PehN, a Polygalacturonase Homologue with a Low Hydrolase Activity, Is Coregulated with the Other Erwinia chrysanthemi Polygalacturonases

Abstract

Erwinia chrysanthemi 3937 secretes an arsenal of pectinolytic enzymes, including at least eight endo-pectate lyases encoded by pel genes, which play a major role in the soft-rot disease caused by this bacterium on various plants. E. chrysanthemi also produces some hydrolases that cleave pectin. Three adjacent hydrolase genes, pehV, pehW, and pehX, encoding exo-poly-α-d-galacturonosidases, have been characterized. These enzymes liberate digalacturonides from the nonreducing end of pectin. We report the identification of a novel gene, named pehN, encoding a protein homologous to the glycosyl hydrolases of family 28, which includes mainly polygalacturonases. PehN has a low hydrolase activity on polygalacturonate and on various pectins. PehN action favors the activity of the secreted endo-pectate lyases, mainly PelB and PelC, and that of the periplasmic exo-pectate lyase PelX. However, removal of the pehN gene does not significantly alter the virulence of E. chrysanthemi. Regulation of pehN transcription was analyzed by using gene fusions. Like other pectinase genes, pehN transcription is dependent on several environmental conditions. It is induced by pectic catabolic products and is affected by growth phase, catabolite repression, osmolarity, anaerobiosis, nitrogen starvation, and the presence of calcium ions. The transcription of pehN is modulated by the repressor KdgR, which controls almost all the steps of pectin catabolism, and by cyclic AMP receptor protein (CRP), the global activator of sugar catabolism. The regulator PecS, which represses the transcription of the pel genes but activates that of pehV, pehW, and pehX, also activates transcription of pehN. The three regulators KdgR, PecS, and CRP act by direct interaction with the pehN promoter region. The sequences involved in the binding of these three regulators and of RNA polymerase have been precisely defined. Analysis of the simultaneous binding of these proteins indicates that CRP and RNA polymerase bind cooperatively and that the binding of KdgR could prevent pehN transcription. In contrast, the activator effect of PecS is not linked to competition with KdgR or to cooperation with CRP or RNA polymerase. This effect probably results from competition between PecS and an unidentified repressor involved in peh regulation.

The enterobacterium Erwinia chrysanthemi is a well-characterized phytopathogen which causes soft-rot disease of various plants. The typical soft-rot maceration of plant tissues involves the depolymerization of the pectin in plant cell walls. The cleavage of this polysaccharide, an α-1,4-galacturonan (or polygalacturonate), is accomplished by a variety of microbial or fungal enzymes produced, in particular, by plant pathogens and by saprophytic microorganisms. Variations in the structure of the galacturonan chain occur in the extent of esterifications by O-acetyl and methoxyl groups, as well as in some modifications, such as xylose substitution or insertion in the polysaccharide backbone of single 1,2-linked α-l-rhamnose (27, 37, 46).

Two types of enzymes are involved in pectin depolymerization, lyases and hydrolases. Pectate lyases play an important role in the soft-rot disease caused by E. chrysanthemi (3). Pectin or pectate lyases cleave, by β-elimination, the internal glycosidic bonds in high- and low-methoxylesterified polygalacturonate, respectively. Polygalacturonases act by hydrolysis of the pectic glycosidic bonds. Depolymerases can also differ in the random or terminal attack of the polymer. Endo-enzymes cut at random sites within the pectic chain to give a mixture of oligomers of various sizes. In contrast, the attack of exo-enzymes is limited to one end of the polymer, producing only one product. For instance, the E. chrysanthemi exo-poly-α-d-galacturonosidase PehX liberates dimers from the nonreducing end, while the exo-pectate lyase PelX liberates unsaturated dimers from the reducing end (40). The existence of additional pectinolytic enzymes specific to each type of modification is highly suspected and has been confirmed, for example, by the recent identification of a fungal endo-xylogalacturonan hydrolase (47).

E. chrysanthemi strain 3937 produces at least eight endo-pectate lyases, PelA, PelB, PelC, PelD, PelE, PelI, PelL, and PelZ (23, 28, 41, 45), two exo-pectate lyases, PelX and PelW (38, 40), and three exo-poly-α-d-galacturonosidases, PehV, PehW, and PehX (26). While the endo-cleaving enzymes are all secreted in the external medium by the Out system, the exo-enzymes are intracellular; PelX, PehV, PehW, and PehX are periplasmic, and PelW is cytoplasmic (38, 40).

The transcription of the pectinase genes is tightly controlled by environmental conditions (16). In the presence of pectin, the formation of catabolic products by a basal level of pectinase activity strongly induces the transcription of all the pectinase genes. This induction involves a derepression mechanism. The interaction of KdgR with intracellular pectin catabolites relieves the binding of this transcriptional repressor from its binding sites situated in the vicinity of the promoters (25). KdgR controls most of the pectinase genes, the outC-N operon involved in pectinase and cellulase secretion, and all the genes necessary for the catabolism of pectin-derived oligomers. Transcription of the pectinase genes is also subject to catabolite repression mediated by the cyclic AMP receptor protein (CRP) activator, which is involved in the global control of sugar catabolism (30). The transcription of several pectinase genes is also controlled by PecS, a transcriptional regulator of the MarR family. PecS represses the transcription of the endo-pectate lyase genes, of the cellulase gene celZ, of the secretion operon outC-N, and of the ind operon involved in the synthesis of a blue pigment, indigoidine (31). In contrast, PecS acts as an activator of the transcription of the three polygalacturonase genes, pehV, pehW, and pehX (26). Until now, the signal triggering PecS control had not been identified, but recent data suggest that it exerts a regulatory influence over some responses to oxidative stress (31a).

To identify the genes involved in pectin degradation in E. chrysanthemi 3937, we isolated mutations on the basis of their induction by pectic derivatives. Such mutations were obtained by selection of Mu-lac insertions generating polygalacturonate-inducible (PGI) lacZ transcriptional fusions (19). Analysis of some PGI mutations allowed us to identify novel pectinase-encoding genes: paeY, which encodes a pectin acetyl esterase (39), and pelI, which encodes an endo-pectate lyase (41). In this paper, we report the characterization of a novel PGI fusion, situated under a gene named pehN because it encodes a protein homologous to endo-polygalacturonases. We constructed uidA transcriptional fusions to analyze the expression of this gene in various conditions and in the presence of regulatory mutations. In addition, we tested the direct interaction of the pectinase regulators KdgR, PecS, and CRP with the promoter region of the pehN gene.

MATERIALS AND METHODS

Bacterial strains and genetic techniques.

The bacterial strains of E. chrysanthemi and Escherichia coli and the plasmids used in this study are listed in Table 1. The pehN::uidA fusions were transduced into various strains using the φEC2 generalized transducing phage (29) and selection for kanamycin resistance. Plasmid pULB110 (48) was used to generate R-prime derivatives containing a uidA-Km (kanamycin resistance cassette) insertion by mating the appropriate E. chrysanthemi mutant, containing pULB110, with E. coli strain HB101. Selection of kanamycin-resistant transconjugants was performed on Luria-Bertani (LB) plates supplemented with kanamycin and streptomycin. Plasmid pULB110 was also used for chromosomal mapping of the pehN and pehX mutations. Various polyauxotrophic strains were used as recipients in conjugation with a donor strain containing the plasmid pULB110.

TABLE 1.

 Bacterial strains and plasmids

Strain or plasmid	Relevant characteristics	Reference or origin
E. chrysanthemi (3937 derivatives)
A350	lmrTc lacZ2	Laboratory collection
PGI40	lmrTc lacZ2 pehN::lacZ, Apr	19
A1077	lmrTc lacZ2 kdgR::Mu Cmr	Laboratory collection
A1524	lmrTc lacZ2 pecS::Mu Cmr	Laboratory collection
A2174	lmrTc lacZ2 pecT::Cmr	Laboratory collection
A2507	lmrTc lacZ2 crp::Cmr	Laboratory collection
A3377	lmrTc lacZ2 pehN::uidA, Kmr	This work
A3392	lmrTc lacZ2 kdgR::Mu CmrpehN::uidA, Kmr	This work
A3393	lmrTc lacZ2 pecS::Mu CmrpehN::uidA, Kmr	This work
A3394	lmrTc lacZ2 pecT::Cm pehN::uidA, Kmr	This work
A3395	lmrTc lacZ2 crp::Cm pehN::uidA, Kmr	This work
Plasmids
pULB110	RP4::Mu3A, Apr Tcr Kms	48
pR10T	Lac+ pULB110 derivative, pehN::lacZ+	This work
pBR325	Apr Tcr Cmr	F. Bolivar
pT149	pBR325 derivative with 11-kb PstI fragment from pR10T, pehN::lacZ+	This work
pT170	pBR325 derivative with a 4.8-kb PstI-HindIII fragment from pT149	This work
pBS Cm	Bluescript KS+, Cmr	Stratagene
p1863	pBSCm derivative with a 6-kb SmaI-HindIII fragment, KmrpehN+	This work
p1878	pBSCm derivative with a 1.8-kb HpaI-BamHI fragment from p1863, pehN+	This work
pT7-7	φ10 promoter, gene 10 translational start, Apr	44
p2480	pT7-7 derivative with a 1.4-kb PCR fragment, pehN+	This work

Media and growth conditions.

Cells were grown in LB or M63 medium that could be modified to test specific growth conditions, as previously described (17). Carbon sources were added at 2 g liter⁻¹. E. chrysanthemi and E. coli cells were usually incubated at 30 and 37°C, respectively. 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; and streptomycin, 20 μg ml⁻¹.

Enzyme assays.

Polygalacturonase assays were performed in 0.1 M sodium acetate buffer, pH 6, using 2.5 g of polygalacturonate (grade II; Sigma Chemical Co.) liter⁻¹. The reaction was started by the addition of enzyme. At intervals, 50-μl samples were taken, and the reaction was stopped by adding each sample to 1 ml of 0.5 M Na₂CO₃. The amount of reducing sugars liberated was determined by using the neocuproin reagent (42). One unit of enzyme is defined as the amount that generates 1 μmol of reducing sugar in 1 h. The assay was also performed using other substrates: lemon pectins with different degrees of methylation, apple pectin and sugar-beet pectin (from Copenhagen Pectin), and modified hairy regions of pectin (36).

Pectate lyase activity was measured at 37°C by monitoring the appearance of unsaturated products at 230 nm (26). The standard assay mixture consisted of 50 mM Tris-HCl (pH 8.5), 0.1 mM CaCl₂, and 0.5 g of polygalacturonate liter⁻¹. For analysis of the synergistic effects of PehN, the assay mixture was incubated in the presence of PehN for 15 min before addition of the purified pectate lyase.

β-Glucuronidase activity was measured by monitoring the cleavage of p-nitrophenyl-β-d-glucuronide at 405 nm (17).

Overproduction of PehN protein.

The pehN gene was overexpressed using the T7 promoter/T7 RNA polymerase system (44). A 1.4-kb fragment corresponding to the pehN open reading frame (ORF) was inserted into the pT7-7 expression vector (see below). To detect the PehN protein, the resulting plasmid, pN2480, was introduced into E. coli K38/pGP1.2, which contains the T7 RNA polymerase gene under the control of the cI857 thermosensitive repressor. Plasmid-encoded proteins were labeled with [³⁵S]cysteine-methionine after thermal induction of the T7 polymerase (44). After separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), proteins were detected by autoradiography.

Plasmid pN2480 was also introduced into E. coli BL21(DE3), which contains a chromosomal copy of the T7 RNA polymerase gene under the control of the lacUV5 promoter (43). The BL21(DE3)/pN2480 cells were grown at 30°C in LB supplemented with ampicillin. At an optical density at 600 nm (OD₆₀₀) of 0.5 to 0.6, the synthesis of T7 RNA polymerase was induced by addition of 1 mM isopropyl-β-d-thiogalactopyranoside, and the bacterial RNA polymerase was blocked by the addition of rifampin (200 μg ml⁻¹). Cells were grown for an additional 2 h and harvested by centrifugation for 2 min at 12,000 × g. The release of periplasmic proteins was performed by osmotic shock (4). SDS-PAGE was performed on slab gels (4% stacking gel and 12% separating gel) using the Mini-Protean II system (Bio-Rad). Proteins were stained with Coomassie blue G-250.

Molecular biology techniques.

Preparation of plasmid and chromosomal DNA, restriction digestions, ligations, DNA electrophoresis, and transformations were carried out as previously described (35). Deletions for nucleotide sequencing were generated with restriction endonucleases, and the sequences were performed by Genome Express SA (Grenoble, Switzerland).

To clone the DNA fragment corresponding to the pehN gene, we designed PCR primers corresponding to the 5′ and 3′ ends of the pehN ORF. Contiguous XbaI and NdeI sites were added to the forward primer (PHN1), and a BamHI site was added to the reverse primer (FPHN). The sequences of PHN1 and FPHN are GCTCTAGACATATGCATAAAGGTGTGGCG and CCCGGATCCTTATCGGGTAACGTTTGCCC, respectively (restriction sites are underlined, and the initiation and termination codons are in italics). Chromosomal DNA from strain 3937 was used as the template with these primers. The PCR product was purified (QIAquick PCR purification kit; Qiagen), digested by XbaI and BamHI, and ligated to the vector pBSCm digested by the same enzymes. One of the resulting plasmids (pN2431) was digested by NdeI and BamHI to insert the pehN gene into the pT7-7 vector, which contains the T7 RNA polymerase promoter, phi10, and the translation start for T7 gene 10 protein (44). Primers PHN1 and FPHN were also used to amplify the pehN gene using chromosomal DNA extracted from other E. chrysanthemi strains.

The pehN::uidA transcriptional fusion was constructed by introduction of the uidA-Km cassette (1) into the SalI restriction site located in pehN (Fig. 1). A plasmid in which the cassette is correctly oriented was introduced in E. chrysanthemi by electroporation, and the fusion was introduced into the chromosome by marker exchange recombination after successive cultures in low-phosphate medium (17).

FIG. 1.

Physical map of the E. chrysanthemi 3937 pehN gene. The restriction maps of the main subclones and of the region sequenced are given. The arrows indicate the position and the transcription direction of the genes. The grey region indicates the Mu dI1681 DNA, while the thin line corresponds to E. chrysanthemi chromosomal DNA. The flags show the sites of insertion of the uidA-Km cassettes. The thick line indicates the DNA fragment used as the pehN promoter region for in vitro experiments.

Primer extension analysis.

Total RNA was extracted from the E. chrysanthemi kdgR mutant A1077 by the frozen phenol method (24). RNA concentration was estimated spectrophotometrically and by electrophoresis on a denaturing formaldehyde-1% agarose gel. For primer extension, 30 and 100 μg of total RNA were annealed in S1 hybridization solution with about 6 × 10⁴ cpm of ³²P-end-labeled oligonucleotide (GTGGAAAAGTCGGCGGACTCGGCG). The primer is complementary to the sequence from +77 to +53 downstream from the pehN translation start site.

Gel shift experiments.

The 0.41-kb SacII-NsiI DNA fragment containing the promoter region of the pehN gene was cloned into the SacII and HindIII sites of the pBluescript vector (Ap^r, KS⁺). One strand was labeled by incorporating [α-³²P]dATP (3,000 Ci mmol⁻¹) with the Klenow fragment of DNA polymerase at the EcoRI end of the EcoRI-HindIII fragment. The other strand was labeled by incorporating [α-³²P]dCTP (3,000 Ci mmol⁻¹) at the HindIII end of this fragment. The labeled fragments were further purified using the Qiaquick gel extraction kit (Qiagen). Gel shift assays were conducted, as previously described, with the KdgR, CRP, and PecS purified proteins (26). The labeled DNA probe (50 000 cpm) and the purified regulator (20 to 200 nM) were incubated for 30 min at 30°C in 20 μl of buffer (Tris-HCl, 10 mM, pH 7 or 7.8) containing 70 mM KCl, 1 mM dithiothreitol (DTT), 100 μM cyclic AMP (cAMP), 4 μg of acetylated bovine serum albumin, and 1 μg of poly(dI-dC)-(dI-dC). The reaction mixtures were subjected to electrophoresis on a 4% nondenaturing polyacrylamide gel in 10 mM Tris-HCl (pH 7.0 or 7.8)-100 μM cAMP. Bands were detected by autoradiography. The apparent dissociation constant (K_d) is the protein concentration at which half of the DNA forms complexes with the protein (26). For K_d determination, autoradiograms of gel shift assays performed with a large concentration range of the regulator were subjected to densitometric analysis to calculate the ratio of free probe versus total DNA.

DNase I footprinting.

DNase I footprinting was performed as previously described (26). About 100,000 cpm of DNA probe, labeled 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 U). Digestion was performed at 30°C for 1 min and stopped by the addition of the same volume of 0.1 M EDTA (pH 8). 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.

Pathogenicity tests.

Thirty plants of Saintpaulia ionantha were inoculated after wounding of a leaf with 50 μl of bacterial suspension (10⁸ bacteria). Results of infections were scored daily for 10 days, and the symptoms were classified in three groups: no symptoms, local necrosis limited to the leaf, and systemic infection of the plant. Potato tubers were inoculated with sterile pipette tips containing 5 μl of bacterial suspension (10⁷ bacteria), inserted into the tuber parenchyma to a depth of 10 mm. Twenty tubers were inoculated with each strain and incubated in a dew chamber. After 2 days, tubers were sliced vertically through the inoculation point, and the weight of decayed tissue was taken as characteristic of disease severity. Chicory leaves were slightly wounded prior to inoculation. Twenty leaves were infected for each strain, using 10⁶ bacteria per inoculation site. After incubation in a dew chamber for 24 h, the length of rotted tissue was measured to estimate disease severity.

Nucleotide sequence accession number.

Sequence data reported in this paper (′mcp-pehN-ybaK′ of E. chrysanthemi) will appear in the EMBL, GenBank, and DDBJ nucleotide sequence data bases under accession number AJ292044.

RESULTS

Identification of the pehN gene and determination of its nucleotide sequence.

We cloned the polygalacturonate-inducible lacZ fusion of strain PGI40 by isolation of an R-prime plasmid selected for the Km^r marker of the Mu dI1681 insertion. Plasmid R10T contained, in addition to the Mu insertion, a fragment of adjacent chromosomal DNA of about 18 kb. We subcloned an 11-kb PstI DNA fragment into pBR325, which conferred the Lac⁺ phenotype (pT149). Restriction analysis confirmed that this fragment contained the S end of the Mu dI1681 phage, including the lacZ gene. We then subcloned the 4.8-kb PstI-HindIII fragment situated at the junction of the Mu insertion (pT170) (Fig. 1). Determination of the nucleotide sequence of this junction and comparison with sequences present in databases revealed that the Mu insertion was situated in a region encoding the N-terminal end of a protein presenting about 30% identity with bacterial polygalacturonases.

To isolate the wild-type pehN gene, we first inserted a uidA-Km cassette into the HpaI site situated 0.4 kb upstream from the Mu insertion (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 into pBSCm enabled us to isolate a HindIII-SmaI fragment of 6 kb (p1863) conferring the Km^r phenotype. This region was then reduced to a 1.8-kb HpaI-BamHI fragment (p1878) (Fig. 1).

Determination of the nucleotide sequence of a 2.1-kb SacII-BamHI fragment (Fig. 1) revealed the presence of a unique complete ORF which began with an ATG codon at position 411, separated by 5 nucleotides from a potential ribosome-binding site, GGGG, and ended with TAA at position 1782. The Mu insertion of mutant PGI40 was situated inside this ORF (at position 644). The pehN gene encodes a 457-amino-acid protein with a molecular mass of 49,461 Da, including a typical amino-terminal signal sequence with a potential cleavage site between two Ala residues at positions 22 and 23 of the protein sequence. The PehN mature protein contains 435 amino acids and has a calculated molecular mass of 47,333 Da and a calculated pI of 7.9.

Centered 69 nucleotides after the pehN translational stop was a GC-rich imperfect inverted repeat (CGGCGGCAGAGGGCAA CACT G-1 nt-CC GTGGAA CCCTCTGCCGCCG [nucleotides that match in this structure are underlined], positions 1831 to 1873), followed by a run of T residues, typical of a rho-independent transcription terminator. The pehN gene was preceded by a partial potential ORF transcribed on the same DNA strand ending at position 116. DNA homology searches revealed homology between the product of this ORF (mcp, Fig. 1) and the C-terminal end of methyl-accepting chemoreceptor proteins from various bacteria, including the E. coli Tsr, Tar, Trg, and Tap proteins, with identities of 61, 58, 56, and 54%, respectively. The pehN gene was followed by a partial potential ORF, transcribed on the other DNA strand and ending at position 1888, which was homologous to an E. coli gene of unknown function, ybaK (75% identity at the protein level) (Fig. 1).

PehN protein belongs to family 28 of glycosyl hydrolases.

The deduced amino acid sequence of PehN is homologous to the polygalacturonases of bacterial, fungal, and plant origin. PehN presents the highest similarity to the endo-polygalacturonases PehA of Erwinia carotovora subsp. carotovora (13, 22, 34) (37% identity), PglA of Ralstonia solanacearum (14) (30% identity), Peh of Agrobacterium vitis (12) (29% identity), and Peh of Burkholderia cepacia (10) (29% identity). PehN is also homologous to the exo-poly-α-d-galacturonosidases PehV, PehW, and PehX of E. chrysanthemi (11, 26) (29% identity), PehY of Yersinia enterocolitica (21) (28% identity), and PehB of Ralstonia solanacearum (15) (28% identity) (Fig. 2). These exo-acting enzymes have an N-terminal extension of about 100 residues compared to the endo-acting polygalacturonases.

FIG. 2.

Comparison of the active sites of PehN and various bacterial polygalacturonases. The PehN sequence from residues 209 to 328 is aligned with the corresponding region of the endo-polygalacturonases Peh of Erwinia carotovora subsp. carotovora (P18192) (13, 22, 34), PglA of Ralstonia solanacearum (P20041) (14), Peh of Agrobacterium vitis (U73161) (12), Peh of Burkholderia cepacia (U85788) (10), and the exo-polygalacturonases PehX of Erwinia chrysanthemi (P15922) (11, 26), PehY of Yersinia enterocolitica (3089553) (21), and PehB of Ralstonia solanacearum (U60160) (15). The residues conserved in more than 50% of the sequences are given below the alignment, and the four groups of amino acids supposed to be involved in catalysis are underlined. The arrowhead points to the conserved His which is replaced by Arg in PehN.

All these proteins belong to family 28 of glycosyl hydrolases (5). This family covers several enzyme specificities, since it includes several polygalacturonases, the endo-xylogalacturonan hydrolase XghA of Aspergillus tubigensis, and three rhamnogalacturonases, RHG-A of Aspergillus aculeatus and RHG-A and RHG-B of Aspergillus niger. The level of identity of PehN to XghA is 24%, whereas the level of similarity to rhamnogalacturonases is very low (and located mainly in the N-terminal part of PehN). Four amino acid groups (NTD, DD, HG, and RIK), supposed to be involved in the catalytic mechanism (2), are well conserved in the family 28 polygalacturonases. In PehN, these regions are conserved with the exception of the His residue, which is replaced by an Arg residue (position 273) (Fig. 2). A similar exception was observed for the three rhamnogalacturonases, and in these enzymes the corresponding His is replaced by Gly. It is tempting to suggest that these variations are linked to an adaptation of the catalytic mechanism to different substrate specificities.

We verified that the Arg residue in PehN did not result from a mutation in the cloned DNA region. Indeed, a single point mutation, CAC to CGC, could have been responsible for this modification. We amplified the pehN gene by PCR using chromosomal DNA of the wild-type E. chrysanthemi strains 3937, B374, EC16, and ENA49 and primers corresponding to the ends of the pehN ORF. A unique 1.4-kb fragment corresponding to the complete pehN gene was obtained with the three strains 3937, B374, and ENA49, while a 0.3-kb shortened fragment was obtained with strain EC16. This strain was previously shown to contain several chromosomal deletions (7). Sequencing of the 3′ end of the three complete pehN fragments confirmed that the GCG codon encoding Arg is present in strains 3937, B374, and ENA49. Comparison of these DNA sequences indicated that pehN is highly conserved among the three strains (more than 98% identity).

Characterization of PehN protein.

The pehN gene was cloned in the pT7-7 vector, and the PehN protein was overproduced in a recombinant E. coli strain. Analysis by SDS-PAGE of the plasmid-encoded proteins revealed the presence of a 48-kDa protein located mainly in the periplasm (data not shown). A periplasmic extract of strain BL21(DE3)/p2480 contained more than 50% PehN. This extract showed very low hydrolase activity when assayed for polygalacturonase activity. In similar conditions, the specific activities of PehN, PehV, PehW, and PehX with polygalacturonate as the substrate were 0.7, 9, 14, and 840 nmol h⁻¹ (μg of protein)⁻¹, respectively. The PehN activity was not significantly affected by the addition of 0. 1 mM Ca²⁺, 10 mM EDTA, or 100 mM NaCl. The optimum pH for the reaction was about 6. A similar low level of hydrolase activity was detected with lemon pectin, apple pectin, and sugar beet pectin as the substrate. No activity was detected with modified hairy regions of pectin as the substrate.

Synergism among pectinases displaying different modes of cleavage might facilitate pectin depolymerization. We tested whether the action of PehN could favor that of the E. chrysanthemi depolymerases. First, we tested the effect of PehN on the activity of the endo-pectate lyases PelA, PelB, PelC, PelD, and PelE (Fig. 3). In the presence of PehN, the pectate lyase activity of PelB and PelC showed a significant and reproducible increase of about 40% (Fig. 4). With the isoenzymes PelD and PelE, only a small increase was detected, 10 to 16%. No significant effect was observed for PelA. These data suggest that the products liberated by PehN are better substrates for some endo-pectate lyases, mainly PelB and PelC. Previous data indicated that the specific activity of PelB and PelC is seven- to eightfold higher on oligomers (hexamers) than on polygalacturonate, while the activity of PelA, PelD, and PelE is not increased on oligomers compared to the polymer (33; unpublished data). Thus, the low hydrolysis of polygalacturonate by PehN could favor PelB and PelC action.

FIG. 3.

Effect of PehN action on the activity of other E. chrysanthemi pectinases. The activity observed in the absence of PehN was arbitrarily defined as 100%. The initial rate of the reaction was determined in the absence and presence of 0.5 μg of PehN extract, which was added 15 min before the enzyme assayed. For the pectate lyase assays, about 0.02 U of each enzyme was used. For the polygalacturonase assays, 400 U was used for PehV and PehW, while 1,500 U was used for PehX. The results reported are the averages of at least three assays, and the standard deviation is indicated.

FIG. 4.

Mapping of the pehN transcript by primer extension analysis. The reactions were carried out with 30 and 100 μg of RNA from the E. chrysanthemi kdgR mutant A1077 (lanes 1 and 2, respectively). The arrow indicates the transcription initiation site.

The activity of the exo-pectate lyase PelX was also clearly increased in the presence of PehN (Fig. 3). In contrast, the hydrolase activity of the exo-poly-α-d-galacturonosidases PehV, PehW, and PehX was not affected by the PehN action. Since PelX attacks the substrate from the reducing end while PehX attacks the polymer from the nonreducing end, this result suggests that PehN liberates oligomers with a normal reducing end, which can be recognized by PelX, but with a blocked nonreducing end, which is not available for PehX action.

Expression of E. chrysanthemi pehN gene.

A pehN::uidA transcriptional fusion was constructed by insertion of a uidA::Km cassette into the SalI site located in pehN (Fig. 1). After recombination into the E. chrysanthemi chromosome, the expression of the fusion was followed during bacterial growth. The expression of the fusion increased by fivefold when the cells entered the late exponential growth phase (data not shown). This increase coincided with that of pectate lyase production.

The expression of the pehN::uidA fusion was also analyzed after growth under various conditions (Table 2). When bacteria were grown in medium containing glycerol as the carbon source, the pehN::uidA mutant showed a significant basal level of expression. In the presence of polygalacturonate, the transcription of pehN was stimulated fourfold (Table 2). In contrast, in the presence of a readily utilizable carbon source such as glucose, a twofold decrease was observed. Variation of the growth temperature (25, 30, and 37°C) did not significantly affect pehN expression. Conditions such as semianaerobic growth, nitrogen starvation, and high medium osmolarity also affected the expression of pehN but by less than twofold (Table 2). As previously observed for pehV, pehW, and pehX of E. chrysanthemi (26), addition of CaCl₂ clearly reduced pehN expression when bacteria were grown in medium containing polygalacturonate but not with other carbon sources (glycerol, glucose, or galacturonate) (Table 2 and data not shown).

TABLE 2.

 Expression of pehN::uidA transcriptional fusion in E. chrysanthemi 3937^a

Mutation	Growth conditions	β-Glucuronidase sp act (nmol min−1 [mg of protein]−1)	Pectate lyase sp act (μmol min−1 [mg of protein]−1)
None	Glycerol	49	0.03
	Polygalacturonate	201	14
	Glucose	22	0.03
	Glucose + polygalacturonate	72	1.26
	Glycerol, 25°C	52	0.07
	Glycerol, 37°C	48	0.01
	Glycerol, oxygen limitation	84	0.25
	Glycerol, nitrogen starvation	27	0.01
	Glycerol + 0.3 M NaCl	36	0.22
	Glycerol + 0.2 mM CaCl2	49	0.09
	Polygalacturonate + 0.2 mM CaCl2	82	11
kdgR	Glycerol	389	1.19
	Polygalacturonate	418	9
pecS	Glycerol	25	1.45
	Polygalacturonate	127	18
pecT	Glycerol	47	1.63
	Polygalacturonate	185	24
crp	Glucose	18	0.01
	Glucose + polygalacturonate	27	0.01

^aThe results reported are the average of at least 3 independent experiments with standard deviations corresponding to less than 20%, except for very low pectate lyase activities (< 0.05) for which standard deviations could reach 50%.

The pehN gene fusion was transduced into strains mutated in genes encoding the main regulators of pectinase production, KdgR, PecS, PecT, and CRP. Inactivation of the kdgR gene strongly increased both pectate lyase production and expression of the pehN fusion (Table 2). While the pecS mutation increased pectate lyase production, it decreased pehN expression (Table 2), as previously observed for the genes pehV, pehW, and pehX (26). A pecT mutation did not affect expression of the pehN fusion (Table 2). The crp mutation strongly decreased both pectate lyase production and pehN transcription (Table 2). These data indicated that pehN transcription is repressed by KdgR and activated by PecS and CRP.

Identification of pehN transcriptional regulatory sites.

We first determined the transcriptional start site of pehN by primer extension (Fig. 4). This site is located 31 nucleotides upstream from the pehN translational start codon. It is preceded by a potential sigma 70-type promoter, with a −35 region (CCGTTT) separated by 17 nucleotides from a −10 region (TACATT) (Fig. 4 and 5). In vitro experiments were then used to study interactions between the 0.41-kb SacII-NsiI DNA fragment containing the pehN promoter region and the different regulators involved in its transcription, PecS, KdgR, and CRP.

FIG. 5.

Organization of the promoter region of the E. chrysanthemi 3937 pehN gene. The regions protected in DNase footprinting by the binding of the various proteins (PecS, KdgR, CRP, and CRP-RNA polymerase) are indicated. The consensus of the binding sites of the regulators KdgR, CRP, and ? (the putative specific peh regulator) are shown below the sequence, with the less-conserved nucleotides in lowercase letters (R = G or A, Y = C or T). The underlined sequences correspond to the main restriction sites, the transcriptional start determined by primer extension (+1), the corresponding −10 and −35 promoter sites, and the putative Shine-Dalgarno sequence (SD). The pehN translational start codon is shown in bold letters.

Gel shift assays demonstrated that KdgR interacted specifically with the pehN promoter region (Fig. 6A). The binding of KdgR produced a single complex at low KdgR concentrations, while a more retarded complex appeared at high KdgR concentrations. The KdgR binding affinity was high, with an apparent dissociation constant (K_d) of 5 nM. Footprinting analysis (Fig. 7A, lanes 2 and 3) revealed two distinct regions protected by KdgR, extending from positions −88 to −51 and −8 to +33 relative to the transcription start site (Fig. 5). These two regions contained sequences homologous to the consensus of the KdgR binding site determined previously (25). The KdgR concentration required for the protection of the KdgR binding site, centered at position +11 (KdgR1), was 10-fold lower than that needed to protect the KdgR binding site centered at position −67 (KdgR2) (10 nM versus 100 nM) (Fig. 7A, lanes 2 and 3), clearly indicating that the affinity of KdgR for the KdgR1 site is greater than that for the KdgR2 site.

FIG. 6.

Analysis of binding of transcription factors on the pehN promoter region in gel shift assays. The indicated amounts of the purified proteins were incubated with the end-labeled 0.41-kb SacII-NsiI fragment corresponding to the 5′ untranslated end of pehN. (A) Binding of KdgR and CRP. (B) Binding of PecS and KdgR.

FIG. 7.

Analysis of binding of transcription factors on the pehN promoter region by DNase I footprinting. (A) Binding of KdgR, CRP, and RNA polymerase (RNAP). Lane 1, control digestion; lanes 2 to 14, DNase I digestion in the presence of the protein(s) indicated above. (B) Binding of PecS and KdgR. Lane 1, control digestion; lanes 2 to 8, DNase I digestion in the presence of the protein(s) indicated above. (C) Binding of PecS, CRP, and RNA polymerase. Lanes 1 and 15, control digestion; lanes 2 to 14, DNase I digestion in the presence of the protein(s) indicated above.

Gel shift experiments showed that cAMP-CRP interacts with the pehN promoter region (Fig. 6A). cAMP-CRP binding gave rise to one complex with a K_d of about 20 nM. DNase I footprinting experiments revealed a protected region of 36 bp, including two DNase I-hypersensitive sites and extending from positions −56 to −21 (Fig. 7A, lanes 6 and 7). This region encompassed a sequence homologous to the consensus of the CRP binding site (20) centered at position −41.5 relative to the transcription start site (Fig. 5). Thus, the pehN gene has a class II CRP-dependent promoter, characterized by CRP binding sites situated at position −41.5, so that CRP and RNA polymerase bind on the same face of the DNA helix (6). The transcription of pehN appears to involve a classical direct mechanism of activation by the cAMP-CRP complex.

Gel shift assays demonstrated that PecS interacts specifically with the regulatory regions of pehN (Fig. 6B), with a K_d of about 50 nM. Using DNase I footprinting, a protected area extending from positions −140 to −84 was detected (Fig. 7B, lanes 2 to 4, and Fig. 7C, lanes 2 to 4).

Simultaneous binding of transcription factors in the pehN promoter region.

DNase I footprinting performed in the presence of RNA polymerase alone revealed no clearly protected area (Fig. 7A, lanes 11 and 12). The combination of RNA polymerase and the activation complex cAMP-CRP gave rise to the protection of a wide region, extending from −78 to + 20 (Fig. 7A, lanes 8 to 10), which largely encompassed the promoter and the CRP binding site (Fig. 5). This area was detected at low protein concentrations (10 nM CRP and 50 nM RNA polymerase) which did not allow protection by each individual protein. These results indicate a synergistic binding of RNA polymerase and cAMP-CRP to the pehN promoter.

Since the organization of different proximate regulatory sites allows physical interactions between the regulators, we analyzed the in vitro interactions between the pehN promoter region and pairs of regulators. The addition of subsaturating concentrations of KdgR and CRP in gel shift assays resulted in the formation of several complexes including one, two, or three proteins (data not shown). At saturating concentrations of CRP and KdgR (50 and 200 nM, respectively), we observed only the large complexes including two KdgR molecules and the cAMP-CRP complex (Fig. 6A). Footprinting experiments conducted with KdgR and cAMP-CRP confirmed that the two proteins could interact simultaneously with their respective binding sites on the pehN promoter region (Fig. 7A, lanes 4 and 5).

Footprinting experiments conducted with KdgR, cAMP-CRP, and RNA polymerase showed that the protection of the KdgR1 site was similar to that observed in the presence of KdgR only, while the protection of the region including the KdgR2 site corresponded to that obtained with CRP and RNA polymerase (Fig. 7A, lanes 13 and 14). Thus, in the presence of CRP and RNA polymerase, KdgR interacted only with its high-affinity binding site. This result suggests that the repression exerted by KdgR on the pehN promoter results mainly from its binding to the KdgR1 site.

Gel shift assays performed in the presence of both PecS and KdgR revealed the formation of highly retarded complexes including both PecS and KdgR (Fig. 6B). Thus, these two proteins are able to bind simultaneously to the pehN promoter region. We verified that simultaneous binding of PecS and KdgR did not modify their respective affinities for the pehN promoter (data not shown). The region protected in the presence of KdgR and PecS corresponds to the addition of the regions protected by each protein separately (Fig. 7B). Thus, no competition for the binding of PecS and KdgR, either on the KdgR1 or on the KdgR2 site, was observed on the pehN promoter.

The combination of PecS and cAMP-CRP was also examined by band shift assays (data not shown) and DNase I footprinting (Fig. 7C, lanes 11 and 12). These experiments revealed that PecS and CRP could bind simultaneously to the pehN promoter region without any competition or synergy. DNase I footprinting performed with PecS and RNA polymerase (Fig. 7C, lanes 5, 7, and 8) showed that the presence of PecS has no effect on the binding of RNA polymerase. DNase I footprints were also performed using the factors PecS, cAMP-CRP, and RNA polymerase (data not shown), giving similar results. The presence of PecS did not affect the binding of the cAMP-CRP/RNA polymerase transcription complex.

Analysis of the pehN mutant.

In strain A3377, insertion of the uidA-Km cassette inactivated the pehN gene. Growth of the pehN mutant A3377 was compared to that of the parental strain A350 when polygalacturonate was used as the sole carbon source. For both strains, the final yields of growth were about 10⁹ bacteria per mg of polygalacturonate, and the doubling time during exponential growth was 120 ± 10 min. Thus, mutation of the pehN gene had no visible effect on polygalacturonate utilization.

We compared the maceration provoked by strains A3377 and A350 on chicory leaves and potato tubers. On chicory leaves, the rotted regions observed 24 h after inoculation measured 43 ± 10 mm and 40 ± 11 mm for strains A3377 and A350, respectively. The weight of macerated tissues observed on potato tubers 2 days after inoculation was 590 ± 101 and 663 ± 84 mg, respectively. We also compared the pathogenic behavior of the E. chrysanthemi pehN mutant on potted plants of S. ionantha with that of the wild-type strain. After infection of 30 plants with strains 3937 and A3377, we followed the appearance of symptoms over a period of 10 days. We observed no significant difference in the progress of the disease between the pehN mutant and the wild-type strain (data not shown).

We performed chromosomal localization to determine whether pehN is situated in the vicinity of the pehV-W-X cluster. The Km^r marker of the pehN::uidA fusion cotransferred with the ade-3, thy-1, and gal-1 markers of the E. chrysanthemi chromosome (18). The frequencies of cotransfer between pairs of markers indicated that the pehN gene is located between ade-3 and thy-1. In contrast, the Km^r marker of the pehX::uidA fusion was localized near the markers ile-1 and arg-1. Three-point analysis indicated that pehX is not located between these two markers but in the vicinity of ile-1. Using the φEC2 phage, a cotransduction of 34% between pehX and ile-1 was obtained.

DISCUSSION

In this paper, we describe the isolation of a novel E. chrysanthemi gene which has the characteristics of genes involved in pectin degradation: (i) pehN transcription is controlled by the classical regulators of pectinase genes and (ii) the protein PehN is homologous to polygalacturonases of family 28 of glycosyl hydrolases. Analysis of its enzymatic activity revealed that PehN has a low hydrolase activity with various pectins as the substrate. This low activity suggests that the true substrate of PehN is not homogalacturonan but a related compound. The simplest hypothesis is that PehN recognizes a rare modification in the pectin chain, either the insertion of a different sugar in the chain, such as a single rhamnose residue, or the presence of a modified galacturonate residue, substituted by glycosylation or esterification with a rare constituent.

The fact that PehN acts in synergy with endo-pectate lyases, mainly with PelB and PelC, which prefer oligomers rather than long chains as a substrate (33), confirms that PehN is able to cleave pectin. Moreover, the fact that PehN increases PelX action but not PehX action suggests that the products liberated by PehN have a normal reducing end, accessible to PelX, but a blocked nonreducing end, not accessible to PehX (40). The nonreducing ends liberated by PehN could have some modification on the terminal or penultimate residue (e.g., presence of a rhamnose residue in the chain, of glycosylation, or of esterification). These data are consistent with the hypothesis that PehN specifically cleaves a glycosidic linkage situated in the vicinity of a rare modification of the pectic polymer.

A precise identification of the products resulting from PehN action could help to clarify the PehN specificity. Preliminary experiments suggest that these products are quite long oligomers (larger than the hexamer), since they could not be separated by thin-layer chromatography (data not shown). Analysis of a pehN mutant indicates that, despite its synergy on endo-pectate lyases, PehN did not appear to be essential for E. chrysanthemi virulence. However, the PehN characteristics suggest that its role could be restricted to some tissues which have a special pectin composition.

Expression of the pehN gene was found to be subject to the three main conditions affecting transcription of all pectinase genes: induction in the presence of pectin, growth phase, and catabolite repression (16). The regulatory proteins KdgR and CRP are involved in control of pehN transcription, as observed for the other pectinase genes. The activation of pehN by CRP is due to a cooperative binding between RNA polymerase and the cAMP-CRP complex. A similar effect was observed for the genes pehX and pelD (26, 32). Footprinting experiments demonstrated that KdgR has two binding sites situated in the vicinity of the pehN promoter. KdgR did not inhibit the binding of the transcription machinery, cAMP-CRP and RNA polymerase, but its interaction with the high-affinity site centered at position +11 relative to the transcription start might inhibit RNA polymerase progress. The KdgR binding sites usually overlap the −35 region of the promoter of the controlled genes, but a more distal position was also observed in the case of the pelD and peh genes of E. chrysanthemi (26, 32). Such a position could allow a more rapid transcription when derepression occurs, since the transcription machinery is already correctly situated to initiate transcription.

In addition, pehN transcription is subject to controls specific to the pehV, pehW, and pehX genes, inhibition by Ca²⁺ and activation by the regulator PecS (26). The inhibition by Ca²⁺ addition does not depend on the known regulators KdgR, PecS, or CRP, since the calcium effect is retained in kdgR, pecS, and crp mutants (data not shown). The activation of pehX by PecS was shown to result from competition with the KdgR repressor for the occupation of overlapping sites (26). The mechanism is different in the case of pehN, since we observed that PecS and KdgR binding is independent. We verified that the activator effect of PecS did not originate from cooperative binding with an element of the transcription machinery, cAMP-CRP, or RNA polymerase. Thus, it is tempting to speculate that the activation of pehN by PecS could result from interference with the binding of an unidentified regulator. This regulator could also act on the promoter regions of the pehV, pehW, and pehX genes, which are coregulated with pehN.

In the related bacterium E. carotovora, the endopolygalacturonase gene pehA is also calcium regulated (9). This gene is specifically controlled by the two-component regulatory system PehR-PehS, which mediates calcium regulation (8). A similar control by a PehR homologue could exist in E. chrysanthemi, and PecS binding could interfere with PehR binding.

Comparison of the PecS-protected regions of the four positively regulated genes (pehN, pehV, pehW, and pehX) revealed the presence of a 17-nucleotide conserved sequence, corresponding to the consensus TTTTTTCCGRATRAAAC (where R = G or A) (Fig. 5). This sequence does not correspond to the PecS binding site, since no homology with this consensus is found in the PecS-protected region of the other PecS-controlled genes. Nevertheless, this sequence could correspond to the binding site of a regulator specific to the four peh genes, and its position implies competition with PecS for binding (Fig. 5). In the case of pehV, pehW, and pehN, for which the KdgR binding site is situated outside the PecS-protected region, the effect of PecS could be limited to interference with the specific peh regulator. In the case of pehX, PecS could compete for binding both with KdgR and with the specific peh regulator. The identification of this potential regulator is now necessary to clarify the mechanism of PecS activation on peh transcription.