Divergent members of a soybean ( Glycine max L.) 4-coumarate:coenzyme A ligase gene family Primary structures, catalytic properties, and differential expression Christian Lindermayr 1 , Britta Mo¨ llers 1 , Judith Fliegmann 1 , Annette Uhlmann 1 , Friedrich Lottspeich 2 , Harald Meimberg 3 and Ju¨ rgen Ebel 1 1 Botanisches Institut der Universita ¨ t, Mu ¨ nchen, Germany; 2 Max-Planck-Institut fu ¨ r Biochemie, Martinsried, Germany; 3 Institut fu ¨ r Systematische Botanik der Universita ¨ t, Mu ¨ nchen, Germany 4-Coumarate:CoA ligase (4CL) is involved in the formation of coenzyme A thioesters of hydroxycinnamic acids that are central substrates for subsequent condensation, reduction, and transfer reactions in the biosynthesis of plant phenyl- propanoids. Previous studies of 4CL appear to suggest that many isoenzymes are functionally equivalent in supplying substrates to various subsequent branches of phenylpropa- noid biosyntheses. In contrast, divergent members of a 4CL gene family were identified in soybean (Glycine max L.). We isolated three structurally and functionally distinct 4CL cDNAs encoding 4CL1, 4CL2, and 4CL3 and the gene Gm4CL3. A fourth cDNA encoding 4CL4 had h igh s imi- larity with 4CL3. The recombinant proteins expressed in Escherichia coli possessed highly d ivergent catalytic e ffi- ciency with various hydroxycinnamic acids. Remarkably, one isoenzyme (4CL1) was able to convert sinapate; thus the first cDNA en coding a 4CL that accepts highly substituted cinnamic acids is available for further studies on branches of phenylpropanoid metabolism that p robably lead to the precursors of lignin. Surprisingly, the activity levels of the four isoenz ymes and steady-state levels of their transcripts were differently affected after elicitor treatment of soybean cell cultures with a b-glucan elicitor of Phytophthora sojae, revealing t he down-regulation of 4CL1 vs. up -regulation of 4CL3/4. A similar regulation of the transcript levels of the different 4CL isoforms was observed in soybean seedlings after infection with Phytophthora sojae zoospores. Thus, partitioning of cinnamic acid building units between phenylpropanoid branch pathways in s oybean could be regulated at t he level of catalytic specificity and the level of expression of the 4 CL isoenzymes. Keywords: 4-coumarate:CoA ligase; differential regulation; heterologous expression; plant defence; soybean (Glycine max L.). Phenylpropanoid compounds are major constituents of higher plants. They can serve as flower pigments, UV protectants, d efence chemicals,signalling c ompounds, a llelo - pathic agen ts, and as building units of the phenolic support polymer, lignin. Their synthesis is r egulated both by developmental processes and by environmental cues and it proceeds via the general phenylpropanoid pathway and subsequent specialized branches of phenylpropanoid meta- bolism. Central t o many of t he biosynthetic pathways is the activation of differently substituted cinnamic acids to the corresponding CoA t hioesters. This r eaction is catalyzed by 4-coumarate:CoA ligase (4CL; EC 6.2.1.12), a member of general phenylpropanoid metabolism. The central position of 4CL, linking the general with specialized branches of phenylpropanoid metabolism, led to the suggestion that 4CL could play a pivotal role in regulating the flux of the activated CoA ester intermediates into subsequent biosynthetic pathways. This idea was substantiated by the observation that isoenzymes of 4CL in soybean (Glycine max), petunia ( Petunia hybrida), pea (Pisum sativum), oat ( Avena sativa), and poplar (Populus · euramericana) d isplayed different substrate a ffinities and/or tissue distribution [1–5]. In contrast, other plants apparently contain only a single 4CL isoenzyme or isoforms that exhibit similar substrate specificities [6–10]. In these cases, the r ing-modifications o n the cinnamic acid derivatives, which precede the partitioning into different pathways, may proceed at the level of the activated esters, as well as the aldehydes and alcohols, as proposed recently [11–14]. Therefore, the physiological relevance of the occurrence o f multiple 4CL in the former plants remains largely unknown. 4CL genes have been studied in a large variety of plants, where the y c omprise s mall gene families in most cases. I n a number o f p lants, including parsley (Petroselinum crispum), loblolly pine (Pinus taeda), and potato (Solanum tuberosum), the g enes encode id entical or very similar p roteins [ 7,15,16], whereasinotherplants,suchastobacco(Nicotiana Correspondence to J. Ebel, Botanisches Institut der Universita ¨ t Mu ¨ nchen, Menzinger Strasse 67, D-80638 Mu ¨ nchen, Germany. E-mail: j.ebel@botanik.biologie.uni-muenchen.de Abbreviation: 4CL, 4-Coumarate:coenzyme A ligase. Enzyme: 4 -coumarate:CoA ligase ( EC 6.2.1.12). Note: the nucleotide sequence data reported w ere deposited under GenBank accession nos AF279267 for Gm4CL1 cDNA, A F002259 for 4CL14 (Gm4CL2 cDNA), AF002258 for 4CL13 ( Gm4CL3 genomic), and X69955 for 4CL16 ( Gm4CL 4 cDNA). (Received 2 3 October 2 001, revised 2 8 December 2001, accepted 9 January 2002) Eur. J. Biochem. 269, 1304–1315 (2002) Ó FEBS 2002 tabacum), Arabidopsis thaliana,aspen(Populus tremuloides), hybrid poplar (Populus trichocarpa · P. deltoides), and soybean structurally divergent isoforms have been identified [9,17–20]. In only a few plants have functionally divergent 4CL gene family members been correlated with specific phenylpropanoid branch pathways, e.g. with plant tissues actively producing t ypical phenylpropanoids, o r w ith p ath- ways that are a ffected by environmental factors. In aspen, Pt4CL1 has been associated with lignin biosynthesis because of substrate preference and expression of the corresponding gene in lignifying xylem tissues [17]. Con- versely, aspen Pt4CL2 is thought to be involved in the biosynthesis of phenylpropanoids in lignin-deficient epider- mal layers. In Arabidopsis, three functionally divergent At4CL forms have been hypothesized to be involved in different phenylpropanoid biosyntheses of lignifying and lignin-free tissues as well as to exhibit different physiological roles against environmental challenges [ 18]. An unresolved question concerns the ability of 4CL to catalyse the activation of sinapate. Sinapoyl-CoA was previously proposed to be a p recursor for syringyl units of angiosperm lignin. Nevertheless, r ecent findings indicate that the activated esters of sinapate, f erulate, and 5-hydroxyferulate are not likely to participate in monolig- nol biosynthesis [11,12,14]. However, 4CL isoforms from soybean, petunia, pea, and poplar have been fo und to convert s inapate to sinapoyl-CoA [1–3,5], whereas enzymes from many other plants apparently lack this activity (see, for example, [7,17–19]). Reactions obviating the CoA- activation of highly ring-modified cinnamic acids have been described, including cytochrome P450-dependent hydroxylations [11,21] as well as O-methylations that operate on caffeoyl-CoA and 5-hydroxyconiferyl aldehyde [12,22,23]. Despite its central position at a branch po int of phenylpropanoid metabolism in plants, the precise func- tion of 4CL isoforms in p roviding CoA ester precursors for t he synthesis o f different classes of phenolic compounds with specialized functions remains, thus, largely controver- sial. In soybean, the production of phenylpropanoid com- pounds comprises one of t he b iochemical d efence reactions that are activated upon challenge with the oomycete pathogen Phytophthor a sojae or treatment with a b-glucan elicitor derived from the pathogen. The phenolic com- pounds that accumulate around infection sites or in elicitor- treated cell c ultures inc lude pterocarpan phytoalexins [24], isoflavone conjugates [25,26], and wall-bound phenylpro- panoid compounds [27]. Phytoalexin accumulation is pre- ceded by the induced expression of many of the enzymes involved in the biosynthetic p athway, including 4CL [20]. Previous biochemical [1] and molecular studies [20] indicated that 4CL in soybean is enco ded by a small gene family. In this study, we report on the isolation and functional assignment of four soybean cDNAs as well as one of the encoding 4CL genes. Pronounced differences in catalytic efficiencies o f t he e ncoded isozymes f or differently substituted cinnamic a cid substrates were found. Com bined with differential expression patterns of the isoforms and corresponding transcripts in different tissues of seedlings as well as in both elicited cell cultures and infected seedlings, these studies substantiate earlier conc lusions t hat t he 4CL isoenzymes in soybean serve different physiological functions. Phylogenetic c omparison based on amino a cid sequences extends the recent classification [18] of 4CL isoforms within angiosperms. EXPERIMENTAL PROCEDURES Plant material Soybean seeds (Glycine max L. cv. H arosoy 63) w ere from R. I. Buzzell and V. Poysa (Agriculture Canada, Research Station, Harrow, Canada); G. max L. cv. 9007 from Pioneer Hi-Bred (Buxtehude, Germany). Seedlings were grown on vermiculite under aseptic conditions as described previously with minor modifications [28]. For infection e xperiments, the taproots of 3-d ay-old seedlings were treated with a suspension of % 10 4 zoospores in 200 lL sterile distilled water by dip inoculation; control see dlings were placed in water [28]. Cell suspension cultures of soybean (G. max L. cv. Harosoy 63) were grown in the dark as described previously [29] and treated with Phytophthora sojae crude elicitor (80 lg glucose equivalentsÆmL )1 medium) obtained by partial a cid hydrolysis of purified cell walls of the oomycete [30], as described previously [31]. 4CL activity assay Protein extracts were prepared from cell suspension cultures of soybean according t o previously reported procedures [29]. Enzyme activity was determined spectrophotometri- cally acco rding to th e method of Knobloch and Hahlbrock [1]. The analysis of activity levels of individual 4CL forms in isoenzyme mixtures in cell culture extracts was based on relative conversion rates (V values) for differently substi- tuted cinnamic acids a t substrate c oncentrations of 500 l M according to Knobloch and Hahlbrock [1]. The change in absorbance caused by CoA-ester production was monitored at 311 nm for cinnamic acid, 333 nm for 4-coumaric acid, 346 nm for caffeic ac id, ferulic acid, and 3,4-dimetho xycin- namic acid, and at 352 nm for sinapic acid [32]. Whereas 4-coumarate served as a s ubstrate f or all four isoenzymes, ferulate was a substrate for isoenzymes 1 and 2 under the conditions used, and 3,4-dimethoxycinnamate was conver- ted exclusively by isoenzyme 1. By measuring relative V-values, the activity level of each 4CL i soenzyme could be estimated indirectly by using the above i ndicated substrates and according to the scheme given in Table 1. Because of highly similar conversion rates of differently substituted cinnamic acids, isoenzymes 3 and 4 could not be distinguished. For substrate affinity measurements of the recombinant proteins, cinnamic acid was t ested in a concentration range of 0.1–4 m M whereas 2.5–1000 l M was used for all other substrates. The procedure for the indirect evaluation of isoenzyme activities in crude plant extracts was validated by mixing the recombinant enzymes (4CL1/4CL2/4CL3 with activity ratios of 1 : 1 : 1, 10 : 10 : 1, and 1 : 10 : 10, respectively, based on p-coumaric acid conversion) and subsequent estimation of isoenzyme activities usin g t he factors given in Table 1 . The calculated activity measures matched the predicted values (data not shown). Protein content was measured according to Bradford [33] with BSA as standard. P rotein extracts were stored at )20 °C. Ó FEBS 2002 4-Coumarate:CoA ligase gene family from Glycine max (Eur. J. Biochem. 269) 1305 Immunoblotting Protein extracts were separated by SDS/PAGE on 10% polyacrylamide gels [34]. The proteins were blotted onto nitrocellulose membranes and blocked with 1% (w/v) nonfat powdered milk and 1% (w/v) BSA. The blot was incubated with an antiserum raised against parsley 4CL [35] at a dilution of 1 : 10000 for 1 h, followed by incubation with goat anti-(rabbit IgG) Ig conjugated to alkaline phosphatase (Sigma) and cross-reacting protein b ands were visualized using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium as substrates. Protein purification and analysis 4CL1 was purified from nontreated soybean cell cultures according to Knobloch and Hahlbrock [1] with some modifications. The following buffers were used in various steps of p artial 4CL1 purification: buffer A , 0.2 M Tris/HCl pH 8.0, 14 m M 2-mercaptoethanol, 0.2 m M phenylmethane- sulphonyl fluoride, 30% (v/v) glycerol; buffer B, 0.05 M Tris/HCl pH 7.9, 0.1 m M dithiothreitol, 0.2 m M phenyl- methanesulphonyl fluoride, 30% (v/v) glycerol. All steps were carried out at 4 °C. Froze n soybean cells (300 g) were thawed and homogenized with 150 g quartz sand and 150 m L buffer A in a chilled mortar, stirred for 20 min with 0.1 g Dowex 1 · 2 (Serva; equilibrate d w ith buffer A) p er gram of cells, a nd centrifuged to remove Dowex 1 · 2and cell debris. 4CL1 activity was precipitated from the super- natant with (NH 4 ) 2 SO 4 (38–72% saturation), dissolved in buffer B, a nd the extract desalted by chromatography on a Sephadex G-25M column (Pharmacia). For separating 4CL isoforms, the protein fraction was applied to a Q Sepharose Fast Flow column (2.6 · 20 cm) w hich had been equili- brated with the same buffer. The proteins were eluted with a linear gradient of 0 –0.4 M KCl in buffer B and 10 mL fractions w ere collected. Fractions which showed 4CL1 activity were combined and loaded onto a Cibacron Blue 3G-A column (Pharmacia). After washing the column with 10 mL 0.6 M KCl in buffer B, proteins were eluted with 2 M KCl in buffer B . Fractions with 4CL1 activity were pooled and d esalted b y c hromatography on Sephadex G -25M. Finally, 4CL1 was separated f rom 4CL2 by anion exchange chromatography on a R esource Q column (Pharmacia) using a linear gradient of 0–0.4 M KCl in buffer B . Protein fractions with enriched 4CL1 activity were separated on SDS/PAGE and the protein band corresponding to 4CL1 was used for sequence analysis. Microsequencing of the N- terminus and two internal oligopeptides obtained after proteolytic digestion with Lys-C resulted in three pep- tide sequences: the N-terminal sequence consisted of APSPQEIIF, s equence S 1 o f ( K)GYLNDPEA, and sequence S2 of (K)ARLVITQSAYVEK. DNA and RNA methods Standard protocols were used for restriction enzyme digestion, RNA and DNA blots [36]. Total RNA from cell suspension cultures and s eedlings of G. max L. was isolated according to [37]. DNA was prepared according t o [38]. Gene-specific hybridization probes have been generated by either amplification of Gm4CL1 with the oligonucleotide primers S1 and S2 (Table 2), or by restriction of the cDNAs releasing a 1.0-kb SalI–KpnI-fragment from Gm4CL2 and a 0.9-kb SacI fragment from pQE-31/Gm4CL3, respectively. For Southern blot hybridization, the complete open reading frame of Gm4CL1 cDNA, a HindIII fragment of pQE-30/Gm4CL2, and a BamHI–HindIII fragment o f pQE-31/Gm4CL3 were prepared as hybridization probes. Table 1. Scheme for in direct calculations of 4CL isoenzyme activities. Isoenzyme Activity calculated for substrate Calculation procedure using various initial substrates 4CL1 Ferulate 0.6 · 4CL1 activity for 3,4-dimethoxycinnamate 4-Coumarate 1.1 · 4CL1 activity for 3,4-dimethoxycinnamate 4CL2 Ferulate Total activity for ferulate minus 4CL1 activity for ferulate 4-Coumarate 1.4 · 4CL2 activity for ferulate 4CL3/4CL4 4-Coumarate Total activity for 4-coumarate minus 4CL1 and 4CL2 activities, respectively, for 4-coumarate Table 2. Sequences of oligonucleotides. Restrict ion sites contained in the oligonucleotides are underlined. Designation Sequence 4CL1-GSP1 5¢-GTTGCGTAGGACGAGCAT-3¢ 4CL1-GSP2 5¢-CGGATGCCGATTTTGTGGAGG-3¢ 4CL1-KpnI5¢-GCT GGTACCGCACCTTCTCCACAAG-3¢ 4CL1-S1 5¢-TCYGGRTCRTTNAGRTADCCTTTCAT-3¢ 4CL1-S2 5¢-TBACNCARTCNGCNTAYGTBGARAA-3¢ 4CL3-HindIII 5¢-GTTCT AAGCTTTTAAGGCGTCTGAGTGGC-3¢ 4CL13–3¢KpnI5¢-AGTTTCAGGGTCAACAACCCTG-3¢ 4CL13-EcoRI 5¢-CTC GAATTCATGACAACGGTAGCTGCTTCTC-3¢ 4CL14-BamHI 5¢-CTC GGATCCATGGCTGATGATGGAAGCAG-3¢ 4CL14-GSP1 5¢-TCAGCGTCACCGTTATCCTC-3¢ 4CL14-GSP2 5¢-GTGAGAAATGGAGATGCTGC-3¢ 4CL16-GSP1 5¢-TGTTCCGGAGAGCCTCCTC-3¢ 4CL16-GSP2 5¢-CAACGGAAGCACGCATAGGAGCAC-3¢ 4CL16-SphI5¢-CACC GCATGCATAACTCTAGCTCCTTCTCTTG-3¢ Seq1 5¢-GTAAAACGACGGCCAGT-3¢ 1306 C. Lindermayr et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Hybridization conditions were according to [39]. DNA fragments were labelled with [a- 32 P]dCTP (Amersham Pharmacia) using the r andom prime p rocedure (Prime- a-Gene, Promega). Membranes were washed at moderate stringency for 20 min twice in 1 · NaCl/Cit, 0.1% (w/v) SDS at 42 °C o r at high stringency c onditions for 30 min in 0.5 · NaCl/Cit, 0.1% (w/v) SDS at 65 °C, and a utoradio- graphed between intensifying screens at )80 °C. When necessary, mem branes were s tripped by incubation in 0.1% SDS, 5 m M EDTA at 95 °C for 30 min. Cloning and subcloning of genomic DNA Genomic DNA was isolated from hypocotyls and primary leaves of 15-day-old G. max L. cv. Harosoy 63 seedlings, partially digested with MboI, fractionated by size, and cloned into BamHI-cut kEMBL3. The DNA library was screened with randomly labelled Gm4CL14 and Gm4CL16 cDNAs t hat h ad be en isolated previously [20]. Positive plaques were purified by four round s o f s creening, and one 9.5-kb DNA clone (kEMBL/Gm4CL3) was isolated. Res- triction maps of the genomic clone kEMBL/Gm4CL3 and subclones w ere c onstructed by single and multiple enzyme digestions of the clones ligated into pBluescriptIIKS and SK vectors [40]. cDNA synthesis and selection RNA from nontreated soybean cell cultures was used for RT/PCR using the peptide-deduced oligonucleotide primers 4CL1-S1 and 4CL1-S2 (Table 2). RT/PCR was performed with an increasing annealing t emperature (52.5 °C+ 0.1 °CÆcycle )1 ) for 25 cycles, followed by 10 cycles using 55 °C, resulting in a 0.8-kb fragment. This DNA fragment was used as a gene-specific 4CL1 probe for screening a cDNA library synthesized from enriched mRNA of nontreated soybean cell cultures. One Gm4CL1 cDNA was detected, plaque-purified and i solated a nd was shown to be almost full length. Completion of cDNA sequences 5¢-RACE was used to complete the open reading frames of the partial cDNAs enco ding Gm4CL1 (see above), Gm4CL2, and Gm4CL4 [20] (Table 3). Amplification was performed in the presence of the respective nested gene- specific primers (Table 2) and the universal amplification primer UAP (Gibco/BRL) after reverse transcription of soybeancellcultureRNA(1 lg each) using the gene-specific oligonucleotides Gm 4CL1-GSP1, Gm4CL14-GSP1, and Gm4CL16-GSP1, respectively, and t ailing with t erminal transferase and dCTP according to the manufacturer’s instructions (Gibco/BRL). The following annealing condi- tions were used during PCR: increasing from 54.5 to 5 8 °C during 35 cycles with Gm4CL1-GSP2 + UAP, 55 °Cfor 30 cycles using Gm4CL14-GSP2 + UAP after four rounds of unidirectional a mplification at 58 °C in the presence of solely the U AP oligonucleotide, and 60 °C with t he primers Gm4CL16-GSP2 + UAP. The partial c DNA clone Gm 4CL13, encoding the 4CL3 isozyme, was completed by RT/PCR using the genomic sequence for the generation of oligonucleotide primers (Gm4CL13–3¢KpnIandGm4CL13-EcoRI, Table 2). PCR was performed with an annealing t emperature of 55 °Cfor 30 cycles. The resulting 5¢ fragments have been cloned and sequenced on both strands. In all cases, they were shown to be identical in the overlapping portions when compared to the respective partial cDNAs. Construction of E. coli expression plasmids For heterologous expression, either the pTrcHis (Invitro- gen), or the pQE (Qiagen) vector series were used. A summary of recombinant plasmids is given i n Table 3. The introduction of the Gm 4CL1 cDNA into the expression vector required the elimination of the initiator codon of the 5¢-RACE product: this was accomplished by introducing a KpnI restriction site. This modification was achieved by PCR u sing the 4CL1-Kpn I oligonucleotide and the vector-binding oligonucleotide Seq1 as primers. The modified 5 ¢-RACE product was inserted into pZL1 con- taining the incomplete Gm4CL1 cDNA using KpnIand SalI. The complete open reading fram e was transferred into pQE-30 using KpnIandHindIII. For the construction of pBluescriptKSII/Gm4CL2, the respective 5¢ fragment was released from the c loning vector by SalI-restriction and inserted into the single SalIsiteofthe partial cDNA (pBluescriptKSII/Gm4CL14). Deletion of the 5¢ noncoding region was achieved by PCR using 4CL14- BamHI and 4CL14-GSP1 a s p rimers and pBluescriptKSII/ Gm4CL2 as template. The amplified product w as digested with BamHI/SalI and cloned together with the SalI/ Table 3. Summary o f recombinant plasmids. Plasmid Insert pZL1/Gm4CL1 Fusion of partial Gm4CL1 recovered from cDNA library screening and 5¢-end fragment recovered from 5¢-RACE; initiator codon was eliminated by introducing a KpnI restriction site pQE-30/Gm4CL1 Gm4CL1 cDNA with modified initiator codon (see pZL1/Gm4CL1) pBluescriptKSII/ Gm4CL14 Partial Gm4CL2 cDNA [20] pBluescriptKSII/ Gm4CL2 Fusion of partial Gm4CL2 cDNA and 5¢-end fragment recovered from 5¢-RACE pQE-30/Gm4CL2 Full-length Gm4CL2 cDNA; 5¢-noncoding nucleotides were eliminated by introducing a BamHI restriction site directly upstream of the initiator codon pTZ19R/Gm4CL13 Partial Gm4CL3 cDNA [20] pTZ19R/Gm4CL3 Fusion of partial Gm4CL3 cDNA and 5¢-end fragment amplified by PCR pTrcHisB/Gm4CL3 Full-length Gm4CL3 cDNA pQE-31/Gm4CL3 Full-length Gm4CL3 cDNA pTZ19R/Gm4CL16 Partial Gm4CL4 cDNA [20] pTZ19R/Gm4CL4 Fusion of partial Gm4CL4 cDNA and 5¢-end fragment recovered from 5¢-RACE; initiator codon was eliminated by introducing a SphI restriction site pQE-30/Gm4CL4 Full-length Gm4CL4 cDNA with modified initiator codon (see pTZ19R/Gm4CL4) kEMBL/Gm4CL3 Genomic clone of Gm4CL3 (9.5 kb) Ó FEBS 2002 4-Coumarate:CoA ligase gene family from Glycine max (Eur. J. Biochem. 269) 1307 HindIII-restricted insert of pBluescriptKSII/Gm4CL14 into pQE-30, previously cut w ith BamHI and Hin dIII. The resulting construct pQE-30/Gm4CL2 contained an upstream in-frame extension of 12 codons, including six histidine codons. The 4CL3 5 ¢ fragment w as ligated into pTZ19R/ Gm4CL13 using KpnI resulting in the f ull-length clone pTZ19R/Gm4CL3. The complete open reading frame was transferred both into the vector pTrcHisB and into pQE-31, yielding upstream in-frame extensions o f 45 a nd 20 codons, respectively, including six h istidine codons each. In spite of different N-terminal extensions, the two expressed 4CL3 proteins showed the same e nzymatic characteristics. For completion of t he Gm4CL4 cDNA, the initiator codon of th e 5¢-RACE product was eliminated by PCR using 4CL16-SphI a nd 4CL16-GSP2 a s primers. The modified 5¢-RACE fragment was inserted into pTZ19R/ Gm4CL16 using SphIandSalI resulting in pTZ19R/ Gm4CL4. For heterologous expression the complete open reading frame was amplified by PCR using 4CL16-SphIand 4CL3-HindIII as primers a nd pTZ19R/Gm4CL4 as tem- plate. The PCR product was inserted into pQE-30 using SphIandHindIII restriction sites. Mutagenized and PCR-amplified fragments inserted into the expression vectors were controlled b y sequencing. Expression in E. coli and isolation of recombinant proteins E. coli strain SG13009 harbouring the plasmids pQE-30/ Gm4CL1 or pQE-30/Gm4CL4, as well as the strain M15, harbouring pQE-30/Gm4CL2 or pQE-31/Gm4CL3, were grown in Luria–Bertani medium in the presence of 100 lgÆmL )1 ampicillin and 2 5 lgÆmL )1 kanamycin. E. coli JM109, harbouring pQE-30/Gm4CL2 or pTrcHisB/ Gm4CL3 were grown in the presence of ampicillin only. Cultures were grown until A 600 % 0.5 was reached, induced with 1.5 m M isopropyl-b- D -thiogalactopyranoside, and incubated for 4 h at 37 °C. After centrifugation, the bacterial cells were resuspended i n an appropriate volume of buffer [50 m M Tris/HCl pH 8.0, 14 m M 2-mercaptoeth- anol, and 30% (v/v) glycerol] a nd disrupted by sonication. After removing cellular debris by centrifugation (20 000 g, 20 min), the crude protein extracts were used for enzyme activity tests. For 4CL3, extracts were concentrated by the addition of solid (NH 4 ) 2 SO 4 to 75% saturation. The precipitate was collected b y centrifugation, dissolved i n buffer (see above), a nd the protein fraction was passed through a Sephadex G-25M column. The recombinant proteins were purified by immobiliz ed metal c helate affinity chromatography using the Ni-NTA Metal affinity matrix (Qiagen) according to the instructions of the manufacturer. Adsorbed proteins were eluted from the affinity matrix with buffer (see above) containing 50 m M imidazole. Analysis of DNA and protein sequences Double-stranded DNA was s equenced on both strands using the dideoxy chain-termination method [41] and a sequenase kit 2.0 (Amersham) or with an ABI Sequencer using BigDye Terminator chemistry (Botanisches Institut, LMU Mu ¨ nchen). Computer analysis was carried out with the PCgene program fro m IntelliGenetics (Geneva), Chromas from Technelysium (Queensland, Australia), and the BioEdit Sequence Alignment Editor [ 42]. The following 4CL sequences were used for the protein sequence alignment (GenBank accession nu mbers given in parentheses): Arabidopsis thaliana 4CL1 (U18675), A. thaliana 4CL2 (AF106086), A. thaliana 4CL3 (AF106088), G. max 4CL1 (AF279267), G. max 4CL2 (AF002259), G. max 4CL3 (AF002258), G. max 4CL4 (X69955), Lithospermum erythrorhizon 4CL1 (D49366), L. erythrorhizon 4CL2 (D49367), Lolium perenne 4CL1 (AF052221), L. perenne 4CL2 (AF052222), L. pe renne 4CL3 (AF0 52223), N icotiana tabacum 4CL (D43773), N. tabacum 4CL1 (U50845), N. tabacum 4CL2 (U50846), Oryza sativa 4CL1 (X52623), O. sativa 4CL2 (L43362), Petroselinum crispum 4CL1 (X13324), P. crispum 4CL2 (X13325), Pinus taeda 4CL1 (U12012), P. taeda 4CL2 (U12013), Populus hybrida 4CL1 (AF008184), P. hybrida 4CL2 (AF008183), Populus tremuloides 4CL1 (AF041049), P. tremuloides 4CL2 (AF041050), Rubus idaeus 4CL1 (AF239687), R. idaeus 4CL2 (AF239686), R. idaeus 4CL3 (A F239685), Solanum tuberosum 4CL1 (M62755), S. tube rosum 4CL2 (AF150686), Vanilla planifolia 4CL (X75542). The alignment of 4CL amino acid sequences was generated using CLUSTAL W 2.0 and corrected by hand. The resulting data matrix was subsequently analysed u sing PAUP version 4.0 [43]. The length of the protein sequences varied between 535 (P. tremuloides 4CL1) and 636 residues (L. erythrorhizon 4CL1). For distance and phylogenetic calculations, overhanging positions were excluded. All heuristic searches were carried out with the following settings: RANDOM addition (10 replicates), TBR branch-swapping, MULPARS, STEEPEST DESCENT, COLLAPSE and ACCTRAN optimization and character states specified as unordered and equally weighted. I n the data matrix all gap characters (–) were scored as missing data (?). Bootstrap values [44] were calculated from 1000 replicates. The resulting data matrix consisted of 557 characters of which 130 were c onstant, 427 were variable, and 324 were potentially informative for phylogenetic analyses. Pair-wise differences varied between 0.02% (4CL1 and 4CL2 of P. taeda) and 45 .85% (L. erythrorhizon 4CL2 vs. O. sativa 4CL1) with an average pair-wise distance of 31.2%. For comparison purposes, t he corresponding nucleotide sequences were aligned and evaluated in the same manner (data not shown). RESULTS The prime objective of t he current investigation was to extend the molecular survey of 4CL isoenzymes from soybean by: (a) completing the existing cDNAs [20]; and (b) by isolating l acking members of the g ene family. Moreover, we aimed to disclose details of the catalytic capacity of the isoforms as well as of the d ifferential regulation of t he isozymes in re lation to specialized branc hes of phenylpro- panoid metabolism. Molecular analysis of the 4CL gene family in soybean Soybean cell cultures have been reported to contain the 4CL1 and 4CL2 isoforms [1]. Of these, 4CL1 is capable of catalysing the activation of the broadest variety of 1308 C. Lindermayr et al. (Eur. J. Biochem. 269) Ó FEBS 2002 substituted cinnamates including sinapate. This catalytic property is shared by only the minority of 4CL isoforms studied to date from any plant. As the cDNA encoding 4CL1 was apparently missing from the pool of partial cDNA clones isolated earlier [20] (see below), the isolation of 4CL1 was attempted by purification from soybean cell cultures as source. The separation of the isoenzymes was achieved by anion exchange chromatography using Resource Q a nd verified by using 3,4-dimethoxycinnamate as substrate which is converted to the CoA ester by 4CL1 only [1]. Microsequencing of the purified 4CL1 established peptide sequences which facilitated the cloning of the corresponding cDNA from a cDNA library generated from untreated soybean cell cultures. For the isolation of full-length cDNAs encoding the isozymes 4CL2, 4CL3, and 4 CL4, respectively, t he 5¢-RACE w as used to yield the 5¢ ends of the partial clones Gm4CL14, Gm4CL13, and Gm4CL16 [20] (for details see Experimental procedures). In summary, four full-length cDNAs were obtained, encoding the soybean 4CL isozymes 1, 2, 3, and 4 which displayed divergent levels of similarity to each other ( Table 4). For example, 4CL3 and 4CL4 s hare a high identity at the deduced amino acid level (94%), whereas in a ll othe r cases the i dentity b etween the d educed 4CLisoformsismuchlower(% 60%). A phylogenetic reconstruction of the k nown plant 4CLs revealed earlier that two major 4CL classes have evolved within the angiosperms [18]. The addition of 4CL sequences deposited i n the databases s ince the earlier rep ort as well as of those p resented here into the protein alignment and the subsequent calculation of the most parsimonious phylo- genetic tree (Fig. 1) confirmed the previous observation of the e volution of two major 4CL groups. According to the earlier designation, soybean 4 CL1 and 4CL2 are members of the class I cluster, whereas 4CL3 and 4 CL4 belong to the more divergent class II cluster (Fig. 1 , upper and lower branch of the phylogram, r espectively). It was s hown previously that 4CL genes can be regulated at the transcriptional level by both, infection of soybean seedlings with Phytophthora sojae z oospores and e licitation of soybean cell cultures [20]. One gene encoding an inducible isoform of the soybean 4CL has been isolated by screening a genomic library with a fragment c omprising 700 bp of the Gm4CL16 cDNA (partial clone of Gm4CL4). The deter- mination of the complete sequence revealed that the clone represented the gene corresponding to the Gm 4CL3 cDNA, which is 93% identical to Gm4CL4. The 4CL3 gene of soybean was characterized by six exons ranging in s ize from 68 to 1068 bp which are flanked by five introns of 1893, 117, 102, 93, and 170 bp. The exon/intron splice junctions revealed not only strong similarity to plant junctions in general [45] but also to both 4CL genes each in parsley [6] andinpotato[7],the4CL1 gene in rice [46], and the three 4CL genes in Arabidopsis [18]. I t is interesting to note t hat, for th e increased number of 4CL genes analysed s o f ar, t he number and the positions of the introns is increasingly variable. The number of introns in the coding region that are unrelated with regard to s equence and size range from three (pine), four (parsley, potato, rice) and five (soybean) to six (Arabidopsis). Southern analysis of genomic DNA from cell cultures was used to verify the number of 4CL genes in soybean. Hybridization of triplicate blots containing restricted DNA with gene-specific p robes for 4CL1 and 4CL2, respectively, or with a probe which was not able to distinguish be tween the 4CL3 and 4CL4 genes, resulted in the d etection of distinct sets of fragments (Fig. 2). The hybridization patterns for 4CL1 and 4CL2 could not be fully explained by the existence of the respective restriction sites in the Table 4. Comparison of the soybean 4CL cDNAs and encoded iso- enzymes. The identity matrix w as calculated in pair-wise alignments using BioEdit [42] and is giv en in e ach case as p ercentage identity of the amino acid (up per line) an d t he n ucleotide seq uence (open reading frame only, l ower line, italic). Isoenzyme 4CL1 4CL2 4CL3 4CL4 Identity matrix 4CL1 – 63 58 58 65 62 61 4CL2 – – 61 63 60 62 4CL3 – – – 94 93 Amino acid residues 546 547 570 562 Molecular mass (kDa) 59.4 60.2 61.8 61.0 St4CL1 St4CL2 Nt4CL1 Nt4CL Nt4CL2 Pc4CL1 Pc4CL2 Vp4CL Le4CL1 Gm4CL2 Ri4CL1 Popt4CL1 Poph4CL1 Poph4CL2 Ri4CL2 At4CL1 At4CL2 Gm4CL1 Gm4CL4 Gm4CL3 Popt4CL2 Ri4CL3 Le4CL2 At4CL3 Lp4CL1 Os4CL2 Lp4CL2 Lp4CL3 Os4CL1 Pt4CL1 Pt4CL2 50 changes 62 2 1 2 1 107 35 23 22 23 16 25 14 22 1 13 19 11 7 13 57 31 31 26 58 62 63 21 26 26 42 39 30 60 39 46 97 51 52 55 32 19 17 45 11 48 57 85 69 53 36 52 58 45 42 60 94 1 100 53 57 78 76 100 87 100 100 5 2 60 100 98 56 70 74 83 61 1 00 100 100 100 99 100 Fig. 1. Heuristic maximum parsimony analysis of 31 4CL protein sequences depicted as phylogram. The phylogenetic analysis was calculated using the pine 4CL sequences as the outgroup. T he protein alignment resulted in one most parsimoniuos tree with a minimal length of 2155 steps and a consistency index of 0.599 (RI, 0.6 25). The branch lengths are annotated above the corresponding branches, bootstrap values for 1000 replicates are indicated below the branch length only at branches supported by bootstrap analysis. Ó FEBS 2002 4-Coumarate:CoA ligase gene family from Glycine max (Eur. J. Biochem. 269) 1309 Gm4CL1 and Gm 4CL2 cDNAs. However, the s ingle fragments detected for 4CL1 by BamHI or EcoRI restric- tion, respectively, and the double signal for 4CL2 after restriction with EcoRI, which is co mpatible with the presence of one EcoRI site in the cDNA sequence, suggested the existence of single genes encoding each of these isozymes (Fig. 2 ). The highly similar Gm4CL3 and Gm4CL4 cDNAs were represented b y a more complex genom ic hybridization pattern (Fig. 2) which nevertheless could be explained by the existence of single genes. Expression pattern of 4CL genes in soybean seedlings The spatial distribution of 4CL expression was studied in soybean seedlings (G. max L. cv. 9007). Expression of 4CL3/4 mRNA at low levels was confined t o roots and hypocotyls, while 4CL1 and 4CL2 mRNA amounts were highest in hypocotyls and stems a nd also in young roots (Fig. 3 ). Only low levels of the latter two mRNAs were observed in 12-day-old roots. In s hoot tips and leaves, no mRNA representing any of t he four 4CL isoforms c ould be detected under the experimental c onditions used. Differential regulation of 4CL transcript and enzyme activity levels Treatment of soybe an cells with Phytophthora sojae crude elicitor resulted in differential changes of the activities of the 4CL isoenzymes. At two growth stages of the cell suspen- sion culture, representing cultures at 1 day after inoculation (stage I) and at the end of the linear growth phase (stage II), respectively [29], t he activity of 4CL3/4 strongly increased, starting from a low basal level and reaching the highest level at about 10 h following the start of treatment (Fig. 4A; results f or stage I not shown). Conversely, the activity level of 4CL1 was strongly redu ced following elicitor treatment of the cells. After 12 h of treatment, the residual 4CL1 activity represented only % 10% of the level found in untreated control cells. Only minor changes occurred for the activity level of 4CL2 when compared with that of untreated control cells. The differential expression of the 4CL gene family members i n soybean ce ll cultures was also a ssessed b y RNA blot analysis. Northern analyses showed that 4CL1 and 4CL2 mRNA but not 4CL3/4 mRNA could be readily detected in untreated control cells (Fig. 4B). When chal- lenged w ith elicitor, differential changes in mRNA levels occurred that reflected those found for isoenzyme activity levels (Fig. 4A). An analysis similar t o that s hown for e licitor-treated cell cultures was performed with roots of 3-day-old soybean seedlings following infection with zoo spores of Phytophtho- ra sojae (Fig. 4C). Under the experimental conditions used, the levels of 4CL1 and 4CL2 mRNAs were low and remained unaffected. In c ontrast, 4CL3/4 mRNA levels were detectable in untreated tissues and increased strongly after infection (Fig. 4C). Functional analysis of recombinant 4CL To elucidate t he biochemical functions of the soybean 4CL gene family members, the substrate specificity of the heterologously expressed 4CL isoenzymes was examined and c ompared w ith that of the previously isolated iso- enzymes [ 1]. The four cDNAs were expressed in E. coli as 10 kb B E HB E HB E 4CL3/44CL2 Gm Gm Gm 4CL1 H 8.0 6.0 4.0 3.0 2.5 2.0 1.5 Fig. 2. Southern blot analysis of soybean genomic DNA. DNA samples (20 lg each) were digested with t he restriction endonucleases BamHI (B), EcoRI (E), a nd HindIII (H), separated in t riplicate on a 0.6% agarose gel and transferred t o a nylon membrane. The complete open reading frame of the Gm4CL1 cDNA, a HindIII-fragment of pQE-30/ Gm4CL2, a nd a BamHI/HindIII-fragment of pQE-31/Gm4CL3 were used as hybridization probes using high stringency conditions. Posi- tions of DNA sta ndards are given o n the rig h t. 1.8 kb 1 2 3 4 5 6 7 8 4CL3/4 4CL2 Gm Gm Gm 28S rRNA 4CL1 1.8 kb 1.9 kb 2.3 kb Fig. 3. Spatial expression pattern of the 4CL mRNAs in soybean seedlings. Total RNA (20 lg each), isolated from different plant tis- sues, was separated on 1.2% agarose gels, blotted onto n ylon mem- branes, and hybridized with gene-specific 4CL probes. Blots w ere washed under high stringency condition s. Hybridization with a 28S rRNA probe d em onstrated equ al l oading. R NA was isolated from 3- (1) and 12-day (2) -old ro ots, from hypocotyls (3), first (4) and second (5) internodium, f rom shoot tips (6), and from young (7) and old (8) leaves detached from 21-day-old plants. The sizes of the hybridizing RNA species are denoted on t he right. 1310 C. Lindermayr et al. (Eur. J. Biochem. 269) Ó FEBS 2002 inducible fusion proteins containing N-terminal His 6 -tags. After immobilized metal a ffinity chromatography, t he four proteins revealed the expected relative molecular masses i n SDS/polyacrylamide gels. Immunoblot analysis demonstra- ted that the recombinant proteins interacted with an antiserum raised against parsley 4CL [35], whereas no 4CL-like protein cross-reacted w ith the antiserum in b acter- ial e xtracts c ontaining the empty expression vector (Fig. 5). The recombinant isoenzymes were tested for their relative abilities t o u se differently substituted cinnamic acids as substrates (Table 5). The affinities for the cinnamic acid substrates were determined using Lineweaver–Burk plots. The recombinant 4CL1 showed simple Michaelis–Menten kinetics in the presence of 4-coumarate, caffeate, ferulate, sinapate and 3,4-dimethoxycinnamate, whereas it showed very low activity towards cinnamate. The recombinant 4CL2 was able to convert cinnamate, 4-coumarate, caffeate, and f erulate but not sinapate an d 3 ,4-dimethoxycinnamate. As summarized in Table 5, t he K m and r elative V max values for t hese two 4CL isoforms closely resembled those found previously for partially purified ligase 1 and 2 [1]. Similar experiments with r ecombinant 4CL3 and 4CL4 demonstra- ted t hat 4-coumarate and caffeate were most efficiently converted to the CoA ester, cinnamate and especially ferulate were converted with very low efficiency, whereas sinapic and 3,4-dimethoxycinnamic acid were not accepted as substrates. The K m values for the catalytic a ction of the four heterologously produced soybean 4 CL isoforms for Fig. 4 . Elicitation of soybean cell cultures and seedlings. (A) Soybean cell cultures of growth s tage II [ 29] were treated with Phy tophthora sojae b-glucans (80 lg glucose equivalentsÆmL )1 , filled symbols) or water (open symbols) for the periods indicated. Changes in 4CL isoenzyme activities were assayed i n crude protein e xtracts a nd d iscriminated by calculating the c ontribution of t he i soenzymes t o t he ove rall 4 CL activity by us ing isoenzyme-specific substrates. Data points shown are mean values of two independent experiments showing similar results. (B) Differential expression of individual members of t he 4CL gene family was assayed in soybean cell cultures after elicitor treatment. Soybean cell cultures were treated as i n ( A) and harvested at t he t imes i ndicated. To tal RNA (20 lg each) fro m elici tor-treated ( E) and untreated (C) c ell c ultures w as separated on a 1.2% agarose gel a nd blotted o nto a nylon membrane. The b lot was hybridize d with gene-specific 4CL probes corresponding to Gm4CL1, Gm4CL2, and Gm4CL3/4 and washed at high stringency. (C) Differential expression of 4CL mRNAs in roots of soybean seedlings upon infection with Phy tophthora so jae was analysed b y Northern b lottin g as described in (B). Roots of soybean (cv. Harosoy 6 3) seedlings were treated with Phytophthora sojae (race 1) zoospores by dip inoculation (E) a nd harvested a t t he times indicated. C ontrol seedlings were placed in sterile water (C). Hybridization with soybean 28S rRNA was used to confirm equal l oadin g. The sizes of t he hybridizing RNA species are shown on the right. Fig. 5. Immunoblot analysis of recombinant soybean 4CL. The four isoforms were expressed as His 6 -tagged fusion proteins and purified by immobilized m etal chelate a ffinity chromatography. The purified proteins were separated by SDS/PAGE and transferred to nitro- cellulosic filters. For immunodetection, antiserum raised against parsley 4CL [35] combined with goat a ntirabbit IgG conjugated to alkaline phosphatase was used. Purified protein extract from b acteria carrying the empty e xpression vector (pQE-30 ) served as a control. The relative molecular masses of protein standards are shown on the right. Ó FEBS 2002 4-Coumarate:CoA ligase gene family from Glycine max (Eur. J. Biochem. 269) 1311 4-coumarate and caffeate were similar t o those r eported for many purified plant 4CL, whereas differences of the substrate s pecificity (V max /K m ) b etween members of the soybean 4CL protein family appeared to be pronounced and w ere not previously rep orted in several of the other plant 4CL analysed so far. 4CL1 accepted the broadest range of hydroxylated and O-methylated cinnamic acids (highest relative V max /K m value for ferulate followed by 4-coumarate and sinapate). 4CL2 used a n intermediate range of substituted cinnamic a cids, while 4CL3 and 4CL4 displayed the highest selectivity towards these acids. All ligases exhibited low affinity for cinnamate (Table 5). A major r esult of the comparative substrate studies thus was the detection of distinct differences in the selectivity of the soybean 4CL isoforms for the various ring-substituted a nd unsubstituted cinnamic acids. DISCUSSION The r esults of our studies demonstrate t hat 4CL in soybean is encod ed by a small g ene family consisting of at least four members. The recombinant proteins expressed from these genes show pronounced differences in the catalytic efficiency for metabolically important ring-substituted cinnamic acids. Expression studies in cell cultures and in seedlings revealed differential regulation of the f our 4CL genes, supporting earlier notions on different physiological functions of membe rs of the 4CL gene family in phenylpropanoid branch pathways. One of our goals in the present studies was to generate full-length 4CL cDNAs for every single member of the gene family. The nucleotide and a mino acid sequences de duced from full-size Gm4CL2, Gm4CL3, and Gm4CL4 cDNAs confirmed the level of identity between the three soybean 4CL genes as d etermined earlier for t he partial cDNAs [ 20]. The present w ork adds a further, a nd presumably l ast, member to the 4CL family in soybean. A positive detection in our earlier work [20] w as not possible g iven that RNA samples from elicitor-treated cells were used. As demon- strated in this work, these RNA samples were an inadequate source for 4 CL1 cDNA isolation due to the repression of 4CL1 mRNA expression in response to elicitor (Fig. 4 A,B). A phylogenetic reconstruction (Fig. 1) illustrates the relationships of 4CL isoforms within the soybean gene family as well as within the a ngiosperms. As noted earlier, two classes o f 4CL proteins have ev olved [18]. R emarkably, there is neither an exclusive bias towards distribution according to lineage nor according to function. Gene duplications took place throughout the evolution of these plant enzymes which partly led to highly divergent struc- tures (for example Gm4CL3/4 vs. Gm4CL1 or 2; At4CL3 vs. At4CL1 or 2) which then eventually developed to serve different environmental needs. Within the class II cluster, for example, the soybean 4CL3 and 4 are the only i soforms which are activated strongly in elicited or infected tissues, whereas the Arab idopsis isoform 3 shows no regulation in response to pathogen c hallenge but to UV i rradiation [18]. Table 5. Substrate spe cificity of rec ombinant soybean 4CL expressed in E. coli. The K m and V max values of recombinant 4CL1, 4CL2, 4 CL3 and 4CL4 were determin ed using Lineweaver–Burk plots with a t l east five data points. Each acid was assayed at the lo ng-wave a bsorbance m aximum o f its CoA ester [32] in the spectrophotometrical test. Relative V max values were obtained by setting V max of 4-coumarate for each isoform to 100%. The enzymatic characteristics of the isolated ligases 1 and 2 given in parenthesis were adopted from Knobloch and Hahlbrock [1]. NC., No conversion. Isoform Substrate K m (l M ) Relative V max (% of coumarate) Relative V max /K m (l M )1 ) Gm4CL1 Cinnamate 4400 (1300) 9 (3) 2.0 · 10 )3 4-Coumarate 22 (32) 100 (100) 4.54 Caffeate 33 (40) 40 (56) 1.21 Ferulate 8 (9) 57 (56) 7.13 Sinapate 11 (11) 35 (46) 3.21 3,4-Dimethoxycinnamate 83 (100) 75 (89) 0.91 Gm4CL2 Cinnamate 1700 (4500) 50 (23) 0.03 4-Coumarate 42 (17) 100 (100) 2.38 Caffeate 13 (14) 37 (87) 2.85 Ferulate 140 (130) 71 (96) 0.51 Sinapate NC (NC) – (–) – 3,4-Dimethoxycinnamate NC (NC) – (–) – Gm4CL3 Cinnamate 1100 45 0.04 4-Coumarate 9 100 11.12 Caffeate 50 74 1.48 Ferulate 3100 25 8.1 · 10 )3 Sinapate NC – – 3,4-Dimethoxycinnamate NC – – Gm4CL4 Cinnamate 260 20 0.08 4-Coumarate 10 100 10.00 Caffeate 34 50 1.47 Ferulate 1300 30 0.02 Sinapate NC – – 3,4-Dimethoxycinnamate NC – – 1312 C. Lindermayr et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Substrate specificity of all four recombinant soybean 4CL isoenzymes was a nalysed for a series of cinnamic a cids bearing phenyl ring substitutions that are typical for phenylpropanoid compounds of higher plants. A major conclusion from this part of the results is that, rather unexpectedly, the substrate specificity of only two recom- binant 4CL isoforms, Gm4CL1 and Gm4CL2, closely matched that of the known 4CL isoforms [1], whereas for Gm4CL3 and 4 there was no known counterpart. A lthough structural constraints in fusion prote ins g enerated from cDNAs could affect substrate conversion, such an effect appears to be unlikely for the observed catalytic properties of recombinant 4CL3 o r 4CL4. I t appears more likely that 4CL3 and 4 had not been identified in earlier work because proteins had been extracted from unstressed tissues [1] o r a differentiation b etween 4CL2 and 4 CL3 or 4 in unfraction- ated protein extracts was not possible due to the lack of knowledge of their catalytic properties (Table 5) [20]. The analysis of the s ubstrate specificity of t he recombin- ant soybean 4CL expressed in E. coli (Table 5) revealed pronounced differences in the ability of the isoenz ymes to utilize differently r ing-substituted cinnamic acids. The four members of soybean 4CL thus represent enzymes with broad (4CL1), intermediate (4CL2), and more restricted substrate specificity ( 4CL3 and 4CL4). Soybean 4CL1 thus far is the only plant isoenzyme capable of activating sinapate, a presumed p recursor f or the syringyl monolignol formation, for whic h a c DNA is available. Alternatively, woody angiosperms have been described to obviate the need for the activation o f highly ring-substituted cinnamic acids as precursors for monolignol biosynthesis by using ring substitution-specific methylations acting on already activa- ted m etabolites [12]. Loss-o f-function experiments i n a lfalfa (Medicago sativa) likewise indicated no necessity for CoA- ligase isozymes converting already highly substituted cin- namic acids [14]. The existence of the soybean 4CL1 isoform, using ferulate and sinapate with very high efficiency, thus indicates an even higher flexibility in the metabolic grid responsible for the distribution of phenyl- propanoids as presently thought [47]. The molecular c haracterization of 4CL isoforms expres- sing pronounced differences in the substrate specificity may facilitate studies on the a ctive site of this class of enz ymes to identify amino acids that are of functional i mportance. This may include amino acid motifs such as a putative AMP- binding domain [48], but also am ino a cids that are responsible for a broad or narrow specificity towards ring- substituted cinnamates. The central position of 4CL in phenylpropanoid branch pathways therefore makes t his enzyme a potentially valuable target for pathway or product engineering in higher plants. Attempts towards this goal have recently been reported for 4CL2 from Arabido psis thaliana [49,50]. Another particularly striking observation is the differen- tial expression of the four 4CL isoforms. Based o n enzyme activity measurements, 4CL1 and 4CL2 are both e xpressed in the unstresse d cell culture whereas based on RNA blot analyses and activity measurements 4CL3 and 4CL4 appear to be the major elicitor-induced forms being not expressed in untreated cells. As the cloned cDNAs for 4CL3 and 4CL4 cross-hybridized under the conditions used, it is not possible to analyse separately the t ranscript levels of Gm4CL3 and Gm4CL4. Even though some uncertainty remains about the relative proporti on of the e xpressed t ranscript levels corres- ponding to the two closely related genes, 4CL3 or the closely related 4CL4 protein very probably represent the highly elicitor-induced enzyme. By contrast, the expression of 4CL1 in the soybean cell culture is r educed by elicitor treatment, a behaviour which is reported only rarely for enzymes c ommitted to the biosynthesis of plant p rotective compounds. A consequence of this differential elicitor responsiveness c ould b e that the overall product profile of CoA esters of cinnamic acids in soybean c ells is shifted after elicitation due to the large differences in substrate preference of the 4 CL isoform 1 vs. 3 or 4. A similar c onsequence might a pply to the soybean seedling after infection. Again transcript levels of predominantly isoforms 3 or 4 are enhanced close to the infection sites with a time-course comparable to that observed for o ther enzymes of phenyl- propanoid metabolism [51]. A genomic clone from soybean containing a complete copy of one of the genes encoding an inducible 4CL isoenzyme, Gm4CL3, was isolated. Although promoter analyses for this gene have not yet been carried out, the presence of a putative T ATA box and other boxes (A, E, L, P, data not shown), which are conserved among several plant g enes in phenylpropanoid m etabolism ( phenylalan ine ammonia-lyase, 4CL, caffeoyl-CoA O-methyltransferase) [19,52–54], indicate common principles of gene regulation under various metabolic conditions. Pathogen attack or elicitor treatment in soybean affects various metabolic activities, i ncluding different branches of phenylpropanoid metabolism. Phenylpropanoid responses are temporally and spatially coord inated [26,28,55]. They lead t o the massive deposition o f cell wall phenolics, release of isoflavones from conjugates, and the production of the soybean phytoalexins, glyceollins. All together, these Fig. 6 . Scheme illustrating the central position of hydroxycinnamate CoA esters and 4CL isoforms in the biosynthesis of various phenyl- propanoid metabolites in soybean under different developmental and environmental con ditions. Ó FEBS 2002 4-Coumarate:CoA ligase gene family from Glycine max (Eur. J. Biochem. 269) 1313 [...]... residues of 4-coumarate:coenzyme A ligase allows the rational design of mutant enzymes with new catalytic properties J Biol Chem 276, 26893– 26897 51 Habereder, H., Schroder, G & Ebel, J (1989) Rapid induction of ¨ phenylalanine ammonia-lyase and chalcone synthase mRNAs during fungus infection of soybean (Glycine max L.) roots or elicitor treatment of soybean cell cultures at the onset of phytoalexin synthesis... ligase, an enzyme involved in the resistance response of soybean (Glycine max L.) against pathogen attack Plant Physiol 102, 1147–1156 Schoch, G., Goepfert, S., Morant, M., Hehn, A. , Meyer, D., Ullmann, P & Werck-Reichhart, D (2001) CYP9 8A3 from Arabidopsis thaliana is a 3¢-hydroxylase of phenolic esters, a missing link in the phenylpropanoid pathway J Biol Chem 276, 36566–36574 Pakusch, A & Matern,... strategies of soybean against the fungus Phytophthora megasperma f sp glycinea: a molecular analysis Trends Biochem Sci 13, 23–37 25 Graham, T.L & Graham, M .Y (1991) Cellular coordination of molecular responses in plant defense Mol Plant-Microbe Interact 4, 415–422 26 Morris, P.F., Savard, M.E & Ward, E.W (1991) Identification and accumulation of isoflavonoids and isoflavone glucosides in soybean leaves... leaves and hypocotyls in resistance responses to Phytophthora megasperma f sp glycinea Physiol Mol Plant Pathol 39, 229–244 27 Graham, M .Y & Graham, T.L (1991) Rapid accumulation of anionic peroxidases and phenolic polymers in soybean cotyledon tissues following treatment with Phytophthora megasperma f sp glycinea wall glucan Plant Physiol 97, 1445–1455 28 Hahn, M.G., Bonhoff, A & Grisebach, H (1985) Quantitative... evolutionary classes in angiosperms Plant J 19, 9–20 Hu, W.-J., Kawaoka, A. , Tsai, C.-J., Lung, J., Osakabe, K., Ebinuma, H & Chiang, V.L (1998) Compartmentalized expression of two structurally and functionally distinct 4-coumarate: CoA ligase genes in aspen (Populus tremuloides) Proc Natl Acad Sci USA 95, 5407–5412 Uhlmann, A & Ebel, J (1993) Molecular cloning and expression of 4-coumarate:coenzyme A ligase, ... 113, 65–74 Allina, S.M., Pri-Hadash, A. , Theilmann, D .A. , Ellis, B.E & Douglas, C.J (1998) 4-Coumarate:coenzyme A ligase in hybrid poplar Properties of native enzymes, cDNA cloning, and analysis of recombinant enzymes Plant Physiol 116, 743–754 Ehlting, J., Buttner, D., Wang, Q., Douglas, C.J., Somssich, I.E & ¨ Kombrink, E (1999) Three 4-coumarate:coenzyme A ligases in Arabidopsis thaliana represent... a certain degree the substitution pattern of subsequent phenylpropanoid branches that require suitably ring-substituted cinnamoyl CoA esters as substrates (Fig 6) However, phenyl ring modification involving hydroxylation and O-methylation can basically occur by different pathways, namely by modifications at the free acid level, by substitutions at the level of conjugated intermediates, such as CoA esters,... esters, and at the level of the aldehyde and alcohol intermediates of monolignol synthesis [21,47,56] The metabolic interconversions of cinnamic acids could add to the complexity of the final phenylpropanoid products, their cellular localization, and the dynamics of their synthesis In any case, the coordinated regulation of 4CL3 or 4 with all other known enzymes of phytoalexin biosynthesis in soybean [24]... Quantitative localization of the phytoalexin glyceollin I in relation to fungal hyphae in soybean roots infected with Phytophthora megasperma f sp glycinea Plant Physiol 77, 591–601 29 Hille, A. , Purwin, C & Ebel, J (1982) Induction of enzymes of phytoalexin synthesis in cultured soybean cells by an elicitor from Phytophthora megasperma f sp glycinea Plant Cell Reports 1, 123–127 30 Sharp, J.K., Valent,... (4CL) gene family Plant Physiol 112, 193–205 Brodelius, P.E & Xue, Z.-T (1997) Isolation and characterization of a cDNA from cell suspension cultures of Vanilla planifolia encoding 4-coumarate:Coenzyme A ligase Plant Physiol Biochem 35, 497–506 Humphreys, J.M., Hemm, M.R & Chapple, C (1999) New routes for lignin biosynthesis defined by biochemical characterization of recombinant ferulate 5-hydroxylase, a . Divergent members of a soybean ( Glycine max L. ) 4-coumarate:coenzyme A ligase gene family Primary structures, catalytic properties, and differential. given in parentheses): Arabidopsis thaliana 4CL1 (U1867 5), A. thaliana 4CL2 (AF10608 6), A. thaliana 4CL3 (AF10608 8), G. max 4CL1 (AF27926 7), G. max 4CL2 (AF00225 9),