INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Microorganisms generally respond to changes in environmental conditions through the actions of specific systems which detect physical or chemical changes and develop coordinated cellular responses to adapt to new conditions. Particularly, microorganisms can resist toxic compounds by various responses which are activated upon exposure to stress. Most of the time, detoxification involves either active efflux of the toxic compound from the cell by highly specific systems (3, 25) or enzymatic conversion of the toxic compound into a less toxic form (32). For some microorganisms, weak acids are considered to be the major natural toxic compounds. At low pH, they strongly inhibit growth by decreasing internal pH (29, 42). Phenolic acids, also called substituted hydroxycinnamic acids, are abundant in the plant kingdom because they are involved in the structure of plant cell walls (19) and are released by hemicellulases produced by several fungi and bacteria (13). Surprisingly, phenolic acids are not potentially toxic to all microorganisms. Some Pseudomonas strains (24, 33), as well as Acinetobacter calcoaceticus (38), are able to use them as the sole source of carbon for growth. They also serve as a signal and induce vir gene expression in the plant-associated Agrobacterium tumefaciens (27, 30). Nevertheless, they display antimicrobial activity against these three bacteria at a concentration above 0.5 mM (27), as well as acting against many other bacteria and fungi (5, 14, 42). Very little is known about the mechanisms evolved by microorganisms to counteract phenolic acid toxicity. Chambel et al. (11) showed that Saccharomyces cerevisiae induced the expression of the H⁺-ATPase pumps in response to inhibitory concentrations of cinnamic acid. In a previous work (4), we demonstrated that the ubiquitous lactic acid bacterium Lactobacillus plantarum exhibits inducible p-coumaric acid decarboxylase (PDC) activity, which converts p-coumaric acid into 4-vinyl phenol, a less toxic compound. We also showed that L. plantarum PDC activity confers a selective advantage for growth in p-coumaric acid-supplemented medium and therefore proposed that PDC synthesis could be considered as a stress response induced by phenolic acid toxicity.

Several microorganisms such as S. cerevisiae (14), Brettanomyces anomalus (21), L. plantarum, and Pediococcus pentosaceus (7) have been reported to decarboxylate phenolic acids into 4-vinyl derivatives, which could then be reduced to 4-ethyl derivatives. These volatile phenols are valuable intermediates in the biotechnological production of new flavor and fragrance chemicals, but they are also regarded as sources of phenolic off-flavors in many beers and wines, due to their characteristic aroma and their low threshold detection (4). To date, only three bacterial phenolic acid decarboxylases (PADs) have been purified, characterized, and cloned: a ferulate decarboxylase (FDC) from Bacillus pumilus (41), a PDC from L. plantarum (8), and a PAD from B. subtilis (10). Although they exhibit 66% amino acid sequence identity, the purified enzymes have different structures, biochemical characteristics, and substrate specificities (10). They also differ from the phenylacrylic decarboxylase of S. cerevisiae (14). Unlike the fungal PADs of S. cerevisiae and B. anomalus, which are constitutively expressed at a low level (about 1 to 10 nmol · min¹ · mg¹) (14, 21), the PADs of L. plantarum and B. subtilis have substrate-inducible decarboxylase activities of about 0.5 µmol · min¹ · mg¹ in the presence of their respective substrates. Transcriptional analyses showed that pdc and pad mRNA could not be detected in uninduced cell extracts, in agreement with the absence of PAD activity in the same extracts. Our results also indicated that expression of these two genes is transcriptionally activated up to 6,000-fold in the presence of phenolic acids (8, 10). These regulatory systems involving phenolic acids which are considered as natural compounds as opposed to classical chemical inducers, could constitute a useful tool for the study of gene expression in lactic acid bacteria and other gram-positive bacteria.

In order to improve our understanding of phenolic acid biodegradation, we have screened bacteria which encounter phenolic acids in their environment and which are able to metabolize these compounds. In the course of our screening, we found a strain of P. pentosaceus, a lactic acid bacterium isolated from wine, which was able to decarboxylate p-coumaric acid and could then be involved in aroma changes and alterations in vegetable fermented products. In this paper, we describe the cloning of the corresponding padA gene encoding a PAD, and we report the first cloning and characterization of a pad transcriptional regulator, named padR, which forms an autoregulated operonic structure with the padA gene.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and culture conditions. P. pentosaceus (strain PP1) was isolated in the laboratory from an aging Pinot noir red Burgundy wine and was identified with the API 50CHL system (BioMérieux, Marcy l'Etoile, France). The strain was grown in MRS medium (17) at 30°C without agitation. Escherichia coli strain TG1 was used as a host for construction of the genomic library and for subsequent cloning steps and was grown aerobically in Luria-Bertani (LB) medium (36) or agar medium at 37°C. Plasmid pTZ19R (35) was used as a vector for the library and for the subcloning steps. Plasmid pJDC9 (12) was used for subcloning steps. When appropriate, ampicillin or erythromycin (100 mg/liter) was added to the medium.

DNA manipulation, sequencing, and computer analysis. Standard molecular procedures described by Sambrook et al. (36) were used. Double-stranded DNA from recombinant plasmids was purified with the Qiagen plasmid kit (Tip 100; Qiagen, Hilden, Germany) and sequenced by the dideoxy chain termination method (37) with the ThermoSequenase cycle sequencing kit (Amersham, Life Science, Inc., Cleveland, Ohio). Both strands were sequenced by using specific synthetic primers (Gibco-BRL, Gaithersburg, Md.). Computer analyses of the sequences were carried out with PC/GENE software (Intelligenetics).

Preparation and screening of the P. pentosaceus genomic library with a pdc-specific probe from L. plantarum. Total DNA from P. pentosaceus was completely digested by HindIII, and the resulting DNA fragments were ligated to HindIII-digested pTZ19R treated with bacterial alkaline phosphatase (Gibco-BRL). The ligation mixture was transferred into E. coli TG1 cells by electroporation, and up to 1,500 recombinant clones containing plasmids with 1- to 7-kb DNA inserts were stored at 70°C in microtitration plates. In order to synthesize a pdc-specific probe, PCR was performed in an automated Hybaid DNA thermocycler by standard procedures with genomic DNA from L. plantarum as the template and the two oligonucleotides LPPDC3 (5'-CACTTGATGACTTTCTCGGCAC-3') and LPPDC8 (5'-CTTCAACCCACTTTGGGAAG-3') (8). The 300-bp PCR product was purified by agarose gel electrophoresis and extraction, by using the Jet-Sorb kit (Genomed, Bioprobe, Montreuil, France). The purified fragment was sequenced to confirm its identity and was radiolabeled with [-³²P]dATP (NEM, Boston, Mass.) by random priming (Gibco-BRL kit). Colony hybridization was carried out at 60°C for 5 h, followed by 5 h at 50°C, using standard procedures as previously described (8). Clones that hybridized with the pdc probe were detected by exposure of the membranes to Kodak BIOMAX MS films.

Isolation of total RNA from P. pentosaceus and from recombinant E. coli strains. For P. pentosaceus, cells were grown in 600 ml of MRS medium to an optical density at 600 nm of 0.7, and the culture was divided in two: a noninduced subculture and an induced subculture to which 2.4 mM p-coumaric acid was added. These cultures were incubated for 120 min at 37°C. During this period, 100-ml samples were quickly removed and refrigerated in ice-water. Total RNAs were extracted and quantified as previously described (8). The RNA integrity was checked by standard denaturing agarose gel electrophoresis. For E. coli recombinant clones, cells were grown in 100 ml of LB medium with the appropriate antibiotic, and 30 ml of culture was treated as described above to obtain total RNA.

Northern blot and primer extension analysis. Total RNAs were separated in denaturing formaldehyde agarose gels and transferred to nylon membranes by using the Pharmacia vacuum system. Hybridization was performed at 60°C with an [-³²P]dATP radiolabeled probe synthesized in a PCR mixture. Determination of the sizes of the transcripts was done by using an RNA ladder (0.24 to 9.5 kb; Gibco-BRL) as the standard. Primer extension analysis was done with two antisense primers, PPPAD5 and PPPAD11, located in the 5' region of the padA gene. Reverse transcription was realized at 42°C with Superscript II reverse transcriptase (RT) (Gibco-BRL) as previously described (10). Three microliters of loading denaturing buffer was added to 3 µl of the reaction mixture. The mixture was denatured at 80°C for 3 min and loaded onto a 6% polyacrylamide gel in parallel with sequencing reactions with the padA DNA as the template and the same primers. For comparison of band intensity, autoradiography results were scanned and digitized. Band intensity was quantified with Bio1D software (Vilber Lourmat).

RT-PCR. To remove any contaminating DNA, 1 µg of total RNA was incubated at room temperature with 1 U of RNase-free DNase I (Gibco-BRL). Residual DNase I was inactivated at 65°C for 10 min. DNase I-treated RNA was subjected to reverse transcription into cDNA with Superscript II RT (Gibco-BRL) as recommended by the manufacturer. Ten percent of the total cDNA was then PCR amplified with Taq DNA polymerase (Appligene) by using the primers PPPAD4 and PP33, and 1/10 of the PCR products were run on a 1% (wt/vol) agarose gel stained with ethidium bromide (0.5 mg/liter). As a control, PCR of DNase-treated RNA was performed with the same primers to check for any DNA contamination.

Preparation of cell extracts, enzyme assays, and protein electrophoresis. Cells of P. pentosaceus and recombinant E. coli strains, grown in MRS and LB medium, respectively, were harvested and washed by centrifugation and then resuspended in phosphate buffer to test PAD activity. Cell extracts were obtained by disrupting concentrated cell suspensions (10 g [dry mass] per liter) with the French press at 1.2 × 10⁸ Pa. PAD activity was assayed by monitoring the disappearance of absorption peaks of the different substrates and the simultaneous appearance of new peaks corresponding to vinyl derivatives as previously described (4, 7, 16). The total protein concentration was determined with a protein assay kit (Bio-Rad, Richmond, Calif.) with bovine serum albumin as the standard, and the specific activity was expressed as micromoles or nanomoles of substrate degraded per minute per milligram of protein. The protein extracts containing PAD activity were resolved by denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12% resolving gel) as previously described (9) with molecular weight markers (SDS-PAGE standards, low range; Bio-Rad) as standards.

Nucleotide sequence accession number. The sequence of the 3,032-bp DNA fragment containing the padA and the padR genes has been deposited in the EMBL nucleotide sequence database under accession no. AJ276891.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of PAD activity in P. pentosaceus. A preliminary experiment revealed that P. pentosaceus was able to decarboxylate p-coumaric acid used to supplement MRS medium. An approximately equimolar concentration of 4-vinyl phenol was found in the growth medium, indicating that this derivative did not accumulate in the cells (data not shown), as previously shown for L. plantarum (8). Resting cells and crude cell extracts were obtained from P. pentosaceus cultures, either induced with 1.2 mM p-coumaric acid or uninduced, and were tested for PAD activity on p-coumaric acid. No PAD activity was detected in the uninduced cells and corresponding cell extracts. Induced cells displayed decarboxylase activity on p-coumaric acid, whereas no detectable activity could be found in the corresponding cell extract. Since we previously showed that the L. plantarum PDC enzyme required ammonium sulfate or NaCl to decarboxylate ferulic acid in vitro (4), 200 g of ammonium sulfate per liter was added to induced cell extracts of P. pentosaceus, which proved able to restore PAD activity.

Cloning and sequencing of the PAD gene from P. pentosaceus. Based on the high degree of conservation among known PADs (8, 10, 41) and because P. pentosaceus is phylogenetically closed to L. plantarum (18), a rapid strategy was adopted to test whether a DNA probe from L. plantarum pdc could be used to screen a P. pentosaceus genomic library. A preliminary Southern hybridization experiment at 50°C showed that a DNA probe encompassing the first 300 bp of the L. plantarum pdc gene hybridized weakly but specifically with one band of about 6 kb from P. pentosaceus total DNA digested with HindIII (data not shown). The same probe was used to screen a P. pentosaceus genomic library in E. coli. One clone from the genomic library clearly hybridized with the pdc probe and was designated TG1(pTPADP1). Genetic and biochemical characterization indicated that pTPADP1 contained a 6-kb-long insert which conferred a PAD activity of 100 nmol · min¹ · mg¹ on p-coumaric acid. Various subfragments of the full-length insert were subcloned in pTZ19R to localize the padA gene by measuring the PAD activity of each subclone (Fig. 1). The results indicated that the PAD-encoding gene was likely interrupted by the EcoRI site located at the junction between TG1(pTPADP4) and TG1(pTPADP5) subclones (Fig. 1).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Physical map of the padAR locus and delineation of subcloned fragments. The parent plasmid pTPADP1 was isolated from the P. pentosaceus genomic library. Restriction sites and primers (small horizontal arrows) used to obtain the different subclones are shown. The ORFs identified by sequencing are indicated by large arrows. PAD activity measured on p-coumaric acid in E. coli is indicated to the right of each subclone (U, micromoles of p-coumaric acid degraded per minute).

View larger version (54K): [in this window] [in a new window]   FIG. 2.   Nucleotide and deduced amino acid sequences of the pad cluster. The localization and orientation of primers PPPAD5, PPPAD11, and PP91, used to identify the transcriptional start site of padA and padR, and primers PPPAD4 and PP33, used to amplify cDNA from padA mRNA, are indicated by horizontal arrows. The transcriptional start site of the padAR operon, determined by primer extension analysis, is indicated by a vertical arrow, and the corresponding 10 and 35 boxes are underlined. The putative ribosome binding sites (RBS) are shaded. Stop codons are indicated by asterisks. The two convergent arrows located under the sequence indicate the putative rho-independent transcriptional terminator of the padAR operon. The dotted convergent arrows indicate the region of dyad symmetry (inverted repeats), which could be the target of the PadR repressor.

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Comparison of the deduced amino acid sequence of the padA gene of P. pentosaceus (PAD-PP) with the sequences of B. pumilus FDC (FDC-BP; accession no. X84815), L. plantarum PDC (PDC-LP; accession no. U63827), and B. subtilis PAD (PAD-BS; accession no. AF017117). The sequences were aligned by using the Clustal program. Identical and similar amino acids are indicated by asterisks and dots, respectively. Conserved boxes are shaded. The numbers on the right correspond to the amino acid position in the protein sequence.

The padA gene belongs to an operon which is transcriptionally regulated by phenolic acids. Northern blot hybridization experiments were undertaken in order to study the transcriptional regulation of the padA gene. RNA samples from uninduced and p-coumaric-induced cultures were prepared to determine the size and level of the padA mRNA at different sampling times after addition of the inducer (Fig. 4A). No transcript was detected in the lane corresponding to the mRNA extracted from uninduced cells. A single transcript of approximately 1,200 nt was detected in the RNA extracted from cells induced with p-coumaric acid and was maximal after 10 min of induction. The level of padA transcript had significantly decreased after 40 min of incubation and was no longer detectable after 2 h. At these two later time points, the inducer was totally degraded. Primer extension experiments were performed with primers PPPAD5 (Fig. 4B) and PPPAD11 (data not shown), using RNA from p-coumaric-induced cells harvested after 10 min of induction. Identical results were obtained with both primers, allowing the identification of G residue, located 42 nt upstream from the start codon, as the transcription start site (Fig. 4B). No primer extension product was detected with RNA from the uninduced culture as the template (Fig. 4B). The size of the padA transcript of approximately 1,200 bp indicated that padA and padR appeared to be transcribed as a single transcription unit (Fig. 2). To confirm this hypothetical operonic organization, RT-PCR was carried out with mRNA prepared from P. pentosaceus grown with and without p-coumaric acid (2.4 mM) by using the primers PPPAD4 and PP33, located within the padA and padR genes, respectively (Fig. 2). A PCR product of the expected size (619 bp) (Fig. 5) and sequence (data not shown) was obtained with mRNA from an induced culture as the template, supporting the operonic arrangement of padA and padR. A very weak amplification product was detected with RNA from uninduced cells, confirming substrate-mediated regulation of the operon. No PCR products were detected in control reactions that were designed to detect chromosomal DNA contamination. To further confirm the operonic organization of the two genes, primer extension experiments were performed with total RNA from uninduced and induced P. pentosaceus cells with antisense primer PP91 internal to padR (Fig. 2). No primer extension product was generated when RNA from uninduced cells was used as the template. With RNA from induced cells, primer extension generated a band which identified a transcriptional start site identical to that of padA (data not shown). Taken together, these results clearly indicate that padA and padR are transcribed as a bicistronic unit and are subjected to transcriptional regulation which involves substrate-mediated induction.

View larger version (89K): [in this window] [in a new window]   FIG. 4.   Transcriptional analysis of the padAR operon. (A) Northern blot analysis with a padA-specific probe of total RNA purified from P. pentosaceus cells harvested after 0 min (lane 1), 5 min (lane 2), 10 min (lane 3), 20 min (lane 4), 40 min (lane 5), and 120 min (lane 6) following the addition of 2.4 mM p-coumaric acid. The arrow indicates the position of the transcript, and molecular size markers are given in the left lane. (B) Mapping of the 5' end of the padA mRNA by primer extension analysis using primer PPPAD5 with total P. pentosaceus RNA from uninduced cells (NI) and cells induced with 2.4 mM p-coumaric acid (I). The products of reverse transcription were loaded in parallel with DNA sequencing reaction mixtures (lanes A, C, G, and T) initiated with the same primer on padA DNA template. The sequence shown to the left is the complementary strand, and the 5' end of the padA mRNA is indicated by an arrow.

View larger version (49K): [in this window] [in a new window]   FIG. 5.   RT-PCR of the padAR region using primers PPPAD4 and PP33 with P. pentosaceus total RNA purified from uninduced cells (NI) and cells induced with 2.4 mM p-coumaric acid (I). Negative controls with no RT (RT) included are shown on the left. Classical PCR using the same primers and with P. pentosaceus chromosomal DNA added as a positive control is shown in lane C. The 100-bp DNA Ladder Plus (MBI Fermentas, Amherst, Mass.) was used as a molecular weight marker (L).

Expression of PAD activity in E. coli is inhibited by the product or products of one or more genes present on pTPADP1. E. coli TG1(pTPADP1) conferred a PAD activity of 0.1 µmol · min¹ · mg¹ on p-coumaric acid, regardless of the presence of phenolic acid in the culture medium (data not shown). No PAD activity was found in the control strain E. coli TG1(pTZ19R). In addition, SDS-PAGE analysis of cell extracts from E. coli TG1(pTPADP1) and a control strain gave identical profiles (data not shown). These results clearly distinguish the P. pentosaceus recombinant PAD from the recombinant PDC of L. plantarum, which displays an inducible activity of 10 µmol · min¹ · mg¹ on p-coumaric acid, and the recombinant PAD from B. subtilis, which metabolizes p-coumaric, ferulic, and caffeic acids with a specific activity of about 2.5 µmol · min¹ · mg¹ in E. coli. Further hypotheses could explain the apparent low and constitutive activity conferred by the P. pentosaceus padA gene in E. coli: the padA promoter is poorly recognized by the E. coli RNA polymerase, the mRNA is unstable, the translation is low, the enzyme is unstable, and the regulation of PAD activity is somewhat altered in E. coli. To investigate the role of padR, plasmid pJPADP6 was constructed, which contained a 970-bp insert encompassing the padA gene and its promoter region, including 280 bp upstream of the 5' end of the padA transcript (Fig. 1). Cell extracts of E. coli TG1(pJPADP6) and TG1(pJDC9) were prepared and analyzed by SDS-PAGE. A strong protein band of 25 kDa was detected in the recombinant E. coli strain carrying the padA gene and was absent in the control (Fig. 6). The observed protein band, which has a molecular mass corresponding to that of the deduced protein from the padA gene, correlated with a high PAD specific activity of 50 µmol · min¹ · mg¹ on p-coumaric acid, measured in cell extracts of TG1(pJPADP6). This activity was 500-fold higher than the maximal PAD activity found in E. coli TG1(pTPADP1) carrying the original insert. Our results indicate that the 6-kb pTPADP1 insert contains one or more genes involved in regulating the expression of PAD activity.

View larger version (73K): [in this window] [in a new window]   FIG. 6.   SDS-PAGE of crude cell extracts from E. coli TG1 carrying various subclones of the padAR locus. Lanes: 1, molecular mass standards (SDS-PAGE standards; Bio-Rad); 2, crude extract from E. coli TG1(pJDC9); 3, crude extract from E. coli TG1(pJPADP6); 4, crude extract from E. coli TG1(pJPADP7); 5, crude extract from E. coli TG1(pJPADP8); 6, crude extract from E. coli TG1(pJPADP9). Molecular size markers are indicated on the left.

padR encodes a transcriptional repressor of the pad operon. Two plasmids, pJPADP7 and pJPADP8, were constructed by amplifying pTPADP3 DNA as the template with PAD9 and U-primer, and PAD8 and R-primer, respectively (Fig. 1). Cell extracts from E. coli TG1(pJPADP8) exhibited a low PAD activity similar to that of E. coli TG1(pTPADP3), while E. coli TG1(pJPADP7) had a PAD activity 500-fold higher, identical to that of E. coli TG1(pJPADP6) (Fig. 1). Moreover, the PAD protein was detected by SDS-PAGE analysis in E. coli TG1(pJPADP7) cell extracts, but not in the cell extract from E. coli TG1(pJPADP8) (Fig. 6), indicating that the region downstream from padA was responsible for the low PAD activity. Plasmid pJPADP9 containing padA and padR was constructed (Fig. 1), and E. coli TG1(pJPADP9) cell extracts were analyzed by SDS-PAGE (Fig. 6) and PAD activity measurement (Fig. 1). The results indicate that PadR is responsible for the low PAD activity in E. coli.

View larger version (72K): [in this window] [in a new window]   FIG. 7.   Mapping of the 5' end of padA mRNA by primer extension with primer PPPAD5 with total RNA from E. coli TG1(pJPADP9) (lane 1) and TG1(pJPADP6) (lane 2). The RT product from E. coli TG1(pJPADP6) was diluted 20-fold before loading. The products of reverse transcription reactions were loaded in parallel with DNA sequencing reaction mixtures (lanes A, C, G, and T) initiated with the same primer on the padA DNA template. The arrow indicates the 5' end of the padA mRNA.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our studies indicate that the P. pentosaceus PadR protein represses transcription of the padAR operon, thereby regulating its own synthesis. Recently, we have cloned by random insertional mutagenesis, the padR gene from B. subtilis (J.-F. Cavin, V. Dartois, and C. Diviès, unpublished data), which corresponds to a gene named yfiO in the B. subtilis genome sequence (28) and the deduced product of which displays the highest amino acid identity with PadR (37%). This last result is consistent with our finding and supports our results concerning the function of the PadR protein in P. pentosaceus.

To our knowledge, PadR is not a member of any known class of transcriptional regulators, but a search in public databases revealed significant homology with four different proteins, suggesting that we have identified a new class of bacterial regulatory proteins. One of them, RVI176C of Mycobacterium tuberculosis, is a protein of unknown function which displays 27% identity with PadR (15). The three other proteins have been identified as potential transcriptional regulators. OrfA of Listeria monocytogenes (27% amino acid identity with PadR) has been described as a putative negative regulator of the hly gene, coding for the pore-forming listeriolysin O, implicated in L. monocytogenes pathogenicity (26). The AphA protein of Vibrio cholerae (27% amino acid identity with PadR) is required for the expression of the ToxR virulence regulon and plays a role in the regulatory cascade that activates expression of the tcpPH operon (39).

Although the PADs of P. pentosaceus, B. subtilis, and L. plantarum (i) display a high amino acid identity of approximately 70%, (ii) exhibit similar activity levels on their respective substrates, and (iii) are all transcriptionally regulated, the genetic organization and regulation of their genes are distinct. The padAR genes of P. pentosaceus are transcribed as an operon, while the padA genes of L. plantarum and B. subtilis are monocistronic (8, 10). However, the padR gene of B. subtilis is not located in the vicinity of the padA gene, because 2,622 kbp separates the two genes (28). Identification of the L. plantarum padR gene is currently in progress in our laboratory. In B. pumilus, FDC expression was shown to be substrate regulated (16), but elucidation of fdc transcriptional regulation awaits further analyses. Interestingly, analysis of the nucleotide sequence of the fdc region (accession no. X84815) revealed a divergently transcribed 575-bp ORF located upstream from fdc and coding for a putative 185-amino-acid polypeptide which displays 57% identity with B. subtilis PadR and 36% identity with P. pentosaceus PadR.

The classical mode of action of known repressors, such as LacI in E. coli (1), involves binding as a dimer or a tetramer to specific DNA sequences which exhibit dyad symmetry. Analysis of the P. pentosaceus padAR promoter region revealed the existence of a perfect inverted repeat, TTTATGTTG-4N-CAACATAAA, which could be the target site for PadR binding. Interestingly, this motif was also partially found in the promoter region of all three padA genes from L. plantarum, B. subtilis, and B. pumilus (Fig. 8), where it is systematically located downstream from the 10 box near the transcription start site (8, 10, 41). It has been demonstrated that the position of the operator site within a promoter sequence determines the repression efficiency and that repressors which bind over the 10 box and start site occlude the most critical initiation region (34). Site-targeted modification of this conserved motif in the P. pentosaceus padAR promoter, combined with mobility gel shift assays and footprinting experiments using purified PadR, will be performed to identify the DNA binding site of PadR.

View larger version (19K): [in this window] [in a new window]   FIG. 8.   Comparison of the promoter sequence from the padAR operon of P. pentosaceus (pad-PP) with the promoter sequences of the L. plantarum pdc gene (pdc-LP), B. subtilis pad gene (pad-BS), and B. pumilus fdc gene (fdc-BP). The putative 10 boxes are underlined. The transcription start sites are indicated by a vertical arrow, when determined. Convergent arrows indicate regions of dyad symmetry, and the conserved motif is highlighted in boldface.

Inactivation of a repressor often involves a conformational change due to the direct binding of the inducer (2, 6). Although PAD activity was clearly induced by p-coumaric or ferulic acid in P. pentosaceus, addition of 3 mM p-coumaric or ferulic acid in a culture of E. coli TG1(pJPADP1), which displayed a low level of PAD, did not increase the level of decarboxylase activity. These results indicate either that phenolic acids remained unable to induce a conformational change of PadR in E. coli, or that phenolic acids act through an additional effector that is absent in E. coli. Such an effector could consist of a two-component system containing a sensor protein kinase, which would detect phenolic acids and activate a response regulator. Most of the time, the response regulator is a transcriptional activator (22). In some instances, however, information was shown to be transduced to a transcriptional repressor, as was demonstrated for LuxO in Vibrio harveyi, which is the final acceptor of a phosphorelay cascade and represses transcription of the lux operon in its phosphorylated form (23). Interestingly, only two systems induced by phenolic acids have been described in the literature: the two-component regulatory system involved in vir gene expression in A. tumefaciens and consisting of the VirA and VirG proteins (31), and the transcriptional activator PobR from A. calcoaceticus involved in the expression of pobA, the structural gene for p-hydroxybenzoate hydroxylase (20).

Taken together, our results led us to propose a model for the regulation of the padAR operon in P. pentosaceus (Fig. 9). In the absence of phenolic acid in the medium, the PadR repressor binds to the region of dyad symmetry within the padAR promoter (Fig. 8). This prevents transcription of the padAR operon, leading to little or no PAD synthesis and to background levels of PadR production. When phenolic acids are present in the medium, they induce inactivation of the PadR repressor by a mechanism which remains to be characterized. This allows transcription of the padAR operon and results in PAD and PadR synthesis. Thus, the PadR repressor remains in its inactive form as long as all available phenolic acids have not been converted by the PAD enzyme into 4-vinyl derivatives, compounds which are less toxic to lactic acid bacteria (4). When all available phenolic acid is degraded, PadR switches to the active form and turns off transcription of the padAR operon. The lactic acid bacterium P. pentosaceus seems to have evolved an efficient system to detoxify phenolic acids into 4-vinyl derivatives by organizing padA, which converts the toxic compound, and padR, which regulates PAD activity, into a single transcriptional unit. Genetic and biochemical studies are currently in progress in order to elucidate the mechanism of PadR inactivation by phenolic acids.

View larger version (43K): [in this window] [in a new window]   FIG. 9.   Model for the transcriptional regulation of the padAR operon in P. pentosaceus. (A) In the absence of phenolic acid, the level of pad transcripts is low and could only be detected by RT-PCR, while PAD activity remained undetectable. (B) Addition of phenolic acid causes inactivation of the PadR repressor, possibly through an additional effector or sensor which mediates the conformational change or modulation of PadR. This allows transcription of the padAR operon and synthesis of the PAD enzyme. Toxic phenolic acids are decarboxylated into vinyl derivatives, which are less toxic and can diffuse outside the cell. Exhaustion of phenolic acids results in PadR switching to its active form and repressing padAR transcription. For p-coumaric acid, R1 = OH and R2 = H.