clones were determined by theADIS service unit (Max-Planck-Institut fuÈ r ZuÈ chtungsforschung,KoÈ ln, Germany) on ABI PRISM 377 DNA sequencers (PE AppliedBiosystems, Foster City, California, USA) using BigDye Termina-tor chemistry.

Nucleic acid methods

Standard molecular biology procedures were performed accord-ing to Sambrook et al. (1989). Genomic DNA was isolated startingfrom 1 g of Arabidopsis Col-0 leaves using the `plant DNAisolation kit' (Boehringer, Mannheim, Germany) according tothe manufacturer's protocol. For DNA blot analyses, 10 mg of totalgenomic DNA per lane was digested with appropriate restrictionenzymes, separated on 0.8% agarose gels, and blotted onto nylonmembranes (Amersham Pharmacia Biotech, Freiburg, Germany).Probes were 32P-labelled using the `random primed labelling kit'(Boehringer). After hybridization, blots were washed undermoderate (23 SSC, 0.5% SDS, 65°C) or high-stringency conditions(0.23 SSC, 0.1% SDS. 65°C). Total RNA was isolated with the `totalRNA Maxi Kit' (Qiagen, Hilden) according to the manufacturer'sprotocol. For RNA blot analyses, 20 mg of total RNA per lane wereseparated on 1.2% agarose gels containing formaldehyde andblotted onto nylon membranes. Filters were washed under high-stringency conditions (0.23 SSC, 0.1% SDS, 65°C) after hybridiza-tion. Equal loading was monitored by ethidium bromide stainingof the gels.

The Arabidopsis C4H probe was generated by PCR usingspeci®c primers (forward: 5¢-GCCGACGATTTTCTCACCGG-3¢ andreverse: 5¢-TCGTAGAACGAACCATTTAAAGAC-3¢) deduced fromthe published C4H sequence (Bell-Lelong et al., 1997). Syntheticoligonucleotides were from MWG (Ebersbach, Germany). Apreviously isolated Arabidopsis genomic PAL clone (Trezziniet al., 1993) was used as probe.

Quantitative RT±PCR

The concentration of total RNA was determined after dilution toapproximately 50 ng ml±1 using the ¯uorescence RNA stain Ribo-Green (Molecular Probes, Eugene, Oregon, USA) in order toensure appropriate concentrations of template RNA, and subse-quently diluted to 10 ng ml±1. For quantitative RT±PCR, 100 ng oftotal RNA was used as template. RT±PCR was performed usingthe `Titan-one-tube-RT±PCR system' (Boehringer) according to themanufacturer's protocol with the following temperature pro®le:50°C for 30 min; 94°C for 2 min; 25 cycles with 94°C for 15 sec, 64°C

for 30 sec and 68°C for 30 sec; and 68°C for 10 min. The gene-speci®c primers D3F (5¢-CCCTTACTATTTGAATTTACATTAC-3¢)and D1R (5¢-CTTAGTGTCCAACATCTATTGCGTCAAAGGATC-3¢)amplify a 340 bp product of the At4CL3 mRNA. D3F was deducedfrom the cDNA such to span an exon±intron junction andtherefore does not bind to genomic DNA templates. RT±PCRproducts obtained after 25 cycles were separated on polyacryla-mide gels and stained with the ¯uorescence DNA dye SybrGreen I(Molecular Probes). PCR products were visualized by scanning thegel with a ¯uorescence imager (STORM860, Molecular Dynamics,Sunnyvale, California, USA) and quanti®ed using ImageQuantsoftware.

Expression of At4CL proteins in E. coli

All three At4CL cDNAs were expressed in E. coli (M15) using theQIAexpressionist expression vectors of the pQE50 (including aHis6-tag) and pQE30 (without His6-tag) series (Qiagen). Thecomplete At4CL1 cDNA was cloned in-frame into the pQE31 andpQE51 vectors. This results in proteins containing 26 additional N-terminal amino acids derived from the 5¢ untranslated region andpart of the polylinker of the pQE vectors. The At4CL2 cDNA wascloned into the pQE30 and pQE50 vectors. However, since theoriginal At4CL2 5¢ untranslated region contains two in-frame stopcodons, it was deleted using PCR. This resulted in an expressedprotein with ®ve additional vector-derived N-terminal aminoacids. The 5¢ untranslated region of At4CL3 was deleted using aPstI site at position +6 in the coding region and the PstI site in thepolylinker of the pBluescript vector and was subsequently clonedinto both the pQE32 and pQE52 expression vectors. The resultingproteins therefore lack three of the original N-terminal aminoacids but contain six additional amino acids derived from theexpression vector.

Proteins were extracted from exponentially growing bacteria (in100 ml LB medium) 3 h after induction of 4CL expression with2 mM IPTG (A600 = 0.7). Bacterial pellets were resuspended in 10 mlbuffer (50 mM Na-phosphate, pH 7.8; 300 mM NaCl; 1 mg ml±1

lysozyme), incubated on ice for 30 min and the bacterial cellswere disrupted by soni®cation. Following centrifugation(20 000 g), the crude protein extracts were supplemented with b-mercaptoethanol (2 mM) and immediately used for enzymeactivity tests. His6-tagged 4CL proteins were puri®ed using Ni-NTA spin columns (Qiagen) according to the manufacturer'sprotocol. Puri®ed proteins were separated by SDS±PAGE, blottedonto nitrocellulose membranes, and detected using an antiserumraised against the parsley 4CL (Ragg et al., 1981).

Enzyme assays

4CL activity was measured at room temperature using a spectro-photometric assay to detect the formation of CoA esters ofvarious cinnamic acid derivatives (Knobloch and Hahlbrock,1977). The change in absorbance was monitored at wavelengthsof 311, 333, 369, 345 and 352 nm according to the absorptionmaxima for cinnamoyl-CoA, 4-coumaroyl-CoA, caffeoyl-CoA,feruloyl-CoA and sinapoyl-CoA, respectively (StoÈ ckigt and Zenk,1975). Phenolic substrates (Sigma, Deisenhofen, Germany) wereused at concentrations ranging from 0.1 to 4 mM (cinnamate) andfrom 0.005 to 0.8 mM (all other substrates). The kinetic constants,Km and Vmax, for the phenolic substrates were determined at ®xedconcentrations of all other substrates by linear regression of vagainst v/s (Eadie±Hofstee plot). Similar values were obtainedfrom Hanes plots (s/v against s). Each plot contained at least six



points. For comparison of activities in different bacterial extracts,the apparent Vmax values were normalized by setting Vmax for 4-coumarate to 100%.

Sequence comparisons

An alignment of published 4CL amino acid sequences wasgenerated using the PILEUP program within the GCG programpackage, version 8.1 (Devereux et al., 1984) with manual optimiza-tion. Based on this alignment, a maximum parsimony analysiswas performed (Fitch, 1977) using the PAUP 3.1.1 program(Smithsonian Institution, Washington DC, USA). The mostparsimonious tree was found using the heuristic search optionwith the `tree bisection reconnection' (TBR) branch swappingalgorithm (Swoffort and Olsen, 1990). For statistical analysis, 500bootstrap replications (Felsenstein, 1985) were analysed. Thefollowing plant 4CL sequences were used (GenBank accessionnumbers given in brackets): Arabidopsis thaliana 4CL1 (U18675),soybean 4CL14 (X69954), soybean 4CL16 (X69955), Lithosper-mum erythrorhizon 4CL1 (D49366), Lithospermum erythrorhizon4CL2 (D49367), tobacco 4CL (D43773), tobacco 4CL1 (U50845),tobacco 4CL2 (U50846), rice 4CL1 (X52623), rice 4CL2 (L43362),parsley 4CL1 (X13324), parsley 4CL2 (X13325), loblolly pine 4CL1(U12012), loblolly pine 4CL2 (U12013), hybrid poplar 4CL1(AF008184), hybrid poplar 4CL2 (AF008183), aspen 4CL1(AF041049), aspen 4CL2 (AF041050), potato 4CL1 (M62755),potato 4CL2 (AF150686) and vanilla 4CL (X75542).

Acknowledgements

We thank Dr W. Knogge and Professor K. Hahlbrock for criticalcomments on the manuscript, and Professor K. Hahlbrock also forhis continuous interest and support of this project. We also thankDr E. Weiler (Ruhr-UniversitaÈ t, Bochum, Germany) for providingthe allene oxide synthase cDNA, Dr Eric B. Holub (HorticultureResearch International, East Malling, UK) for providing the P.parasitica isolate Noks1, Dr T. Debener (Bundesanstalt fuÈ rZuÈ chtungsforschung, Ahrensburg, Germany) for help in estab-lishing the pathosystem in our laboratory, and Dr H. Kranz for thesupply of RNA of UV-treated plants. The excellent technicalassistance of Eva SchloÈ sser is also acknowledged. This work wassupported in part by the Deutsche Forschungsgemeinschaft (Ko1192/4-1) and by a research grant from the National Sciences andEngineering Research Council of Canada to C.J.D.

References

Allina, S.M., Pri-Hadash, A., Theilmann, D.A., Ellis, B.E. and

Douglas, C.J. (1998) 4-coumarate:coenzyme A ligase in hybrid

poplar. Plant Physiol. 116, 743±754.Becker-AndreÂ , M., Schulze-Lefert, P. and Hahlbrock, K. (1991)

Structural comparison, modes of expression, and putative cis-

acting elements of the two 4-coumarate: CoA ligase genes in

potato. J. Biol. Chem. 266, 8551±8559.Bell-Lelong, D.A., Cusumano, J.C., Meyer, K. and Chapple, C.

(1997) Cinnamate-4-hydroxylase expression in Arabidopsis.

Plant Physiol. 113, 729±738.Cormack, R.S. and Somssich, I.E. (1997) Rapid ampli®cation of

genomic ends (RAGE) as a simple method to clone ¯anking

genomic DNA. Gene, 194, 273±276.Dangl, J.L., Holub, E.B., Debener, T., Lehnackers, H., Ritter, C. and

Crute, I.R. (1992) Genetic de®nition of loci involved in

Arabidopsis±pathogen interactions. In Methods in ArabidopsisResearch (Koncz, C., Chua, N.-H. and Schell, J., eds). Singapore:World Scienti®c Publishing Co, pp. 393±418.

Devereux, J., Haeberli, P. and Smithies, O. (1984) Acomprehensive set of sequence analysis programs for theVAX. Nucl. Acids Res. 12, 387±395.

Dixon, R.A. and Paiva, N.L. (1995) Stress-induced phenyl-propanoid metabolism. Plant Cell, 7, 1085±1097.

Dong, X. (1998) SA, JA, ethylene, and disease resistance in plants.Curr. Opin. Plant Biol. 1, 316±323.

Douglas, C.J. (1996) Phenylpropanoid metabolism and ligninbiosynthesis: from weeds to trees. Trends Plant Sci. 1, 171±178.

Douglas, C., Hoffmann, H., Schulz, W. and Hahlbrock, K. (1987)Structure and elicitor or u.v.-light-stimulated expression of two4-coumarate:CoA ligase genes in parsley. EMBO J. 6, 1189±1195.

Felsenstein, J. (1985) Con®dence limits on phylogenies: anapproach using the bootstrap. Evolution, 39, 783±791.

Fitch, W.M. (1977) On the problem of discovering the mostparsimonious tree. Am. Nat. 111, 223±257.

Hahlbrock, K. and Grisebach, H. (1979) Enzymic controls in thebiosynthesis of lignin and ¯avonoids. Annu. Rev. Plant Physiol.30, 105±130.

Hartmann, U., Valentine, W.J., Christie, J.M., Hays, J., Jenkins,G.I. and Weisshaar, B. (1998) Identi®cation of UV/blue light-response elements in the Arabidopsis thaliana chalconesynthase promoter using a homologous protoplast transientexpression system. Plant Mol. Biol. 36, 741±754.

Hauffe, K.D., Paszkowski, U., Schulze-Lefert, P., Hahlbrock, K.,Dangl, J.L. and Douglas, C.L. (1991) A parsley 4CL-1 promoterfragment speci®es complex expression patterns in transgenictobacco. Plant Cell, 3, 435±443.

Holub, E.B., Beynon, J.L. and Crute, I.R. (1994) Phenotypic andgenotypic characterization of interactions between isolates ofPeronospora parasitica and accessions of Arabidopsis thaliana.Mol. Plant±Microbe Interact. 7, 223±239.

Hu, W.-J., Kawaoka, A., Tsai, C.-J., Lung, J., Osakbe, K., Ebinuma,H. and Chiang, V.L. (1998) Compartmentalized expression oftwo structurally and functionally distinct 4-coumarate:CoAligase genes in aspen (Populus tremuloides). Proc. Natl Acad.Sci. USA, 95, 5407±5412.

Kajita, S., Hishiyama, S., Tomimura, Y., Katayama, Y. and Omori,S. (1997) Structural characterization of modi®ed lignin intransgenic tobacco plants in which the activity of 4-coumarate:coenzyme A ligase is depressed. Plant Physiol.114, 871±879.

Knobloch, K.H. and Hahlbrock, K. (1977) 4-Coumarate: CoA ligasefrom cell suspension cultures of Petroselinum hortense Hoffm.Arch. Biochem. Biophys. 184, 237±248.

Kombrink, E. and Somssich, I.E. (1995) Defense responses ofplants to pathogens. Adv. Bot. Res. 21, 1±34.

Koopmann, E., Logemann, E. and Hahlbrock, K. (1999) Regulationand functional expression of cinnamate 4-hydroxylase fromparsley. Plant Physiol. 119, 49±55.

Korfhage, U., Trezzini, G.F., Meier, I., Hahlbrock, K. and Somssich,I.E. (1994) Plant homeodomain protein involved intranscriptional regulation of a pathogen defense-related gene.Plant Cell, 6, 695±708.

Kunkel, B.N. (1996) A useful weed put to work: genetic analysis ofdisease resistance in Arabidopsis thaliana. Trends Genet. 12,63±69.

Laudert, D., Pfannschmidt, U., Lottspeich, F., HollaÈnder-Czytko,H. and Weiler, E.W. (1996) Cloning, molecular and functionalcharacterization of Arabidopsis thaliana allene oxide synthase



(CYP 74), the ®rst enzyme of the octadecanoid pathway tojasmonates. Plant Mol. Biol. 31, 323±335.

Lee, D. and Douglas, C.J. (1996) Two divergent members of atobacco 4-coumarate:coenzyme A ligase (4CL) gene family.cDNA structure, gene inheritance and expression, andproperties of the recombinant proteins. Plant Physiol. 112,193±205.

Lee, D., Ellard, M., Wanner, L., Davis, K.R. and Douglas, C.J. (1995)The Arabidopsis thaliana 4-coumarate:CoA ligase (4CL) gene:stress and developmentally regulated expression andnucleotide sequence of its cDNA. Plant Mol. Biol. 28, 871±884.

Lee, D., Meyer, K., Chapple, C. and Douglas, C.J. (1997) Antisensesuppression of 4-coumarate:coenzyme A ligase activity inArabidopsis leads to altered lignin subunit composition. PlantCell, 9, 1985±1998.

Logemann, E., Parniske, M. and Hahlbrock, K. (1995) Modes ofexpression and common structural features of the completephenylalanine ammonia-lyase gene family in parsley. Proc. NatlAcad. Sci. USA, 92, 5905±5909.

Parker, J.E., Coleman, M.J., SzaboÁ , V., Frost, L.N., Schmidt, R.,

van der Biezen, E.A., Moores, T., Dean, C., Daniels, M.J. and

Jones, J.D.G. (1997) The Arabidopsis downy mildew resistancegene RPP5 shares similarity to the toll and interleukin-1receptors with N and L6. Plant Cell, 9, 879±894.

Ragg, H., Kuhn, D. and Hahlbrock, K. (1981) Coordinatedregulation of 4-coumarate:CoA ligase and phenylalanineammonia lyase mRNAs in cultured plant cells. J. Biol. Chem.256, 10061±10065.

Reinold, S. and Hahlbrock, K. (1997) In situ localization ofphenylpropanoid biosynthetic mRNAs and proteins in parsley(Petroselinum crispum). Bot. Acta, 110, 431±443.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular

Cloning. A Laboratory Manual, 2nd edn. Cold Spring Harbor:Cold Spring Harbor Laboratory Press.

Schmelzer, E., KruÈ ger-Lebus, S. and Hahlbrock, K. (1989)Temporal and spatial patterns of gene expression aroundsites of attempted fungal infection in parsley leaves. PlantCell, 1, 993±1001.

StoÈ ckigt, J. and Zenk, M.H. (1975) Chemical syntheses andproperties of hydroxycinnamoyl-coenzyme A derivatives. Z.Naturforsch. 30c, 352±358.

Swoffort, D.L. and Olsen, G.J. (1990) Phylogeny reconstruction. InMolecular Systematics (Hillis, D.M. and Moritz, C., eds).Sunderland, Massachussets: Sinauer Associates, Inc. pp. 411±501.

Trezzini, G.F., Horrichs, A. and Somssich, I.E. (1993) Isolation ofputative defense-related genes from Arabidopsis thaliana andexpression in fungal elicitor-treated cells. Plant Mol. Biol. 21,385±389.

Tsuji, J., Jackson, E.P., Gage, D.A., Hammerschmidt, R. andSomerville, S.C. (1992) Phytoalexin accumulation inArabidopsis thaliana during the hypersensitive reaction toPseudomonas syringae pv syringae. Plant Physiol. 98, 1304±1309.

Uhlmann, A. and Ebel, J. (1993) Molecular cloning and expressionof 4-coumarate:coenzyme A ligase, an enzyme involved in theresistence of soybean (Glycine max) against pathogeninfection. Plant Physiol. 102, 1147±1156.

Ye, Z.H., Kneusel, R.E., Matern, U. and Varner, J.E. (1994) Analternative methylation pathway in lignin biosynthesis inZinnia. Plant Cell, 6, 1427±1439.

Zhang, X.-H. and Chiang, V.L. (1997) Molecular cloning of 4-coumarate:coenzyme A ligase in loblolly pine and the roles ofthis enzyme in the biosynthesis of lignin in compression wood.Plant Physiol. 113, 65±74.

GenBank database accession numbers AF106084 (At4CL1 genomic), AF106085 (At4CL2 genomic), AF106086 (At4CL2

cDNA), AF106087 (At4CL3 genomic), AF106088 (At4CL3 cDNA).



three 4-coumarate:coenzyme a ligases in arabidopsis thaliana represent two evolutionarily divergent...

Documents