Histidine kinases play important roles in the perception and signal transduction of hydrogen peroxide in the cyanobacterium, Synechocystis sp. PCC 6803


(fax +81 557 85 5205; e-mail murata@nibb.ac.jp).


Oxidative stress caused by reactive oxygen species and, in particular, to hydrogen peroxide (H2O2) has a major impact on all biological systems, including plants and microorganisms. We investigated the H2O2-inducible expression of genes in the cyanobacterium Synechocystis sp. PCC 6803 using genome-wide DNA microarrays. Our systematic screening of a library of mutant lines with defects in histidine kinases (Hiks) by RNA slot-blot hybridization and DNA-microarray analysis suggested that four Hiks, namely, Hik33, Hik34, Hik16 and Hik41, are involved in the perception and transduction of H2O2 signals that regulate the gene expression of 26 of the 77 H2O2-inducible genes with induction factors higher than 4.0. Among the four Hiks, Hik33 was the main contributor and was responsible for 22 of the 26 H2O2-inducible genes under the control of the Hiks. By contrast to Hik33, PerR encoding putative peroxide-sensing protein is involved in the regulation of only nine H2O2-inducible genes.


The production of reactive oxygen species (ROS), such as hydrogen peroxide (H2O2) and superoxide anions, is an inevitable consequence of aerobic life. It has been suggested that the presence of ROS results in damage to proteins, cleavage of nucleic acids, the peroxidation of unsaturated fatty acids (Halliwell and Gutteridge, 1999) and, in particular, in the inhibition of protein synthesis (Nishiyama et al., 2001, 2006). Photosynthetic organisms generate ROS during photosynthesis. However, concentrations of ROS are maintained at harmless levels by the action of various antioxidant systems and enzymes, such as catalases, peroxidases and superoxide dismutases in plants (Blokhina et al., 2003; Mullineaux et al., 2000) and catalase-peroxidase and thioredoxin peroxidase in cyanobacteria (Tichy and Vermaas, 1999; Yamamoto et al., 1999). However, their actions might not be sufficient to protect cellular components against sudden large increases in concentrations of ROS that can result from the distortions of metabolism caused by changes in environmental conditions. Enhanced protection against H2O2, over and above the actions of catalases and peroxidases, involves the detection of increases in concentrations of H2O2 and subsequent appropriate regulation of the expression of large numbers of genes, which confer the ability to tolerate the oxidative stress.

In Escherichia coli, a LysR-type transcription factor, OxyR, and a ‘[2Fe-2S] center’ transcription factor, SoxR, participate in the perception of signals arising from H2O2 and superoxide radicals, respectively (Kim et al., 2002; Pomposiello and Demple, 2001; Zheng et al., 1998). In Bacillus subtilis, a Fur-type transcription factor, PerR, regulates the expression of a group of H2O2-responsive genes (Helmann et al., 2003). In yeasts two-component systems, such as Sln1-Skn7 in Saccharomyces cerevisiae and Mak2-Mak3 in Schizosaccharomyces pombe, are involved in the response to H2O2 and were found to alter gene expression (Buck et al., 2001; Lee et al., 1999; Li et al., 2002). However, in plants such ROS-sensing systems are not well understood.

The genome of Synechocystis sp. PCC 6803 (hereafter Synechocystis) includes a gene for a homolog of PerR (Li et al., 2004), but there are no other genes that encode proteins homologous to the above-mentioned transcription factors. By contrast, considerable evidence suggests that histidine kinases (Hiks) act as receptors and/or transducers of environmental stress signals in Synechocystis. For example, pathways for the transduction of signals as a result of cold stress (Mikami et al., 2002; Suzuki et al., 2000), osmotic stress (Mikami et al., 2002; Paithoonrangsarid et al., 2004), salt stress (Marin et al., 2003; Shoumskaya et al., 2005) and strong light (Hsiao et al., 2004; Tu et al., 2004) are regulated by a variety of Hik-containing two-component systems in this microorganism.

In the present study, using DNA microarrays, we examined the effects of the peroxide signal caused by H2O2 on the genome-wide expression of genes in Synechocystis in an attempt to define factors that are involved in the signal transduction of H2O2-dependent oxidative stress. We identified four Hiks that appear to be involved in the perception of elevated levels of H2O2 and in the subsequent regulation of expression of a large number of H2O2-inducible genes.


Mutants of Hiks

Histidine kinases are important in the perception and signal transduction of environmental signals in bacteria, cyanobacteria, fungi and plants (West and Stock, 2001). The genome of Synechocystis includes 47 putative genes for Hiks (Kaneko et al., 1996, 2003), some of which might be involved in the signal transduction of H2O2 stress for the regulation of the gene expression at the level of transcript. To examine the possibility that Hiks which have already been characterized might be involved in the perception of H2O2 stress, we examined cell lines with mutations in Hik33 (involved in signaling of cold stress, osmotic stress and salt stress; Murata and Suzuki, 2006; Suzuki et al., 2000, 2001), and Hik34, Hik16 and Hik41 (in hyperosmotic and salt signaling; Marin et al., 2003; Murata and Suzuki, 2006; Paithoonrangsarid et al., 2004; Shoumskaya et al., 2005).

As indicated in Figure 1a, which shows the structure of the Synechocystis genome in the vicinity of the sites of the various mutations, the hik33, hik34 and hik16 genes seem to exist as individual respective transcriptional units, whereas the hik41 gene is located in a gene cluster. Thus, mutation of hik41 might be expected to affect the expression of the genes located downstream of this gene. However, DNA-microarray analysis indicated that the expression of these downstream genes was unaffected by mutation of the hik41 gene (see below). Analysis by PCR (Figure 1b) indicated that replacement of the wild-type gene by the respective mutated gene was complete in all lines of ΔHik mutant cells.

Figure 1.

 Insertional mutagenesis of histidine kinase (hik) and perR genes involved in the perception of hydrogen peroxide (H2O2) signals in Synechocystis sp. PCC 6803.
(a) Sites of insertion of either a spectinomycin- or a chloramphenicol-resistance gene cassette in the wild-type genome in each mutant. Solid bars in each figure correspond to 1 kb. Primers for PCR are indicated by small black arrowheads.
(b) Analysis by PCR of cells with mutations in either hik or perR genes. Each pair of columns shows results for wild-type (WT) and mutant cells.

Oxidative stress caused by H2O2 induces the expression of a large number of genes

We examined the effects of various concentrations of H2O2 on the growth of Synechocystis. In the presence of 0.25 mm H2O2, cell growth was unaffected during a 24-h incubation. By contrast with 0.35 mm H2O2 cell growth was strongly repressed. Therefore, we chose 0.25 mm as the concentration of H2O2 for investigations of the responses of Synechocystis to H2O2 in terms of gene expression.

We examined the responses of gene expression in Synechocystis to incubation with 0.25 mm H2O2 for 20 min using DNA microarrays (Figure 2a) and found that the expression of 225 genes was induced with induction factors higher than 2.0, whereas the expression of 320 genes was decreased with induction factors lower than 0.5. The H2O2-inducible genes included the hspA gene for a small heat-shock protein; the gifA and gifB genes for subunits of glutamine synthetase-inactivating factor; the hliA, hliB and hliC genes for three high-light-inducible proteins; the nblA and nblA2 genes for phycobilisome-degradation proteins; the dnaJ gene for heat-shock protein 40; and the perR gene for peroxide-dependent transcription factor (Table 1). The full data set is available at http://www.genome.jp/kegg/expression/.

Figure 2.

 DNA microarray analysis of changes in patterns of expression of H2O2-regulated genes in (a) wild-type, (b) ΔHik33, (c) ΔHik34, (d) ΔHik16, (e) ΔHik41, and (f) ΔPerR cells upon the incubation in the presence of 0.25mM H2O2 for 20 min. Each assay with the DNA microarry was performed twice. Each point represents the average of results of two independent experiments. Single experimental data was obtained by an average of the values due to duplicated spots of each gene in DNA microarray. Dashed lines indicate the induction factor 4.0. Gray circles indicate that the H2O2-inducible genes whose induction factor in wild-type cells were higher than 4.0.

Table 1.   Genes whose H2O2-inducible expression was regulated either by a histidine kinases (Hik) or by PerR. Cells of the exponential phase of growth (O.D.730nm = 0.2–0.3) were incubated in the presence of 0.25 mm H2O2 for 20 min. Each value indicates the ratio of the relative level of expression of the indicated gene in stressed cells to that in unstressed cells. The numbering of open reading frames (ORFs) and the annotation of gene products correspond to that in the CyanoBase (http://www.kazusa.or.jp/cyanobase/). This table lists the H2O2-inducible genes with induction factors equal to or higher than 4.0 in wild-type cells. The total list can be accessed at http://www.genome.ad.jp/kegg/kegg2.html
ORFNameProductLevels of induction after a 20-min incubation with 0.25 mm H2O2
  1. The bold lettering indicates genes that were used as probes for the screening of the Hik knockout library by RNA slot-blot hybridization to identify other sensor Hiks that are not mentioned above. Underlining indicates genes whose expression was regulated by both Hik33 and PerR. Bold-faced numbers depict genes whose expression by H2O2 stress was significantly changed by the mutation of a Hik or PerR with the RE value (see Results) lower than 15%. Values are averages and ± error ranges of the results of two independent experiments.

Gene whose H2O2-inducible expression was abolished in ΔHik33 cells
 slr1544 Hypothetical protein131.3 ± 11.21.5 ± 0.5116.1 ± 6.788.3 ± 16.488.9 ± 5.834.0 ± 4.0
 ssr2016pgr5Ferredoxin:plastoquinone reductase120.2 ± 25.81.3 ± 0.658.3 ± 2.672.3 ± 11.473.6 ± 9.424.0 ± 0.3
 ssr2595hliBHigh light inducible protein49.2 ± 5.40.9 ± 0.134.3 ± 0.344.8 ± 12.245.2 ± 5.620.4 ± 3.4
 ssl2542hliAHigh light inducible protein29.7 ± 7.50.9 ± 0.017.3 ± 0.614.9 ± 0.512.7 ± 0.38.4 ± 0.7
 ssl0453nblA2Phycobilisome degradation protein NblA29.3 ± 5.53.8 ± 1.826.8 ± 7.27.5 ± 0.622.0 ± 5.44.4 ± 0.5
 slr1687 Hypothetical protein27.9 ± 4.53.6 ± 0.314.3 ± 0.913.3 ± 2.715.2 ± 1.817.9 ± 2.7
 ssl0452nblA1Phycobilisome degradation protein NblA18.1 ± 4.32.4 ± 0.318.0 ± 4.97.2 ± 0.813.1 ± 1.73.3 ± 0.8
 slr1291ndhD2NADH dehydrogenase I chain M18.0 ± 2.31.2 ± 0.114.2 ± 1.06.2 ± 1.37.8 ± 0.71.0 ± 0.1
 sll1483 Periplasmic protein15.2 ± 6.30.5 ± 0.013.0 ± 0.25.9 ± 1.89.0 ± 1.85.6 ± 1.9
 ssl1633hliCCAB/ELIP/HLIP superfamily12.4 ± 2.21.2 ± 0.27.9 ± 1.35.6 ± 1.13.5 ± 0.17.9 ± 2.7
 sll0330 Sepiapterine reductase9.7 ± 2.61.3 ± 0.332.7 ± 8.16.6 ± 1.45.6 ± 1.19.5 ± 3.6
 slr0611sdsSolanesyl diphosphate synthase9.4 ± 1.51.9 ± 0.16.1 ± 0.45.6 ± 0.27.0 ± 0.65.3 ± 0.4
 ssl2162 Hypothetical protein8.6 ± 1.91.1 ± 0.24.3 ± 0.63.8 ± 0.75.4 ± 0.94.0 ± 0.1
 ssl3446 Hypothetical protein8.1 ± 1.92.0 ± 1.35.6 ± 0.16.0 ± 0.34.4 ± 1.39.7 ± 1.0
 slr0228ftsHCell division protein FtsH homolog6.9 ± 1.50.9 ± 0.16.0 ± 0.23.7 ± 0.53.9 ± 0.33.7 ± 0.0
 slr0270 Hypothetical protein6.5 ± 1.91.4 ± 0.25.5 ± 0.62.7 ± 0.24.7 ± 0.43.9 ± 0.5
 sll0157 Hypothetical protein6.4 ± 1.00.9 ± 0.16.2 ± 1.22.0 ± 0.13.8 ± 0.54.6 ± 1.1
 sll1541 Lignostilbene-alpha, beta- dioxygenase5.3 ± 0.90.8 ± 0.13.9 ± 0.42.7 ± 0.12.1 ± 0.13.2 ± 0.6
 sll0992 Esterase5.0 ± 1.01.6 ± 0.12.5 ± 0.32.1 ± 0.32.8 ± 0.02.2 ± 0.0
 slr0518abfBArabinofuranosidase4.3 ± 1.61.1 ± 0.42.1 ± 0.22.7 ± 0.12.8 ± 0.22.8 ± 0.1
 slr1747 Hypothetical protein4.1 ± 0.50.9 ± 0.22.4 ± 0.22.4 ± 0.02.1 ± 0.32.0 ± 0.2
 sll0086 Arsenical pump-driving ATPase4.1 ± 0.40.9 ± 0.24.7 ± 0.12.2 ± 0.32.8 ± 0.63.7 ± 0.7
Gene whose H2O2-inducible expression was abolished in ΔHik34 cells
 sll0430htpGHtpG, heat shock protein 904.8 ± 0.65.2 ± 0.30.8 ± 0.07.8 ± 2.52.2 ± 0.116.2 ± 4.7
 slr1285hik34Histidine kinase Hik344.0 ± 1.88.9 ± 4.60.7 ± 0.011.4 ± 0.56.1 ± 2.315.9 ± 0.5
Gene whose H2O2-inducible expression was abolished in ΔHik16 and ΔHik41cells
 slr0967 Hypothetical protein26.5 ± 4.719.4 ± 2.827.1 ± 3.54.8 ± 0.51.1 ± 0.035.6 ± 7.5
 sll0939 Hypothetical protein13.3 ± 3.410.3 ± 6.214.5 ± 2.52.7 ± 4.10.9 ± 0.036.6 ± 11.5
Gene whose H2O2-inducible expression was abolished in ΔPerR cells
 sll1621ahpCAhpC13.4 ± 6.026.7 ± 4.118.0 ± 2.816.2 ± 0.813.7 ± 1.41.0 ± 0.1
 slr0589 Hypothetical protein12.1 ± 3.722.6 ± 0.015.7 ± 0.014.5 ± 0.812.5 ± 0.51.0 ± 0.1
 slr1738perRTranscription factor PerR12.0 ± 2.210.6 ± 5.416.3 ± 0.35.4 ± 0.711.0 ± 2.91.7 ± 0.2
 sll1620 Hypothetical protein5.4 ± 0.65.7 ± 3.610.3 ± 1.33.1 ± 0.35.6 ± 0.91.3 ± 0.1
 sll1550 Probable porin4.5 ± 1.12.6 ± 0.32.0 ± 0.32.9 ± 0.32.4 ± 0.21.4 ± 0.1
 slr0587 Hypothetical protein4.5 ± 0.42.8 ± 0.62.2 ± 0.02.7 ± 0.22.4 ± 0.31.4 ± 0.0
Gene whose H2O2-inducible expression was not abolished in either of ΔHik and ΔPreR cells
 sll1515gifBGlutamine synthetase inactivating factor IF1781.3 ± 4.588.3 ± 0.990.3 ± 3.364.0 ± 14.372.8 ± 4.464.7 ± 9.0
 sll1514hspASmall heat-shock protein76.8 ± 17.457.5 ± 7.228.6 ± 0.062.6 ± 4.038.2 ± 3.4122.7 ± 49.2
 sll0528 Hypothetical protein74.4 ± 27.327.5 ± 1.068.7 ± 8.269.5 ± 8.096.2 ± 3.476.1 ± 3.2
 sll1549 Salt-enhanced periplasmic protein39.4 ± 9.321.4 ± 11.312.8 ± 2.412.4 ± 0.214.1 ± 1.524.3 ± 10.0
 ssl1911gifAGlutamine synthetase inactivating factor IF733.2 ± 7.825.5 ± 7.433.6 ± 3.816.4 ± 5.425.9 ± 4.212.4 ± 3.9
 sll0846 Hypothetical protein29.6 ± 8.315.4 ± 1.89.2 ± 2.016.2 ± 2.711.9 ± 0.729.8 ± 11.1
 ssl3044 Hydrogenase component23.4 ± 2.65.8 ± 2.012.2 ± 0.112.2 ± 3.417.4 ± 1.125.2 ± 7.7
 slr0093dnaJDnaJ protein, heat shock protein 4022.1 ± 5.721.4 ± 5.912.3 ± 2.618.6 ± 1.218.9 ± 1.124.8 ± 1.8
 slr1674 Hypothetical protein18.9 ± 0.526.6 ± 5.416.9 ± 0.016.3 ± 1.114.9 ± 1.416.3 ± 2.5
 slr1641clpB1ClpB protein13.7 ± 8.26.3 ± 1.35.0 ± 0.125.9 ± 3.33.4 ± 0.339.7 ± 1.2
 slr1603 Hypothetical protein13.2 ± 1.915.4 ± 2.49.6 ± 1.114.7 ± 0.512.5 ± 1.618.2 ± 0.1
 sll0185 Hypothetical protein11.5 ± 1.53.2 ± 0.210.0 ± 1.46.9 ± 1.87.1 ± 0.36.3 ± 0.8
 slr1318fecEABC-type iron(III) dicitrate transport system9.1 ± 2.14.0 ± 0.73.1 ± 0.72.5 ± 0.14.8 ± 0.14.4 ± 0.5
 sll0248isiBFlavodoxin8.9 ± 2.310.3 ± 6.09.0 ± 2.19.3 ± 0.34.3 ± 1.420.6 ± 5.1
 sll0247isiAIron-stress induced chlorophyll- binding protein8.2 ± 5.422.0 ± 4.97.9 ± 0.310.6 ± 0.214.8 ± 3.923.2 ± 11.9
 slr1686 Hypothetical protein8.2 ± 2.42.4 ± 1.19.1 ± 1.25.6 ± 0.88.0 ± 2.614.2 ± 0.3
 sll0990 Formaldehyde dehydrogenase (glutathione)8.1 ± 1.52.9 ± 1.26.5 ± 0.12.1 ± 0.24.6 ± 0.83.4 ± 0.9
 slr1295futA1ABC-type iron transport system7.9 ± 1.25.4 ± 0.55.5 ± 0.23.8 ± 0.56.7 ± 0.06.1 ± 0.8
 ssr2333 Hypothetical protein7.9 ± 1.24.1 ± 2.46.7 ± 0.62.4 ± 0.02.9 ± 0.57.4 ± 2.2
 ssr0692 Hypothetical protein7.1 ± 1.37.7 ± 1.56.3 ± 0.12.6 ± 0.53.1 ± 0.15.2 ± 1.1
 slr1484 Hypothetical protein7.0 ± 1.34.0 ± 0.43.4 ± 0.35.4 ± 0.62.4 ± 0.25.7 ± 0.8
 slr1963 Water-soluble carotenoid protein6.7 ± 2.49.3 ± 0.15.6 ± 0.15.6 ± 0.712.4 ± 0.17.7 ± 1.2
 sll0170dnaK2DnaK protein 2, heat shock protein 706.5 ± 1.611.4 ± 0.64.5 ± 0.111.4 ± 1.78.3 ± 0.712.8 ± 2.4
 slr0008ctpAcarboxyl-terminal processing protease6.4 ± 0.65.5 ± 0.73.9 ± 0.04.1 ± 0.94.5 ± 0.26.1 ± 0.6
 slr1675hypA1Putative hydrogenase expression/ formation protein6.3 ± 2.47.7 ± 2.03.4 ± 0.211.0 ± 0.24.5 ± 0.79.5 ± 0.9
 sll0306sigBRNA polymerase group 2 sigma factor6.2 ± 0.815.1 ± 0.76.9 ± 0.314.3 ± 0.48.2 ± 0.414.1 ± 0.8
 slr1119 Hypothetical protein5.8 ± 1.27.6 ± 0.25.0 ± 0.92.8 ± 0.15.0 ± 0.17.6 ± 1.5
 slr1413 Hypothetical protein5.7 ± 2.34.0 ± 0.23.2 ± 0.16.5 ± 0.42.0 ± 0.16.0 ± 0.9
 sll0249 Hypothetical protein5.7 ± 2.47.7 ± 0.13.5 ± 0.812.7 ± 0.23.0 ± 0.216.6 ± 0.3
 sll0688 Hypothetical protein5.7 ± 1.42.9 ± 1.83.0 ± 0.32.0 ± 0.23.0 ± 1.04.0 ± 0.4
 ssl0832 Hypothetical protein5.6 ± 0.33.4 ± 0.53.7 ± 0.42.2 ± 0.33.1 ± 0.52.8 ± 0.5
 sll1158 Hypothetical protein5.5 ± 1.85.4 ± 3.22.3 ± 0.53.5 ± 0.33.5 ± 0.711.2 ± 1.3
 slr0818 Hypothetical protein5.2 ± 0.92.5 ± 0.23.0 ± 0.46.0 ± 0.73.2 ± 0.22.2 ± 0.4
 slr0095 O-methyltransferase5.1 ± 1.23.7 ± 2.24.8 ± 0.65.5 ± 0.42.8 ± 0.88.3 ± 0.2
 ssl1533 Hypothetical protein5.0 ± 0.52.0 ± 0.43.8 ± 0.42.1 ± 0.12.2 ± 0.12.3 ± 0.4
 slr0513futA2ABC-type iron transport system4.9 ± 0.46.8 ± 1.44.8 ± 0.62.8 ± 0.14.5 ± 0.510.0 ± 0.1
 slr1851 Hypothetical protein4.7 ± 0.55.4 ± 1.22.2 ± 0.12.6 ± 0.32.5 ± 0.32.0 ± 0.1
 sll2012sigDRNA polymerase sigma factor4.7 ± 0.52.3 ± 0.45.1 ± 0.45.3 ± 0.56.2 ± 1.04.9 ± 0.0
 ssl2501 Hypothetical protein4.4 ± 0.12.7 ± 0.34.0 ± 0.12.8 ± 0.32.5 ± 0.23.8 ± 0.1
 slr1259 Hypothetical protein4.3 ± 1.52.8 ± 0.52.7 ± 0.33.2 ± 0.52.4 ± 0.25.0 ± 0.8
 slr0233trxMThioredoxin M4.2 ± 0.44.0 ± 0.23.5 ± 0.12.3 ± 0.43.6 ± 0.32.3 ± 0.3
 sll1774 Hypothetical protein4.1 ± 0.64.0 ± 0.14.5 ± 0.04.1 ± 0.04.9 ± 0.14.1 ± 0.0
 slr0852 Hypothetical protein4.1 ± 0.84.8 ± 2.14.0 ± 0.33.2 ± 0.13.0 ± 0.67.0 ± 0.9
 sll0788 Hypothetical protein4.0 ± 1.38.7 ± 0.88.2 ± 0.53.6 ± 0.28.3 ± 0.215.4 ± 7.4
 slr1476pyrBAspartate carbamoyltransferase catalytic chain4.0 ± 0.64.3 ± 0.83.8 ± 0.22.9 ± 0.03.3 ± 0.41.9 ± 0.4

Hiks are involved in the perception of H2O2 signals

Table 1 lists the H2O2-inducible genes the expression of which was affected in ΔHik33, ΔHik34, ΔHik16 and ΔHik41 mutant cells. To quantitatively evaluate the impact of each mutation of Hiks, the effectiveness of each mutation on the H2O2-inducible expression of a gene, RE value (Shoumskaya et al., 2005), was calculated by a following method, RE = [(induction factor in ΔHik cells − 1.0)/(induction factor in wild-type cells − 1.0)] × 100 (%). In this equation, 1.0 was subtracted from each induction factor because the absence of a change in expression corresponds to an induction factor of 1.0. We defined the H2O2-inducible genes, the expression of which was strictly affected by the mutation of a Hik, as those that had RE values of <15%. H2O2-inducible genes with RE values of <15% are listed in Table 1.

Among 77 genes whose expression was strongly induced by H2O2 in wild-type cells, with induction factors higher than 4.0, 22 H2O2-inducible genes were either totally or almost totally unresponsive to H2O2 in ΔHik33 mutant cells. By contrast, mutation of Hik34, Hik16 or Hik41 abolished the H2O2 induction of only two genes. It appeared, moreover, that Hik16 and Hik41 regulated the H2O2-inducible expression of the same two genes.

We also investigated the response of ΔPerR cells to 0.25 mm H2O2 in order to determine whether the genes regulated by Hik33, Hik34, Hik16 and Hik41 differed from those regulated by PerR. Analysis by PCR (Figure 1b) indicated that replacement of the wild-type gene by the respective mutated gene was complete in ΔPerR mutant cells. Mutation of PerR eliminated the inducibility by H2O2 of the nblA1, nblA2 and ndhD2 genes, which were also under the control of Hik33 and the inducibility by H2O2 of six other genes, which were not controlled by any of the Hiks that we examined (Table 1). These findings suggested that signals caused by H2O2, mediated by both Hik33 and PerR, might be necessary to regulate the H2O2-inducible expression of the nblA1, nblA2 and ndhD genes. By contrast, the expression of the other six genes, which included the ahpC gene for alkyl hydroperoxide reductase and the perR gene, was regulated solely by PerR. In Bacillus subtilis, PerR, the ortholog of Synechocystis PerR, works as a major regulator of gene expression under H2O2 stress (Helmann et al., 2003). However, our results indicated that the signal transduction pathway that includes Hiks, such as Hik33, plays an important role in the regulation of the expression of a large number of H2O2-inducible genes in Synechocystis.

The H2O2-induced expression of a large number of genes (45 of 77 H2O2-inducible genes) was similar to that in wild-type cells in all of the above-mentioned Hik and PerR mutant lines (Table 1). To determine whether the expression of these 45 genes might be controlled by other Hiks, we screened our Hik ‘knockout’ library by slot-blot hybridization with probes derived from the gifB, hspA, isiA and sll0846 genes, the expression of which in response to 0.25 mm H2O2 was not regulated by any of Hik33, Hik34, Hik16, Hik41 and PerR. However, we failed to identify a role of any other Hik in the response to H2O2 stress on the expression of these genes. Our results suggest a possibility that an unknown H2O2 sensor, which is neither a Hik nor PerR, regulates the expression of these 45 genes.

Northern blotting analysis of H2O2-inducible genes the expression of which was under the control of either Hiks or PerR

We also performed northern blotting analysis of the H2O2-inducible expression of genes in wild-type, ΔPerR and four kinds of Hik-mutant cells to confirm the results of the DNA microarray (Figure 3). We selected ssr2016, slr0967 and ndhD2 genes as representatives of genes under the control of Hik33, Hik16 plus Hik41 and PerR plus Hik34, respectively (Table 1). The H2O2-induced expression of the ndhD2 gene was under the control of both Hik33 and PerR (Table 1). As shown in Figure 3a, the H2O2-induced expression of ssr2016 and slr0967 was either diminished or significantly reduced in mutant cells of ΔHik33, and both ΔHik16 and ΔHik41, respectively. The induced expression of the ndhD2 gene was remarkably decreased in both ΔHik33 and ΔPerR. The results of northern blotting were consistent with those obtained with the microarray method, confirming the reliability of the result with DNA microarray.

Figure 3.

 Northern blotting analysis of changes in gene expression.
(a) Hydrogen peroxide (H2O2)-induced expression of genes that were regulated under the control of histidine kinases (Hiks) and PerR. Wild-type and mutant cells at the exponential phase of growth (OD730nm = 0.2–0.3) were incubated in the presence of 0.25 mm H2O2 for 20 min.
(b) Changes in gene expression caused by the mutation of either a Hik or PerR. Wild-type and mutant cells were grown under normal conditions. The coding region of each gene was amplified by PCR and purified for the use as a probe. Each experiment was repeated twice and essentially the same results were obtained.

Positive and negative regulation of gene expression

To determine whether the H2O2-induced expression of genes by various Hiks and PerR is regulated in either a positive or a negative manner, we examined the effects of inactivation of individual Hiks and PerR on the expression of genes in the absence of exogenously applied H2O2 (Figure 4 and Table 2). Separate inactivation of Hik33, Hik34, Hik16, Hik41 and PerR induced changes in the expression of some genes. A very small number of these genes belonged to the group of genes whose expression was strongly induced by H2O2 via a mechanism that involved either Hik34 or PerR. The expression of the htpG gene, which, in the presence of 0.25 mm H2O2, was controlled by Hik34, was enhanced upon inactivation of Hik34. This was confirmed by northern blotting analysis (Figure 3b). These observations suggest that Hik34 might regulate the expression of the htpG gene in a negative manner. A previous study (Suzuki et al., 2005) demonstrated that Hik34 transiently regulates the expression of heat-shock genes during acclimation to high-temperature conditions. However, the mechanism for the transient expression by Hik34 is unknown. Although mutation of Hik34 also enhanced the expression of slr1634 and slr1808 (hemA) genes (Table 2), exogenous H2O2 did not enhance the expression of these genes (Table 1). These observations suggest that such genes might be regulated by other kinds of stress in a negative manner.

Figure 4.

 DNA microarray analysis of changes in patterns of gene expression caused by the mutation of histidine kinases (Hiks), (a) Hik33, (b) Hik34, (c) Hik16, (d) Hik41 and PerR (e) under normal growth conditions.
Each assay was repeated twice and the values represent the averages of results from two independent experiments. Each point represents the average of duplicate results for each gene in the DNA microarray. Dashed lines indicate the ranges of experimental deviation in induction factor between 2.0 and 0.5 (Kanesaki et al., 2002; Suzuki et al., 2001).

Table 2.   Genes whose expression was enhanced by the mutation of a histidine kinases (Hik) or PerR. Wild-type and the mutant cells were grown to an OD730nm of 0.2–0.3 under control conditions. Each value indicates the ratio of the level of expression of the transcript in mutant cells to that in wild-type cells. The numbering of open reading frames (ORFs) corresponds to that given in CyanoBase (http://www.kazusa.or.jp/cyanobase/). Values are averages and ± error ranges of the results of two independent experiments.
ORFNameProductChanges in expression
  1. Underlining indicates genes whose induction factors caused by H2O2 stress were higher than 4.0 (see Table 1).

Genes whose expression was induced by mutation of Hik33
 sll1471cpcG2Phycobilisome rod-core linker polypeptide6.2 ± 1.2
 slr1535 Hypothetical protein3.1 ± 0.1
 sll1694pilA1Cyanobacterial pilin2.5 ± 0.5
 sll1695pilA2Cyanobacterial pilin2.0 ± 0.5
Genes whose expression was induced by mutation of Hik34
 sll0430htpGHeat-shock protein2.9 ± 0.2
 slr1634 Hypothetical protein2.7 ± 1.0
 slr1808hemATransfer RNA-Gln reductase2.2 ± 0.7
Genes whose expression was induced by mutation of Hik16
 slr0288glnNGlutamate-ammonia ligase2.2 ± 0.1
 sll1453nrtDABC-type nitrate/nitrite transport system2.1 ± 0.0
Genes whose expression was induced by mutation of Hik41
 slr1920 Hypothetical protein5.2 ± 0.8
 sll1695pilA2Cyanobacterial pilin2.6 ± 0.1
 slr1753 Hypothetical protein2.1 ± 0.3
 sll0381 Hypothetical protein2.1 ± 0.2
Genes whose expression was induced by mutation of PerR
 sll1621ahpCAlkylhydroperoxide reductase40.4 ± 3.4
 slr0589 Hypothetical protein28.5 ± 5.0
 ssl2982ycf61DNA-directed RNA polymerase4.7 ± 0.4
 slr0376 Hypothetical protein4.6 ± 1.9
 sll0381 Hypothetical protein3.8 ± 1.4
 slr1738perRTranscription factor PerR3.5 ± 0.5
 sll1688thrCThreonine synthase2.6 ± 0.3
 slr1737 Hypothetical protein2.5 ± 0.0

Mutation of PerR by insertion of the Cmr cassette into the middle position of the gene enhanced the expression of ahpC, slr0589 and perR genes, the expression of which was strongly induced by H2O2 via a mechanism that involved PerR (Table 1). The negative regulation of expression of ahpC and perR genes by PerR has been previously studied by Li et al. (2004) and Kobayashi et al. (2004). Our data in Table 1 and 2 indicated that the enhanced expression of the ahpC gene, perR, and slr0589 genes by H2O2 may be regulated by the derepression of transcription by PerR. However, the expression of other similarly H2O2-inducible genes under the control of PerR, such as sll1620, sll1550 and slr0587, was unaffected by the inactivation of PerR under unstressed conditions (Table 2). These findings suggest that PerR is also able to act as a positive regulator of these H2O2-inducible genes. Therefore it may be likely that PerR of Synechocystis regulates the expression of some of the H2O2-inducible genes in a negative manner, but that it may regulate the expression some other genes in a positive manner.

In either E. coli or B. subtilis, H2O2-inducible expression of most of the H2O2-inducible genes was regulated by transcription factors, such as either OxyR or PerR, respectively, which have active thiol groups and directly sense the redox state in the cytosol. However, in Synechocystis OxyR does not exist and PerR regulates a limited number of genes. In this organism, two-component systems may play a most important role for the H2O2-signal transduction.

Inactivation of Hik33, Hik16 and Hik41 also resulted in changes in the expression of several genes in each case. However, these genes were not members of the group of H2O2-responsive genes (Table 1). It seems likely that Hik33, Hik16 and Hik41 might respond to other kinds of environmental stress and might negatively regulate the expression of these genes. Upon exposure of cells to such environmental stress, inactivation of Hik33, Hik16 or Hik41 would be expected to enhance the expression of the above-mentioned genes under normal conditions. However, it appears that these Hiks do not act negatively to regulate the H2O2-inducible expression of any genes.


H2O2-inducible genes and their functions

Using DNA microarrays, we demonstrated that the genome-wide pattern of gene expression in Synechocystis is dramatically changed by exogenous H2O2. Our results suggest that the response to peroxidative stress is effectively regulated at the level of gene expression in this microorganism. Peroxidative stress as a result of 0.25 mm H2O2 remarkably enhanced the expression of 77 genes in wild-type cells, with induction factors higher than 4.0. The genes can be divided into six groups with respect to the functions of the encoded proteins.

The first group includes genes for the regulation of the intracellular redox state, such as the ahpC gene for an antioxidative enzyme and the pgr5 gene for the regulation of the cyclic electron transport of photosystem I. The ahpC gene, identified in E. coli, encodes a homolog of a peroxiredoxin in the thiol-specific antioxidant (TSA)/AhpC family, which can scavenge H2O2 and alkylhydroxyperoxide (Jeong et al., 2000). In Synechocystis, two enzymes catalyze the reduction of H2O2 to H2O, namely, catalase (Tichy and Vermaas, 1999) and thioredoxin peroxidase (Yamamoto et al., 1999), which are encoded by the katG and tpx genes, respectively. These proteins seem to be expressed constitutively in cells that are grown in the light under conditions that allow intracellular concentrations of H2O2 to remain at low and harmless levels (Tichy and Vermaas, 1999; Ushimaru et al., 2002; Yamamoto et al., 1999). However, the activity of scavenging enzymes might be insufficient to meet the challenge presented by exogenous H2O2 at 0.25 mm. Therefore, expression of the ahpC gene might be necessary for the scavenging of ROS generated by reactions with various metal ions, such as the Fenton reaction. The H2O2-inducible expression of the ahpC gene was controlled by PerR. The pgr5 gene encodes a protein with substantial similarities of amino acid sequence to a component of the antimycin A-sensitive ferredoxin:plastoquinone reductase (FQR) in Arabidopsis thaliana (Munekage et al., 2002). The pgr5 mutant of Synechocystis is sensitive to strong light, suggesting that FQR might be important in controlling the intracellular redox state under adverse conditions (Yaremenko et al., 2005). The H2O2-inducible expression of the pgr5 gene was controlled by Hik33. These results also suggest that the Hik33-dependent and PerR-dependent pathways are important for the control of the intracellular redox state.

The second group includes genes for proteins related to the response to iron deficiency, such as fecE, futA1, futA2, isiA and isiB, that are not under the control of any of the Hiks or PerR. FutA1 and FutA2 are components of an ABC-type transport system, which is a major transporter of ferric ions in Synechocystis (Katoh et al., 2001). The FecE protein is a homolog of a component of the ABC-type ferric dicitrate transporter in E. coli (Staudenmaier et al., 1989). The isiA and isiB genes encode the chlorophyll-binding protein CP43’ (Burnap et al., 1993) and flavodoxin, respectively (Straus, 1994). Expression of these genes is also induced under iron-deficient conditions in Synechocystis (Kobayashi et al., 2004; Vinnemeier et al., 1998). The metabolism of iron and several iron-containing proteins might be closely related to cellular responses to oxidative stress (Hantke, 2001; Touati, 2000). Regulation of the expression of genes in this group might be important for acclimation to peroxidative stress.

The third group includes genes for the regulation of transcription, such as sigB and sigD for sigma factors, which were not under the control of any of Hiks and PerR (Table 1). The H2O2-inducible expression of these genes was also observed by Li et al. (2004). These findings may suggest that regulation of the transcription of certain genes is important for acclimation to peroxidative stress.

The fourth group includes genes for heat-shock proteins, such as hspA, dnaJ, dnaK2 and htpG. The enhanced expression of heat-shock genes suggests that chaperones might be important for acclimation of Synechocystis to peroxidative stress. We found previously that protein synthesis de novo in Synechocystis is strongly inhibited by exogenous H2O2 (Nishiyama et al., 2001). Enhanced expression of genes for chaperones might be necessary for unhindered translation. Apart from induction by H2O2 of the htpG gene, which was controlled by Hik34, the H2O2-induced expression of heat-shock proteins was regulated by an as-yet uncharacterized mechanism. The expression of these heat-shock genes was induced transiently by heat shock (Inaba et al., 2003: Suzuki et al., 2005), strong-light stress (Hihara et al., 2003), UV irradiation (Huang et al., 2003), hyperosmotic stress (Paithoonrangsarid et al., 2004) and salt stress (Shoumskaya et al., 2005). These findings indicate that the induction of the heat-shock proteins is very rapid and a general response to various kinds of stress.

The fifth group includes genes for proteases, such as ftsH, clpB1 and ctpA, and for proteins of phycobilisome degradation, such as nblA1 and nblA2. The expression of these genes is also induced by UV irradiation (Huang et al., 2003). In Synechocystis, H2O2 inhibits de novo synthesis of the D1 protein in photosystem II (Nishiyama et al., 2001), and thus decreases the rate of turnover of the D1 protein in photosystem II. The induction expression of genes for proteases, including FtsH, which are located in thylakoid membranes and are involved in the degradation and rapid turnover of the D1 protein (Itzhaki et al., 1998; Lindahl et al., 2000; Nixon et al., 2005), may be necessary for the rapid turnover of D1 under peroxidative stress. The expression of nblA1 and nblA2 genes was controlled by both Hik33 and PerR, whereas that of ftsH genes was controlled only by Hik33.

The sixth group includes 35 genes for hypothetical proteins of unknown function (Table 1). Among these 35 genes, four genes, namely, slr0967, slr0589, sll0157 and slr1259, the induction of which by H2O2 was regulated by Hik16 plus Hik41, PerR, Hik33 and an as-yet uncharacterized mechanism, respectively, are conserved in A. thaliana and in eight species of cyanobacteria, the genomes of which have been sequenced (Cyanobase; http://www.kazusa.or.jp/cyanobase/). These genes might be important for the acclimation of plants and cyanobacteria to peroxidative stress. By contrast, 27 of the 35 genes are conserved in other cyanobacteria but not in A. thaliana, and four genes, such as ssl2162, ssl1533, ssl2501 and slr0587, are found only in Synechocystis. These results suggest that a number of hypothetical proteins, the functions of which are unknown, might be involved in the acclimation of cyanobacterial cells to peroxidative stress.

Genes whose expression was decreased by incubation with H2O2 includes the genes for phycobilisome components, photosystem-I components and a number of hypothetical proteins (the whole set of data is available from http://www.genome.jp/kegg/expression/). The expression of these genes was also decreased by UV irradiation (Huang et al., 2003) and strong-light stress (Hihara et al., 2001). Rapid reduction of photosystem I and phycobilisome components may suggest that the regulation of the photosynthetic activity by decreasing the expression of these genes is one of the primary responses for the acclimation to many kinds of stress.

Peroxidative stress-inducible genes in other organisms

In E. coli, the expression of 140 genes (about 3% of the total number of genes) is induced by 1 mm H2O2 with induction factors of either 4.0 or higher (Zheng et al., 2001). In A. thaliana, the expression of 113 genes is induced by 20 mm H2O2 with induction factors of either 1.5 or higher (Desikan et al., 2001). Comparing the H2O2-inducible genes in E. coli with those in Synechocystis, we failed to identify any homologous genes whose expression was induced by H2O2 in both microorganisms. However, the expression of the hspA, htpG, dnaJ, slr0442 and sll1615 genes, which correspond to hsp17, hsp83, AA04366 (EST accession no.), N37850 (EST accession no.) and the ras gene for GTP-binding protein in A. thaliana, respectively, is induced by H2O2 in both Synechocystis and A. thaliana. As only a small number of genes are induced in common, it is possible that the responses to H2O2 differ among species.

Regulation by Hiks of the H2O2-inducible expression of genes

As noted above, mechanisms for the H2O2-induced regulation of the expression of genes seem to differ among species. In this study, we found that Hiks are involved in the expression of a large number of H2O2-inducible genes in Synechocystis. Figure 5 shows a hypothetical scheme for the perception of H2O2 stress signals and the subsequent regulation of gene expression. Localization of Hik molecules either in the cytoplasm or within the membrane was predicted by the smart program (http://smart.embl-heidelberg.de/) and the sosui program (http://bp.nuap.nagoya-u.ac.jp/sosui/) for individual Hiks. The amino acid sequence of Hik33 indicates that it has two transmembrane regions, a histadine kinases, adenylyl cyclases, methyl binding proteins and phosphatases (HAMP)-linker, a period circadian protein, Ah receptor nuclear translocator protein, single-minded protein (PAS) domain and a Hik domain (Mikami et al., 2002). Hik34 has a Hik domain (Marin et al., 2003) and a proline- and leucine-rich region at its amino terminus. Hik16 has seven membrane-spanning domains, a cGmp-specific phosphodiesterases (GAF) domain and a Hik domain (Marin et al., 2003). Hik41 has a receiver domain and a Hik domain at its amino and carboxyl termini, respectively (Marin et al., 2003). From the results in Table 1 and the domain structures of Hik16 and Hik41, we propose that Hik16 is located in the membrane, perceives the signal caused by H2O2 and transfers the signal, as a phosphate group, to the receiver domain of Hik41, which is located in the cytosol (Marin et al., 2003).

Figure 5.

 A hypothetical scheme for perception and transduction of hydrogen peroxide (H2O2) stress signals in Synechocystis sp. PCC 6803.
Putative signal cascades are represented by bold arrows. Representative genes, the expression of which is selectively induced by H2O2, are listed in boxes. The numbers enclosed by squares indicate the total numbers of H2O2-inducible genes with induction factors higher than 4.0 and with RE values of less than 15%. The contribution of histidine kinase 33 (Hik33) and PerR in the expression of the ndhD2,nblA1 and nblA2 genes is suggested by the finding that a mutation of either of these components eliminated the H2O2-induced expression of these genes (see Table 1).

It has been suggested that PAS and GAF domains might bind co-factors, such as heme and cGMP, respectively (Galperin et al., 2001). Although the PAS domain of Hik33 does not contain the cysteine residues that are necessary for binding to a heme molecule, these functional domains might be involved in the perception of signals caused by H2O2 by an as-yet unknown mechanism. It has been suggested that the integrity of the periplasmic region of Sln1, a Hik in S. cerevisiae, might be essential for sensing osmotic stress and oxidative stress (Reiser et al., 2003). As the structure of Sln1 is similar to that of Hik33 (Mikami et al., 2002), it is possible that the periplasmic region of Hik33 might also be related to the perception of peroxidative signals in Synechocystis.

Tu et al. (2004) and Hsiao et al. (2004) reported that Hik33 negatively regulates the expression of hli genes under control conditions. Their conclusion was based on the changes in the global expression of genes under normal conditions, and the complementation of ΔHik33 cells of Synechocystis sp. Pcc. 7942 with the nblS gene from Synechococcus with respect to the light-induced expression of the hli genes assayed by the northern blotting analysis. As the complementation test was not performed by the genome-wide expression of genes, it is likely that their ΔHik33 cells contain a mutation in addition to the mutation of the hik33 gene (Murata and Los, 2006).

Our results demonstrated that Hiks and PerR regulate the H2O2-induced expression of a number of genes. It is likely that as-yet unknown mechanisms are also involved in the H2O2-regulated expression of genes. The data presented here will form the basis for future research on the entire pathway of H2O2 signal transduction.

Experimental procedures

Strains and culture conditions

The wild-type strain of Synechocystis sp. PCC 6803 was kindly donated by J. G. K. Williams (Du Pont de Nemours, Wilmington, DE, USA). Mutations in Hiks were generated as described previously (Suzuki et al., 2000; http://www.kazusa.or.jp/cyanobase/Synechocystis/mutants/). In four of the resultant mutant lines, the wild-type genes for Hik33, Hik34, Hik16 and Hik41 were disrupted by the insertion of a spectinomycin-resistance gene cassette (Paithoonrangsarid et al., 2004; Suzuki et al., 2000). A mutation in the gene for PerR was generated by the insertion of a chloramphenicol-resistance gene cassette at the Eco47III site of the perR gene in a DNA fragment that included the perR gene and had been amplified by PCR using primers slr1738F (5′-ACATATGCCCTTAAACAGAGAAGAAATT-3′) and slr1738R (5′-TGGATCCCTAATCGCCCCAGTTTTCACA-3′) and cloned into pT7Blue (Novagen Inc., Madison, WI, USA).

Wild-type and mutant cells were grown at 34°C in BG-11 medium (Stanier et al., 1971) buffered with 20 mm Hepes-NaOH (pH 7.5) under continuous illumination by light at 70 mE m−2 sec−1 from incandescent lamps, as described previously (Wada and Murata, 1989). Wild-type and mutant cells at the exponential phase of growth (OD730 nm = 0.2–0.3) were used for all experiments.

DNA-microarray analysis and northern blotting analysis

A 50-ml suspension of cells, either before or after incubation for 20 min in the presence of H2O2, was poured into an equal volume of an ice-cold mixture of phenol and ethanol (1:10, w/v). Then total RNA was extracted, as described previously (Los et al., 1997), and treated with RNase-free Dnase I (Nippon Gene, Tokyo, Japan) to remove any contaminating genomic DNA. Genome-wide analysis of gene expression was performed with Synechocystis DNA microarrays, as described previously (Suzuki et al., 2001; Kanesaki et al., 2002). The control experiment with wild-type cells (Kanesaki et al., 2002) demonstrated that the range of experimental errors in the induction factor was <2.0 and >0.5, as indicated by the dashed lines in Figures 2 and 4. Each DNA microarray experiment was duplicated and the average of the values obtained from two arrays was calculated. Northern blotting analysis was performed as described previously (Kanesaki et al., 2002; Los et al., 1997).


This work was supported, in part, by Grants-in-Aid for Scientific Research (no. 14086207 to NM and no. 16013249 to IS) from the Ministry of Education, Science, Sports and Culture of Japan. It was also supported by the Program for Cooperative Research on the Stress Tolerance of Plants of the National Institute for Basic Biology, Japan.