A two-component signal transduction system involved in nickel sensing in the cyanobacterium Synechocystis sp. PCC 6803



In the cyanobacterium Synechocystis sp. PCC 6803, genes for Ni2+, Co2+, and Zn2+ resistance are grouped in a 12 kb gene cluster. The nrsBACD operon is composed of four genes, which encode proteins involved in Ni2+ resistance. Upstream from nrsBACD, and in opposite orientation, a transcription unit formed by the two genes rppA and rppB has been reported previously to encode a two-component signal transduction system involved in redox sensing. In this report, we demonstrate that rppA and rppB (here redesigned nrsR and nrsS respectively) control the Ni2+-dependent induction of the nrsBACD operon and are involved in Ni2+ sensing. Thus, expression of the nrsBACD operon was not induced by Ni2+ in a nrsRS mutant strain. Furthermore, nrsRS mutant cells showed reduced tolerance to Ni2+. Whereas the nrsBACD operon is transcribed from two different promoters, one constitutive and the other dependent on the presence of Ni2+ in the medium, the nrsRS operon is transcribed from a single Ni2+-inducible promoter. The nrsRS promoter is silent in a nrsRS mutant background suggesting that the system is autoregulated. Purified full length NrsR protein is unable to bind to the nrsBACD-nrsRS intergenic region; however, an amino-terminal truncated protein that contains the DNA binding domain of NrsR binds specifically to this region. Our nrsRS mutant, which carries a deletion of most of the nrsR gene and part of the nrsS gene, does not show redox imbalance or photosynthetic gene mis-expression, contrasting with the previously reported nrsR mutant.


Increasing interest has been developed during the last years on the biological roles of nickel and the structure of nickel metalloenzymes (Maroney, 1999). Ni atoms form part of the active site of a number of enzymes including glyoxalases I, peptide deformylases, methyl-CoM reductase and ureases, as well as some superoxide dismutases and hydrogenases (Ermler et al., 1998). However, at high concentration, Ni2+ ions induce different types of harmful effects including generation of free radicals, inhibition of enzyme activity and DNA damage, leading to genetic instability, developmental defects and cancer (Von Burg, 1997; Costa, 1991, 1998). Because of these opposite effects, organisms have to carefully regulate the intracellular concentration of Ni2+.

Microbial Ni2+ uptake is mediated by non-specific transport systems for divalent cations and by high affinity specific systems (for two recent reviews, see Nies, 1999; Eitinger and Mandrand-Berthelot, 2000). Two types of high affinity transporters have been identified: (i) multicomponent ABC (ATP-binding cassette) transport systems (such as the NikABCDE system of Escherichia coli) (Navarro et al., 1993); and (ii) one-component permeases (such as NixA, UreH, HupN and HoxN) (Eitinger and Friedrich, 1991; Fu et al., 1994; Mobley et al., 1995) which are integral membrane proteins with eight transmembrane spanning helices. Most of the one-component permeases have been shown to share a conserved His-X4-Asp-His sequence in the second transmembrane helix involved in Ni2+ ligation. In the case of the Nik system, the periplasmic NikA protein is also involved in Ni2+ binding but the amino acid determinant of the interaction has not yet been identified. Beside these uptake systems, various Ni2+ export systems have also been described. The best known nickel resistance systems belong to bacteria of the genus Ralstonia. The Cnr (cobalt, nickel) and the Ncc (nickel, cobalt, cadmium) systems are membrane protein complexes that carry out Ni2+ efflux driven by proton/cation antiport (Collard et al., 1993; Liesegang et al., 1993; Schmidt and Schlegel, 1994; Grass et al., 2000). Both systems are closely related to the cobalt/zinc/cadmium resistance system Czc from Ralstonia sp. CH34 strain. In this system, CzcA is the transmembrane protein that carries out the antiport. CzcB is a membrane fusion protein that may span the peri-plasmic space and CzcC seems to be attached to the outer membrane (Nies, 1992; 1995; Rensing et al., 1997).

In the unicellular cyanobacterium Synechocystis sp. PCC 6803 (hereafter referred to as Synechocystis), a nickel resistance operon (nrsBACD) formed by four open reading frames (ORFs) has been described previously by us (García-Domínguez et al., 2000). NrsB and NrsA proteins are homologues to CzcB and CzcA, respectively, and they very probably form a membrane-bound protein complex catalysing Ni2+ efflux by a proton/cation antiport. NrsC is not homologous to proteins encoded by the czc or related operons, and its role in Ni2+ export is unknown. Finally, NrsD is a membrane protein belonging to the major facilitator superfamily of transport proteins. NrsD is highly homologous to NreB from Achromobacter xylosoxidans (Grass et al., 2001). Expression of NreB in E. coli confers resistance to Ni2+, suggesting that this permease is able to carry out Ni2+ export. Interestingly the carboxy terminal part of NrsD contains 14 histidine residues involved in Ni2+ binding (García-Domínguez et al., 2000). The nrsBACD operon is integrated into the Syne-chocystis metal resistance cluster which includes nine genes involved in Zn2+, Co2+ and Ni2+ resistance (Fig. 1) (Thelwell et al., 1998; Rutherford et al., 1999; García-Domínguez et al., 2000). Two ORFs, rppA (sll0797) and rppB (sll0798), are found 118 bp upstream from the nrsBACD operon and transcribed in the opposite direction (Fig. 1). Whereas rppA encodes a response regulator, rppB encodes a sensor histidine kinase, suggesting that both proteins form a classical two-component signal transduction system (Stock et al., 2000). A role of these two proteins on redox sensing has been postulated previously, based on the analysis of an insertional mutant of the response regulator rppA. However, the same phenotype was not obtained in a mutant of the sensor histidine kinase rppB (Li and Sherman, 2000).

Figure 1.

ORF organization of the metal-resistance gene cluster from Synechocystis. The nrsBACD operon was described in García-Domínguez et al. (2000), the corR (coaR) and corT (also called coaT) genes were described in Rutherford et al., 1999 and García-Domínguez et al. (2000), and the zia genes were described in Thelwell et al. (1998). Cyanobase database ORFs numbers are also included (Kaneko et al., 1996).

In this work, we demonstrate that the signal transduction system formed by the rppA and rppB products is controlling the expression of the nrsBACD operon, and is involved in Ni2+ sensing. A rppAB mutant with a deletion of most of the rppA gene and part of the rppB gene did not show any kind of redox imbalance, suggesting that the previously reported phenotype is the consequence of unspecific cross-talk. Therefore, we propose the re-designation of rppAB as nrsRS.


Ni2+-dependent expression of nrsR and nrsS genes

The stop codon of the nrsR (sll0797) gene overlaps with the start codon of the nrsS (sll0798) gene, strongly suggesting that they form a single transcription unit. This was further demonstrated by the size of the mRNA transcript detected in Northern blotting experiments (about 2000 nucleotides). Metal-dependent expression of the nrsRS operon was analysed by Northern blotting. For this, a probe of the nrsRS operon was used to hybridize total RNA obtained from mid-log Synechocystis cells grown in BG11C medium and exposed for 1 h to 17 μM of either ZnSO4, CdCl2, CoCl2, CuSO4, NiSO4 or MgCl2. Control cells were not exposed to added metals. As shown in Fig. 2A, the nrsRS transcript was strongly induced in the presence of Ni2+ and weakly induced in the presence of Co2+. Therefore, the nrsRS operon follows the same pattern of expression as the nrsBACD operon (Fig. 2A). A time-course experiment demonstrated that nrsRS and nrsBACD transcripts showed different inducibility. Thus, the amount of the nrsBACD transcript increased about 10 fold, 4 h after Ni2+ addition, whereas the amount of the nrsRS mRNA increased only threefold after Ni2+ addition (Fig. 2B).

Figure 2.

Metal-dependent regulation of nrsBACD and nrsRS.

A. Northern blot analysis of nrsBACD and nrsRS expression from mid-log phase Synechocystis sp. PCC 6803 cells were exposed to 17 μM of different metals for 1 h. Control cells were not exposed to added metals.

B. Time-course expression of nrsBACD and nrsRS after 17 μM NiSO4 addition. Radioactive signals from a time-course northern blot analysis of nrsBACD and nrsRS transcripts were quantified using an Instant Imager Electronic Autoradiography apparatus (Packard Instrument Company). Levels of nrsBACD (○) and nrsRS signals (•) were normalized to the rnpB signal as indicated in Experimental procedures. It should be noted that 100% corresponds to the maximal signal of hybridization for each probe, and that signals from different probes can not be compared. Average of two independent determinations is shown.

The nrsRS system controls the Ni2+-dependent expression of the nrsBACD operon

To identify the signal transduced by the nrsRS two-component system, a nrsRS mutant was generated by replacing most of the nrsR sequence and part of the nrsS sequence by a kanamycin resistance cassette (C.K1) in both orientations (Fig. 3A and B). Similar results were obtained for both mutants and, therefore, only data about the mutant with the kanamycin resistance gene (npt) in the opposite orientation with respect to the nrsRS genes are shown, except when indicated. The nrsRS::C.K1 Synechocystis strain was viable and its growth rate in BG11C medium was similar to that of the wild-type strain (data not shown). As expression of the nrsRS operon is induced in the presence of Ni2+ and Co2+, growth of nrsRS::C.K1 mutant was examined in Ni2+ and Co2+ supplemented BG11C medium. Normal growth was observed in Co2+-containing medium (data not shown); however, nrsRS::C.K1 cells were clearly less tolerant to the presence of Ni2+ in the medium (Fig. 3C). The effect of a number of other metals on the growth of wild-type Synechocystis and nrsRS::C.K1 cells was also tested, but none of them had a differential effect on the growth of nrsRS::C.K1 cells with respect to the effect observed for wild-type cells (data not shown).

Figure 3.

Loss of Ni2+ induction of the nrsBACD operon and nickel-reduced tolerance of nrsRS deletion mutant.

A. Schematic representation of the nrs genomic regions in the wild type and in the nrsRS::C.K1 mutant strains.

B. Southern blot analysis of Synechocystis wild type and nrsRS::C.K1 mutant. Genomic DNA was digested with HincII and hybridized using the DNA fragment indicated in panel A as a probe. The size of the hybridizing bands is indicated.

C. Ni2+ tolerance of wild-type and nrsRS::C.K1 strains. Ten-fold culture dilutions were spotted on BG11C plates supplemented with the indicated concentration of NiSO4 and photographed after 10 days of growth.

D. Northern blot analysis of nrsBACD expression in wild type or in nrsRS::C.K1 mutant cells exposed to 17 μM Ni2+, 17 μM Co2+ for 2 h, or in the absence of metal added.

How can the nrsRS system affect Ni2+ tolerance? One obvious possibility is that NrsR and NrsS are controlling the expression of the nickel resistance operon (nrsBACD). Thus, Northern blotting experiments demonstrated that the Ni2+-dependent inducibility of the nrsBACD operon was completely abolished in the nrsRS::C.K1 mutant (Fig. 3D).

Analysis of the nrsRS-nrsBACD intergenic region

The 118 bp-nrsRS-nrsBACD intergenic region should contain the promoter and the regulatory sequences responsible for the response to Ni2+. To identify the promoters of both nrsRS and the nrsBACD operons, transcription start-points were determined by primer extension (Fig. 4A). In the absence of Ni2+, one only transcription start site was found 49 bp upstream of the nrsB ATG start codon. This transcription start-point was unaltered in the nrsRS::C.K1 mutants. However, in the presence of Ni2+, two different mRNAs were found. One starts in the same position as the transcript found under non-induced conditions. The second, much more abundant than the previous one, starts 33 bp upstream of the nrsB ATG start codon. The inducible transcription start site was absent in the nrsRS::C.K1 mutant. These data indicate that the nrsBACD operon has two promoters; one constitutive and a second one inducible by the nrsRS system in the presence of Ni2+. Putative –10 boxes in the form TTTCAT and CAGACT were found 7 and 6 bp upstream of the Ni2+-independent and the Ni2+-dependent transcription start-points respectively. No obvious –35 boxes were detected at the appropriated positions (Fig. 4C).

Figure 4.

Primer extension analysis of the nrsBACD and nrsRS transcripts. Primer extension analysis of the nrsBACD (A) and nrsRS (B) transcripts from Synechocystis wild type, nrsRS::C.K1(+) and nrsRS::C.K1(–) mutant cells exposed to 17 μM Ni2+ for 2 h (+) or control cells (–).To avoid polar effects, insertional mutants with the npt gene of the C.K1 cassette in the same orientation [nrsRS::C.K1(+)] or in opposite orientation [nrsRS::C.K1(–)] than the nrsRS genes were used. Sequencing ladders generated with the same oligonucleotides used for primer extension are also shown.

C. Schematic representation and sequence of the nrs intergenic region. Transcripts stars points are marked with an arrow. Putative –10 boxes based on the transcription start site are boxed. Direct repeats are shaded. The translation start codons are indicated in boldface type.

Primer extension analysis demonstrated the existence of one only transcription start point for the nrsRS transcription unit 10 bp upstream of the nrsR translation start codon (Fig. 4B). As shown in RNA blotting experiments, primer extension products from the nrsRS promoter were more abundant in samples from cells exposed to Ni2+, than in non-exposed cells. Ni2+-dependent transcription from the nrsRS promoter was abolished in the nrsRS::C.K1 mutants indicating that expression of nrsRS genes is autoregulated. A putative –10 box in the form CGTTAT was found 5 bp upstream of the transcription start point (Fig. 4C).

NrsR binds to the nrsRS-nrsBACD intergenic region

NrsR belongs to the PhoB/OmpR subfamily of response regulators (Martinez-Hackert and Stock, 1997). Most response regulators consist of multiple domains: an N-terminal receiver domain that contains the aspartic residue that is phosphorylated by the histidine kinase and a C-terminal output domain that often binds DNA and activates transcription. To verify whether NrsR is able to interact with the nrsRS-nrsBACD intergenic region, we purified a recombinant amino-terminal His-tagged version of NrsR expressed in E. coli (Fig. 5A and B). DNA binding was tested by gel retardation assays using a fragment encompassing the 118 bp nrsRS-nrsBACD intergenic region. Full-length NrsR was not able to bind to this region (Fig. 5C), even in the presence 100 mM of acetyl phosphate. Similar results were obtained using a carboxy terminal His-tagged NrsR protein (data not shown). In a number of cases, it has been proposed that the unphosphorylated receiver domain functions as an intramolecular repressor of the DNA binding and transactivation activities of the N-terminal domain (Baikalov et al., 1996; Ames et al., 1999). This repressive effect is naturally released upon phosphorylation. A repressive role has been further supported by the fact that deletion of the N-terminal receiver domain leads to a constitutive DNA binding activity (Perez-Martin and de Lorenzo, 1996; Ellison and McCleary, 2000). Therefore, we generated a truncated form of NrsR deleted from the first 117 amino acids of the N-terminal part (Fig. 5A and B). This truncated version, named NrsRΔN, bound specifically to the nrsRS-nrsBACD intergenic region (Fig. 5C). The NrsRΔN-dependent band shift was severely diminished in the presence of a 10-fold excess of the same unlabelled fragment and it was unaffected by the presence of an excess of an unrelated DNA fragment (data not shown).

Figure 5.

NrsRΔN binds to the nrsRS-nrsBACD intergenic region.

A. Schematic representation of NrsR and NrsRΔN recombinant proteins domain organization.

B. SDS–PAGE of purified NrsR and NrsRΔN. Lane M, molecular mass markers.

C. Band-shift assay of the nrsRS-nrsBACD intergenic region with increasing quantities (from 0.07 to 5.6 μM) of purified NrsR (left) and NrsRΔN (right).

The nrsRS::C.K1 Synechocystis strain showed normal expression of photosynthetic genes

Li and Sherman have previously reported the inactivation of the rppA (nrsR) and rppB (nrsS) genes by insertional mutagenesis (Li and Sherman, 2000). Surprisingly, whereas ΔrppA cells showed a pleiotropic phenotype characterized by reduced chlorophyll and phycobiliprotein content and mis-expression of photosynthetic genes, ΔrppB cells behaved as wild-type cells. Thus, expression of the psbA (gene encoding the D1 protein of photosystem II) and nblA (gene encoding a polypeptide involved in degradation of phycobiliproteins) was strongly upregulated (around 5- and 100-fold respectively) in ΔrppA cells compared with wild-type cells. Li and Sherman suggest that RppA (NrsR) is controlling the stoichiometry between photosystem I and photosystem II. Our nrsRS::C.K1 Synechocystis strain is a double mutant lacking both NrsR and NrsS products; however, we did not observe any effect on growth or reduction in the chlorophyll (31 ± 2.10–3μg Chl μg–1 total protein for the nrsRS::C.K1 cell versus 33 ± 3.10–3μg Chl μg–1 total protein for the wild-type cells) and phycobiliprotein (PC) content (306 ± 35.10–3μg PC μg–1 total protein for the nrsRS::C.K1 cell versus 352 ± 11.10–3μg PC μg–1 total protein for the wild-type cells). Furthermore, Northern blot experiments using psbA and nblA probes demonstrated that both genes were equally expressed in nrsRS::C.K1 and wild-type cells (Fig. 6). Therefore, our results contrast with the data published for the ΔrppA mutant strain. Taking into account our results, one simple explanation could be that the pleiotropic phenotype reported for the ΔrppA mutant was the consequence of a high intracellular accumulation of Ni2+ and that this phenotype was not observed in the nrsRS::C.K1 mutant because of differences in the Ni2+ concentration of our BG11C medium. To test this hypothesis, expression of nblA and psbA genes was analysed in the presence of a growth-inhibitory Ni2+ concentration. Three different mutants presenting low tolerance to Ni2+ were analysed: nrsRS::C.K1, nrsA:C.K1 and nrsD::C.K1 (García-Domínguez et al., 2000). As shown in Fig. 6, neither nblA nor psbA gene expression was affected by the Ni2+ concentration in any of the analysed strains.

Figure 6.

nblA and psbA expression in nrs mutants. Northern blot analysis of nblA and psbA expression in Synechocystis wild type, nrsRS::C.K1, nrsA::C.K1 and nrsD::C.K1 (García-Domínguez et al., 2000) cells under normal growth conditions or after treatment with 17 μM Ni2+ for 6 h. The filters were stripped and hybridized with a rnpB gene probe as control.


We have reported previously that Synechocystis contains a metal resistance gene cluster formed by nine genes involved in Ni2+, Co2+ and Zn2+ resistance (García-Domínguez et al., 2000). The Co- and Zn-dependent P-type ATPases, encoded by the genes corT (also called coaT) and ziaA, are regulated in a metal-dependent way by transcription factors also encoded by genes of the cluster (corR/coaR and ziaR) (Thelwell et al., 1998; Rutherford et al., 1999; García-Domínguez et al., 2000) (see Fig. 1). However, genes controlling the Ni2+-dependent induction of the nickel resistance system (nrsBACD) were unknown. In this study, we demonstrate that ORFs sll0797 (nrsR) and sll0798 (nrsS) also form part of the same gene cluster and constitute a two-component signal transduction system involved in Ni2+ sensing and regulation of the nrsBACD operon. Thus, the nrsRS::C.K1 mutant displays a reduced tolerance to Ni2+ and it is unable to induce the nrsBACD operon in the presence of Ni2+. Whereas the carboxy terminal part of the NrsS protein shares homology with the histidine kinases of the PhoR subfamily, the amino terminal half appears to be a periplasmic domain, as predicted by two possible transmembrane helices (amino acids 14–34 and 187–214). This putative periplasmic domain shows significant sequence similarity with the alpha subunit of the methyl CoM reductase (MCR) from several Methanobacteria (Ermler et al., 1997) (Fig. 7). MCR is the final step enzyme of the methanogenesis pathway and it catalyses the reduction of methyl-CoM to methane. Interestingly, one of the prosthetic groups of the enzyme is a tetrapyrrole ring of coenzyme F430 which co-ordinates a nickel atom. In addition to the four tetrapyrrole nitrogens, nickel is also co-ordinated by a Gln residue of the MCR alpha subunit (marked in Fig. 7) (Ermler et al., 1997) which is located in the region that aligns with the NrsS periplasmic domains. These data raise the interesting possibility that the Ni2+-sensing domain of NrsS is phylogenetically related to the MCR alpha subunit. Whether the periplasmic domain of NrsS contains a tetrapyrrole ring involved in Ni2+ binding is currently unknown. NrsR is a response regulator of the OmpR/PhoB subfamily (Martinez-Hackert and Stock, 1997). PhoB-like factors contain two functional domains, an N-terminal phosphorylation domain (receiver domain) and a C-terminal DNA-binding/transactivation domain. In several cases, it has been demonstrated that phosphorylation of the receiver domain induces dimerization and binding to the DNA, whereas unphosphorylated proteins are unable to bind DNA (Fiedler and Weiss, 1995; Baikalov et al., 1996; Da Re et al., 1999). It has also been shown that deletion of the receiver domain yields proteins that can bind to the DNA constitutively, suggesting that the unphosphorylated receiver domain functions as an intramolecular repressor of the DNA binding; (Perez-Martin and de Lorenzo, 1996; Ellison and McCleary, 2000). In agreement with this view, recombinant NrsR purified from E. coli was unable to bind to the nrsRS-nrsBACD intergenic region; however, a mutant protein deleted of the first 117 amino acids bound specifically to this region. PhoB binds DNA tandemly to two direct repeated sequences placed around 10 bp upstream of the –10 box (Okamura et al., 2000). Upon sequence examination, two direct repeats in the form GA(T/A)TTTCA, separated by 3 bp, were found 12 bp upstream of the –10 box of the inducible nrsBACD promoter (Fig. 4B). In the same region, but in the opposite strand, two direct repeats also separated by 3 bp, and related to the previous one (GAAA(T/A)TC(A/T), were found 22 bp upstream of the –10 box of the nrsRS promoter (Fig. 4C). The long distance between the activator site and the nrsRS–10 box might account for the lower inducibility of the nrsRS promoter with respect to the nrsBACD promoter (see Fig. 2B). Thus, this sequence analysis suggests that four binding sites for NrsR monomers may exist in the nrsRS-nrsBACD intergenic region. Therefore, our current model suggests that the presence of Ni2+ in the medium stimulates the kinase activity of NrsS which transfers a phosphate group to NrsR. Phosphorylated NrsR binds (probably as a dimer) to the nrsRS-nrsBACD intergenic region activating the transcription of nrsBACD genes and positively autoregulating its own synthesis.

Figure 7.

Sequence alignment of the NrsS N-terminal periplasmic domain with the alpha subunit of the methyl CoM reductases from different methanobacteria. metvo, Methanococcus voltae; barker, Methanosarcina barkeri; metka, Methanopyrus kandleri. Glutamine residue involved in nickel co-ordination is marked with a vertical arrow. Identical amino acids are marked with (*), conservative changes are marked with ‘:’ or ‘.’ as defined by CLUSTALX (Thompson et al., 1997).

Our results concerning the role of the NrsR and NrsS proteins contrast with those of Li and Sherman (2000). Their nrsR insertional mutant (ΔrppA) presents a redox imbalance characterized by altered levels of phycobiliproteins and chlorophylls, upregulation of genes encoding photosystem II proteins, downregulation of genes encoding photosystem I proteins and a 100-fold induction of the nblA gene (NblA is a small polypeptide involved in degradation of phycobiliproteins). Strikingly, the nrsS mutant (ΔrppB) showed wild-type phenotype. To explain this result, the authors suggest that RppA may be phosphorylated by another histidine kinase. Our nrsRS::C.K1 mutant harbours a replacement of most of the nrsR sequence and part of the nrsS sequence by a kanamycin resistance cassette and, therefore, is a double mutant. However, this strain showed normal growth rate, normal levels of phycobiliproteins and chlorophylls and normal expression of psbA and nblA genes. Therefore, the phenotype observed by Li and Sherman (2000) in the ΔrppA mutant cannot be attributed to the lack of NrsR, but to the expression of NrsS in the absence of its natural substrate (NrsR). Thus, a cross-talk between NrsS and a response regulator involved in redox control may be responsible for the phenotype observed in the ΔrppA mutant.

The results shown here demonstrate conclusively that the NrsRS two-component system is controlling the nickel-dependent expression of the nrsBACD operon. How nickel sensing is carried out and how the signal is transduced is currently being investigated.

Experimental procedures

Bacterial strains and growth conditions

Synechocystis sp. strain PCC 6803 was grown photoautotrophically at 30°C in BG11 (Rippka et al., 1979) medium supplemented with 1 g per litre of HCO3Na (BG11C) and bubbled with a continuous stream of 1% (v/v) CO2 in air under continuous fluorescent illumination (50 μmol of photons per m2 per second, white light). For plate cultures, BG11C liquid medium was supplemented with 1% (wt/vol) agar. Kanamycin was added to a final concentration of 50–200 μg ml–1 when required. BG11C medium was supplemented with different concentrations of ZnSO4, CdCl2, CoCl2, CuSO4, NiSO4 and MgCl2 when indicated.

Escherichia coli DH5α (Bethesda Research Laboratories), grown in Luria-Bertani broth (LB broth) medium as described in Sambrook et al. (1989) was used for plasmid construction and replication. Escherichia coli BL21(DE3) grown in LB broth medium was used for expression of NrsR and NrsRΔN His-tagged proteins. Escherichia coli was supplemented with 100 μg ml–1 ampicillin or 50 μg ml–1 kanamycin when required.

Insertional mutagenesis of Synechocystis genes

Loci sll0797 and sll0798 were inactivated by replacing a 535 bp HpaI–EcoRI fragment by a kanamycin resistance cassette (C.K1) (Elhai and Wolk, 1988). For this, a genomic DNA fragment was polymerase chain reaction (PCR)-amplified with oligonucleotides NIW1-NIW2 from the cosmid cs1377 (provided by Kazusa DNA Research Institute) and cloned into pGEM-T (Promega) to generate pNIQ6. The targeting vector was generated by inserting the C.K1 cassette into HpaI–EcoRI-digested pNIQ6, in the same orientation as the sll0797 gene (pNIQ9+) or in the inverse orientation (pNIQ9–). Synechocystis was transformed with these plasmids to generate the nrsRS::C.K1(+) and nrsRS::C.K1(–) mutant strains. Correct integration and complete segregation of the mutant strain was tested by Southern blotting. For this, total DNA from the cyanobacteria was isolated as described previously (Cai and Wolk, 1990). DNA was digested with HincII and electrophoresed in 0.7% agarose gels in a Tris-borate-EDTA buffer system (Sambrook et al., 1989), then DNA was transferred to nylon Z-probe membranes (Bio-Rad). DNA probes were 32P-labelled with a random-primer kit (Pharmacia) using [α-32P]-dCTP (3000 Ci mmol–1).

RNA isolation and Northern blot analysis

Total RNA was isolated from 25 ml samples of Synechocystis cultures at the mid-exponential phase (3–5 μg chlorophyll ml–1). Extractions were performed by vortexing cells in the presence of phenol–chloroform and acid-washed baked glass beads (0.25–0.3 mm diameter; Braun) as described previously (García-Domínguez and Florencio, 1997).

For Northern blotting, 15 μg of total RNA was loaded per lane and electrophoresed in 1.0% agarose denaturing formaldehyde gels, and transferred to nylon membranes (Hybond N-plus). Prehybridization, hybridization and washes were performed as described in the Amersham instruction manual.

Probes for Northern blot hybridization were PCR-synthesized using the following oligonucleotides pairs: NIA3-NIA4 (García-Domínguez et al., 2000) for the nrsB probe; NIW1-NIW2 for the nrsRS probe; NBLA1-NBLA2 for nblA probe; and PSBA1-PSBA2 for the psbA probe (Table 1). As a control, in all the cases the filters were stripped and reprobed with a HindIII–BamHI 580 bp probe from plasmid pAV1100 that contains the constitutively expressed RNase P RNA gene (rnpB) from Synechocystis (Vioque, 1992). DNA probes were 32P-labelled with a random-primer kit (Pharmacia, Sweden) using [α-32P]-dCTP (3000 Ci mmol–1). To determine cpm of radioactive areas in Northern blot hybridizations, whole band radioactivity was quantified using an InstantImager Electronic Autoradiography apparatus (Packard Instrument Company). The signal was then normalized by calculating the ratios between the mRNA signal and rnpB signal for each lane, to avoid loading effects. Then, maximal ratio was considered as 100% induction for each mRNA.

Table 1. Oligonucleotides used in this work.

Primer extension analysis of nrsRS and nrsBACD transcripts

Oligonucleotides NIW3 and NIA1, end-labelled with T4 polynucleotide kinase and [γ-32P]-dATP (3000 Ci mmol–1) following standard procedures (Sambrook et al., 1989), were used for primer extension analysis of nrsRS or nrsBACD promoters respectively. For annealing a 10 μl mixture containing 0.15 M HCl, 10 mM Tris HCl pH 8.0, 1 mM EDTA, 20 μg of total RNA and about 2 pmoles of oligonucleotide (106 cpm) were prepared. The annealing mixture was heated for 2 min at 90°C in a water bath and cooled slowly to 50°C. For extension, a 10 μl mixture was prepared with half of the annealing mixture: 10 mM DTT, 0.5 mM each dNTP, 2 μg of Actinomycin D, 50 mM Tris HCl pH 8.3, 75 mM KCl, 3 mM MgCl2 and 100 U of Superscript™ II RNase H-Reverse Transcriptase (Gibco BRL). The mixture was incubated for 45 min at 45°C, and the reaction was stopped by adding 4 μl of formamide-loading buffer. Half of the reaction was electrophoresed on a 6% polyacrylamide sequencing gel together with a sequencing reaction of the nrsRS or nrsBACD promoter region, using the NIW3 or NIA1 oligonucleotide respectively.

Cloning and purification of NrsR and NrsRΔN His-tagged proteins

The complete nrsR ORF was PCR-amplified with oligonucleotide TCP1 (which introduces a NdeI restriction site) and oligonucleotide TCP5 (which introduces a BamHI restriction site) (Table 1), digested with NdeI–BamHI and cloned into the NdeI–BamHI-digested pET28a (Novagen) to generate pTCP3. The NrsR protein expressed from this plasmid contains a six-histidine tag in the amino terminus. nrsRΔN was cloned by amplifying a 386 bp (from amino acid 118 to the end of the protein) using oligonucleotide TCP4 (which introduces a NdeI restriction site) and oligonucleotide TCP5 (which introduces a BamHI restriction site), digested with NdeI–BamHI and cloned into the NdeI–BamHI-digested pET28a (Novagen) generating pTCP2. The NrsRΔN protein expressed from this plasmid contains a six-histidine tag in the amino terminus. NrsR and NrsRΔN were expressed in E. coli BL21 from the plasmids pTCP3 and pTCP2 respectively. Then, 200 ml of culture was grown in L broth to an optical density at 600 nm of 0.6, induced with 1 mM isopropyl-β-D-thiogalactopyranoside for 2.5 h, harvested by centrifugation and resuspended in 8 ml of Tris-HCl pH 8.0 50 mM, KCl 50 mM, 10% glycerol, 0.1% Triton X-100, 5 mM imidazol (buffer A) supplemented with 1 mM phenylmethylsulphonyl fluoride. Cells were broken by sonication, and insoluble debris was pelleted by centrifugation at 18 000 g for 15 min. The supernatant was then applied to a Ni2+-charged, His·Bind, beads column (Novagen) (1 ml bead volume), washed with buffer A and eluted with the same buffer containing 200 mM imidazol. Imidazol was removed dialysing the samples against buffer A without imidazol. The purity of the samples was about 90%, as determined by scanning densitometry from the Coomassie blue-stained gel.

Gel retardation assays

The probe was PCR-synthesized using oligonucleotides NIP1-NIP2 (Table 1), which introduce StyI and NcoI restriction sites, respectively, from cosmid cs1377 (provided by Kazusa DNA Research Institute), and the resulting DNA was digested with either of the enzymes and end-labelled with [α-32P]-dCTP (3000 Ci mmol–1) using Klenow fragment. The binding reaction was carried out in a final volume of 25 μl containing 4 ng of labelled DNA and 4 μg salmon sperm DNA in 20 mM Tris-HCl pH 8.0, 150 mM KCl, 10 mM spermidine, 10 mM DTT, 1 mM EDTA, 10% glycerol and different amounts of partially purified NrsR and NrsRΔN. The mixtures were incubated for 25 min at 4°C and loaded on a non-denaturing 6% polyacrylamide gel. Electrophoresis was carried out at 4°C and 280 V in 10 mM Na2HPO4 pH 6.0. Gels were transferred to a Whatman 3 MM paper, dried and autoradiographed.


We thank Kazusa DNA Research Institute and Dr S. Tabata for providing cs1377 cosmid DNA. We thank Marika Lindahl for critical reading of the manuscript. Luis López-Maury is the recipient of a fellowship from the Spanish Ministerio de Educación Cultura y Deporte. This work was supported by grant PB97–0732 from DGESIC and by Junta de Andalucía (group CV1–0112).