Physiology, metabolism

Molecular characterization of the phenolic acidmetabolism in the lactic acid bacteria

Lactobacillus plantarum

Lise BARTHELMEBS, Charles DIVI ÈS, Jean-François CAVIN *

Laboratoire de Microbiologie, UMR INRA ENSBANA, Université de Bourgogne,1 esplanade Erasme, 21000 Dijon, France

Abstract — The lactic acid bacteria Lactobacillus plantarumdisplays substrate-inducible decar-boxylase activities on p-coumaric, caffeic and ferulic acids. Purification of the p-coumaric aciddecarboxylase (PDC) was performed. Sequence of the N-terminal part of the PDC led to the cloningof the corresponding pdcgene. Expression of this gene in Escherichia colirevealed that PDC displayeda weak activity on ferulic acid, detectable in vitro in the presence of ammonium sulfate. Transcrip-tional studies of this gene in L. plantarum demonstrated that the pdctranscription is phenolic acid-dependent. A mutant deficient in the PDC activity, designated LPD1, was constructed to study phe-nolic acid alternate pathways in L. plantarum. LPD1 mutant strain remained able to metabolizeweakly p-coumaric and ferulic acids into vinyl derivatives or into substituted phenyl propionicacids. These results indicate thatL. plantarumhas a second acid phenol decarboxylase enzyme andalso displays inducible acid phenol reductase activity. Finally, PDC activity was shown to confer aselective advantage for LPNC8 grown in acidic media supplemented with p-coumaric acid, comparedto the LPD1 mutant devoid of PDC activity.

phenolic acid / phenolic acid decarboxylase / phenolic acid reductase / Lactobacillus plantarum

Résumé — Caractérisation moléculaire du métabolisme des acides phénol chez Lactobacillusplantarum. La bactérie lactique Lactobacillus plantarum est capable de décarboxyler les acides p-cou-marique, férulique et caféique. La purification de l’acide p-coumarique décarboxylase (PDC) a été réa-lisée. La détermination de la séquence N-terminale de cette enzyme a permis de cloner le gène pdccorrespondant. L’expression de ce gène chez Escherichia colirévèle que la PDC possède une faibleactivité sur l’acide férulique, détectable in vitro uniquement en présence de sulfate d’ammonium.L’étude des ARN messagers du gène pdc chezL. plantarum montre que la transcription de ce gènedépend des acides phénols. La construction d’un mutant déficient pour l’activité PDC, nommé LPD1a été réalisée afin d’étudier les métabolismes secondaires des acides phénols chez L. plantarum. Cemutant LPD1 reste capable de dégrader les acides phénols en dérivés de type vinyl phénol ou enacides phényl propioniques. Ces résultats indiquent que L. plantarumpossède une seconde acide

Lait 81 (2001) 161–171 161© INRA, EDP Sciences, 2001

* Correspondence and reprintsTel.: (33) 3 80 39 66 72; fax: (33) 3 80 39 66 40; e-mail: cavinjf@u-bourgogne.fr

L. Barthelmebs et al.

1. INTRODUCTION

Phenolic acids, of which ferulic andp-coumaric acids are the main constituents,represent 1–5% (m/w dry weight) of tropi-cal grasses, of maize and wheat bran and ofsugarbeet pulp [22]. These acids are cova-lently bound to polysaccharides (xylan andlignin) in plant cell walls to ensure cell wallrigidity [12]. A wide range of microorgan-isms display cinnamoyl esterase activitieswhich have been shown to release pheno-lic acids from plant cell wall polymers [9].Under their free form, phenolic acids areconsidered differently according to themicroorganisms. Phenolic acids haveinhibitory effect on the growth of mostmicroorganisms [4, 34]. It is the reason whythey are used as antimicrobial agents in thefood industry [25]. Phenolic acids also serveas a signal and induce virulence gene expres-sion in the plant-associated Agrobacteriumtumefaciens[20]. Some Pseudomonasstrains as well as Acinetobacter calcoaceti-cusare able to use these acids as the solesource of carbon for growth [17, 24, 28, 30].Moreover, ferulic acid is exploited to pro-duce value-added aromatic compounds, suchas vanillin. This bioconversion is performedby two fungi, Aspergillus niger and Pycno-porus cinnabarinus[21].

In other microbial systems, phenolic acidsare metabolized into volatile phenols bytwo different pathways. Most often, theyare first decarboxylated into 4-vinyl deriva-tives and then reduced into 4-ethyl deriva-tives. Phenolic acid decarboxylases (PAD)have been characterized in the yeasts Sac-charomyces cerevisiaeand Brettanomycesannomalus [10, 15] and also in the bacteria

Bacillus pumilus [33], Lactobacillus plan-tarum [6] andBacillus subtilis[8]. A secondpathway has been proposed for Lactoba-cillus pastorianuswhere caffeic andp-coumaric acids are first reduced into sub-stituted phenyl propionic acids and thendecarboxylated into 4-ethyl derivatives [31].Some of these volatile phenols, particularlyvinyl and ethyl guaiacol (generated fromferulic acid) are useful aromatic chemicals[19] or contribute naturally to aroma in wine[16] and other fermented foods and bever-ages. On the contrary, some volatile phe-nols, such as vinyl and ethyl phenol (fromp-coumaric acid) are most often consideredphenolic off-flavors and are responsible foralterations in organoleptic properties [15].

The interest in improving our under-standing of phenolic acid biodegradation ismultiple. Firstly, as has been shown forS. cerevisiae, PAD activity may confer aselective advantage upon microorganismsduring growth on plant, where PAD expres-sion could constitute a stress responseinduced by phenolic acid [10, 18]. Secondly,phenol derivatives are valuable intermediatesin the biotechnological production of newflavor and flagrance chemicals [19]. Thirdly,they are regarded as a source of phenolicoff-flavors in many beers and wines, due totheir characteristic aroma and their lowthreshold detection [16, 29].

Several strains of lactic acid bacteria iso-lated from wine were tested for their activ-ities to metabolize ferulic and p-coumaricacids [5]. It was shown that ferulic andp-coumaric acids were strongly decarboxy-lated by growing cultures ofL. plantarumand Pediococcus pentosaceus.The decar-boxylase activity was only detected for

162

phénol décarboxylase ainsi qu’une activité acide phénol réductase. Enfin, l’activité PDC confère àL. plantarum un avantage sélectif pour sa croissance en milieu contenant des acides phénols, com-paré à la souche LPD1 dépourvue d’activité PDC.

acide phénol / acide phénol décarboxylase / acide phénol réductase / Lactobacillus plantarum

Phenolic acid metabolism in L. plantarum

DNA was prepared using the methoddescribed by Posno et al. [23]. PCR productswere purified with the Jet Pur Kit(Genomed, Bioprobe, Montreuil, France)and sequenced by the dideoxy chain termi-nation method [27] with the Thermoseque-nase radiolabeled terminator cycle sequenc-ing kit (Amersham Life Science Inc.,Cleveland, Ohio, USA) in accordance withthe recommendations of the manufacturer.PCR amplification was performed using0.1 µg DNA template with 0.5 U of TaqDNA polymerase (Appligene) under stan-dard conditions, in an automatic HybaidDNA thermocycler (Hybaid Ltd., Tedding-ton, United Kingdom). E. coliand L. plan-tarumstrains were transformed by electro-poration as described by Dower et al. [14]and Aukrust and Nes [1], respectively.

2.3. Preparation of whole cellsuspensions, cell-free extractsand assay of phenolic aciddegradation

Cells of L. plantarumgrown in MRSmedium and E. coli grown in LB mediumwere harvested by centrifugation, washedwith 25 mmol.L–1 potassium phosphatebuffer (pH 6.0) and resuspended in the samebuffer at a final concentration of 5 g (dryweight) of cells per liter. Reactions werestarted by adding p-coumaric or ferulicacids. During incubation, samples were cen-trifuged and supernatants were diluted50-fold in Stop buffer (20 mmol.L–1 Tris-HCl, 3 g.L–1 SDS to stop activity, pH 6)prior to analysis. For cell-free extract prepa-ration, cells were harvested as describedabove and disrupted with a French press at1.2 × 108 Pa. Kinetic reactions were startedby addition of the substrate and sampleswere diluted in Stop buffer. Phenolic acidconsumption and derivative production wasmonitored by UV-spectrophotometry scan-ning (from 180 to 330 nm using quartzcuvettes in a Beckman DU600 spectropho-tometer) as previously described [2, 5, 13].

bacteria grown with these substrates sug-gesting that this activity was inducible.Moreover, growth of L. plantarumandP. pentosaceusstrains was not particularlyaffected by presence of ferulic andp-coumaric acids in growth medium, indi-cating that decarboxylation of these com-pounds may confer resistance to inhibitoryeffects of phenolic acids. The ubiquitouslactic acid bacteria, L. plantarum, used asa malolactic starter in wine and as a lacticstarter for many vegetable fermentations,was selected as a model for studying phe-nolic acid metabolism. This bacterium dis-played substrate-inducible decarboxylaseactivity on p-coumaric, ferulic, and caffeicacids. Purification of the L. plantarump-coumaric acid decarboxylase (PDC)enables the biochemical characterization ofthis enzyme, and the cloning of the corre-sponding pdc gene. A food grade mutantstrain, deficient in PDC activity, was con-structed to modify the aromatic property ofthe strain, to study the metabolic pathwayof phenolic acids, and the influence of PDCactivity on the growth in the presence ofphenolic acids.

2. MATERIALS AND METHODS

2.1. Microorganisms and cultureconditions

The L. plantarumLPNC8 and LPD1were grown at 37 °C in MRS medium with-out shaking [11]. E. coliTG1 was grown at37 °C on LB medium under shaking con-ditions [3]. Antibiotics were used in thefollowing concentrations: erythromycin100 mg.L–1 for E. coli and 5 mg.L–1 forL. plantarum.

2.2. DNA manipulationand transformation procedures

Standard procedures described bySambrook et al. [26] were used for DNAmanipulation. L. plantarumchromosomal

163

L. Barthelmebs et al.

3. RESULTS

3.1. Phenolic acid metabolismin L. plantarum, purificationand characterization of thep-coumaric acid decarboxylase

L. plantarum is a lactic acid bacteria whichis able to decarboxylate p-coumaric and fer-ulic acids into 4-vinyl phenol and4-vinyl gaiacol respectively; however, wedid not know if this bacteria had one or twodifferent phenolic acid decarboxylases(PADs). Previous experiments revealed thatL. plantarum induced with 3 mmol.L–1

p-coumaric acid displayed decarboxylaseactivity on p-coumaric acid (0.6 µmol.min–1.mg–1). While L. plantarum inducedwith 3 mmol.L–1 ferulic acid displayed a sim-ilar activity on p-coumaric acid and also aweak activity (0.01 µmol.min–1.mg–1) on fer-ulic acid [7]. These results indicate that twodistinct PADs are present in L. plantarum.

The enzyme induced by p-coumaric acid,which metabolized p-coumaric acid, waspurified and designated p-coumaric aciddecarboxylase (PDC) [7]. SDS-page analysisof this purified PDC revealed the presenceof single band of 23.5 kg.mol–1. Determina-tion of PDC molecular mass indicated thatthis enzyme of 93 kg.mol–1 was a homote-tramer. Full enzyme activity was obtained in25 mmol.L–1 phosphate buffer with an opti-mal pH of 6.0 and temperature of 30 °C. Onlyp-coumaric and caffeic acids were decar-boxylated by the PDC in the conditions testedin vitro [7]. This result indicated that L. plan-tarumshould produce a second PAD, spe-cific for ferulic acid and induced by this sub-strate. The N-terminal amino-acid sequenceof the purified PDC was obtained.

3.2. Cloning of the pdcgene,transcriptional analysisand characterizationof the recombinant PDCin Escherichia coli

By using degenerate primers designatedfrom the first 19 N-terminal amino-acid

sequence of PDC, a 57-bp fragment wasamplified fromL. plantarumby PCR, wascloned in E. colivector and was sequenced[6]. Its deduced amino-acid sequence provedto be identical to the 19 amino-acids of thePDC enzyme. This nucleotide sequence wasused as a probe to screen the L. plantarumgenomic library. Only one clone, designatedpJPDC1 with a 2.3-kb insertion, hybridizedwith this probe. An open reading frame(ORF), with a coding capacity of 522-bpwas detected and encoding for a 174 amino-acid deducted product, which N-terminalamino acid sequence was identical to theN-terminal extremity of the purified PDC[6]. The primary structure of the deducedpolypeptide displays 64% identity with theFDC from Bacillus pumilus [33]. This 522-bpORF was identified as the L. plantarum pdcgene. A DNA probe encompassing about300 nucleotides from the pdc gene ofL. plantarumwas used to screen a B. subtilisand a P. pentosaceusgenomic library andled to the cloning of the genes pad and padAof, respectively, B. subtilis [8] and P. pen-tosaceus (Barthelmebs L., Lecomte B.,Diviès C., Cavin J.-F., submitted for publi-cation).

Although these four PADs exhibited 64%of amino-acid sequence identity, the puri-fied enzymes had different characteristics.First, PDC was a homotetramer while FDCand BSPAD were homodimers. These twopurified enzymes displayed about the samehigh activity (500 nmol.min–1.mg–1) forp-coumaric and ferulic acids, contrary to

the PDC and the PPPAD which displayed noactivity on ferulic acid. Moreover, thesequence alignment of the four enzymesshowed that the main differences betweenthese three proteins are located in the N-and C-terminal parts (Fig. 1). Nevertheless,no homology was found between thesePADs with the other decarboxylases, includ-ing the phenylacrylic acid decarboxylase(PAD1) cloned from Saccharomyces cere-visiae, which was able to decarboxylate,with a weak activity, p-coumaric and fer-ulic acids.

164

Phenolic acid m

etabolism in L

. pla

nta

rum

165

Figure 1.Comparison between the sequence of the B. subtilisPAD (PAD-BS), the B. pumilusFDC (FDC-BP), the L. plantarumPDC (PDC-LP) and theP. pentosaceusPAD (PAD-PP). The sequences were aligned using the Clustal program. Identical residues are boxed and shaded. The numbers on theright correspond to the amino-acid position in the protein sequence.

L. Barthelmebs et al.

Transcriptional analysis of the pdcgenewas carried out with RNA from L. plan-tarum grown with or without 1.2 mmol.L–1

p-coumaric acid in the medium. A pdc tran-script of 700 nucleotides was detected inthe induced RNA while no transcript wasdetected in the uninduced RNA, even byNorthern blotting or by primer extensionexperiments. These results indicate that pdctranscription is acid phenol-dependent. Thesame results were obtained for the tran-scription of the pad from B. subtilis [8]and the padA from P. pentosaceus (Barthel-mebs L., Lecomte B., Diviès C., Cavin J.-F.,submitted for publication).

P-coumaric and ferulic acid metabolismwas tested in whole cells and correspond-ing cell-free extract of recombinant E. coliTG1 (pJPDC1) strain grown with or withoutp-coumaric acid as inducer. No PDC activ-ity was found in the control E. coli TG1(pJDC9). Induced and uninduced wholecells and cell-free extracts displayed similardecarboxylase activity of 5 to 12 µmol.min–1.mg–1 on p-coumaric acid. In inducedand uninduced whole cells, ferulic acid wasdecarboxylated at a rate of about 40 nmol.min–1.mg–1, while the corresponding cell-free extract displayed no detectable activ-ity on ferulic acid in phosphate buffer. Reac-tion conditions were then modified byindependently varying pH and temperature,and by adding glycerol or salts to kineticbuffer. Ferulic acid decarboxylase activitywas stimulated in cell-free extracts supple-mented with 20% ammonium sulfate or 20%NaCl with an optimum activity of about30 nmol.min–1.mg–1, indicating that PDCdisplays a low ferulic acid decarboxylaseactivity under those conditions.

3.3. Phenotypic analysisof the L. plantarumstraindeficient in PDC activity

In order to investigate alternate pathwaysfor phenolic acid degradation, a L. plan-tarum strain deficient in PDC activity was

constructed [2]. The procedure used led to acompletely stable chromosomal deletionwithin the pdc gene through a two-stephomologous recombination process. In short,a copy (A-C) of the pdcgene lacking 200-bpinternal region (B) (Fig. 2), was cloned on avector. A frame shift in the deleted copycreated a stop codon, which caused the syn-thesis of a truncated polypeptide of 51 aminoacids. Contrary to the E. coli TG1 (pJPDC1)clone, no PDC activity was detected oneither substrate in whole cells or cell-freeextracts of the E. coli TG1 strain bearingthis plasmid. This plasmid, which is suit-able for generating the L. plantarum mutant,was introduced in L. plantarumby electro-poration. Campbell-type integration of thevector in the chromosome (involving the Aor C region) resulted in a wild-type pheno-type. Secondary excision through intra-chromosomal recombination within theother region of homology (C or A) led to adeletion of the internal B region of the pdcgene. The mutant strain was named LPD1.Southern blotting and DNA sequencingwere performed on the LPD1 DNA to con-firm the deletion [2].

In order to characterize the phenolic acidmetabolism of LPD1 mutant, growing

166

Figure 2. Physical map of the pdclocus in: (a) thewild-type strain L. plantarumand (b) the mutantstrain LPD1. Horizontal arrows represent the twoORFs and their orientation. The start sites areindicated by a vertical arrow and the stop codonsby a T. The position and orientation of theprimers are indicated by short horizontal arrows.

Phenolic acid metabolism in L. plantarum

3.4. Influence of p-coumaricand ferulic acid concentrationon the growth of wild typeand LPD1 mutant strainsat different pH

Increasing concentrations of p-coumaricand ferulic acids (0, 1.2, 3, and 6 mmol.L–1)were tested for their effect on growth of thewild-type and LPD1 mutant strains in MRSbroth at pH 6.5. For the wild-type strain,addition of 1.2 or 3 mmol.L–1 p-coumaricacid and 1.2 mmol.L–1 ferulic acid in themedium had no apparent effect on growth(Figs. 1a, 2a and 3). Addition of 6 mmol.L–1

p-coumaric acid and 3 or 6 mmol.L–1 fer-ulic acid increased the lag phase, but thefinal biomass reached a level similar to thatof the control culture without p-coumaricor ferulic acids. The growth of LPD1 mutantstrain was not different to that of the wild-type strain in presence of increasing con-centrations of ferulic acid (Figs. 2b and 3).LPD1 growth was weakly reduced with1.2 mmol.L–1 p-coumaric acid but was sig-nificantly affected with 3 or 6 mmol.L–1

p-coumaric acid in the medium, with a finalbiomass reaching only 60 and 80% com-pared to that of the control without acid(Figs. 1b and 3).These results seem to indi-cate that p-coumaric was more toxic thanferulic acid for L. plantarum cells devoidof their major decarboxylase activity. Inorder to investigate the influence of aciduptake on growth rate, cultures were per-formed at pH 4.5 in MRS broth with thethree p-coumaric and ferulic acid concen-trations previously used (Fig. 4). At pH 4.5,the growth of the wild-type strain remainedunaffected in the presence of 1.2 and3 mmol.L–1 p-coumaric and 1.2 mmol.L–1

ferulic acids, was partially reduced with3 mmol.L–1 p-coumaric and 1.2 mmol.L–1

ferulic acids and was totally inhibited with6 mmol.L–1 ferulic acid (Figs. 1a, 2a and 4).On the other hand, LPD1 growth wasstrongly inhibited in the presence of 1.2 and3 mmol.L–1 p-coumaric acid and totallyinhibited with 6 mmol.L–1 p-coumaric acid

cultures of LPD1 were divided into 5 sam-ples and induced with 1.2 or 3 mmol.L–1

p-coumaric acid and ferulic acid or noninduced. Whole resting cell suspensionswere prepared in 25 mmol.L–1 phosphatebuffer (pH 6.0) and tested for p-coumaricand ferulic acid metabolism. No acid degra-dation was detected in the uninduced wholecells. Cells induced with ferulic acid (1.2and 3 mmol.L–1) displayed a low decar-boxylase activity of about 12 nmol.min–1.mg–1 on p-coumaric and ferulic acids, indi-cating that L. plantarumpossesses a secondPDC. This activity was not detected in cellsinduced with 1.2 mmol.L–1 p-coumaric acid,although the two substrates were degradedinto a product which could be either substi-tuted phenyl propionic acids or 4-ethylderivatives. Production of 4-ethyl deriva-tives from phenolic acids requires the priorformation of 4-vinyl derivatives or substi-tuted phenyl propionic acids. As 4-vinylderivatives were never detected in kineticexperiments, phenolic acids were likelyreduced into substituted phenyl propionicacids, and subsequently decarboxylated into4-ethyl derivatives. Cells induced with3 mmol.L–1 p-coumaric seem to display thetwo activities of decarboxylase and reduc-tase. Since a reduced cofactor is generallyrequired for enzymatic reduction,20 mmol.L–1 glucose was added to wholecells prior to starting the kinetics, in order tostimulate glycolysis and increase the poolof reduced cofactors. This addition stimu-lated a reductase activity of 10 nmol.min–1.mg–1 on the two phenolic acids forwhole cells induced by p-coumaric acid (1.2or 3 mmol.L–1) and by 1.2 mmol.L–1 fer-ulic acid. Decarboxylase activity was nev-ertheless detected in whole cells induced by3 mmol.L–1 ferulic acid. Taken together, ourresults indicate the presence, in L. plan-tarum, of a second phenolic acid decar-boxylase (named PDC2), greatly inducedby ferulic acid and of a putative phenolicacid reductase activity (named PAR)induced by p-coumaric and ferulic acids inthe presence of glucose.

167

L. Barthelmebs et al.

(Figs. 1b and 4). LPD1 growth was similarto L. plantarum wild-type growth with thesame concentrations of ferulic acid (Figs. 2band 4). Taken together, these results indi-cated that the toxicity of p-coumaric andferulic acids is enhanced under low initialpH condition. Nevertheless, PDC activityseems to confer resistance to p-coumaricacid toxicity.

4. DISCUSSION

Study of phenolic acid metabolism inL. plantarumwild-type and mutant straindeficient in PDC activity, revealed that thislactic acid bacteria possesses at least threesubstrate inducible enzymatic activities ableto metabolize phenolic acids. PDC enzymedisplays the higher activity on p-coumaric

acid, and a 50-fold lower activity on ferulicacid, detectable in vitro only when ammo-nium sulfate or NaCl were added in the reac-tion buffer. L. plantarum is, to our knowl-edge, the first microorganism which displaystwo distinct and functional PAD activities.In B. subtilis, a yclB gene which deducedpolypeptide displayed 35% amino acidsequence identity with the PAD1 fromS. cerevisiae which was identified bygenome sequencing [10, 32]. Nevertheless,the functionality of the YclB protein wasnot proved. The analysis of the LPD1 mutantstrain also revealed that L. plantarumcon-verts phenolic acids into substituted phenylpropionic acids, therefore indicating thatL. plantarumappears to have a weakp-coumaric and ferulic acid reductase activ-ity, induced by both substrates and mostlyactive when glucose is added. This pathway

168

Figure 3. Growth of L. plantarumwild-type (a) and LPD1 (b) strains at (1) different p-coumaric acidconcentrations (in mmol.L–1), r: 0; ■: 1.2; m: 3; ●: 6 and at (2) different ferulic acid concentrations(in mmol.L–1), r: 0; ■: 1.2; m: 3; ●: 6, at pH 6.5.

Phenolic acid metabolism in L. plantarum

making this regulatory system a potentialtool for gene expression studies in lacticacid bacteria or other gram-positive bacteria.

REFERENCES

[1] Aukrust T., Nes I.F., Transformation of Lacto-bacillus plantarumwith the plasmid pTV1 byelectroporation, FEMS Microbiol. Lett. 52(1988) 127–132.

[2] Barthelmebs L., Diviès C., Cavin J.-F., Knock-out of the p-coumarate decarboxylase gene fromLactobacillus plantarumreveals the existenceof two other inducible enzymatic activitiesinvolved in phenolic acid metabolism, Appl.Environ. Microbiol. 66 (2000) 3368–3375.

[3] Bertani G., Studies on lysogenesis. I. The modeof phage liberation by lysogenic Escherichiacoli, J. Bacteriol. 60 (1951) 293–300.

[4] Borneman W.S., Akin D.E., van Eseltine W.P.,Effect of phenolic monomers on ruminal bac-teria, Appl. Environ. Microbiol. 52 (1986)1331–1339.

has been previously shown in Lactobacil-lus pastorianus[31] and also in cells of S.cerevisiae, grown under argon condition[19].

We have also demonstrated that func-tional PDC enzyme clearly confers a selec-tive advantage on the wild-type strainL. plantarum for growing in the presenceof p-coumaric acid, while growth of theLPD1 mutant at acidic pH is strongly inhib-ited by p-coumaric acid. Therefore, PDCsynthesis in L. plantarumappears to be themost efficient cellular response to quicklyconvert p-coumaric acid into a less toxicderivative. Mechanism of induction of thePDC synthesis is currently in progress. Theabsence of detectable PDC activity with theabsence of detectable corresponding mRNAin the uninduced cells indicate that the pdctranscription is phenolic acid-dependant,

169

Figure 4. Growth of L. plantarumwild-type (a) and LPD1 (b) strains at (1) different p-coumaric acidconcentrations (in mmol.L–1), ◆: 0; ■: 1.2; m: 3;●: 6, and at (2) different ferulic acid concentrations(in mmol.L–1), ◆: 0; ■: 1.2; m: 3; ●: 6, at pH 4.5.

L. Barthelmebs et al.

[5] Cavin J.-F., Andioc V., Etiévant P.X., DivièsC., Ability of wine lactic acid bacteria to metab-olize phenol carboxylic acids, Am. J. Enol. Vitic.44 (1993) 76–80.

[6] Cavin J.-F., Barthelmebs L., Diviès C., Molec-ular characterization of an inducible p-coumaricacid decarboxylase from Lactobacillus plan-tarum: Gene cloning, transcriptional analysis,overexpression in Escherichia coli, purificationand characterization, Appl. Environ. Microbiol.63 (1997) 1939–1944.

[7] Cavin J.-F., Barthelmebs L., Guzzo J., vanBeeumen J., Samyn B., Travers J.-F., Diviès C.,Purification and characterization of an induciblep-coumaric acid decarboxylase from Lacto-bacillus plantarum, FEMS Microbiol. Lett. 147(1997) 291–295.

[8] Cavin J.-F., Dartois V., Diviès C., Gene cloning,transcriptional analysis, purification and char-acterization of phenolic acid decarboxylase fromBacillus subtilis,Appl. Environ. Microb. 64(1998) 1466–1471.

[9] Christov L.P., Prior B.A., Esterases of xylan-degrading microorganisms: production, prop-erties and significance, Enzyme Microb. Tech-nol. 15 (1993) 460–475.

[10] Clausen M., Lamb C.J., Megnet R., DoernerP.W., PAD1 encodes phenylacrylic acid decar-boxylase which confers resistance to cinnamicacid in Saccharomyces cerevisiae, Gene 142(1994) 107–112.

[11] de Man P.J., Rogosa M., Sharpe M., A mediumfor the cultivation of Lactobacilli, J. Appl. Bac-teriol. 23 (1960) 130–135.

[12] de Vries R.P., Michelsen B., Poulsen C.H.,Kroon P.A., van den Heuvel R.H.H., FauldsC.B., Williamson G., van den HomberghJ.P.T.W., Visser J., The faeA genes fromAspergillus nigerandAspergillus tubingensisencode ferulic acid esterases involved in degra-dation of complex cell wall polysaccharides,Appl. Environ. Microbiol. 63 (1997) 4638–4644.

[13] Degrassi G., Polverino de Laureto P., BruschiC.V., Purification and characterization of feru-late and p-coumarate decarboxylase fromBacil-lus pumilus,Appl. Environ. Microbiol. 61 (1995)326–332.

[14] Dower W.J., Miller F., Ragsdale C.W., Highefficient transformation of Escherichia colibyhigh voltage electroporation, Nucl. Acids Res. 16(1988) 6127–6145.

[15] Edlin D.A.N., Narbad A., Gasson M.J.,Dickinson J.R., Lloyd D., Purification and char-acterization of hydroxycinnamate decarboxy-lase from Brettanomyces anomalus, EnzymeMicrobiol. Technol. 22 (1998) 232–239.

[16] Etiévant P.X., Issanchou S.N., Marie S., DucruetV., Flanzy C., Sensory impact of volatile phenolson red wine aroma: Influence of carbonic mac-eration and time of storage, Sci. Aliments 9(1989) 19–33.

[17] Gasson M.J., Kitamura Y., McLauchlan W.R.,Narbad A., Parr A.J., Parsons E.L.H., Payne J.,Rhodes M.J.C., Walton N.J., Metabolism of fer-ulic acid to vanillin, J. Biol. Chem. 273 (1998)4163–4170.

[18] Goodey A.R., Tubb R.S., Genetic and bio-chemical analysis of the ability of Saccha-romyces cerevisiae to decarboxylate cinnamicacids, J. Gen. Microbiol. 128 (1982) 2615–2620.

[19] Huang Z., Dostal L., Rosazza J.P.N., Microbialtransformation of ferulic acid by Saccharomycescerevisiae and Pseudomonas fluorescens,Appl.Environ. Microbiol. 59 (1993) 2244–2250.

[20] Lee Y.W., Jin S., Sim W.S., Nester E.W.,Genetic evidence for direct sensing of pheno-lic compounds by the VirA protein of Agrobac-terium tumefaciens, Proc. Natl. Acad. Sci. USA92 (1995) 12245–12249.

[21] Lesage-Meessen L., Delattre M., Haon M.,Thibault J.F., Ceccaldi B.C., Brunerie P., AstherM., A two-step bioconversion process forvanillin production from ferulic acid combin-ing Aspergillus nigerandPycnoporus cinnabar-inus, J. Biotechnol. 50 (1996) 107–113.

[22] McSweeney C., Dulieu A., Webb R.I., del DotT., Blackall L.L., Isolation and characterizationof a Clostridium sp. with cinnamoyl esteraseactivity and unusual cell envelope ultrastruc-ture, Arch. Microbiol. 172 (1999) 139–149.

[23] Posno M.R., Leer J., van Luik N., van GiezenM.J.F., Heuvelmans P.T.H.M., Lokman B.C.,Pouwels P.H., Incompatibility of Lactobacillusvectors with replicons derived from small cryp-tic Lactobacillusplasmids and segregationalinstability of the introduced vectors, Appl. Env-iron. Microbiol. 57 (1991) 1822–1828.

[24] Priefert H., Rabenhorst J., Steinbuchel A.,Molecular characterization of genes of Pseu-domonas sp. strain HR199 involved in biocon-version of vanillin to protocatechuate, J. Bac-teriol. 179 (1997) 2595–2607.

[25] Ramos-Nino M.E., Clifford M.N., Adams M.R.,Quantitative structure activity relationship forthe effect of benzoic acids, cinnamic acids andbenzaldehydes on Listeria monocytogenes,J. Appl. Bacteriol. 80 (1996) 303–310.

[26] Sambrook J., Fritsch E.F., Maniatis T., Molec-ular cloning: a laboratory manual, 2nd ed., ColdSpring Harbor, NY, USA, 1989.

[27] Sanger F., Nicklen S., Coulson A.R., DNAsequencing with chain-terminating inhibitors,Proc. Natl. Acad. Sci. USA. 74 (1977) 5463–5467.

[28] Segura A., Bünz P.V., D’Argenio D.A., OrnstonL.N., Genetic analysis of a chromosomal regioncontaining vanAand vanB, genes required forconversion of either ferulate or vanillate to pro-tocatechuate in Acinetobacter, J. Bacteriol. 181(1999) 3494–3504.

170

Phenolic acid metabolism in L. plantarum

[32] Yamane K., Kumano M., Kurita K., The 25degrees-36 degrees region of the Bacillus subtilischromosome: determination of the sequence ofa 146 kb segment and identification of 113genes, Microbiology 142 (1996) 3047–3056.

[33]. Zago A., Degrassi G., Bruschi C.V., Cloning,sequencing, and expression in Escherichia coliof the Bacillus pumilusgene for ferulic aciddegradation, Appl. Environ. Microbiol. 61(1995) 4484–4486.

[34]. Zaldivar J., Ingram L.O., Effect of organic acidson the growth and fermentation of ethanolo-genic Escherichia coliLY01, Biotechnol. Bio-eng. 66 (1999) 203–210.

[29] Thurston P.A., Tubb R.S., Screening yeaststrains for their ability to produce phenolic off-flavours: A simple method for determining phe-nols in wort and beer, J. Inst. Brew. 87 (1981)177–179.

[30] Venturi V., Zennaro F., Degrassi G., OkekeB.C., Bruschi C.V., Genetics of ferulic acid bio-conversion to protocatechuic acid in plant-growth-promoting Pseudomonas putidaWCS358, Microbiology 144 (1998) 965–973.

[31] Whiting G.C., Carr J.G., Metabolism of cin-namic acid and hydroxy-cinnamic acids by Lac-tobacillus pastorianusvar. quinicus, Nature 184(1959) 1427–1428.

171

To access this journal online:www.edpsciences.org