Comparative genomics of the KdgR regulon in Erwinia chrysanthemi 3937 and other gamma-proteobacteria

Dmitry A. Rodionov1, Mikhail S. Gelfand1,2 and Nicole Hugouvieux-Cotte-Pattat3

1 State Scientific Centre GosNIIGenetika, Moscow, 117545, Russia
2 Institute for Problems of Information Transmission, Russian Academy of Sciences, Bolshoy Karetny per. 19, Moscow GSP-4, 127994, Russia
3 Unité de Microbiologie et Génétique – Composante INSA, UMR CNRS-INSA-UCB 5122, bat Lwoff, 10 rue Dubois, Domaine Scientifique de la Doua, 69622 Villeurbanne Cedex, France

Correspondence
Dmitry Rodionov
rodionov{at}genetika.ru


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
In the plant-pathogenic enterobacterium Erwinia chrysanthemi, almost all known genes involved in pectin catabolism are controlled by the transcriptional regulator KdgR. In this study, the comparative genomics approach was used to analyse the KdgR regulon in completely sequenced genomes of eight enterobacteria, including Erw. chrysanthemi, and two Vibrio species. Application of a signal recognition procedure complemented by operon structure and protein sequence analysis allowed identification of new candidate genes of the KdgR regulon. Most of these genes were found to be controlled by the cAMP-receptor protein, a global regulator of catabolic genes. At the next step, regulation of these genes in Erw. chrysanthemi was experimentally verified using in vivo transcriptional fusions and an attempt was made to clarify the functional role of the predicted genes in pectin catabolism. Interestingly, it was found that the KdgR protein, previously known as a repressor, positively regulates expression of two new members of the regulon, phosphoenolpyruvate synthase gene ppsA and an adjacent gene, ydiA, of unknown function. Other predicted regulon members, namely chmX, dhfX, gntB, pykF, spiX, sotA, tpfX, yeeO and yjgK, were found to be subject to classical negative regulation by KdgR. Possible roles of newly identified members of the Erw. chrysanthemi KdgR regulon, chmX, dhfX, gntDBMNAC, spiX, tpfX, ydiA, yeeO, ygjV and yjgK, in pectin catabolism are discussed. Finally, complete reconstruction of the KdgR regulons in various gamma-proteobacteria yielded a metabolic map reflecting a globally conserved pathway for the catabolism of pectin and its derivatives with variability in transport and enzymic capabilities among species. In particular, possible non-orthologous substitutes of isomerase KduI and a new oligogalacturonide transporter in the Vibrio species were detected.


Abbreviations: CRP, cAMP-receptor protein; DK-I, 5-keto-4-deoxyuronate; DK-II, 2,5-diketo-3-deoxygluconate; GA, galacturonate; KDG, 2-keto-3-deoxygluconate; MCP, methyl-accepting chemotaxis protein; OGA, oligogalacturonate; PGA, polygalacturonate


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Pectin is one of the signals detected by Erw. chrysanthemi that indicates the presence of plant tissues, triggering induction of pectinolysis and possibly of other associated virulence factors. Investigation of gene expression demonstrated that transcription of all genes involved in pectin catabolism is induced in the presence of pectin or its derivatives, such as polygalacturonate (PGA) and galacturonate (GA) (Hugouvieux-Cotte-Pattat et al., 1992). This included genes encoding various types of pectinases, i.e. pectin acetylesterases (paeX, paeY), pectin methylesterases (pemA, pemB), pectate lyases (pelA, pelB, pelC, pelD, pelE, pelI, pelL, pelW, pelZ, pelX) and polygalacturonases (pehN, pehV, pehW, pehX), and also proteins necessary for secretion of pectinases (the outCM operon), transporters of pectic oligomers (kdgM, togT, togMNAB) and intracellular enzymes involved in the cleavage of dimers (ogl) and the catabolism of unsaturated monomers (kduI, kduD, kdgK) (for a review, see Robert-Baudouy et al., 2000).

Induction of several genes in the presence of pectin indicated a co-ordinated regulation of these genes. KdgR, which belongs to the IclR family of transcriptional regulators, has been characterized as being responsible for this regulation (Nasser et al., 1992). In a kdgR mutant, the expression of all these genes, except pelL, increased, indicating that they are repressed in vivo by KdgR. In vitro analysis demonstrated that KdgR directly interacts with the promoter regions of the in vivo-controlled genes/operons (Nasser et al., 1994). KdgR-binding sites usually overlap with or are close to the promoters. These observations suggest that the KdgR protein and the RNA polymerase compete for adjacent binding sites on DNA, explaining how KdgR binding prevents gene expression. Physiological and biochemical studies indicated that an intermediate of pectin catabolism, 2-keto-3-deoxygluconate (KDG), is the main inducing molecule which interacts in vivo and in vitro with KdgR, provoking dissociation of KdgR from its operators. In vivo data indicate that two other unsaturated monomers formed during pectin catabolism, 5-keto-4-deoxyuronate (DK-I) and 2,5-diketo-3-deoxygluconate (DK-II), are also able to act as inducers by interaction with KdgR. From current data, the KdgR repressor directly controls at least 13 operons that constitute the KdgR regulon (Hugouvieux-Cotte-Pattat et al., 1996). A genetic screen using lacZ transcriptional fusions indicated that as much as 1 % of the Erw. chrysanthemi genes (about 50 genes) could be induced in the presence of pectin (Hugouvieux-Cotte-Pattat & Robert-Baudouy, 1989). Identification of some of these genes has confirmed the presence of known pectinase genes, but has also revealed new pectin-inducible loci, such as the recently identified rhiTN operon involved in the catabolism of a pectin-related plant polysaccharide, rhamnogalacturonan (Hugouvieux-Cotte-Pattat, 2004). These observations show that KdgR has a wide range of targets and its role may not be restricted to pectinolysis.

KdgR homologues were also identified in other plant-pathogenic enterobacteria, Erwinia carotovora subsp. carotovora, Erw. carotovora subsp. atroseptica and Erw. amylovora (Liu et al., 1999; Thomson et al., 1999). In animal-related enterobacteria, such as Escherichia coli, the action of KdgR was found to be restricted to the control of expression of genes involved in the catabolism of KDG (kdgK, kdgA and kdgT). These results indicate that KdgR is a regulatory protein conserved in the Enterobacteriaceae. Moreover, the KdgR proteins appeared to be functionally interchangeable between species. For instance, KdgR of E. coli is able to repress in vivo transcription of the Erw. chrysanthemi pelD gene (James & Hugouvieux-Cotte-Pattat, 1996), whereas KdgR of Erw. chrysanthemi is able to bind in vitro to regulatory regions of pectinase genes from Erw. carotovora and vice versa (Thomson et al., 1999). Thus, conservation of the KdgR regulator is accompanied by conservation of its specific binding signals.

Comparative genomics is a powerful approach for the prediction of gene regulation and the annotation of the bacterial genome (Gelfand et al., 2000; Gelfand, 2003). Previous in silico analysis of the KdgR regulon revealed several novel KdgR-regulated genes in gamma-proteobacteria (Rodionov et al., 2000), such as the predicted oligogalacturonide transporter OgtABCD, which was confirmed in an independent experimental study to have the proposed function (renamed as TogMNAB) and to be regulated in vivo by KdgR in Erw. chrysanthemi (Hugouvieux-Cotte-Pattat et al., 2001). Recent availability of many complete genomes of enterobacteria, including the two plant pathogens Erw. carotovora (http://www.sanger.ac.uk/Projects/E_carotovora) and Erw. chrysanthemi (http://www.tigr.org/tdb/mdb/mdbinprogress.html), provides an opportunity to perform more detailed comparative analysis of the KdgR regulon in a variety of bacteria. This has allowed us to identify a large number of new KdgR regulon members. A complete description of the KdgR regulon in enterobacteria and Vibrio species has revealed the main differences in the pectin and KDG degradation pathways in these bacteria. We took advantage of the genetic tools and knowledge obtained in Erw. chrysanthemi to validate some of the data resulting from this comparative analysis.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Datasets and genomic sequences.
Complete genomic sequences of E. coli K-12 and CFT073, Salmonella typhimurium, Yersinia pestis, Vibrio vulnificus and Vibrio parahaemolyticus with their annotation files were obtained from GenBank (Benson et al., 2000). Unannotated contig sequences of Erwinia chrysanthemi and Klebsiella pneumoniae were downloaded from the websites of the Institute for Genomic Research (www.tigr.org) and the Washington University Consortium (www.genome.wustl.edu), respectively. The Erwinia carotovora subsp. atroseptica and Yersinia enterocolitica complete sequence data were produced by the respective sequencing groups at the Sanger Institute and were obtained from ftp://ftp.sanger.ac.uk/pub/.

Identification of DNA-binding motifs.
All previously characterized KdgR-binding sites in Erw. chrysanthemi were collected from the literature (Hugouvieux-Cotte-Pattat et al., 1996). The KdgR search profile was constructed using an alignment of these known sites. Positional nucleotide weights in this profile were derived using the following formula:

{mic1503571E001}
where N(b, k) is the count of nucleotide b in position k in the training sample of aligned sites. The consensus for the KdgR sites is the 21 bp sequence WAWTRAAAYRnYRTTTYAWTW. The score of a candidate site was calculated as the sum of the positional nucleotide weights:

{mic1503571E002}
where L=21 is the length of the KdgR signal. The site score defined by this formula is linearly related to the discrimination energy and can be used to assess the significance of individual sites (Mironov et al., 1999). Then, each genome was scanned with the KdgR profile, and genes with candidate regulatory sites in upstream regions (normally in positions –300 to +50 relative to the translation start) were selected. The cut-off score for putative KdgR-binding sites in closely related enterobacteria was defined as a lowest score within the training set (5·20). In the case of more distant Vibrio species, the KdgR search profile was derived from a set of upstream regions of orthologous KDG genes using the SignalX program (Mironov et al., 2000). The cut-off score for candidate KdgR sites in these genomes was 5·0. To account for possible operon structures, the resulting set of candidate regulon members was supplemented by genes that are likely to be co-transcribed with genes preceded by candidate sites (with an intergenic distance less than 100 bp).

The recognition profiles for the catabolic regulatory proteins CRP (cAMP receptor protein) and FruR were constructed using the same procedure and training sets of 70 known CRP-binding sites and 12 known FruR-binding sites were collected from the literature (data not shown). Consensus sequences for the CRP and FruR sites are WWWTGTGATNNNNATCACAWWW and GCTGAAWCGWTTCAGC, respectively. The search profile for RhaS sites was kindly provided by O. Laikova (Gelfand & Laikova, 2003).

Other computer programs.
The signal recognition procedure and the Smith–Waterman alignment of protein sequences were performed using the Genome Explorer program (Mironov et al., 2000). Orthologous genes in studied gamma-proteobacteria were identified by the bidirectional best hits criterion (Tatusov et al., 2000). Additional protein sequence comparisons and search of distant homologues in protein databases were performed using gapped BLASTP and PSI-BLAST programs (Altschul et al., 1997). If necessary, orthologous or paralogous relationships of proteins were confirmed by construction of phylogenetic trees. The phylogenetic trees were constructed by the maximum-likelihood method implemented in PHYLIP (Felsenstein, 1981). Multiple protein sequence alignments were constructed by CLUSTALX (Thompson et al., 1997). Potential transmembrane segments and signal peptide cleavage sites were predicted using the TMpred (www.ch.embnet.org/software/TMPRED_form.html) and SignalP (www.cbs.dtu.dk/services/SignalP/) servers, respectively (Hofmann & Stoffel, 1993; Nielsen et al., 1997).

Strains, media and growth conditions.
The bacterial strains of Erw. chrysanthemi and the plasmids used in this study are listed in Table 1. The Phi-EC2 generalized transducing phage was used for transduction (Resibois et al., 1984). Erw. chrysanthemi cells were grown at 30 °C in M63 medium (Miller, 1972). Carbon sources, namely glycerol, GA and PGA, were added at 2 g l–1. E. coli cells were grown at 37 °C in LB medium (Miller, 1972). The media were solidified with agar (15 g l–1). 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.


View this table:
[in this window]
[in a new window]
 
Table 1. Bacterial strains, plasmids and oligonucleotides used in this study

 
Chemotaxis was measured by determining the size of haloes observed on semi-solid agar plates containing 4 g agar l–1 in M63 medium (Hugouvieux-Cotte-Pattat et al., 2001). Oligogalacturonides were prepared by degradation of PGA with pectate lyases (Hugouvieux-Cotte-Pattat et al., 2001). For these experiments, 0·2 mM glycerol was added as the carbon source and attractants were used at a final concentration of 1 mM. The diameters of the chemotactic rings were measured after incubation for 24 h at 30 °C.

Recombinant DNA techniques.
Preparation of plasmid or chromosomal DNA, restriction digestions, ligations, DNA electrophoresis and transformations were carried out as described by Sambrook et al. (1989).

PCR primers were designed (24- to 28-mers, Table 1) to clone 0·7–1·3 kb chromosomal DNA containing the entire gene or its 5' end; restriction sites were added at each end to determine the orientation of the DNA insertion in the vector (BamHI or BglII at the 5' end and XbaI at the 3' end). Strain 3937 chromosomal DNA was used as the template. The PCR products were purified (QIAquick PCR purification kit; Qiagen) and directly ligated to the pGEMR-T vector (Promega) which has a protruding T nucleotide at each 3' end.

Genetic fusions were constructed on the cloned genes, by insertion of uidA-Km cassettes (Bardonnet & Blanco, 1992) into a restriction site situated inside the corresponding ORF (Table 1). The orientation of the uidA-Km cassette was determined by restriction analysis. Only plasmids in which uidA and the mutated gene have the same transcriptional direction were retained. Plasmids bearing the uidA-Km insertion were then introduced into Erw. chrysanthemi cells by electroporation. The insertions were integrated into the Erw. chrysanthemi chromosome by marker exchange recombination after successive cultures in low phosphate medium supplemented with Km (Roeder & Collmer, 1985). After verification of the correct recombination of the uidA-Km insertions by PCR, {beta}-glucuronidase activity was measured to estimate the expression of the fused gene. The degradation of p-nitrophenyl-{beta}-D-glucuronide into p-nitrophenol, was followed at 405 nm. Specific activity is expressed as nmol products liberated min–1 (mg bacterial dry wt)–1.


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Characterization of the KdgR regulons in enterobacteria and Vibrio species
Existence of the KdgR orthologue is a prerequisite to the comparative analysis of the KdgR regulons in bacteria. Based on the phylogenetic tree of the IclR family homologues from various bacteria (Table 2 and data not shown), we identified KdgR in all studied enterobacteria: two Erwinia species, Erw. chrysanthemi (ER) and Erw. carotovora (EO), two Yersinia species, Y. pestis (YP) and Y. enterocolitica (YE), K. pneumoniae (KP), S. typhimurium (ST) and E. coli (EC), and in two Vibrio species, V. vulnificus (VV) and V. parahaemolyticus (VP). For EC, strain CFT073 was also considered since it contains more KdgR-controlled genes than strain K-12. A high degree of sequence conservation in the KdgR proteins implies conservation of KdgR-binding signals in all considered species of enterobacteria. Known ER KdgR-binding sites were collected from previous studies and comprised the training set for a 21-bp recognition profile (Fig. 1a). Then, the KdgR profile was used to search for new candidate KdgR-binding sites in the genomes of ER and other enterobacteria. Table 3 lists both previously known and newly identified KdgR-binding sites.


View this table:
[in this window]
[in a new window]
 
Table 2. Percentage identity between the KdgR proteins from enterobacteria and Vibrio sp.

 


View larger version (30K):
[in this window]
[in a new window]
 
Fig. 1. Sequence logos for the KdgR-binding sites in enterobacteria and Vibrio spp. (a) The KdgR motif drawn from a training set of known Erw. chrysanthemi sites; (b) the most significant motif obtained by the signal determination procedure for Vibrio spp.

 

View this table:
[in this window]
[in a new window]
 
Table 3. Predicted KdgR regulon in enterobacteria and Vibrio spp.

Genes marked with an asterisk (*) were named in this study. Divergently located genes are separated by /. Lower case letters in the site sequences indicate positions that do not conform to the consensus. Site scores lower than 5·20 for enterobacteria and lower than 5·00 for Vibrio sp. are underlined and correspond to weak sites. The table contains all candidate KdgR sites with a score higher than the respective cut-offs, and also several weak sites that either are conserved in other species or precede pectin degradation genes. The last column represents the experimental data on regulation: S, in vivo and in vitro functional KdgR-binding sites; R, previously known in vivo regulation by KdgR; EC, regulation by KdgR confirmed by experiments in this study; EN, regulation by KdgR could not be confirmed.

 
Since KdgR orthologues detected in the Vibrio species are less similar to KdgR from enterobacteria (Table 2), we tried to construct a more specific profile of the KdgR-binding sites in VV and VP. Towards this aim, we selected the regions upstream of the orthologues of the ER KdgR-regulated genes in the genomes of VV and VP. As a result, a common 21-bp palindromic signal highly similar to the KdgR site from enterobacteria was obtained (Fig. 1b) and was used for identification of new members of the KdgR regulon in both Vibrio species (Table 3). In contrast to the VV and VP genomes, orthologues of kdgR and of genes involved in pectin/KDG catabolism were not detected in the genome of Vibrio cholerae, arguing for possible recent loss of the complete KdgR regulon in this highly pathogenic bacterium.

Almost all previously known KdgR-regulated genes in ER are subject to catabolic repression by glucose moderated through the CRP (Reverchon et al., 1997). To test whether this global CRP regulation is conserved for other members of the KdgR regulon, we scanned all studied genomes with the CRP profile. The online version of this paper (at http://mic.sgmjournals.org) contains a supplementary table showing the list of all candidate CRP sites found upstream of the KdgR-controlled genes. This analysis suggests that the majority of the pectin degradation and utilization genes in gamma-proteobacteria are under dual regulation by KdgR and CRP (Fig. 2). Moreover, the relative positions of the candidate CRP and KdgR-binding sites agree with the known antagonistic effect of CRP and KdgR on the expression of the pectinolytic genes in ER (Nasser et al., 1997).



View larger version (53K):
[in this window]
[in a new window]
 
Fig. 2. Operon structures and predicted regulatory sites for the KdgR regulons in enterobacteria and Vibrio species.

 
The operon structures of the KdgR-regulated genes and the location of potential KdgR and CRP sites are summarized in Fig. 2. The predicted KdgR regulons of two Erwinia species are particularly large, containing 31 operons for ER and 29 operons for EO (Fig. 2). These regulons include a variety of extracellular pectinolytic enzymes (Pel, Peh, Pem and Pae), a secretion system (Out), several porins (KdgM) and transport systems (TogMNAB, TogT, KdgT, etc.), as well as enzymes for the intracellular catabolism of dimers and monomers (Ogl, KduI, KduD and KdgK). This set of genes agrees with the ability of these plant-pathogenic bacteria to degrade plant pectin and to use the resulting oligomers and monomers as a carbon source for growth. Among other bacteria considered in this study, only the Yersinia and Vibrio species possess pectinolytic enzymes, which are probably periplasmic since these bacteria lack orthologues of the Out system. These bacteria also contain cytoplasmic enzymes initially identified in Erwinia (Fig. 2). Thus, Yersinia and Vibrio are predicted to degrade pectic oligomers entering the periplasm via the KdgM porins, as well as unsaturated monomers. Considering the predicted composition of the KdgR regulon, KP should be able to degrade only short oligomers, dimers or trimers, by cytoplasmic enzymes (Fig. 2). ST and EC strain K-12, with only six predicted members of the KdgR regulon, appear to be able to degrade the unsaturated monomers DK-I, DK-II and KDG, but not pectic oligomers. Notably, EC strain CFT073 has acquired a cluster of genes allowing it to use short oligomers (Fig. 2). Thus, the kduI, kduD and kdgK genes involved in the degradation of unsaturated monomers constitute the core of the KdgR regulon conserved in all studied bacteria with the exception of kduI, which is absent in KP and the Vibrio species (possibly replaced by a non-orthologous enzyme, see below).

KdgR-regulated transporters
In ER, extracellular oligogalacturonides resulting from the pectin degradation first enter the periplasm by the specific porin KdgM (Blot et al., 2002) and then cross the inner membrane using either an ABC transporter, TogMNAB or a GPH transporter, TogT (Hugouvieux-Cotte-Pattat & Reverchon, 2001). Genes required for the transport of oligogalacturonides (kdgM, togMNAB and togT) and their subsequent degradation to monomeric acid sugars in the cytoplasm (pelW and ogl) are present only in the Erwinia, Yersinia, Klebsiella and Vibrio species. Erwinia and Yersinia have from two to four homologues of the oligogalacturonate (OGA)-specific porin KdgM (Fig. 3). A close paralogue of kdgM in ER (kdgN) is preceded by a strong KdgR site. We have not detected candidate KdgR sites upstream of the kdgM gene in ER, although it was previously shown to be under KdgR regulation (Blot et al., 2002). In addition to kdgM and kdgN, EO has two more homologues of OGA-specific porin located in one cluster with the periplasmic pectate lyase gene pelP. The upstream region of the possible kdgM3-kdgM4-pelP operon contains a candidate KdgR site. In both Yersinia species, there are two KdgR-regulated paralogues of the kdgM gene; the first one is located immediately after the pelW-togMNAB cluster, whereas the second one belongs to the possible kdgM2-pelP-sghX operon. Notably, the sghX gene (see YPO3993 for reference), encoding a hypothetical secreted protein weakly similar to various glycosyl hydrolases, was also found as a single KdgR-regulated gene in EO. In addition, the kdgK-kdgA cluster in VV contains the pelP-sghX pair with a candidate KdgR site upstream. Orthologues of SghX were not found in other bacterial genomes. This genetic organization suggests that the function of SghX, a new member of the KdgR regulon that has a N-terminal signal sequence, may be closely related to the periplasmic pectate lyase PelP. KdgR-regulated kdgM homologues were also identified in KP and ST. However, ST lacks both known oligogalacturonide transport systems, TogT and TogMNAB. We propose rhamnogalacturonide specificity for the porin encoded by the KdgR-regulated kdgM homologue in ST, since this gene is located between genes possibly involved in rhamnogalacturonide transport (see below) and genes involved in the L-rhamnose catabolism (Fig. 2).



View larger version (33K):
[in this window]
[in a new window]
 
Fig. 3. Phylogenetic tree of bacterial orthologues of OGA-specific porin KdgM from Erw. chrysanthemi. Proteins are denoted by their genome abbreviations (listed in Methods). Multiple gene paralogues are numbered. Genes predicted to be regulated by KdgR are boxed.

 
Both TogMNAB and TogT, the oligogalacturonide transporters, are present in enterobacteria that are able to degrade pectic oligomers, namely Erwinia, Yersinia and Klebsiella. Only togT is present in the KdgR-regulated locus of the EC strain CFT073, allowing it to transport such oligomers. In contrast, the complete genomes of both Vibrio species, VV and VP, lack orthologues of both known OGA transport systems. In Vibrio, most KdgR-regulated genes are organized in one locus (Fig. 2) encoding porin KdgM, periplasmic pectate lyase PelX and cytoplasmic enzymes for the catabolism of OGAs (Ogl, KduD, KdgK, KdgA, etc.). This locus also contains two highly similar genes encoding hypothetical transporters with 14 candidate transmembrane segments (see VVA1379 and VVA1382). The closest homologues of these two genes are sodium : glucose co-transporters from Eukaryota. To fill the metabolic gap, we tentatively assigned OGA specificity to these two transporters and named them TogX1 and TogX2.

Among enterobacteria, only Erwinia species and EC have the kdgT gene, which is a member of the KdgR regulon encoding a transporter for KDG, DK-I and DK-II (Condemine & Robert-Baudouy, 1987). Interestingly, the kdgT gene of EO is located in one putative operon with a kduI paralogue (72 % similarity) and the rexZ gene encoding a regulator of exoenzyme production (Thomson et al., 1999). In an attempt to identify the apparently missing KDG permease in other enterobacteria, we detected a candidate KdgR-regulated gene, named kdgX (previous ST name yifZ), which is present in Yersinia species, ST and EO (Fig. 2). KdgX has nine predicted transmembrane segments and belongs to the Drug/Metabolite transporter family. One characterized member of this family, RhaT of EC, functions in sugar uptake. Since all identified kdgX genes are preceded by candidate KdgR- and CRP-binding sites, we proposed that the specificity of the KdgX transporter is similar to that of KdgT in ER and EC.

The KdgR-regulated pectinases
Erwinia species possess a variety of extracellular pectinolytic enzymes, most of which are controlled by KdgR (Fig. 2). Both ER and EO contain single pectinase genes pelI, pelX, pehN and pemB. Two intracellular pectinase genes, pelW and paeX, are included in the large gene cluster encoding transporters and enzymes for OGA catabolism. There are several remarkable differences between the two Erwinia species concerning the arrangement of pectinolytic genes. The protein export system for secretion of extracellular pectinases is encoded by the out gene cluster and regulated by KdgR in ER (Condemine et al., 1992). The out cluster in EO is also predicted to belong to the KdgR regulon and it includes the polygalacturonase gene pehX and a new pectate lyase gene, named pelF, which is most similar to the pectate lyase gene pel from Bacillus subtilis (Nasser et al., 1993). While EO has only one pehX gene in the out cluster, the pehX gene of ER forms a cluster with two close paralogues, pehV and pehW, suggesting recent gene duplication in ER. The pectin acetylesterase gene paeY and the methylesterase gene pemA in ER are located in a KdgR-regulated operon with the pectate lyase gene pelD, whereas EO has a possible operon, paeY-pemA, with two upstream KdgR sites. It is noticeable that EO has no orthologue for pelD while the same ER cluster has two pelD paralogues, pelA and pelE, again suggesting gene duplication in ER. Finally, the pelBCZ gene cluster of ER contains two orthologous pectate lyase genes, pelB and pelC, preceding a weakly similar gene, pelZ. The same gene cluster of EO includes three close homologues of pelB/C, named pel1, pel2 and pel3, and a pelZ orthologue. This organization suggests that an ancestor of the pelB/C genes was subject to one duplication event in ER and two duplications in EO. Duplication of pectinase genes seems to be a common phenomenon in Erwinia species and could favour adaptation of these pathogenic bacteria to various plant tissues.

Among other analysed bacteria, only Yersinia and Vibrio species have several pectinolytic enzymes (PelX, PehX, PelP, PemA, PaeX and possibly SghX), which also are members of the KdgR regulon. Although the Out-dependent secretion system is absent in these species, all these proteins contain candidate N-terminal signal sequences, arguing for their periplasmic location.

Other KdgR-regulated genes
A search for candidate KdgR-binding sites in bacterial genomes complemented by operon structure analysis allowed us to detect a number of new members of the KdgR regulon (Table 3). In ER, genes ygjV, tpfX, chmX, ppsA/ydiA (divergent genes), gntD2, yeeO, spiX, yjgK and dhfX are predicted to have strong KdgR-binding sites, i.e. sites with scores higher than 5·20. Most of these genes encode hypothetical proteins of unknown function. The ppsA product was previously characterized as a phosphoenolpyruvate synthase (Niersbach et al., 1992), but its potential KdgR regulation has yet to be described. At this stage, we performed experimental verification of the predicted regulation for each novel candidate member of the KdgR regulon in ER prior to clarification of their role in pectin catabolism by detailed functional, positional and phylogenetic analysis of these genes.

We observed that the presence of a weak KdgR-binding site (score below 5·2, Table 3) can also have a biological significance. Indeed, some previously described genes, namely pehX, pehW, pelI, rhiTN and pehV, that are known to be controlled by KdgR, have KdgR sites with scores below cut-off (4·79, 4·64, 4·62, 4·37 and 4·35, respectively). Thus, we also tested potential KdgR regulation of genes that have a KdgR site conserved in other bacteria (gntDBMNAC, pykF and sotA) or whose function could be related to pectin catabolism or plant infection (indA, pecT, pir and expI).

Construction of uidA transcriptional fusions in novel members of the KdgR regulon
To analyse expression of the selected ER genes, namely chmX, dhfX, gntB, gntD, gntD2, ppsA, pykF, sotA, spiX, tpfX, ydiA, yeeO, ygjV and yjgK, we constructed transcriptional fusions by inserting a uidA-Km cassette into a selected restriction site located in the corresponding ORF. For the genes indA, pecT, pir and expI, we used previously constructed fusions (Table 1). Fourteen selected genes were cloned after PCR amplification (only their 5' end for genes longer than 1 kb). The uidA-Km cassette was inserted into a restriction site situated in the coding region of each independently cloned gene (Table 1). Insertion of the cassette in the correct orientation generates a transcriptional fusion with the uidA gene encoding {beta}-glucuronidase (Bardonnet & Blanco, 1992).

Plasmids containing the insertions were introduced into ER strain 3937. Transformants were submitted to successive cultures in Km-containing low-phosphate medium lacking the antibiotic to which resistance is encoded on the plasmid. Chromosomal uidA-Km insertions were obtained for the 11 genes chmX, dhfX, gntB, ppsA, pykF, sotA, spiX, tpfX, ydiA, yeeO and yjgK. In each case, the correct insertion of the cassette into the chromosome was confirmed by PCR. Attempts to obtain recombination of the insertions in three genes, yjgV, gntD and gntD2, were unsuccessful. Mutations in these genes could be deleterious for the bacterial growth, preventing their isolation.

The growth of the 11 mutants was analysed using GA or PGA as sole carbon source. None of the mutants was affected for growth rate or the final growth yield of cultures with these compounds (data not shown). Thus, genes chmX, dhfX, gntB, ppsA, pykF, sotA, spiX, tpfX, ydiA, yeeO and yjgK are dispensable for PGA or GA catabolism. Since the chmX product could be involved in chemotaxis, we tested chemotactic ability towards GA and oligogalacturonides by determining the diameter of the rings observed on semi-solid agar plates. The size of the rings observed for the chmX mutant and parental strain A350 were similar when the medium was supplemented either with GA or oligogalacturonides (data not shown). This result suggests either that ChmX is not involved in chemotaxis towards GA and oligogalacturonides, or that ER possesses additional chemotaxis receptor proteins that are also specific to these compounds. Identification of two additional chmX homologues preceded by candidate KdgR sites (see below) reinforces the second hypothesis, arguing for the existence of at least three KdgR-regulated proteins that could be involved in chemotaxis towards pectic oligomers or monomers.

Expression of uidA transcriptional fusions in the candidate members of the KdgR regulon
We tested expression of the constructed fusions in the presence of various carbon sources and in different ER genetic backgrounds. The basal level of expression was determined in the presence of glycerol as carbon source. GA and PGA were used as potential inducing compounds. The fusions were transduced into ER strains A1077 and A576 that contain a mutation in kdgR and kdgK, respectively. The kdgR mutation allows for the direct determination of in vivo regulation of the fusion by KdgR. In the kdgR mutant A1077, expression of genes negatively controlled by KdgR, such as those encoding pectate lyases, increased in the absence or presence of the inducer. In the presence of PGA or GA, the kdgK mutation allows for accumulation of the KdgK substrate, KDG, which is the intracellular inducer interacting with KdgR. Accumulation of KDG in a kdgK mutant leads to a very high induction of genes controlled by the KdgR/KDG couple. For instance, the pectate lyase activity greatly increased in the kdgK mutant A576 in the presence of either PGA or GA. Thus, each fusion was assayed in the wild-type background, in the presence of a kdgR or kdgK mutation, in media supplemented with glycerol, GA or PGA. Based on the expression profiles of the tested genes, four classes of genes were defined (Fig. 4).



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 4. Expression of transcriptional fusions in the predicted members of the KdgR regulon. Strains containing a uidA transcriptional fusion in one of the selected genes were grown in minimal medium containing either glycerol (no inducer, white bars) or GA (hatched bars) or PGA (black bars) as a potential inducer. Each fusion was introduced in a wild-type background (WT) and in mutants affected either for the gene kdgK, accumulating the intracellular inducer KDG, or for the regulatory gene kdgR. The {beta}-glucuronidase activities reported are the means of three to five independent experiments and standard deviations are indicated.

 
Seven genes of class 1, namely chmX, dhfX, gntB, pykF, tpfX, yeeO and yjgK, were moderately induced by GA or PGA in the wild-type background, with induction ratios of two- to sixfold. The induction ratios of chmX, dhfX and ygjK clearly increased in the kdgK mutant, reaching 10- to 30-fold. Moreover, they showed a highly derepressed expression in the kdgR mutant, with a {beta}-glucuronidase activity 16- to 26-fold higher than that of the wild-type in the uninduced medium (Fig. 4). The expression of gntB, pykF, tpfX and yeeO remained moderately induced in a kdgK mutant and was derepressed in the kdgR mutant by factors of 3, 20, 6 and 5, respectively. Thus, KdgR clearly represses expression of chmX, dhfX, gntB, pykF, tpfX, yeeO and yjgK. Most previously described genes of the KdgR regulon showed similar results (data not shown) and, on the basis of their expression, the following genes could be considered as members of class 1: kdgT, kduI-kduD, togT, pelW-togMNAB and pelX. It was noticed that all genes of class 1, except pykF and yeeO, have strong KdgR sites with scores between 6 and 5·2 (Table 3). However, some genes with high scoring, mainly extracellular pectate lyase genes pelE, pelA, pelC-pelZ, pelB and pelD (scores between 5·6 and 5·2), do not belong to this class. It is known that expression of the pel genes is controlled by a set of regulators (KdgR, PecS, PecT, CRP, etc.) and the direct effect of KdgR is probably modulated by competition between the regulatory proteins for binding to adjacent sites.

Two genes constituting class 2, spiX and sotA, were not significantly induced by GA or PGA in the wild-type background. However, their transcription was stimulated three- to fivefold in the kdgK mutant in the presence of GA or PGA. Transcription of these two genes was also derepressed in the kdgR mutant; the fusion expression increased about threefold compared to the wild-type strain under non-inducing conditions. These data indicate that spiX and sotA are weakly controlled by the KdgR repressor. The scores of the predicted KdgR sites of spiX and sotA are 5·3 and 4·6, respectively. Previously characterized members of class 2, i.e. genes moderately controlled by KdgR, include pehX, pehW, pehV and rhiT-rhiN (scores between 4·8 and 4·3). These genes are known to be only partially regulated by KdgR and other regulators are involved in their expression; for instance, RhaS is the main activator of rhiT-rhiN transcription (Hugouvieux-Cotte-Pattat, 2004). Additional unidentified regulators could have a major role in the control of spiX and sotA transcription.

Two divergently transcribed genes, ppsA and ydiA, share a common regulatory region containing a strong predicted KdgR-binding site (score 5·6, Table 3). We observed an entirely different expression profile for these genes (Fig. 4) and assigned them to class 3. Both genes were weakly repressed in the presence of GA or PGA in the wild-type background, with repression ratios of about two. This repression became more visible in the kdgK mutant, reaching three- to sixfold. Moreover, both genes showed decreased expression in the kdgR mutant. These results are typical for positive regulation rather than negative control. Thus, KdgR behaves as an activator of ppsA and ydiA expression. Intracellular formation of KDG during pectinolysis provokes dissociation of KdgR from its operators and in such conditions expression of ppsA and ydiA decreases. There are several examples of regulators that could be either activator or repressor (the most classical one being CRP), but this double function was not previously observed in the case of KdgR.

We analysed the role of KdgR in the expression of transcriptional fusions in indA, pecT, pir and expI that have candidate KdgR sites with scores of 4·95, 4·84, 4·76 and 4·65, respectively. The fusions were transduced into either kdgR or kdgK mutant strains and their expression was measured under non-inducing or inducing conditions. The pecT gene appeared to be weakly positively regulated by KdgR and thus could be considered as a member of class 3 (Fig. 4). In contrast, the great variability of expression of indA or expI did not allow us to observe a significant reproducible effect of a kdgR mutation on their expression (data not shown). Expression of the pir gene was clearly independent of KdgR (data not shown). In this case, the site detected as a potential KdgR-binding site could be a Pir-binding site. Indeed, Pir is also a regulator of the IclR family and its binding site was shown to be similar to that of KdgR (Nomura et al., 1999). This observation prompted us to verify that the expression of the genes chmX, dhfX, expI, indA, pecT, ppsA, sotA, spiX, ydiA and yjgK is not affected by Pir (data not shown).

Potential function of genes strongly regulated by KdgR: chmX, dhfX and yjgK
Enterobacteria possess a set of methyl-accepting chemotaxis proteins (MCPs) which are involved in the control of flagellar activity so that the bacterial cells move toward favourable environmental conditions (Stock & Surette, 1996). The periplasmic substrate-binding component TogB of the TogMNAB transport system acts as an oligogalacturonide-specific chemoreceptor in ER (Hugouvieux-Cotte-Pattat et al., 2001). Thus, TogB most probably interacts with an oligogalacturonide-specific MCP of the inner membrane, which could transduce the signal to the motility apparatus. A newly identified KdgR-regulated MCP gene, chmX, is a good candidate for this function. The absence of an observable phenotype of the chmX mutant suggests that additional MCP proteins allow ER to be attracted by pectic oligomers. A large number of MCP homologues observed in the ER chromosome suggests possible redundancy in their substrate specificity. Among 44 predicted MCP genes, two more genes (chmX13 and chmX21) are preceded by candidate KdgR-binding sites with scores only slightly below the cut-off (5·09 and 4·97, respectively; Table 3). Thus, additional experiments are necessary to conclusively assign the role of ChmX, ChmX13 and ChmX21 in chemotaxis towards pectic oligomers or monomers.

In both the Erwinia and Yersinia species, the KdgR regulon includes a hypothetical protein, DhfX, from the dienelactone hydrolase family (see GenBank entry NP_667845 for reference). DhfX has no other orthologues and is weakly similar to an acetyl xylan esterase from Bacillus pumilus and a cephalosporin C deacetylase from B. subtilis. Since KdgR-regulated protein DhfX of ER has a candidate N-terminal signal sequence, we propose that it is a periplasmic esterase acting on pectic oligomers, possibly a novel pectin acetyl esterase, in addition to PaeY and PaeX.

Hypothetical gene yjgK was predicted to be regulated by KdgR in all enterobacteria. In ER, it is among the genes which are strongly controlled by KdgR in vivo. We noticed that the yjgK gene from YE, in addition to the predicted KdgR site, has a candidate binding site for the ExuR repressor, a regulator of the GA catabolism. The yjgK product belongs to the DUF386 family, consisting of conserved hypothetical proteins, typically about 150 aa in length, with no known function. The phylogenetic tree of this family has several distinct branches, three of which, YjgK, YhcH and YiaL, are specific for enterobacteria (data not shown). While yjgK is a single gene in all enterobacteria, two other EC members of the DUF386 family are located in gene clusters involved in the catabolism of N-acetylneuraminic acid (nanATKE-yhcH) and, possibly, 2,3-diketo-L-gulonate (yiaKLMNOPQRS) (Yew & Gerlt, 2002). A more sensitive homology search with PSI-BLAST showed weak similarity of YjgK to EbgC of EC (17 % identity, 21 % similarity). The function of EbgC is not well defined, but it is required for the full activity of the second EC {beta}-galactosidase encoded by the ebgA gene (Elliott et al., 1992). Since YjgK is one of the most conserved members of the KdgR regulon, we suppose that it may be involved in the downstream part of the pectin catabolic pathway, probably being required for full activity of a conserved enzyme, KduD, KduI or KdgK.

The KdgR-regulated genes tpfX and yeeO
In contrast to all other KdgR-regulated genes, tpfX and yeeO lack candidate binding sites for the catabolic repressor protein CRP. Orthologues of the tpfX gene were found only in two other enterobacteria, EO and ST, where they are also predicted members of the KdgR regulon (see STM1931 in ST). Hypothetical protein TpfX belongs to the ThiJ/PfpI family that includes thiamine biosynthesis protein ThiJ from bacteria and intracellular protease PfpI from archaea. Although orthologues of yeeO were found in all enterobacteria, predicted KdgR-binding sites were observed only in Erwinia species, KP and ST. This gene encodes a hypothetical transport protein from the multi-antimicrobial extrusion family (Hvorup et al., 2003). The data available are insufficient to assign a role to tpfX and yeeO.

Genes weakly regulated by KdgR: spiX and sotA
In all enterobacteria, the predicted KdgR regulon includes a hypothetical sugar isomerase gene spiX (see STM1933 in ST for reference). In ER, this gene is expressed at a low level and is weakly controlled by KdgR (Fig. 4). Searching the databases, we identified SpiX orthologues in other bacterial species (the phylogenetic tree is shown in Fig. 5). In the Vibrio species, the spiX gene is located in the KdgR-regulated cluster kduD-ygjV-kdgF-spiX (Fig. 2). Notably, the complete genomes of VV and VP, as well as the unfinished genome of KP, lack the kduI gene involved in the first step of pectic monomer catabolism, isomerization of DK-I to DK-II. In some bacteria from the Bacillus/Clostridium group, the spiX orthologues are located in the cluster, including kdgK, kdgA and kduD (Fig. 5), and there are no kduI homologues in these genomes. However, in Xanthomonas species, we observed the kduI-kduD-spiX gene cluster encoding both the KduI and SpiX isomerases. In contrast to EC strain K-12, EC strain CFT073 possesses an additional KdgR-regulated locus, including the spiX gene (Fig. 2). A search at a low level of stringency allowed us to observe distant homology of SpiX to galactose-6-phosphate isomerase LacB from Streptococcus mutans and ribose-5-phosphate isomerase RpiB from EC. Summarizing all these data for SpiX, we propose that this novel member of the KdgR regulon could function as an additional isomerase, complementing the absence of KduI in some bacterial species.



View larger version (27K):
[in this window]
[in a new window]
 
Fig. 5. Phylogenetic tree of hypothetical sugar isomerases, SpiX. Proteins are denoted by their genome abbreviations (listed in Methods). Genes predicted to be regulated by KdgR are boxed.

 
Another gene weakly regulated by KdgR encodes the sugar efflux transporter SotA (Condemine, 2000). Expression of sotA in ER is activated by CRP and weakly induced by GA. The sotA gene is preceded by a weak KdgR site in both the Erwinia and Klebsiella species, whereas the Yersinia species lack the sotA gene. In ST and EC, this gene is present but has no candidate KdgR site in upstream region (Fig. 2). Conservation of a candidate CRP site upstream of sotA in all enterobacteria is in agreement with the previously proposed broad substrate specificity of this sugar efflux pump. Weak regulation of SotA by KdgR confirmed in ER (Fig. 4) indicates that in plant-pathogenic bacteria, SotA could be more specifically involved in the efflux of intermediates of pectin catabolism that could have a toxic effect if they accumulated intracellularly. Indeed, a strong toxic effect was observed in EC for 6-phospho-KDG (Fuhrman et al., 1998) and growth inhibition was frequently observed in ER mutants accumulating DK-I, DK-II or KDG (unpublished observations).

Function of genes positively regulated by KdgR: ppsA, ydiA and pecT
In both Erwinia species, a strong KdgR site was identified in the common upstream region of two divergently transcribed genes, ydiA and ppsA, encoding a hypothetical conserved protein of unknown function and phosphoenolpyruvate synthase, respectively. Although the ydiA/ppsA gene cluster was identified in all enterobacteria, the predicted KdgR-binding site is not conserved in Yersinia or Salmonella species (Fig. 2). Since expression of both these genes is reduced in ER in the presence of GA or PGA, and in the kdgR mutant, we concluded that they are positively regulated by KdgR. All previously known members of the KdgR regulon are negatively regulated by this transcriptional factor (Hugouvieux-Cotte-Pattat et al., 1996). In EC, the ppsA gene is positively regulated by FruR, a global regulator of the carbon utilization (Negre et al., 1998). Using the FruR site profile, we showed that the candidate FruR-binding site in the ydiA/ppsA regulatory region is conserved in all enterobacteria (Fig. 2). In most bacteria, this region also contains a CRP-binding site (Fig. 2). In EC, expression of ppsA is also negatively regulated by the carbon storage regulator CsrA (Sabnis et al., 1995). Thus, complex regulation of the ppsA gene could take place in Erwinia species, involving several regulators of sugar catabolism: KdgR, FruR, CRP and possibly CsrA (RsmA in EO). The catabolically activated phosphoenolpyruvate synthase PpsA is a key gluconeogenic enzyme in EC (Oh et al., 2002). The metabolic role of the ydiA gene product is not clear; it could also be linked to gluconeogenesis since ydiA is co-localized and probably co-regulated with ppsA in all enterobacteria. We conclude that the role of KdgR, at least in Erwinia, is not restricted to the negative control of the pectin catabolism, but is extended to the positive regulation of gluconeogenesis. The effect of KdgR will be to favour carbon flow through the gluconeogenic pathway when pectin is not metabolized. Thus, during plant infection, KdgR could play a role in coordination of central carbohydrate metabolism by directing the intracellular carbon flux. This role could be even larger since we noticed that among weaker candidate KdgR sites (score 4·75), there is the pykF gene encoding fructose-stimulated pyruvate kinase I. Regulation of pykF was shown to be opposite to that of ppsA, since this gene is involved in glycolysis. In EC, pykF is repressed by FruR and activated by CsrA (Bledig et al., 1996; Sabnis et al., 1995). The pykF upstream regions in ER, EO, KP and EC contain candidate KdgR sites with scores ranging from 4·68 to 4·95. Exactly the same set of enterobacterial genomes is predicted to have a KdgR-regulated ppsA gene. The respective position of candidate binding sites and promoter elements in the pykF upstream regions of these enterobacteria suggests negative regulation of pykF by both KdgR and FruR. Our in vivo analysis confirmed that PpsA and PykF, catalysing reverse reactions of the central carbohydrate metabolism, are regulated by KdgR in the opposite manner.

PecT is a negative regulator of the LysR family involved in the control of the pectate lyase synthesis (Surgey et al., 1996). PecT expression is subject to autoregulation and negatively controlled by the nucleoid-associated protein H-NS (Nasser & Reverchon, 2002). The signal to which PecT responds remains unknown, but it is clear that variations in PecT concentration have drastic effects on the controlled genes. We showed that KdgR contributes to modulation of the PecT intracellular concentration, although the KdgR effect is weaker than that observed previously with the two negative regulators of pecT transcription, PecT and H-NS. The positive regulation exerted by KdgR could be an antirepressor effect rather than a direct activation. The regulatory network involved in the control of the pectate lyase synthesis includes several cross-relations. Identification of an additional link between KdgR and PecT adds a novel complexity between two pathways of this interactive network.

Potential function of other candidates of the KdgR regulon: gntD, gntBMNAC and ygjV
In both Erwinia species and KP, we found a new KdgR-regulated locus, named gntDBMNAC. The short distances between these genes make it likely that they form an operon. In addition, ER has a close paralogue of gntD, a single gene, gntD2, which is also preceded by a strong candidate KdgR-binding site (Fig. 2). While we could not obtain data for gntD and gntD2, we showed that gntB is controlled in vivo by KdgR. The gntD product is similar to two sugar acid dehydratases from EC that are specific to D-glucarate (gudh) and D-galactonate (DgoA). The gntBMNAC genes encode components of an ABC transport system from the oligopeptide permease family, including one substrate-binding protein, two transmembrane proteins and two ATP-binding proteins. The characterized members of this large family transport a variety of substrates, including small peptides, opines, nickel, {alpha}-galactosides and other oligosaccharides (Gage & Long, 1998). A similar gntDBMNAC locus was found in Pseudomonas syringae pv. tomato and Pseudomonas fluorescens. Noteworthy, in contrast to Pseudomonas aeruginosa and Pseudomonas putida, these two plant-associated Pseudomonas species have the KDG kinase gene kdgK and a homologue of the OGA-specific porin kdgM (Fig. 3), although other pectin catabolic genes from ER, including kduI and kduD, were not found in these complete genomes. Moreover, kdgM and gntDBMNAC are divergently transcribed in P. fluorescens, whereas these genes probably form a single transcriptional unit in P. syringae pv. tomato. Considering these data, we propose that the function of the KdgR-regulated locus gntDBMNAC is the catabolism (GtnD) and active transport (GtnBMNAC) of some direct KDG precursor, most probably of plant origin.

The ygjV gene encoding a hypothetical transporter with four predicted transmembrane segments was found immediately downstream of the GA catabolic cluster uxaCBA in all enterobacteria except ST. Candidate KdgR sites upstream of this gene were observed in all enterobacteria, except KP. A KdgR-binding site is located between uxaA and ygjV, immediately after the predicted {rho}-independent transcriptional terminator of the uxa operon. Moreover, we identified several paralogues of ygjV in both Vibrio species possessing the KdgR regulon, three paralogues in VV and two paralogues in VP. In both Vibrio species, one paralogue is located within the uxaBC-kdgKA-ygjV cluster (not regulated by KdgR), whereas another copy belongs to the kduD-ygjV-kdgF-spiX cluster preceded by two candidate KdgR sites (Fig. 2). The Vibrio species also have additional copies of the kdgK and kdgA genes that are members of the KdgR regulon. The duplication of these catabolic genes in Vibrio could be explained by a recent specialization of the paralogues towards catabolism of pectin (regulated by KdgR) or GA, two catabolic pathways converging on KDG formation. YgjV has no orthologues in other genomes and is not similar to other proteins from public databases. The predicted regulation by KdgR and clustering with the uxa genes suggest that genes of the yjgV family could be involved in transport of some intermediates of DK-I and GA catabolic pathways.

Conclusions
In this study we combined bioinformatic and experimental approaches to reconstruct and compare the pectin degradation pathways and the KdgR regulons in various gamma-proteobacteria. Fig. 6 summarizes previously known and newly identified members of the KdgR regulon and shows the main differences between the KdgR-regulated pathways in related gamma-proteobacteria. Two animal-associated bacteria, EC strain K-12 and ST, possess only the core part of this catabolic pathway, allowing them to utilize only monomers DK-I, DK-II and KDG. However, a recently sequenced uropathogenic strain of EC (CFT073) acquired an additional KdgR-regulated locus for transport and catabolism of short oligogalacturonides. KP also seems to use only short oligogalacturonides as KDG precursors. In contrast, Yersinia species possess two periplasmic pectinases, a pectate lyase and a polygalacturonase, and thus could utilize longer oligogalacturonide molecules. The KdgR regulons of two plant-pathogenic Erwinia species are the largest ones and contain an array of genes for the extracellular degradation of polymeric plant pectin and subsequent utilization of the resulting pectin oligomers of various lengths (Fig. 6).



View larger version (44K):
[in this window]
[in a new window]
 
Fig. 6. Reconstruction of the catabolic pathway for pectin and its derivatives in gamma-proteobacteria. The KdgR- and RhaS-regulated genes are highlighted in blue and green, respectively. Solid arrows denote the presence of enzyme/transporter in both species according to the colour scheme shown. Newly identified KdgR-regulated genes are marked by asterisks.

 
Bacteria from another family, Vibrionaceae, also have KdgR regulons, although with several differences (Fig. 2). Both Vibrio species, VV and VP, contain a large cluster of KdgR-regulated genes encoding a porin and different enzymes necessary for the catabolism of OGAs. This locus also contains hypothetical transporters (TogX, Fig. 6) that could be responsible for OGA uptake in these species lacking TogT or TogMNAB homologues. Interestingly, VV has an additional KdgR-regulated locus encoding two homologues of chrondroitinase AC, a potential disaccharide ABC transporter, porin, three hypothetical sulfatases, sulfatase-activating enzyme, a homologue of a sulfate transporter and a homologue of the unsaturated glucuronyl hydrolase Ugl from Bacillus sp. We propose that this locus could be involved in the catabolism of chondroitin sulfate, a sulfated polysaccharide consisting of 1,4-linked derivatives of hexosamine and D-glucuronate. The predicted regulation of these genes by KdgR could be explained by the fact that the action of unsaturated glucuronyl hydrolase on chondroitin disaccharide will produce {Delta}4,5-D-glucuronate, which is spontaneously transformed into DK-I, a KdgR inducer (Hashimoto et al., 1999).

This unexpected observation provides one more example of extension of the KdgR regulon in some bacterial species. For example, the KdgR regulon of Erwinia species is significantly extended to include most of the known pectin degradation enzymes, as well as the Out system for pectinase secretion. In addition, the KdgR regulon in ER includes the rhamnose-regulated operon rhiTN for transport and catabolism of rhamnogalacturonides (Hugouvieux-Cotte-Pattat, 2004). Double regulation of rhiTN by RhaS, activator of the rhamnose catabolism, and KdgR is explained by formation of both rhamnose and DK-I by cleavage of this oligosaccharide. In other enterobacteria (EO, KP and ST) possessing rhiTN or only rhiN, we observed conservation of RhaS- and KdgR-binding sites in their promoter region (Fig. 2). Moreover, a search for other RhaS and KdgR-regulated genes allowed us to identify a new TRAP-type transport system, named rhiABC, in EO and ST (Fig. 2). In EO, the rhiABC locus is preceded by candidate RhaS and KdgR sites. In the ST chromosome rhiABC has only a RhaS site, but it is adjacent to a KdgR-regulated gene, kdgM, itself adjacent to the rhamnose utilization locus rhaT-rhaBAD. RhiABC is a good candidate for the function of rhamnogalacturonide transporter, mainly in ST, in which RhiT is missing. Another example of possible regulon extension was observed in KP, which has a strong candidate KdgR site in the regulatory region of the divergently transcribed garD/garPLRK operons involved in D-galactarate catabolism. However, the significance of KdgR regulation of this pathway is not clear.

The use of comparative analysis allowed us to extend the knowledge about the KdgR regulon in the plant-pathogenic bacterium Erw. chrysanthemi, resulting in identification of ten novel genes preceded by strong KdgR sites (chmX, dhfX, gntD, ppsA, spiX, tpfX, ydiA, yeeO, ygjV and yjgK). Experiments conducted to verify these predictions indicated that seven novel genes, chmX, dhfX, gntB, spiX, tpfX, yeeO and yjgK are indeed negatively controlled by KdgR. Predictions of sites with weaker scores also led us to analyse the expression of some previously identified genes, and we observed that regulator PecT, glycolytic enzyme PykF and sugar efflux transporter SotA are also regulated by KdgR, albeit at a lower level. Demonstration of the KdgR influence on the PecT intracellular level provides a new example of interactions between different regulators in the regulatory network controlling pectate lyase synthesis in Erw. chrysanthemi. Moreover, pecT and two genes with strong KdgR sites, ppsA and ydiA, were found to be positively regulated by KdgR. Noteworthy is the fact that the effect of gene activation has not been described previously for the classical repressor KdgR. Thus, the results presented here clearly demonstrate the interest of comparative genomics for the prediction of gene regulation, reconstruction of metabolic pathways and identification of apparently missing steps, either for transport systems or for enzymic activities.


   ACKNOWLEDGEMENTS
 
The authors are grateful to members of the Lyon Erwinia group, Guy, Sylvie, Vladimir and William, for helpful discussions. We thank A. A. Mironov for providing software for genome analysis and useful discussions, and to O. Laikova for the RhaS recognition profile. This study was partially supported by grants from the Howard Hughes Medical Institute (55000309), Russian Foundation for Basic Research (02-04-49111), the Centre National de la Recherche Scientifique, the Ministère de l'Education Nationale et de la Recherche and the Programme Microbiologie 2003 (ACIM-2-17). This study has been done in part during the visit by D. R. to the Unit of Microbiology and Genetics, INSA-Lyon, France, supported by an exchange grant within the ESF Programme on Integrated Approaches for Functional Genomics.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Altschul, S., Madden, T., Schaffer, A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25, 3389–3402.[Abstract/Free Full Text]

Bardonnet, N. & Blanco, C. (1992). uidA antibiotic resistance cassettes for insertion mutagenesis, gene fusion and genetic constructions. FEMS Microbiol Lett 93, 243–248.[CrossRef]

Benson, D. A., Karsch-Mizrachi, I., Lipman, D. J., Ostell, J., Rapp, B. A. & Wheeler, D. L. (2000). GenBank. Nucleic Acids Res 28, 15–18.[Abstract/Free Full Text]

Bledig, S. A., Ramseier, T. M. & Saier, M. H., Jr (1996). FruR mediates catabolite activation of pyruvate kinase (pykF) gene expression in Escherichia coli. J Bacteriol 178, 280–283.[Abstract]

Blot, N., Berrier, C., Hugouvieux-Cotte-Pattat, N., Ghazi, A. & Condemine, G. (2002). The oligogalacturonate-specific porin KdgM of Erwinia chrysanthemi belongs to a new porin family. J Biol Chem 277, 7936–7944.[Abstract/Free Full Text]

Condemine, G. (2000). Characterization of SotA and SotB, two Erwinia chrysanthemi proteins which modify isopropyl-beta-D-thiogalactopyranoside and lactose induction of the Escherichia coli lac promoter. J Bacteriol 182, 1340–1345.[Abstract/Free Full Text]

Condemine, G. & Robert-Baudouy, J. (1987). 2-Keto-3-deoxygluconate transport system in Erwinia chrysanthemi. J Bacteriol 169, 1972–1978.[Medline]

Condemine, G., Dorel, C., Hugouvieux-Cotte-Pattat, N. & Robert-Baudouy, J. (1992). Some of the out genes involved in the secretion of pectate lyases in Erwinia chrysanthemi are regulated by KdgR. Mol Microbiol 6, 3199–3211.[Medline]

Elliott, A. C., Sinnott, M. L., Smith, P. J., Bommuswamy, J., Guo, Z., Hall, B. G. & Zhang, Y. (1992). The catalytic consequences of experimental evolution. Studies on the subunit structure of the second (ebg) beta-galactosidase of Escherichia coli, and on catalysis by ebgAB, an experimental evolvant containing two amino acid substitutions. Biochem J 282, 155–164.[Medline]

Felsenstein, J. (1981). Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol 17, 368–376.[Medline]

Fuhrman, L. K., Wanken, A., Nickerson, K. W. & Conway, T. (1998). Rapid accumulation of intracellular 2-keto-3-deoxy-6-phosphogluconate in an Entner–Doudoroff aldolase mutant results in bacteriostasis. FEMS Microbiol Lett 159, 261–266.[CrossRef][Medline]

Gage, D. J. & Long, S. R. (1998). Alpha-galactoside uptake in Rhizobium meliloti: isolation and characterization of agpA, a gene encoding a periplasmic binding protein required for melibiose and raffinose utilization. J Bacteriol 180, 5739–5748.[Abstract/Free Full Text]

Gelfand, M. S. (2003). Computational identification of regulatory sites in DNA sequences. In Artificial Intelligence and Heuristic Methods in Bioinformatics, pp. 148–172. Edited by P. Frasconi & R. Shamir. Amsterdam: IOS Press.

Gelfand, M. S. & Laikova, O. N. (2003). Prolegomena to the evolution of transcriptional regulation in bacterial genomes. In Frontiers in Computational Genomics, pp. 195–216. Edited by E. V. Koonin & M. Y. Galperin. Wymondham: Caister Academic Press.

Gelfand, M. S., Novichkov, P. S., Novichkova, E. S. & Mironov, A. A. (2000). Comparative analysis of regulatory patterns in bacterial genomes. Brief Bioinform 1, 357–371.[Medline]

Hashimoto, W., Kobayashi, E., Nankai, H., Sato, N., Miya, T., Kawai, S. & Murata, K. (1999). Unsaturated glucuronyl hydrolase of Bacillus sp. GL1: novel enzyme prerequisite for metabolism of unsaturated oligosaccharides produced by polysaccharide lyases. Arch Biochem Biophys 368, 367–374.[CrossRef][Medline]

Hofmann, K. & Stoffel, W. (1993). TMbase – A database of membrane spanning proteins segments. Biol Chem Hoppe-Seyler 374, 166.

Hugouvieux-Cotte-Pattat, N. (2004). The RhaS activator controls the Erwinia chrysanthemi 3937 genes rhiN, rhiT and rhiE involved in rhamnogalacturonan catabolism. Mol Microbiol 51, 1361–1374.[Medline]

Hugouvieux-Cotte-Pattat, N. & Reverchon, S. (2001). Two transporters TogT and TogMNAB are responsible for oligogalacturonide uptake in Erwinia chrysanthemi 3937. Mol Microbiol 41, 1125–1132.[CrossRef][Medline]

Hugouvieux-Cotte-Pattat, N. & Robert-Baudouy, J. (1989). Isolation of Erwinia chrysanthemi mutants altered in pectinolytic enzyme production. Mol Microbiol 3, 1587–1597.[Medline]

Hugouvieux-Cotte-Pattat, N., Dominguez, H. & Robert-Baudouy, J. (1992). Environmental conditions affect the transcription of the pectinase genes of Erwinia chrysanthemi 3937. J Bacteriol 174, 7807–7818.[Abstract]

Hugouvieux-Cotte-Pattat, N., Condemine, G., Nasser, W. & Reverchon, S. (1996). Regulation of pectinolysis in Erwinia chrysanthemi. Annu Rev Microbiol 50, 213–257.[CrossRef][Medline]

Hugouvieux-Cotte-Pattat, N., Blot, N. & Reverchon, S. (2001). Identification of TogMNAB, an ABC transporter which mediates the uptake of pectic oligomers in Erwinia chrysanthemi 3937. Mol Microbiol 41, 1113–1123.[CrossRef][Medline]

Hvorup, R. N., Winnen, B., Chang, A. B., Jiang, Y., Zhou, X. F. & Saier, M. H. (2003). The multidrug/oligosaccharidyl-lipid/polysaccharide (MOP) exporter superfamily. Eur J Biochem 270, 799–813.[Abstract/Free Full Text]

James, V. & Hugouvieux-Cotte-Pattat, N. (1996). Regulatory systems modulating the transcription of the pectinase genes of Erwinia chrysanthemi are conserved in Escherichia coli. Microbiology 142, 2613–2619.[Medline]

Liu, Y., Jiang, G., Cui, Y., Mukherjee, A., Ma, W. L. & Chatterjee, A. K. (1999). KdgREcc negatively regulates genes for pectinases, cellulase, protease, HarpinEcc, and a global RNA regulator in Erwinia carotovora subsp. carotovora. J Bacteriol 181, 2411–2421.[Abstract/Free Full Text]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Mironov, A. A., Koonin, E. V., Roytberg, M. A. & Gelfand, M. S. (1999). Computer analysis of transcription regulatory patterns in completely sequenced bacterial genomes. Nucleic Acids Res 27, 2981–2989.[Abstract/Free Full Text]

Mironov, A. A., Vinokurova, N. P. & Gel'fand, M. S. (2000). Software for analyzing bacterial genomes. Mol Biol 34, 253–262.

Nasser, W. & Reverchon, S. (2002). H-NS-dependent activation of pectate lyases synthesis in the phytopathogenic bacterium Erwinia chrysanthemi is mediated by the PecT repressor. Mol Microbiol 43, 733–748.[CrossRef][Medline]

Nasser, W., Reverchon, S. & Robert-Baudouy, J. (1992). Purification and functional characterization of the KdgR protein, a major repressor of pectinolysis genes of Erwinia chrysanthemi. Mol Microbiol 6, 257–265.[Medline]

Nasser, W., Awade, A. C., Reverchon, S. & Robert-Baudouy, J. (1993). Pectate lyase from Bacillus subtilis: molecular characterization of the gene, and properties of the cloned enzyme. FEBS Lett 335, 319–326.[CrossRef][Medline]

Nasser, W., Reverchon, S., Condemine, G. & Robert-Baudouy, J. (1994). Specific interactions of Erwinia chrysanthemi KdgR repressor with different operators of genes involved in pectinolysis. J Mol Biol 236, 427–440.[CrossRef][Medline]

Nasser, W., Robert-Baudouy, J. & Reverchon, S. (1997). Antagonistic effect of CRP and KdgR in the transcription control of the Erwinia chrysanthemi pectinolysis genes. Mol Microbiol 26, 1071–1082.[CrossRef][Medline]

Negre, D., Oudot, C., Prost, J. F., Murakami, K., Ishihama, A., Cozzone, A. J. & Cortay, J. C. (1998). FruR-mediated transcriptional activation at the ppsA promoter of Escherichia coli. J Mol Biol 276, 355–365.[CrossRef][Medline]

Nielsen, H., Engelbrecht, J., Brunak, S. & von Heijne, G. (1997). Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng 10, 1–6.[CrossRef]

Niersbach, M., Kreuzaler, F., Geerse, R. H., Postma, P. W. & Hirsch, H. J. (1992). Cloning and nucleotide sequence of the Escherichia coli K-12 ppsA gene, encoding PEP synthase. Mol Gen Genet 231, 332–336.[Medline]

Nomura, K., Nasser, W. & Tsuyumu, S. (1999). Self-regulation of Pir, a regulatory protein responsible for hyperinduction of pectate lyase in Erwinia chrysanthemi EC16. Mol Plant–Microbe Interact 5, 385–390.

Oh, M. K., Rohlin, L., Kao, K. C. & Liao, J. C. (2002). Global expression profiling of acetate-grown Escherichia coli. J Biol Chem 277, 13175–13183.[Abstract/Free Full Text]

Resibois, A., Colet, M., Faelen, M., Schoonejans, E. & Toussaint, A. (1984). PhiEC2, a new generalised transducing phage of Erwinia chrysanthemi. Virology 137, 102–112.[CrossRef]

Reverchon, S., Expert, D., Robert-Baudouy, J. & Nasser, W. (1997). The cyclic AMP receptor protein is the main activator of the pectinolysis genes in Erwinia chrysanthemi. J Bacteriol 179, 3500–3508.[Abstract]

Robert-Baudouy, J., Nasser, W., Condemine, G., Reverchon, S., Shevchik, V. E. & Hugouvieux-Cotte-Pattat, N. (2000). Regulation of pectinase gene expression in Erwinia chrysanthemi. Plant–Microbe Interact 5, 221–268.

Rodionov, D. A., Mironov, A. A., Rakhmaninova, A. B. & Gelfand, M. S. (2000). Transcriptional regulation of transport and utilization systems for hexuronides, hexuronates and hexonates in gamma purple bacteria. Mol Microbiol 38, 673–683.[CrossRef][Medline]

Roeder, D. L. & Collmer, A. (1985). Marker-exchange mutagenesis of pectate lyase isozyme gene in Erwinia chrysanthemi. J Bacteriol 164, 51–56.[Medline]

Sabnis, N. A., Yang, H. & Romeo, T. (1995). Pleiotropic regulation of central carbohydrate metabolism in Escherichia coli via the gene csrA. J Biol Chem 270, 29096–29104.[Abstract/Free Full Text]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning. A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Surgey, N., Robert-Baudouy, J. & Condemine, G. (1996). The Erwinia chrysanthemi pecT gene regulates pectinase gene expression. J Bacteriol 178, 1593–1599.[Abstract]

Tatusov, R. L., Galperin, M. Y., Natale, D. A. & Koonin, E. V. (2000). The COG database: a tool for genome-scale analysis of protein functions and evolution. Nucleic Acids Res 28, 33–36.[Abstract/Free Full Text]

Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. (1997). The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25, 4876–4882.[Abstract/Free Full Text]

Thomson, N. R., Nasser, W., McGowan, S., Sebaihia, M. & Salmond, G. P. (1999). Erwinia carotovora has two KdgR-like proteins belonging to the IclR family of transcriptional regulators: identification and characterization of the RexZ activator and the KdgR repressor of pathogenesis. Microbiology 145, 1531–1545.[Medline]

Yew, W. S. & Gerlt, J. A. (2002). Utilization of L-ascorbate by Escherichia coli K-12: assignments of functions to products of the yjf-sga and yia-sgb operons. J Bacteriol 184, 302–306.[Abstract/Free Full Text]

Received 19 January 2004; revised 25 June 2004; accepted 11 August 2004.



This Article
Abstract
Full Text (PDF)
Supplementary Table
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Rodionov, D. A.
Articles by Hugouvieux-Cotte-Pattat, N.
Articles citing this Article
PubMed
PubMed Citation
Articles by Rodionov, D. A.
Articles by Hugouvieux-Cotte-Pattat, N.
Agricola
Articles by Rodionov, D. A.
Articles by Hugouvieux-Cotte-Pattat, N.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2004 Society for General Microbiology.