Bacteria use two-component signal transduction pathways to sense both extracellular and intracellular environment and to coordinate cellular events according to changing conditions. Adaptation can be either physiological or genetical. Here, we present evidence that a genome reorganization process such as transposition can be controlled by certain environmental cues sensed by a two-component signal transduction system. We demonstrate that transposition-dependent accumulation of phenol-utilizing mutants is severely decreased in Pseudomonas putida defective in a two-component system colRS. Translocation of Tn4652 is decreased both in colR- and colS-defective strains, indicating that signal transduction from a histidine kinase ColS to a response regulator ColR is necessary for the activation of Tn4652 in bacteria starving on phenol. However, overexpression of ColR in a colS-defective strain restores Tn4652 transposition, suggesting that absence of the signal from ColS can be compensated by an elevated amount of ColR. In vitro analysis of purified ColR and ColS proteins evidenced that they constitute a functional phosphorelay. Site-directed mutagenesis revealed that a conserved H221 can be the phosphoryl-accepting residue in ColS and that aspartate residues D8 and D51 of ColR are necessary for the phosphotransfer from ColS to ColR. To our knowledge, Tn4652 is the first bacterial transposon regulated by a two-component system. This finding indicates that transpositional activity can respond to signals sensed and processed by the host.
Most of the bacteria live in unstable environments and their successful survival requires adaptive response to altered conditions. In prokaryotes, two-component systems are principal signal transducers involved in environmental sensing (e.g. Stock et al., 2000). In a typical two-component signalling pathway, an environmental stimulus is detected by a transmembrane histidine sensor kinase (HK) which is capable of autophosphorylation at a conserved histidine residue. The phosphoryl group is then transferred from the histidine to an aspartic acid residue in the conserved regulatory domain of the second component of the pathway, a response regulator (RR). Phosphorylated RR modifies cellular behaviour by means of an effector domain that usually binds the DNA and activates or represses specific genes (e.g. Robinson et al., 2000). Structural studies of proteins involved in two-component systems have revealed a modular architecture of HKs and RRs with well-conserved domains involved in His-Asp phosphorelay (West and Stock, 2001). The effector domains of different RRs, usually possessing a DNA-binding function, are also well conserved. The most variable are sensory domains of HKs reflecting diversity of the stimuli they sense (Galperin et al., 2001).
The number of two-component systems in different bacteria varies largely and may depend on the need of bacteria to adapt to changing environments (Galperin et al., 2001; Wang et al., 2002). Indeed, in order to cope with a broad range of environmental situations, a large number of sensory pathways would be advantageous. The diverse genus Pseudomonas, colonizing numerous environmental niches ranging from soil and water to plants and animals (Spiers et al., 2000), possesses a large number of signalling components. Analysis of complete genome sequences of P. aeruginosa (Rodrigue et al., 2000), P. putida and P. syringae (http://www.tigr.org) has revealed more than 120 members of two-component systems.
Transposon Tn4652, residing in the chromosome of P. putida PaW85, is a 17-kb-long deletion derivative of the toluene degradation transposon Tn4651 (Tsuda and Iino, 1987). Our previous study has shown that activity of transposon Tn4652 depends on physiological state of the host being activated in the stationary phase. As already referred, stationary phase-specific RNA polymerase sigma factor RpoS and histone-like IHF are both implicated in Tn4652 transposition (Ilves et al., 2001; 2004). In this article, we demonstrate that a two-component signal transduction system is involved in regulation of Tn4652 in bacteria starving on phenol. Namely, we found that transposition of Tn4652 in P. putida PaW85 is controlled by the ColR–ColS two-component system. The ColR–ColS system was originally identified as an important component of root-colonizing ability of Pseudomonas fluorescens (Dekkers et al., 1998). So far, no other function for this two-component system has been described. The biochemical parameters and mechanisms of functioning of ColR–ColS system were also unclear. Here, we show that purified ColS and ColR are active in vitro. The active sites involved in autophosphorylation of ColS and signal transduction to ColR were determined by using site-directed mutagenesis. Measurement of Tn4652 transposition in vivo indicated that signal transduction from ColS to ColR is necessary for the activation of transposition in wild-type bacteria. Furthermore, involvement of ColR in target site selection of Tn4652 is hypothesized.
Transposition of Tn4652 is decreased in both colR- and colS-defective P. putida strains
We have found that under selection conditions transposition of P. putida transposon Tn4652 is regulated by a two-component signal transduction system colRS. The movement of native chromosomally located Tn4652 was examined in the P. putida wild-type strain PaW85 and in its colR- and colS-defective derivatives. Transposition of Tn4652 was tested in an assay that detects insertions of Tn4652 into plasmid pEST1332 in front of promoterless phenol degradation genes pheBA (Ilves et al., 2001; 2004). Translocation of Tn4652 can activate the pheA gene because of generation of a fusion promoter at the insertion site (Nurk et al., 1993) resulting in accumulation of phenol utilizing Phe+ mutants on phenol minimal plates. According to polymerase chain reaction (PCR) analysis, more than 95% of Phe+ mutants arise because of transposition of Tn4652 (Ilves et al., 2001). Thus, simple counting of colonies on phenol minimal plates gives information on frequency of transposition of Tn4652. Comparative examination of P. putida wild-type strain PaW85 and its colR- or colS-defective derivatives in a transposition assay revealed severely decreased accumulation of phenol-utilizing colonies in both mutant strains. About 10-fold less Phe+ colonies emerged on selective plates in colR- and colS-minus strains if compared with the wild-type P. putida(Fig. 1A). PCR analysis of Phe+ mutants from colR- or colS-defective strains demonstrated that as in the wild-type strain most of them originated from transposition of Tn4652 (Fig. 1B and C). PCR analysis proved that the Phe+ colonies of colR-minus strain were still colR knockouts, not occasional revertants. Thus, the absence of ColR or ColS decreased transposition substantially, but did not prevent it entirely.
In order to assure that colR- and colS-defective bacteria are not dying during the transposition assay, we estimated the viability of P. putida colR- and colS-minus strains starving on phenol minimal plates. Cell viability measurement on phenol minimal plates was performed as described previously (Ilves et al., 2001). The number of viable cells remained constant for wild-type, colR- and colS-defective P. putida during 6 days of starvation (data not shown). These results suggest that 10-fold decrease in accumulation of Phe+ mutants in colR- and colS-minus strains is not caused by decreased survival of mutant strains; rather, transposition of Tn4652 is downregulated in these strains.
The possible checkpoint of the ColR in regulation of Tn4652
We have previously shown that transposition of Tn4652 is positively affected by host factors σS and IHF (Hõrak and Kivisaar, 1998; Ilves et al., 2001; 2004), and negatively regulated by Tn4652-encoded TnpC (Hõrak and Kivisaar, 1999). Therefore, we examined whether some of these factors known to be involved in regulation of Tn4652 might be controlled by colRS. By using σS- and IHF-specific antibodies, we tested expression of these proteins in wild-type and colR-defective backgrounds in rich liquid medium and also on phenol minimal plates under starvation conditions. However, no background-specific differences in amounts of either σS or IHF were detected (Fig. 2A and B). The expression of tnpC was compared in wild-type and colR-minus backgrounds by using different transcriptional fusions of tnpC gene with a reporter gene β-glucuronidase (GUS). Results obtained revealed that expression of tnpC was also not affected by ColR (data not shown). All abovementioned proteins are involved in regulation of the Tn4652 transposase (TnpA). Therefore, we also tested the amount of TnpA in wild-type and colR-knockout strains. TnpA protein can be quantified on an immunoblot, if the copy number of the tnpA gene is increased by introducing the tnpA into cells on a plasmid (Hõrak and Kivisaar, 1999). As presented in Fig. 2C, the amount of plasmid-encoded TnpA is not changed in colR-minus background. Consequently, sS, IHF, TnpC and TnpA are not potential targets for the colRS two-component system.
Transpositional activity of a mobile element can be controlled by target site selection (reviewed in Craig, 1997). PCR analysis of Phe+ colonies of wild-type, colR- and colS-defective strains revealed that the same insertion sites were used by Tn4652 in these strains (Fig. 1). However, the relative insertion site pattern is changed in the colR- and colS-defective strains. One particular insertion site, characterized by a 372-bp-long PCR product and previously named as RA2 (Nurk et al., 1993), is more frequently used in colR- and colS-defective strains than in the wild-type strain (compare B and C in Fig. 1). Analysis of about 100 Phe+ colonies of wild-type, colR- and colS-defective strains demonstrated that in wild-type bacteria only about 2% of Tn4652 insertions occurred into the RA2 site. At the same time, up to 20% of all transposition events were targeted to the RA2 site in colR- and colS-knockout strains. These data suggest that ColR may somehow be involved in target site selection.
Signal transduction between purified ColS and ColR in vitro
Because the transposition of Tn4652 was decreased in both colR- and colS-minus strains, it is reasonable to assume that signal transduction from ColS to ColR is necessary for the activation of transposition. In order to characterize the biochemical features of ColS and ColR more precisely, the sensor and the RR were overexpressed as his-tagged recombinant proteins and purified. According to the sequence analysis, ColS possesses two putative transmembrane domains, indicating that it is a membrane-bound protein. Several HKs, deprived of N-terminal signal sensing domain but retaining the conserved kinase domain, are known to catalyse autophosphorylation and to be substrates for cognate RRs (Roberts et al., 1994; Zhang and Hulett, 2000). Therefore, to facilitate production of the soluble protein, the truncated version of ColS, Δ158ColS-His6, comprising only the C-terminal potentially cytoplasmic domain of the protein, was purified. The purification yielded more than 90% pure His6-ColR with a calculated molecular weight of 26.3 kDa and about 80% pure Δ158ColS-His6 with a calculated molecular weight of 31.6 kDa (Fig. 3A). Notably, hereafter we designate the recombinant Δ158ColS-His6 and His6-ColR as ColS and ColR.
To assay the functionality of purified ColS and ColR, in vitro phosphorylation and phosphotransfer reactions were performed using [γ-32P]-ATP as a phosphodonor. Results obtained clearly show that ColS is capable of autophosphorylation (Fig. 3B). After addition of the purified RR ColR to the phosphorylated sensor ColS, a rapid transfer of labelled phosphoryl group to ColR occurred (Fig. 3B). These data demonstrate that ColS and ColR constitute a functional signal transduction unit.
Mutagenesis of amino acid residues potentially phosphorylated upon the signal transduction between ColS and ColR
The signal transducing domains of HKs and RRs are well conserved. Therefore, alignment of amino acid sequences can be used to predict the putative phosphoryl-accepting histidine in HK and aspartate in RR. The amino acid sequence analysis of ColS revealed that the invariant histidine H221 might be the site of phosphorylation (Fig. 4B). However, there are several well-conserved aspartic acid residues in the ColR (Fig. 4A). The in silico structure analysis of ColR revealed that aspartic acid residues D8 and D51 may locate in the active site of ColR (Fig.S1 in Supplementary material). In a slightly less conserved region, but close to abovementioned residues, D58 is located in the modelled structure of ColR. The structure analysis of ColS strongly supported the assumption that H221 may be the site of phosphorylation (Fig.S1 in Supplementary material).
To examine the function of H221 of ColS, as well as the role of D8, D51 and D58 of ColR in phosphoryl transfer, site-directed mutagenesis of ColS and ColR was performed. H221 was replaced with valine and the conserved aspartic acid residues in ColR were replaced with alanine. The mutated proteins were purified and tested in a phosphorylation assay. Substitution of H221 abolished the autophosphorylation activity of ColS (Fig. 3B). Notably, the preparation of ColSH221V contained few contaminating proteins from Escherichia coli lysate and a weak band was detectable on autoradiographs when ColSH221V was tested in a phosphorylation reaction. However, this band is located slightly above the band of ColSH221V and appears to represent a larger contaminating protein readily detectable in the E. coli BL21 lysate (Fig. 3C). Furthermore, the mutated ColSH221V was defective in signal transduction – if the wild-type purified ColR was added to the phosphorylation reaction, the RR was not phosphorylated (Fig. 3B). Thus, these results indicate that H221 is obviously the site of autophosphorylation in ColS. Mutagenesis of conserved aspartic acid residues in ColR and subsequent analysis of purified mutant proteins in phosphorylation reactions revealed that substitution at positions 8 (D8A) and 51 (D51A) abolished phosphorylation of ColR (Fig. 3B). At the same time, ColRD58A was phosphorylated although less than the wild-type ColR (Fig. 3B). These data imply that both D8 and D51 are important amino acid residues in the active site of ColR.
Overexpression of either wild-type or mutant ColR in both colR- and colS-defective strains can restore transposition of Tn4652
The data presented above suggest that the phosphorylated form of ColR activates transposition of Tn4652. Therefore, we presumed that if we complement a colR-minus strain either with the wild-type ColR or with a phosphorylation-deficient form of ColR, only the former can complement the transposition defect. To test this, we introduced an extra copy of wild-type colR gene under the control of the Ptac promoter and lacIq repressor into the chromosome of a colR-minus strain. For complementation experiments with mutant variants of ColR, the wild-type sequence of ColR in the lacIq-Ptac-colR expression cassette was replaced with either D8A or D51A substitution. Immunoblot analysis of ColR expression in constructed strains revealed gradual overexpression of ColR alongside with increasing concentrations of IPTG (Fig. 5C). However, already without the induction, the ColR was present at almost wild-type level in lacI q-Ptac-colR (or colRD8A or colRD51A) expression cassette carrying strains, indicating the leakiness of the tac promoter (Fig. 5C). Measurement of transposition of Tn4652 demonstrated that this leaky expression of wild-type ColR was sufficient to restore transposition in a colR-minus background (Fig. 5A). High overexpression of ColR with induction by 0.5 mM IPTG resulted in partial recovery of transpositional activity of Tn4652 (Fig. 5A). Complementation of colR-defective strain with ColRD8A or ColRD51A gave somewhat unexpected results (Fig. 5B, only results for ColRD51A are presented). Namely, phosphorylation-deficient ColRD8A and ColRD51A were able to increase transposition. However, a much higher concentration of mutant ColR was required for the recovery of transposition as compared with complementation with wild-type ColR. Interestingly, high concentrations of ColRD8A and ColRD51A enhanced the activity of Tn4652 even over the wild-type transposition level (see Results with 0.5 mM IPTG, Fig. 5B). PCR analysis verified that the Tn4652 had been transposed in front of the pheA in these Phe+ colonies. These results imply that, in regard to regulation of Tn4652, overexpression of ColR can compensate the defect in signal transfer from ColS to ColR. If so, then transposition defect in a colS-minus strain can also be restored by overexpression of ColR. In order to verify this, the colS-minus strain was complemented with both wild-type and mutant variants of ColR. Results presented in Fig. 6 demonstrate that high overexpression of either wild-type or mutant ColR can restore transposition of Tn4652 in a colS-knockout strain. Note that, differently from the complementation of a colR-minus strain, the leaky expression of ColR was not sufficient to restore the transposition in a colS-minus strain. Again, as in colR-minus strain a high overexpression of ColRD8A or ColRD51A was able to enhance translocation of Tn4652 over the wild-type transposition level (see Results with 0.5 mM IPTG, Fig. 6B). Taken together, the complementation experiments indicate that, in regard to regulation of Tn4652, the requirement for phosphorylation of ColR can be bypassed by increased amount of this protein.
Transposons are intrinsic but often genetically invisible components of most genomes. Indeed, because of the potential to cause different DNA rearrangements, transposition is adjusted to strict control resulting usually in a very low transposition frequency. However, the rate of transposition is not constant and may respond to various cues. Here, we present evidence that activity of Tn4652 can respond to environmental stimuli as transposition of Tn4652 is substantially affected by deficiency of a two-component signal transduction system colRS.
Transposition of Tn4652 declined by about one order of magnitude in both colR- and colS-defective P. putida (Fig. 1). Thus, although ColR is present in colS-defective strain, transposition of Tn4652 is downregulated, indicating that signal transduction from ColS to ColR is necessary for full transpositional activity of Tn4652. However, the transposition of Tn4652 can be restored in a colS-defective strain by overexpression of ColR. Moreover, overexpression of ColRD8A or ColRD51A, which is defective in phosphorylation, can also complement transposition defect in a colR- as well as in a colS-defective strain (Figs 5B and 6B). This suggests that both phosphorylated and unphosphorylated ColR can activate transposition, but higher than the natural level of ColR is needed for the activation of transposition in the case of unphosphorylated ColR. Interestingly, high concentrations of ColRD8A and ColRD51A acted in the regulation of transposition differently from the wild-type ColR. Namely, while medium overexpression of the wild-type ColR restored transposition of Tn4652 in both colR- and colS-minus strains, its high overexpression had some inhibitory effect on transposition (compare results of induction by 0.01 mM and 0.5 mM IPTG, Figs 5A and 6A). Differently from that, high overexpression of ColRD8A and ColRD51A resulted in even increased transposition when compared with transposition frequency in the wild type (Figs 5B and 6B). To explain these at first sight contradictory results, we hypothesize that although the ColR acts in the regulation of Tn4652 primarily as an activator, it may have some negative impact on transposition under certain circumstances. This may take place, for instance, if the activity of ColS is decreased or the level of ColR is increased. Thus, the ColR–ColS system may represent a homeostatic mechanism controlling the activity of Tn4652.
Initially, the colRS two-component signal transduction pathway was characterized in P. fluorescens as a system involved in the ability of bacteria to colonize plant roots (Dekkers et al., 1998). However, impaired colonization ability of a colS-defective strain was observed only after co-inoculation of a colS-mutant strain with a wild-type one. Thus, the ColR–ColS two-component system is necessary in competitive colonization rather than in colonization per se (Dekkers et al., 1998). Another suggestion that the ColR–ColS system may be important in communication of bacteria can be found in the article by Duan et al. (2003). Namely, when P. aeruginosa was screened to find the genes expression of which is modulated in response to its co-cultivation with avirulent oropharyngeal flora strains, the promoter of the colR gene was essentially upregulated in response to co-cultivation (Duan et al., 2003). Thus, the expression of ColR in P. aeruginosa can respond to the presence of heterologous bacteria, i.e. it may be involved in interspecies communication. However, so far, the signal molecule sensed by ColS and the target genes of ColR have not been identified. We showed that previously identified regulators of Tn4652 transposition –σS, IHF and TnpC – are not been potential targets for the colRS two-component system. Therefore, further experiments are necessary to find the target genes of colRS pathway.
Examination of the Tn4652 insertion sites in front of the pheA gene in Phe+ mutants demonstrated that Tn4652 was transposed into the same sites in both wild-type and mutant strains. Thus, the reduced transposition of Tn4652 in colR- and colS-minus strains cannot be explained by exclusion of some insertion sites. Nevertheless, the relative usage of particular sites was changed in colR- and colS-defective strains if compared with the wild-type (Fig. 1). Therefore, we hypothesize that ColR, directly or indirectly, may be involved in regulation of the target site selection of Tn4652.
The biochemical characterization of purified ColS and ColR revealed that they constitute a typical two-component histidine-aspartate phosphorelay system. Site-directed mutagenesis of ColS indicated that an invariant H221 could serve as the amino acid which is autophosphorylated after sensing a signal. In the putative structure of ColS, the H221 locates in the most conserved region of the four-helix bundle subdomain (Fig. S1 in Supplementary material) which resembles histidine phosphotransfer or HPt domains of class II HKs (Dutta et al., 1999) such as CheA (Zhou et al., 1995; Mourey et al., 2001) and ArcB (Kato et al., 1997) from E. coli and Ypd1p from S. cerevisiae (Song et al., 1999; Xu and West, 1999). It is interesting to note that the best-scored putative model of a ColS cytoplasmatic domain, available in ModBase databank (Pieper et al., 2004), has been calculated on the basis of resolved structure of the rat mitochondrial protein kinase BCK (Machius et al., 2001), implying to evolutional conservation of the kinase core fold.
Mutagenesis of conserved aspartate residues in ColR revealed that D8 and D51 are both involved in phosphorylation (Fig. 3B). Both biochemical and structural studies of different RRs have established that phosphorylation occurs at an aspartate positioned at the tip of the third β-strand of the five-stranded β-sheet inside the acidic active site pocket (reviewed in Robinson et al., 2000). D51 locates at the tip of the β3 in the model structure of ColR (Fig.S1 in Supplementary material), indicating that it is obviously the phosphoaccepting residue in signal transduction. D8 most probably participates in coordination of a Mg2+ ion that is required for the phosphoryl transfer.
There are various interpretations of the nature of transposons – from calling them parasites up to considering them as useful entities of the host genome. Indeed, because of the ability to amplify itself by moving from site to site, transposons behave selfish and are very dissimilar to other functional components of the cell. However, transposable elements, like other genome structures, are subjected to different regulatory mechanisms of the host. Thus, although transposition is a selfish reaction in many aspects, it is most probably governed by the host. Our present study suggests that transposition of Tn4652, similarly to other cellular events, responds to environmental stimuli being regulated by a two-component signal transduction pathway. Another mobile DNA element – the phage Mu – is also shown to respond to signal transduction network sensitive to different physiological inputs (Maenhaut-Michel and Shapiro, 1994; Lamrani et al., 1999). Notably, involvement of signal transduction in transposition of Tn4652 and Mu was discovered during selection of mutants capable of growing on a particular carbon source. Therefore, we believe that further investigation of transposition under selective conditions may reveal many other transposons controlled by cellular signal transduction network.
Bacterial strains, plasmids and media
The bacterial strains and plasmids used are described in Table 1. Bacteria were grown on Luria–Bertani (LB) medium (Miller, 1992) or on M9 minimal medium (Adams, 1959) containing 0.2% glucose. Phenol minimal plates with 1.5% Difco agar contained 2.5 mM phenol as the carbon source. When selection was necessary, the growth medium was supplemented with ampicillin (100 µg ml−1), kanamycin (50 µg ml−1), tetracycline (10 µg ml−1) for E. coli and with carbenicillin (1500 µg ml−1), kanamycin (50 µg ml−1), tetracycline (40 µg ml−1) and potassium tellurite (25 µg ml−1) for P. putida. P. putida was incubated at 30°C and E. coli at 37°C. E. coli was transformed with plasmid DNA as described by Hanahan (1983). P. putida was electrotransformed according to the protocol of Sharma and Schimke (1996).
Table 1. Bacterial strains and plasmids.
Strain or plasmid
Genotype or construction
Source or reference
a. Oligonucleotides used for the construction of plasmids are presented in Table 2
pJMT6 containing 2.8 kb NotI fragment with lacIq-Ptac-colR cassette (Apr Telr)
pJMT6 containing 2.8 kb NotI fragment with lacIq-Ptac-colRD8A cassette (Apr Telr)
pJMT6 containing 2.8 kb NotI fragment with lacIq-Ptac-colRD51A cassette (Apr Telr)
Construction of plasmids and strains
Oligonucleotides used in PCR amplifications are listed in Table 2. Library plasmid pACYC184/colRS, which contains the colRS genes in a 6.13 kb BamHI–BglII fragment of P. putida PaW85 genome, was used as a DNA source for all subcloning procedures. For the construction of P. putida PaW85 colR-defective strain, the colR gene-containing 694 bp Eco147I–NcoI fragment from pACYC184/colRS was cloned into SmaI-digested pUC18Not. The colR gene in the obtained plasmid pUCNot/colR was interrupted by insertion of the PCR-amplified Kmr gene (Table 2) into ClaI site inside the colR gene. The resulting colR::Km sequence from pUCNot/colR::Km was inserted as a XbaI–EcoRI fragment into delivery plasmid pGP704L (Pavel et al., 1994) to yield pGP704L/colR::Km. To knock out the colS gene, first, the colS coding region was excised as a 1704 bp PvuII–EcoRI fragment from pACYC184/colRS and inserted into SmaI–EcoRI-cleaved pUC18Not resulting in pUCNot/colS. The colS gene was interrupted by insertion of the PCR-amplified Kmr gene into SacI site inside the colS gene. The colS::Km sequence from pUCNot/colS::Km was cloned into pGP704L resulting in pGP704L/colS::Km. The interrupted colR and colS genes were inserted into the chromosome of P. putida PaW85 by homologous recombination. Plasmids pGP704L/colR::Km and pGP704L/colS::Km were conjugatively transferred from E. coli CC118 λpir (Herrero et al., 1990) into P. putida PaW85 using a helper plasmid pRK2013 (Figurski and Helinski, 1979). The PaWcolR and PaWcolS knockouts were verified by PCR analysis. The lack of ColR protein in PaWcolR was also proved by immunoblot analysis.
In order to simplify purification of ColR and cytoplasmatic domain of ColS, a hexahistidine tag was fused to the N-terminus of ColR and C-terminus of ColS by using designed oligonucleotides. The colR- and colS-containing fragments were amplified from plasmid pACYC184/colRS by using either oligonucleotides colRXNhis and colSkesk or oligonucleotides colSXN and colShistail. The colR-containing PCR product cleaved with XhoI and EcoO109I was cloned into XhoI–SmaI-opened pBluescript KS resulting in pKS/hiscolR. The pKS/Δ158colShis was obtained by cloning the XhoI–BamHI-cleaved colS-containing PCR fragment into XhoI–BamHI-digested pBluescript KS. The intactness of colR and colS in pKS/hiscolR and pKS/Δ158colShis was verified by sequencing. Subsequently, the sequences encoding recombinant proteins were subcloned into NdeI–BamHI-opened pET11c resulting in plasmids pET11c/hiscolR and pET11c/Δ158colShis.
For site-directed mutagenesis of ColR and ColS, the plasmids pKS/hiscolR and pKS/Δ158colShis were used as targets to introduce mutations into coding region of colR and colS respectively. In order to generate DNA fragments with mutations and suitable restriction sites, two sequential PCRs were performed. For the substitution of H221 in ColS for valine, the first PCR was carried out with oligonucleotides colSH221V and colSXN (Table 2). The synthesized fragment served as a reverse primer for the oligonucleotide colSEco147 in a second PCR. The product of second PCR was digested with XhoI and Eco147I, and ligated into the XhoI–Eco147I-opened pKS/Δ158colShis resulting in pKS/colSH221. For the substitution of aspartic acids D51 and D58 in ColR protein for alanine, the first PCRs were carried out with either oligonucleotides colRD51A and colRBcl or colRD58A and colRBcl. The designed substitution-containing PCR products were used as reverse primers in second PCRs with colRXNhis. The PCR products were cleaved with XhoI and BclI, and ligated into XhoI–BclI-digested pKS/hiscolR. For the generation of a DNA fragment with D8A in colR, two partially complementary oligonucleotides, hiscolRD8A and colREcoRVD8A, were annealed after the treatment with Klenow Fragment. The resulting product was digested with XhoI and EcoRV and cloned into XhoI–EcoRV-cleaved pKS/hiscolR. After verification of designed mutations by DNA sequencing, the mutated colR and colS genes were cloned as NdeI–BamHI fragments into NdeI–BamHI-opened pET11c.
In order to overexpress the intact ColR in a colR- or colS-defective strain, first, the colR gene-containing 846 bp Eco147I–Eam1104I fragment was cloned under the control of Ptac promoter and lacI q repressor into the plasmid pBRlacItac. The lacI q-Ptac-colR cassette was excised from plasmid pBRlacItac/colR with BamHI and EcoO109I, and subcloned into BamHI–HincII-opened pUC18Not resulting in pUCNot/taccolR. Finally, the colR expression cassette was inserted as a NotI fragment into the tellurite resistance-encoding minitransposon of the delivery vector pJMT6 resulting in pTel/taccolR. For the construction of ColRD8A and ColRD51A overexpression cassettes, the wild-type sequence of the colR gene in pUCNot/taccolR was replaced with mutant colR sequences that were obtained as 630 bp Mva1269I–NcoI fragments from pKS/hiscolRD8A and pKS/hiscolRD51A respectively. Subsequently, the lacIq-Ptac-colRD8A and lacIq-Ptac-colRD51A cassettes were inserted into the NotI-opened pJMT6. For introduction of different gene expression cassettes into the chromosome of P. putida PaWcolR or PaWcolS, the triparental mating between P. putida strain, E. coli CC118 λpir carrying either pTel/taccolR, pTel/taccolRD8A or pTel/taccolRD51A and a helper plasmid pRK2013-containing E. coli HB101, was performed. Transconjugants were selected on potassium tellurite and kanamycin-containing minimal plates. The presence of the lacI-Ptac-colR(D8A)(D51A) cassette in the chromosome of transconjugants was verified by PCR and inducible expression of ColR was proved by immunoblot analysis.
Transposition of native Tn4652 was examined in an assay that involves starvation of bacteria on phenol minimal plates (Kasak et al., 1997; Ilves et al., 2001). The transposition assay was carried out with P. putida PaW85 wild-type strain, with its colR- and colS-defective derivatives PaWcolR and PaWcolS and with ColR-overexpressing strains PaWStaccolR, PaWStaccolRD8A and PaWStaccolRD51A carrying target plasmid pEST1332. Bacteria were grown overnight in LB medium at 30°C and washed with M9 solution. Approximately 5 × 108 cells from at least three independent overnight cultures of each P. putida strains were plated onto phenol minimal plates and accumulation of mutant Phe+ colonies was monitored upon incubation of plates at 30°C during 7 days. To achieve different expression level of ColR in P. putida PaWStaccolR, PaWStaccolRD8A and PaWStaccolRD51A, phenol minimal plates were supplied with either 0.01 mM or 0.5 mM IPTG (isopropyl-β- d-thiogalactopyranoside), or no IPTG was added. To prove transposition of Tn4652 into pEST1332, random portion of Phe+ mutants (up to 25 per day) was analysed by PCR using oligonucleotides pheA and IRout (Table 2).
Purification of proteins
ColR, ColRD8A, ColRD51A and ColRD58A were purified as fusion proteins containing an N-terminal hexahistidine (his6) tag. For the overexpression of proteins, the E. coli strain BL21(DE3) harbouring different expression plasmids was grown in LB medium at 37°C up to OD580 of ≈ 0.5. The culture was subsequently transferred to 30°C and incubated for 30 min before the induction of ColR expression with 0.5 mM IPTG. After 4 h of induction, cells were pelleted and sonicated in buffer A [50 mM phosphate buffer (pH 7.0), 1 M NaCl]. The cell lysate was centrifuged at 12 000 g for 30 min at 4°C. Imidazole (50 mM) was added to the supernatant before it was mixed with Ni-NTA agarose matrix (Qiagen) and binding of the protein to the agarose was performed in a slowly rotating test tube at 4°C for 1 h. Then the agarose matrix was washed twice with 50 volumes of buffer B [50 mM phosphate buffer (pH 6.2), 1 M NaCl, 50 mM imidazole]. Subsequently the matrix was loaded onto a column and ColR protein was eluted with buffer C [100 mM phosphate buffer (pH 7.0), 300 mM NaCl, 500 mM imidazole, 10% glycerol]. The purified ColR was dialysed against the storage buffer [10 mM Tris-HCl (pH 7.5), 200 mM KCl, 0.5 mM EDTA, 1 mM DTT, 50% glycerol] and stored at −20°C.
ColS and ColSH221V were expressed as N-terminally truncated proteins lacking the first 158 amino acids. The hexahistidine tag was fused with C-terminus of ColS. ColS was overexpressed and purified analogously to ColR. However, differently from ColR, the induction temperature of ColS was 18°C and concentration of imidazole in all buffers used in purification was 25 mM.
Phosphorylation assay was carried out in a reaction volume of 20 µl. Purified ColS (0.3 µg) was pre-incubated in TP buffer [50 mM phosphate (pH 7.6), 50 mM KCl] supplemented with 2 mM MgCl2 at room temperature for 3 min. After addition of 0.5 µCi [γ-32P]-ATP (6000 Ci mmol−1; Amersham) to the reaction, autophosphorylation of ColS was allowed to occur during 20 min. To test autophosphorylation ability of ColSH221V, incubation time was extended up to 1 h. To assay the phosphotransfer between ColS and ColR, 1.5 µg of purified ColR (or its mutant forms) was added to the autophosphorylated ColS. Reactions were stopped after 5 min of incubation by the addition of 2× SDS sample buffer. Proteins were separated using a 8% or 12.5% SDS-PAGE, electrotransferred to a nitrocellulose membrane and exposed to a PhosphorImager screen (Molecular Dynamics).
Immunoblotting of proteins
Cell lysates were prepared from 20 ml of overnight cultures or from starved bacteria collected from several phenol minimal plates. Equal amounts of total protein (30 µg) were used in Western immunoblotting assay. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. For Western blotting, the membranes were probed with antibodies arisen against RpoS, IhfA, TnpA or ColR, followed by treatment with alcaline phosphatase-conjugated goat anti-rabbit or anti-mouse immunoglobulin G. The blots were developed using bromochloroindolyl phosphate/nitro blue tetrazolium (BCIP/NBT).
Homology modelling of ColS and ColR
The models for ColS and ColR (Fig.S1 in Supplementary material) were obtained from the ModBase database (Pieper et al., 2004). The ColS cytoplasmatic kinase domain was modelled using the structure of the rat BCKD kinase (PDB ♯ 1KGZ) as a template. The structural template for ColR model was the RR DrrD from Thermotoga maritima (PDB ♯ 1KGS). For mapping of evolutionarily conserved residues on protein surfaces, the models of ColS and ColR were subjected to analysis in ConSurf at: http://consurf.tau.ac.il (Armon et al., 2001). Further analysis of modelled structures was performed with the program rasmol (Sayle and Milner-White, 1995).
We are grateful to Tiina Alamäe, Niilo Kaldalu and Paula Ann Kivistik for critical reading of the manuscript. We thank Aare Abroi for helping in structure analysis of ColR and ColS and for his helpful discussions. This work was supported by Grants 4481 and 5757 from the Estonian Science Foundation and by Grant HHMI 55000316 from the Howard Hughes Medical Institute International Research Scholars Program.