Analysis of three clustered polygalacturonase genes in Erwinia chrysanthemi 3937 revealed an anti-repressor function for the PecS regulator

Authors

  • William Nasser,

    1. Unité Microbiologie et Génétique — composante INSA, UMR INSA-UCB-CNRS 5577, Bat 406, INSA, 20, Avenue Albert Einstein, F-69621 Villeurbanne Cedex, France
    Search for more papers by this author
  • Vladimir E. Shevchik,

    1. Unité Microbiologie et Génétique — composante INSA, UMR INSA-UCB-CNRS 5577, Bat 406, INSA, 20, Avenue Albert Einstein, F-69621 Villeurbanne Cedex, France
    Search for more papers by this author
  • Nicole Hugouvieux-Cotte-Pattat

    1. Unité Microbiologie et Génétique — composante INSA, UMR INSA-UCB-CNRS 5577, Bat 406, INSA, 20, Avenue Albert Einstein, F-69621 Villeurbanne Cedex, France
    Search for more papers by this author

Nicole Hugouvieux-Cotte-Pattat. E-mail cotte@insa.insa-lyon.fr; Tel. (+33) 472 43 80 88; Fax (+33) 472 43 87 14.

Abstract

Erwinia chrysanthemi 3937 secretes an arsenal of pectinolytic enzymes including several pectate lyases encoded by the pel genes. We characterized a novel cluster of pectinolytic genes consisting of the three adjacent genes pehV, pehW and pehX, whose products have polygalacturonase activity. The high similarity between the three genes suggests that they result from duplication of an ancestral gene. The transcription of pehV, pehW and pehX is dependent on several environmental conditions. They are induced by pectin catabolic products and this induction results from inactivation of the KdgR repressor which controls almost all the steps of pectin catabolism. The presence of calcium ions strongly reduced the transcription of the three peh genes. Their expression was also affected by growth phase, osmolarity, oxygen limitation and nitrogen starvation. In addition, the pehX transcription is affected by catabolite repression and controlled by the activator protein CRP. PecS, which was initially isolated as a repressor of virulence factors, acts as an activator of the peh transcription. We showed that the three regulators KdgR, PecS and CRP act by direct interaction with the promoter regions of the peh genes. Analysis of simultaneous binding of KdgR, PecS, CRP and RNA polymerase indicated that the activator effect of PecS results from a competition between PecS and KdgR for the occupation of overlapping binding sites. Thus, to activate peh transcription, PecS behaves as an anti-repressor against KdgR.

Introduction

The enterobacterium Erwinia chrysanthemi causes soft-rot disease of various plants. The maceration process involves the depolymerization of the pectin in plant cell walls. To degrade pectin, plant pathogenic microorganisms produce a variety of enzymes. Pectin esterases facilitate the subsequent action of two types of depolymerases: pectate lyases, which cleave the glycosidic bond by β-elimination, and polygalacturonases, which catalyse a hydrolytic cleavage. Depolymerases may also differ in the random or terminal attack of the polymer (endo- or exocleaving enzymes respectively). E. chrysanthemi strain 3937 produces the two pectin methylesterases PemA and PemB and the pectin acetylesterase PaeY (Shevchik et al., 1996; Shevchik and Hugouvieux-Cotte-Pattat, 1997). This strain also secretes multiple isoenzymes of endopectate lyases (PelA, PelB, PelC, PelD, PelE, PelI, PelL and PelZ), most of which play an important role in the soft-rot disease (Barras et al., 1994; Lojkowska et al., 1995; Pissavin et al., 1996; Shevchik et al., 1997). The genes encoding these pectinases are located at different loci on the bacterial chromosome, most of them are clustered (pelA–pelE–pelD–paeY–pemA and pelB–pelC–pelZ) but a few genes appear to be isolated (pelI, pelL or pemB). The only hydrolase activity previously characterized in E. chrysanthemi is an exocleaving polygalacturonase that liberates dimers, the exopoly-α-d-galacturonosidase PehX of strain EC16 (He and Collmer, 1990).

The transcription of the E. chrysanthemi pectinase genes is modulated by various environmental conditions, notably induction in the presence of pectin at high cell density and catabolite repression (Hugouvieux-Cotte-Pattat et al., 1996). Some genes are also affected by temperature, nitrogen starvation, anaerobiosis, osmolarity and iron availability (Hugouvieux-Cotte-Pattat et al., 1992; Sauvage and Expert, 1994). Different regulators controlling the transcription of the pectinase genes were characterized. Induction in the presence of pectin results from inactivation of the KdgR repressor after intracellular formation of pectin catabolites (Nasser et al., 1994; Hugouvieux-Cotte-Pattat et al., 1996). PecS, a member of the MarR family of regulators, acts as a transcriptional repressor of different virulence factors, including the pel genes (Reverchon et al., 1994; Praillet et al., 1996). PecT, a member of the LysR family, also represses the expression of some pel genes (Surgey et al., 1996). Until now, the signals triggering PecS or PecT controls have not been identified. The CRP protein, involved in the global control of sugar catabolism in response to the cAMP level, is the main activator of the pel transcription (Reverchon et al., 1997). The ExpR regulator is one of the elements mediating the cell density-dependent regulation of the pel genes (Nasser et al., 1998).

In this paper, we present the characterization of a cluster of three homologous genes in E. chrysanthemi 3937, pehV, pehW and pehX, whose products have polygalacturonase activity. We analysed the individual expression of these genes using transcriptional fusions. We tested the direct interaction of the regulators KdgR, PecS and CRP with the promoter regions of the peh genes. In particular, we analysed the role of PecS, which is a repressor of the pel transcription but an activator of the peh transcription.

Results

Characterization of the pehV–pehW–pehX cluster

Using a genomic library of strain PMV4116, a 3937 derivative with the pelA–pelE–pelD–paeY and pelB–pelC–pelZ gene clusters deleted, we selected three E. coli NM522 transformants that had a weak pectinolytic activity on the pectate agar medium but no pectate lyase activity (Shevchik et al., 1999). Subcloning and sequencing revealed that these plasmids (pPN2, pPN7 and pPN12; Fig. 1) contain the pehX gene, previously identified in the E. chrysanthemi strain EC16 and encoding an exopoly-α-d-galacturonosidase (He and Collmer, 1990). The pehX gene is preceded and followed by partial potential ORFs. As shown in strain EC16, the ORF situated on the 3′ side of pehX and encoded on the other DNA strand is homologous to the rep gene of Escherichia coli, encoding the Rep helicase, a single-stranded DNA-dependent ATPase required for replication of some phages (Daniels et al., 1992). In addition, the partial ORF situated on the 5′ side of pehX and encoded on the same DNA strand is homologous to the pehX gene itself. This gene was designated pehW.

Figure 1.

. Identification of the E. chrysanthemi 3937 pehV–pehW–pehX cluster. The different subclones are indicated and the restriction map of the whole region is given below. The arrows indicate the position and the transcription direction of the genes. The flags show the sites of insertion of the uidA-Km cassettes. The thick lanes indicate the DNA fragments used as peh promoter regions for in vitro experiments.

To isolate the complete pehW gene, we inserted a uidA-Km cassette into the NheI site in pehX (Fig. 1). After recombination of this insertion into the 3937 chromosome, we selected an R-prime plasmid bearing the uidA-Km insertion and adjacent chromosomal DNA. Subcloning enabled us to isolate a 3.4 kb EcoRI/NheI fragment (pACB5; Fig. 1) encoding a weak pectinolytic activity, as detected on pectate agar medium. Sequencing of this DNA insert showed that it contains the complete pehW gene. DNA homology searches revealed that the ORF situated on the pehW 5′ side and encoded on the same DNA strand is, again, homologous to the pehW and pehX genes. This gene was designated pehV.

A similar approach was used to isolate the complete pehV gene. A uidA-Km cassette was inserted into the HindIII site in pehW (Fig. 1) and, after recombination into the 3937 chromosome, we selected an R-prime plasmid bearing the region adjacent to the uidA-Km insertion. Subcloning led to the isolation of the 3.5 kb PstI/HindIII fragment (pJH32; Fig. 1). Sequencing of this DNA insert showed that it contains the complete pehV gene. The ORF situated on the pehV 5′ side and encoded on the same DNA strand is homologous to the gppA gene of E. coli, encoding the guanosine-5′-triphosphate,3′-diphosphate pyrophosphatase (Daniels et al., 1992).

Analysis of the pehV–pehW–pehX sequence and of the encoded proteins

The nucleotide sequence of the 7.4 kb PstI/MunI DNA fragment (Fig. 1) was determined (accession no. AJ132326). The gppA stop codon TGA is present at position 552. The rep stop codon TGA is present at position 7014. The three peh genes begin with ATG and end with TGA codons, which are, respectively, situated at positions 771 and 2589 for pehV, 2971 and 4789 for pehW, and 5156 and 6962 for pehX. The ATG starts of pehV, pehW and pehX are preceded (six nucleotides in each case), respectively, by the potential ribosome-binding sites GAGGA, AAGGA and AAGGA (Fig. 2). Each peh gene is separated from the adjacent genes by GC-rich inverted repeats followed by runs of T residues that might be involved in transcription termination (Fig. 2).

Figure 2.

. The pehV, pehW and pehX promoter regions of E. chrysanthemi 3937. A vertical line indicates identical bases. The dotted arrows indicate the inverted repeats corresponding to potential transcription terminators. The sequences underlined correspond either to the translational start codons with the Shine–Dalgarno sequences (in black) or to the nucleotides conserved in the consensus of the binding sites of KdgR (in blue or violet) and CRP (in green). The transcriptional start sites determined by S1 mapping, with the corresponding −10 and −35 promoter sites, are shown in red. The regions protected in DNase footprinting by the binding of KdgR and PecS are indicated in blue and pink, respectively, with overlapping areas in violet.

According to their nucleotide sequence, pehV, pehW and pehX encode proteins of 606, 606 and 602 amino acids, respectively, with calculated molecular masses of 65 014, 64 554 and 64 217 Da, respectively, including potential amino-terminal signal sequences. The 27-amino-acid signal sequence of PehX from the E. chrysanthemi strain EC16, which is 90% identical to PehX of strain 3937, was confirmed by sequencing the N-terminus of the mature protein (He and Collmer, 1990). Comparison of the protein sequences indicated that there is 67% identity between PehV and PehX, 70% between PehV and PehW, and 83% between PehW and PehX. PehV, PehW and PehX also showed similarity with a large number of polygalacturonases from bacterial, fungal or plant origin belonging to family 28 of the glycosyl hydrolases (Davies and Henrissat, 1995).

The pehV, pehW and pehX genes were independently cloned into pT7 vectors and overexpressed in E. coli BL21(DE3) (Studier and Moffat, 1986). Analysis by SDS–PAGE revealed the presence of overproduced proteins of about 62 kDa in the periplasmic space of E. coli (data not shown), in agreement with the molecular mass deduced from the nucleotide sequence and the presence of a signal sequence. Polygalacturonase assays revealed that the periplasmic extracts containing the overproduced proteins show activities of 1, 5 and 45 μmol min−1 ml−1 for PehV, PehW and PehX extracts respectively.

It was previously demonstrated that PehX liberates digalacturonate from polygalacturonate by attacking the non-reducing end of the substrate (He and Collmer, 1990; Shevchik et al., 1999). We used thin layer chromatography to characterize the reaction end-products of each Peh. Both PehW and PehX catalyse the formation of only one product, corresponding to digalacturonate (data not shown). This product is specific to exodepolymerases cleaving at the penultimate glycosidic bonds. Thus, like PehX, PehW is an exopoly-α-d-galacturonosidase. The activity of PehV was too weak under the conditions used for detection after thin layer chromatography to clearly deduce its mode of cleavage.

We analysed the ability of PehV, PehW and PehX to macerate plant tissues by incubating the enzyme preparations with potato cubes. Equivalent activities (1 μmol min−1 ml−1) of each enzyme caused no visible maceration after 24 h.

Analysis of the virulence and growth of aΔpehV–pehW–pehX mutant

To analyse the role of the pehV, pehW and pehX cluster, a deletion of these three adjacent genes was constructed by replacement of a 5.2 kb KpnI/SacII fragment (Fig. 1) with a uidA-Km cassette. After recombination of this insertion into the bacterial chromosome, we compared the capacity of maceration of the ΔpehV–pehW–pehX mutant A3418 with that of the parental strain A350 on chicory leaves and potato tubers. On chicory leaves, the rotted regions observed 24 h after inoculation measured 40 ± 11 mm with the strain A350 and 28 ± 7 mm with the strain A3418. Two days after inoculation of potato tubers, the weight of macerated tissues was 577 ± 63 mg with the strain A3418 and 663 ± 84 mg with the strain A350. Thus, only a weak reduction of macerating capacity was observed after deletion of the pehV, pehW and pehX cluster.

We also compared the pathogenic behaviour of the ΔpehV–pehW–pehX mutant on potted plants of Saintpaulia ionantha with that of the wild-type strain. After infecting 30 plants with each strain, we followed the appearance of symptoms over 10 days. We observed no significant difference in the progress of the disease between the mutant strain and the wild type (data not shown).

We also compared the growth of the ΔpehV–pehW–pehX mutant A3418 with that of the parental strain A350 when polygalacturonate was used as the sole carbon source. For both strains, the final yield of growth was about 109 bacteria per milligram of polygalacturonate and the doubling time during exponential growth was ≈ 120 min. The sole difference between the two strains concerned the lag phases when precultures were grown in the presence of another carbon source (glycerol or glucose). Whereas the lag phase was about 2 h for A350, it increased to 6 h for A3418. Thus, deletion of the pehV–pehW–pehX cluster only affects the commencement of polygalacturonate catabolism. As previously suggested (He and Collmer, 1990), depolymerases liberating low oligomers, such as the exopoly-α-d-galacturonosidases which liberate dimers, could hasten the formation of intracellular inducers of pectate lyase synthesis.

Expression of the pehV, pehW and pehX genes

pehV::uidA, pehW::uidA and pehX::uidA transcriptional fusions were constructed by insertion of an uidA-Km cassette (Bardonnet and Blanco, 1992) in the corresponding ORFs (Fig. 1). After recombination into the E. chrysanthemi chromosome, the expression of the fusions was monitored during bacterial growth. When the cells entered the late exponential growth phase, the expression of pehV, pehW and pehX increased three- to fivefold. This increase coincided with pectate lyase production (data not shown).

The expression of the pehV::uidA, pehW::uidA and pehX::uidA fusions was analysed after growth under various conditions (Table 1). In minimal medium containing glycerol as the carbon source, pehV::uidA and pehX::uidA showed a significant basal level of expression, whereas that of pehW::uidA was very low. In medium containing polygalacturonate as the carbon source, the transcription of pehV, pehW and pehX was stimulated threefold, 15-fold and 10-fold respectively. In the presence of a readily utilizable carbon source, such as glucose, a threefold decrease was observed for pehX transcription whereas no significant variation was recorded for pehV and pehW. Oxygen limitation weakly increased transcription of the three genes. Increasing the medium osmolarity reduced their transcription. Under conditions of nitrogen starvation, expression of pehV, pehW and pehX was strongly inhibited (Table 1). Variation of the growth temperature (25°C, 30°C and 37°C) did not significantly affect peh expression (data not shown). As described for the Erwinia carotovora polygalacturonase gene (Flego et al., 1997), addition of CaCl2 reduced pehV, pehW and pehX expression in E. chrysanthemi (Table 1). However, this reduction was observed only when bacteria were grown in medium containing polygalacturonate but not with other carbon sources (glycerol, glucose or galacturonate). This effect was visible at very low CaCl2 concentrations but was maximal for concentrations higher than 0.2 mM (data not shown).

Table 1. . Expression of transcriptional fusions in peh genes of Erwinia chrysanthemi 3937. The different compounds were added at the following concentrations: glycerol and glucose, 2 g l−1; polygalacturonate 2.5 g l−1; NaCl, 0.3 M; and CaCl2, 0.2 mM. The results reported are the means of at least three independent experiments with standard deviations corresponding to less than 20%, except for very low activities (< 10 and < 0.05 for β-glucuronidase and pectate lyase activities respectively), for which standard deviations could reach 30%.Thumbnail image of

The gene fusions were transduced into strains mutated in one of the regulatory genes kdgR, pecS, pecT, expR or crp. Inactivation of the kdgR gene strongly increased the expression of the fusions in pehV, pehW and pehX (Table 1). Although the pecS mutation increased pectate lyase production, it decreased the expression of pehV, pehW and pehX (Table 1). The crp mutation decreased expression of the pehX fusion twofold without affecting pehV and pehW (Table 1). In contrast, mutations in pecT or expR did not affect expression of the peh fusions (data not shown). Because in vitro analysis (see below) suggested interference between PecS and KdgR action, we analysed pehX transcription in the strain containing the two regulatory mutations. In the pecS kdgR double mutant, the level of pehX expression was similar to that obtained in the kdgR single mutant (Table 1). Thus, the decreased pehX transcription due to the pecS mutation is abolished in a kdgR mutant. In contrast, the derepression of pehX transcription due to the kdgR mutation is conserved in a pecS mutant.

Identification of the KdgR, CRP and PecS binding sites

We first determined the peh transcriptional start sites by S1 mapping experiments. Because the level of peh transcription is low in the 3937 wild-type strain, mRNA was extracted from the E. chrysanthemi kdgR mutant A1077, which exhibits increased peh transcription (Table 1), and from E. coli recombinant clones containing each peh gene. A single transcription start site was observed in each case (data not shown). The 5′ end of the pehX mRNA was the same for mRNA extracted from E. chrysanthemi or from E. coli. The 5′ ends of the pehV and pehW mRNA could be detected only with mRNA extracted from E. coli. Regions homologous to the −10 and −35 elements of the σ70 promoter were detected upstream from these transcriptional sites (Fig. 2). Although the regulatory regions of pehV, pehW and pehX are similar, the three promoters are situated at different positions relative to the ATG codon and are located in weakly conserved regions.

Gel shift assays and DNase I footprinting were used to study interactions between the promoter regions of the peh genes and the KdgR, PecS and CRP proteins (Figs 3 and 4 for pehX ). KdgR specifically interacted with the regulatory regions of pehV, pehW and pehX and a single KdgR–DNA complex was observed at subsaturating KdgR concentrations (Fig. 3A). The KdgR-binding affinity was similar for all three genes (Kd, 30 nM). The location of the KdgR binding site was determined by footprinting (Fig. 4A, lanes 6 and 7; Fig. 4B lanes 15 and 16). For each gene, the protected area encompassed a sequence homologous to the consensus of the KdgR box previously determined (AATRAAAYRNYRTTTYATT, with R = G or A, and Y = C or T) (Nasser et al., 1994) and overlaps the promoters (Fig. 2).

Figure 3.

. Gel shift experiments using the pehX promoter region and the regulators KdgR and CRP (A) or PecS and KdgR (B). The indicated amounts of purified KdgR, CRP and/or PecS proteins were incubated with the end-labelled 0.50 kb BamHI/NheI fragment corresponding to the 5′ untranslated end of pehX.

Figure 4.

. DNAse I footprint of the pehX promoter region with KdgR, PecS, CRP and RNA polymerase. A. Protection by PecS and/or KdgR of the 0.50 kb BamHI/NheI fragment. Lane 1, control digestion. PecS was used at a concentration of 50 nM (lanes 2 and 4) and 200 nM (lanes 3 and 5). KdgR was used at a concentration of 50 nM (lanes 4 and 6) and 200 nM (lanes 5 and 7). The broken line on the right indicates the region partially deprotected from DNAse I digestion in the presence of both PecS and KdgR. The positions are numbered with respect to the transcription start site. B. Influence of KdgR and PecS on the binding of the CRP–RNA polymerase complex. Lane 1, control digestion. The proteins were added at different concentrations: CRP 50 mM (lanes 9 and 10) or 200 nM (lanes 2–7, 11 and 12); RNA polymerase (RNAP) 75 mM (lane 9), 200 nM (lanes 3–7 and 10–12) or 300 mM (lane 8); KdgR 50 nM (lanes 4 and 15), 100 nM (lane 5) or 200 nM (lanes 6, 7 and 16); PecS 50 nM (lanes 4, 12 and 13), 100 nM (lane 5) or 200 mM (lanes 6, 11 and 14). The asterisks on the right indicate the DNase-hypersensitive sites observed with CRP and RNAP binding. The bold ‘plus’ signs are used to indicate high protein concentrations.

Gel shift experiments (Fig. 3A) showed that cAMP-CRP interacts with the pehX promoter region with a Kd of about 50 nM. The binding of cAMP-CRP produced a single complex with DNA at low CRP concentrations, whereas a second complex appeared at high CRP concentrations. Footprinting experiments performed with CRP on the pehX regulatory region revealed modifications in the digestion pattern, but with DNase I-hypersensitive sites rather than a clearly protected area (Fig. 4B, lanes 1 and 2). Because the synergistic binding of CRP and RNA polymerase has been previously reported (Busby and Ebright, 1997), footprinting experiments were performed in the presence of both CRP and RNA polymerase. The presence of RNA polymerase alone also affected the DNase I digestion pattern, giving rise to hypersensitive sites (Fig. 4B, lane 8). The region protected in the presence of both CRP and RNA polymerase (Fig. 4B, lanes 3, 9 and 10) encompassed the σ70 promoter of pehX and a sequence homologous to the consensus proposed for the E. coli CRP binding site (TGTGA–N6–TCACA) (Kolb et al., 1993) (Fig. 2). The pehX gene has a class II CRP-dependent promoter characterized by CRP binding sites situated near position −41.5, so that CRP and RNA polymerase are situated on the same face of the DNA helix (Ebright, 1993). Thus, the transcription of pehX appears to involve a classical direct activation mechanism.

Gel shift experiments demonstrated that PecS specifically interacts with the regulatory regions of pehV, pehW and pehX (Fig. 3B for pehX). A similar affinity (Kd≈ 25 nM) was observed for the three genes. Footprinting experiments were performed to locate the PecS binding sites (Fig. 4A for pehX). Previous studies failed to identify a consensus for the PecS binding site, and PecS-protected regions of the three positively regulated peh genes (Fig. 2) did not resemble the PecS binding sites previously determined for the PecS-repressed genes celZ, outC and pecS (Praillet et al., 1996, 1997).

Simultaneous binding of the regulatory proteins in the pehX promoter region

KdgR and PecS interact with partially or extensively overlapping regions near the promoter of the three peh genes (Fig. 2). An additional protein, CRP, is able to bind to the pehX promoter region. As this organization allows for physical interactions between the regulators, we analysed the in vitro interactions between the pehX promoter region and several regulators.

We first used CRP and KdgR simultaneously in gel shift assays. In the presence of subsaturating concentrations of KdgR and CRP, we observed a second, low-abundance retarded band that could correspond to the ternary KdgR–CRP–DNA complex (Fig. 3A). Using CRP and KdgR at saturating concentrations, two complexes were also observed: one involving KdgR and the other corresponding to the ternary complex (Fig. 3A). The incomplete formation of the ternary complex could result from an inhibition by KdgR of the CRP binding. Footprinting experiments conducted with KdgR, CRP and RNA polymerase showed a large protected region (Fig. 4, lane 7) similar to that obtained with KdgR alone (Fig. 4, lanes 15 and 16). Thus, in the presence of the three proteins, the preferred binding of KdgR prevents the formation of the transcription initiation machinery.

The simultaneous presence of KdgR and PecS in gel shift assays did not give rise to a more retarded complex, indicating the absence of a ternary complex (Fig. 3B). The gel shift profiles corresponded to that obtained in the presence of PecS alone (Fig. 3B). The mechanism of pehX activation by PecS may be inhibition of the KdgR binding. When DNase I footprinting was performed with both KdgR and PecS, the area protected was shorter than that observed with each protein alone (Fig. 4A) and a partial deprotection of the region overlapping the −10 promoter site was observed.

Footprinting experiments conducted with PecS, CRP and RNA polymerase (Fig. 4B, lanes 11 and 12) showed a clear protected area around the −35 region. This area is clearly smaller than that protected by PecS alone (Fig. 4B, lanes 13 and 14), indicating that PecS, CRP and RNA polymerase are able to bind simultaneously to the pehX promoter region. DNAse I footprinting experiments were then performed in the presence of all the partners: PecS, KdgR, CRP and RNA polymerase (Fig. 4B, lanes 4–6). Although the profile obtained after addition of KdgR, CRP and RNA polymerase (Fig. 4B, lane 7) corresponded to that obtained in the presence of KdgR alone (Fig. 4B, lanes 15 and 16), the profile obtained after addition of KdgR, CRP, RNA polymerase and PecS (Fig. 4B, lanes 4–6) corresponded to that obtained with a mixture of PecS, CRP and RNA polymerase (Fig. 4B, lanes 11 and 12). Thus, DNase I footprinting revealed that the prevention of the binding of the transcription machinery by KdgR is no longer observed when PecS is present. Thus, in the case of the pehX promoter, PecS behaves as an anti-repressor against KdgR.

Discussion

The only polygalacturonase activity previously reported in E. chrysanthemi is the exopoly-α-d-galacturonosidase PehX encoded by the pehX gene in strain EC16 (He and Collmer, 1990). In this study, we showed that E. chrysanthemi 3937 possesses a cluster of three adjacent genes, pehV, pehW and pehX (Fig. 1), encoding enzymes with polygalacturonase activity. The high similarity between these three genes suggests that they arose by duplication of a common ancestor. A similar mechanism was proposed to explain the existence of the adjacent pel genes of the pelADE or pelBC clusters. Thus, duplication could be a general mechanism for the generation of novel pectinolytic genes in E. chrysanthemi. In addition, it is interesting to note that the gppA gene, situated upstream from pehV, and the rep gene, situated downstream from pehX (Fig. 1), are adjacent on the E. coli chromosome (Daniels et al., 1992). Thus, acquisition of the pehV–pehW–pehX cluster by E. chrysanthemi could result from an insertion between the genes gppA and rep in the ancestral enterobacterial genome.

As shown for the other pectinase genes, pehV, pehW and pehX transcription is affected by several environmental stimuli (Table 1). The three peh genes are induced in the presence of pectin catabolic products, at high cell density and during oxygen limitation. They are repressed in the presence of glucose, during nitrogen starvation and at high osmolarity. In addition, as described for the E. carotovora polygalacturonase gene (Flego et al., 1997), the addition of Ca2+ provokes a strong repression of pehV, pehW and pehX expression in E. chrysanthemi. Use of mutants that were affected for genes involved in the production of pectate lyases indicated that KdgR is a repressor of pehV, pehW and pehX transcription whereas CRP is an activator of pehX transcription.

An unusual activator effect was observed for PecS on the three peh genes. PecS was previously described as a repressor of various virulence factors (Reverchon et al., 1994; Praillet et al., 1996). In vitro experiments, such as gel shift assays and DNase I footprinting, using purified KdgR and PecS proteins indicated that each regulator specifically interacts with the peh promoter regions and that the DNA regions covered by their binding widely overlap (Fig. 2). In vitro experiments were then used to analyse the interactions among KdgR, PecS, CRP and RNA polymerase during their binding to the promoter region of pehX. Although KdgR prevents the binding of the transcription initiation machinery (CRP and RNA polymerase), the presence of PecS clearly interfered with the binding of KdgR. Thus, the activator function of PecS on peh transcription could result from its anti-repressor behaviour towards KdgR. Such a hypothesis is reinforced by the failure of PecS to affect peh transcription in a kdgR mutant. A similar anti-repression action was recently suggested for the strain EC16 Pir activator, which probably affects the pelE transcription by competing with KdgR (Nomura et al., 1999). The KdgR repressor controls about 50 genes of E. chrysanthemi that are involved in cell wall degradation (Hugouvieux-Cotte-Pattat et al., 1996). The conditions encountered by the bacteria during plant infection could necessitate the transcription of only one or a few of these genes. The mechanism of anti-repression exerted by different regulatory proteins, such as Pir or PecS, could be an efficient way to increase the transcription of a limited set of genes by specifically counteracting the main repressor KdgR. In addition, the dual role of the PecS regulator (a repressor for the pel genes and an activator of the peh genes) illustrates the complexity of the regulation of the pectinolytic system in E. chrysanthemi.

Experimental procedures

Bacterial strains and plasmids

The bacterial strains and plasmids used in this study are listed in Table 2. The peh::uidA fusions were transduced into various strains using the phi-EC2 generalized transducing phage (Resibois et al., 1984) and selection of kanamycin resistance.

Media and growth conditions

Cells were grown in Luria–Bertani (LB) or M63 media that could be modified to test specific growth conditions, as described previously (Hugouvieux-Cotte-Pattat et al., 1992). E. chrysanthemi and E. coli cells were usually incubated at 30°C and 37°C respectively. Carbon sources were added at 2 g l−1 except for polygalacturonate (grade II, Sigma Chemical Co.), which was added at 4 g l−1. Pectate agar medium (Keen et al., 1984) was used to detect the pectinolytic activity of colonies.

When required, antibiotics were added at the following concentrations: kanamycin (Km), 20 μg ml−1; ampicillin, 50 μg ml−1; chloramphenicol, 20 μg ml−1; streptomycin, 20 μg ml−1.

Enzyme assays and analysis of the polygalacturonase reaction products

Polygalacturonase activity was determined by the increase in reducing groups generated by cleavage of polygalacturonate (0.5 g l−1) in 50 mM sodium acetate, pH 6, and 1 mM EDTA (Collmer et al., 1982). Pectate lyase activity was determined by the formation of unsaturated products that absorb at 230 nm (Shevchik et al., 1997). The β-glucuronidase activity was measured by monitoring the cleavage of p-nitrophenyl-β-d-glucuronide at 405 nm (Hugouvieux-Cotte-Pattat et al., 1992). Pectate lyase and β-glucuronidase assays were usually performed on toluenized extracts of cells grown to early stationary phase.

Thin layer chromatography was used to identify the products obtained after polygalacturonate degradation. The reaction mixtures (10 μl) contained polygalacturonate (2.5 g l−1) in 50 mM sodium acetate, pH 6, and 1 mM EDTA. The chromatograms were developed with a 5:3:2 mixture of n-butanol/water/acetic acid and the products were visualized by treatment with phosphomolybdic acid (Lojkowska et al., 1995).

Molecular biology techniques

Preparation of plasmid or chromosomal DNA, restriction digestions, ligations, DNA electrophoresis and transformations were carried out as previously described (Sambrook et al., 1989). RNA was extracted by the frozen phenol method (Maes and Messens, 1992). Deletions for nucleotide sequencing were generated with restriction endonucleases and the sequences were performed by Genome Express. The nucleotide sequence data reported in this paper will appear in the EMBL, GenBank and DDBJ nucleotide sequence data bases under accession number AJ132326.

The peh::uidA transcriptional fusions were constructed by introduction of the uidA-Km cassettes (Bardonnet and Blanco, 1992) into the EcoRI, HindIII and NheI restriction sites located in pehV, pehW and pehX respectively (Fig. 1). The plasmids in which the cassette is correctly oriented were introduced in E. chrysanthemi by electroporation and the fusions were introduced into the chromosome by marker exchange recombination after successive cultures in low phosphate medium (Hugouvieux-Cotte-Pattat et al., 1992).

The plasmid pULB110 (Van Gijsegem et al., 1985) was used to generate R-prime derivatives containing an uidA-Km insertion by mating the appropriate E. chrysanthemi mutant, containing pULB110, with the E. coli strain HB101. Selection of KmR transconjugants was performed on LB plates supplemented with kanamycin and streptomycin.

Gel shift experiments

The DNA fragments containing the regulatory regions of the peh genes (Fig. 1) were end-labelled with [α-32P]-dCTP or [α-32P]-dATP (3000 Ci mmol−1–1) using the Klenow fragment and purified using the Qiaquick gel extraction kit (Qiagen).

Gel shift assays were conducted as previously described (Praillet et al., 1996; Nasser et al., 1997) with the KdgR, CRP and PecS purified proteins (Nasser et al., 1994, 1997; Praillet et al., 1997). The labelled DNA probe (50 000 c.p.m.) and the purified regulator (20–200 nM) were incubated for 30 min at 30°C in 20 μl of 10 mM tris-HCl buffer (pH 7 or 7.8) containing 70 mM KCl, 1 mM DTT, 100 μM cAMP, 4 μg of acetylated bovine serum albumin and 1 μg poly-(dI–dC)–(dI–dC). The reaction mixtures were then submitted to electrophoresis on a 4% non-denaturing polyacrylamide gel in 10 mM tris-HCl (pH 7 or 7.8) with 100 μM cAMP. Bands were detected by autoradiography.

The apparent dissociation constant (Kd) is the protein concentration for which half of the DNA forms complexes with the protein (Carey, 1988). For Kd determination, autoradiograms of gel shift assays, performed with a large concentration range of the regulator, were submitted to densitometric analysis to calculate the ratio of free probe compared with total DNA.

DNase I footprinting

DNase I footprinting was performed as previously described (Nasser et al., 1997). The E. coli RNA polymerase holoenzyme was purchased from TEBU. About 100 000 c.p.m. of DNA probe, labelled at one end, was incubated for 30 min at 30°C with the purified protein(s) in the gel shift assay buffer. The reaction mixtures were adjusted to 10 mM MgCl2 and 5 mM CaCl2 before the addition of DNase I (5 × 10−3 units). Digestion was performed at 30°C for 1 min and stopped by the addition of 25 μl of stop solution. After phenol/chloroform extraction, DNA fragments were ethanol precipitated and separated by electrophoresis on a 6% polyacrylamide sequencing gel. The digestion profile was revealed by autoradiography.

S1 mapping analysis

The DNA fragments used for S1 mapping (Sambrook et al., 1989) corresponded to the 0.36 kb MunI/ClaI, 0.33 kb MunI/SphI and 0.36 kb BamHI/NarI fragments overlapping the regulatory regions and the ATG initiation codon of pehV, pehW and pehX respectively (Fig. 1). About 150 μg of total RNA was annealed in S1 hybridization solution with about 6 × 104 c.p.m. of the 32P-end-labelled DNA probe. The DNA–RNA duplexes were then submitted to digestion with 100, 250 and 500 U ml−1 of S1 nuclease. The digestion products were separated on 6% polyacrylamide sequencing gels and detected by autoradiography.

Acknowledgements

Appreciation is expressed to V. James for reading the manuscript and to Y. Rhabe for his help in preparing the illustrations. We thank Professor Robert Baudouy for her interest during this work. We gratefully acknowledge the members of this laboratory, particularly S. Reverchon and G. Condemine for their valuable discussions. This work was supported by grants from the Centre National de la Recherche Scientifique and from the Ministère de l'Education Nationale, de l'Enseignement Supérieur, de la Recherche et de l'Insertion Professionnelle.

Ancillary