Novel Putative Photoreceptor and Regulatory Genes Required for the Positive Phototactic Movement of the Unicellular Motile Cyanobacterium Synechocystis sp. PCC 6803

Abstract

Synechocystis sp. PCC 6803 is a unicellular motile cyanobacterium, which shows positive or negative phototaxis on agar plates under lateral illumination. By gene disruption in a substrain showing of positive phototaxis, it was demonstrated that mutants defective in sll0038, sll0039, sll0041, sll0042 or sll0043 lost positive phototaxis but showed negative phototaxis away from the light source. Mutants of sll0040, which is located within the cluster of these genes, retained the capacity of positive phototaxis but to a lesser extent than the parent cells. These genes are homologous to che genes, which are involved in flagellar switching for bacterial chemotaxis. Interestingly, sll0041 (designated pisJ1) is predicted to have a chromophore-binding motif of phytochrome-like proteins and a signaling motif of chemoreceptors for bacterial chemotaxis. It is strongly suggested that the positive phototactic response was mediated by a phytochrome-like photoreceptor and CheA/CheY-type signal transduction system.

(Received September 25, 2000; Accepted October 26, 2000).

Introduction

Photosynthetic organisms must sense and choose their light environment since weak light is not sufficient to support the photosynthesis and excess light damages the photosystem and cellular processes. For this purpose, many cyanobacteria and eukaryotic algae show motility in response to light. The motility of cyanobacteria has been featured by gliding, swimming or twitching in the absence of flagella (Diehn et al. 1977, Häder 1987, Waterbury et al. 1985). The responses to light have been categorized into three different processes: photokinesis, photophobic response and phototaxis (Häder 1987). Photokinesis is a movement whose speed depends on the fluence rate of irradiation. The photophobic response consists of a reversal of movement, temporary stop or a directional change upon a sudden change in the fluence rate. By contrast, phototaxis is a movement oriented with respect to the direction of light but not to the fluence rate. Some cyanobacteria such as Phormidium autumnale, Anabaena variabilisand Synechocystis sp. show a true directed movement of cells toward the light source or a movement away from the light source (Nultsch 1961, Nultsch et al. 1979, Stanier et al. 1971). Generally, these responses finally result in accumulation or escape of cells in a certain light field. Although these photosensory behaviors have been studied for many years, the molecular mechanisms underlying photoperception and signal transduction are still obscure.

Synechocystis sp. PCC 6803 is a unicellular motile cyanobacterium, whose entire genome sequence has been determined (Kaneko et al. 1996). Over 3,000 open reading frames (ORFs) have been detected and some of them were found to bear significant homology at least in part to known genes, which encode receptors for light or chemical compounds, signal transducers and regulators in other organisms. Since Synechocystis sp. PCC 6803 is naturally transformable, the roles of these ORFs can be challenged by targeted disruption, although none of them have been shown to be involved in phototactic motility. On the other hand, there is increasing evidence that the motility of Synechocystis cells is mediated by a pilus structure as in the twitching motility and the type IV pilus structure of Gram-negative non-photosynthetic bacteria (Mattick et al. 1996, Darzins and Russell 1997). These findings were obtained by creating disrupted mutants of pil-related ORFs in Synechocystis sp. PCC 6803 (Bhaya et al. 1999, Bhaya et al. 2000, Okamoto and Ohmori 1999, Yoshihara et al. 2001).

Here, we describe by gene disruption evidence that a series of ORFs is essential for the positive phototactic movement of Synechocystis sp. PCC 6803. These ORFs are homologous to the genes for the flagellar switching in chemotaxis and for the biogenesis of type IV pili in various bacteria. One of them is also homologous to the genes of phytochrome-like photoreceptors in phototrophic organisms. To our knowledge, this is the first line of evidence for genes involved in the photoperception and signal transduction for the phototactic movement in cyanobacteria.

Materials and Methods

The motile strain of the unicellular cyanobacterium Synechocystis sp. PCC 6803 was obtained from the Pasteur Culture Collection (PCC). Clones that showed a positive or negative phototactic movement were isolated from the cells from PCC (see Results and Discussion). Cells were grown in liquid BG11 medium (Stanier et al. 1971) bubbled with air containing 1% (v/v) CO₂ at 31°C under illumination with a white fluorescent lamp at a fluence rate of 50 µE m^–2 s^–1. Kanamycin was included at 20 µg ml^–1 when mutants were screened and maintained.

Phototactic movement was examined on 0.8% (w/v) agar-solidified BG11 (Bacto-Agar, Difco, Detroit, U.S.A.) supplemented with 0.3% (w/v) sodium thiosulfate under lateral illumination with a white fluorescent lamp at 10 µE m^–2 s^–1 for 3–4 d.

The protein database derived from the genome of Synechocystis sp. PCC 6803 was searched for homology using the BLAST program (Altschul et al. 1997). Each target ORF was amplified with a set of primers by polymerase chain reaction (Fig. 1, Table 1), cloned into pT7Blue-T vector (Novagen, Madison, U.S.A.), and then interrupted at a unique restriction site by insertion of Tn5-derived kanamycin-resistant cassette in the same direction as the ORF. These constructs were designed to inactivate not only the target ORF but also to allow transcriptional expression of the downstream gene(s) from the promoter of aminoglycoside 3′-phosphotransferase in the cassette (Kamei et al. 1998). Both positive and negative phototactic cells of Synechocystis were transformed with these DNAs as described previously (Hihara and Ikeuchi 1997). At least four independent transformants were selected and subjected to full segregation, which was confirmed by polymerase chain reaction (not shown).

Results and Discussion

It is prerequisite to know the phototactic properties of the parent cells for gene disruption. We found that colonies derived from single cells obtained from PCC as strain number 6803, showed either positive or negative phototactic movements under lateral illumination at 10 µE m^–2 s^–1. Some colonies reproducibly moved towards the light source, whereas the others moved away from the light source (Fig. 2). Thus, we named the positive and negative phototactic clones as PCC-P and PCC-N, respectively. Although these phototactic properties are stable, we do not know their genetic background at the moment. Closer observation of both colonies revealed that a considerable number of cells gathered together to form the leading tips of colonies, while a few cells remained at the original position (Fig. 2, dotted line). There were also some cells on the way to the tips, indicative of heterogeneity in motility. Moreover, the cells remaining at the original position often began moving after illumination for several days. Although the phototactic movement was such a complex phenomenon, the phototactic orientation of colonies of PCC-P and PCC-N was reproducible and thus suitable for gene disruption analysis. It should be noted that PCC-P and PCC-N grew at similar rates under the experimental conditions. We used both PCC-P and PCC-N for the gene disruption in this study.

In the entire genome of Synechocystis sp. PCC 6803, we could identify only one ORF (sll0041), whose deduced product has a hybrid feature of the chromophore-binding in the phytochrome-related photoreceptors and the highly conserved signaling domain of the methyl-accepting chemotaxis protein, Tsr, for bacterial chemotaxis. This ORF is flanked with a series of ORFs in the same direction, which showed some similarities to chemotaxis genes of flagellated bacteria (Fig. 1). These ORFs were insertionally disrupted in both PCC-P and PCC-N with the kanamycin-resistant cassette, which allows read-through transcription.

Strikingly, the disruption mutants of sll0038, sll0039, sll0041, sll0042, and sll0043 (Msll0038, Msll0039, Msll0041, Msll0042, and Msll0043) derived from the clone PCC-P showed negative phototactic movement in contrast to the positive movement of PCC-P (Fig. 3). These results were obtained from all the independent transformants. In contrast, Msll0040 retained the positive phototactic properties, although its moving capacity appeared to be somewhat affected compared with the parent strain PCC-P. Since sll0040 is located in the middle of the gene cluster, it was suggested that the phenotype of all mutants was not due to a polar effect of the cassette insertion but was the true result of disruption of the target ORFs. It is noteworthy that the disruptants derived from the clone PCC-N showed more or less the same negative phototaxis as PCC-N (data not shown).

In addition to sll0041, a predicted product of the downstream ORF, sll0042, was also homologous to another methyl-accepting chemotaxis protein, Tar (Kaneko et al. 1996). In E. coli, Tsr and Tar are chemotaxis receptors for serine and aspartate, respectively. Both proteins have an N-terminal sensor domain protruded into the periplasmic space and a C-terminal signaling domain retained in the cytoplasm. As shown in Fig. 4, only the signaling domain of Tsr and Tar was conserved in Sll0041 and Sll0042 proteins. It has been established in the flagella-mediated chemotaxis in enteric bacteria that the C-terminal highly conserved domain of the chemoreceptor proteins is essential for both signal transduction to the flagellar motor and methylation of particular amino acid residues by methyltransferase CheR (Liu and Parkinson 1991). The methylation of Tsr or Tar proteins attenuates the sensitivity of their signal transduction system (Kehry and Dahlquist 1982). This is known as adaptation of chemotaxis to the signal and is critical for the true chemotaxis in nature. Notably, one or two of the five methylated residues of Tsr and Tar (Rice and Dahlquist 1991) were also conserved in Sll0041 and Sll0042 (Fig. 4, asterisks). However, a methyltransferase gene homologous to cheR was not detected in the Synechocystis genome, suggesting that Sll0041 and Sll0042 proteins may not be methylated. In accordance with this, typical adaptation to the light signal has not been observed in cyanobacteria, although the well-documented photophobic response may be a kind of adaptation (Häder 1987).

Sll0041 but not Sll0042 harbors a region homologous to the putative chromophore-binding motif for the phytochrome-like RcaE protein of a filamentous cyanobacterium Fremyella diplosiphon (Fig. 4). RcaE has been reported to be a photoreceptor for the complementary chromatic adaptation (Kehoe and Grossman 1996), in which the peripheral antenna of phycobilisome changes from the blue pigment-protein phycocyanin in red light to the red pigment-protein phycoerythrin in green light (Tandeau de Marsac 1977). Recent reconstitution experiments demonstrated that rather diverged motifs related to the plant phytochrome can be assembled with phycocyanobilin or phytochromobilin, which could be photoconverted by red and far red light (Hughes et al. 1997, Yeh and Lagarias 1998). It is of note that Sll0041 has two GAF domains, which are often detected in phototransducing proteins such as phytochrome, Anabaena adenylate cyclase and E. coli FhlA (Aravind and Ponting 1997). One GAF domain corresponded to the putative chromophore-binding region homologous to RcaE, while the other domain at position from 230 to 374 did not show discernable similarity to RcaE. Sll0041 does not have a PAS domain, which has been widely detected as an important signaling module in various sensors that monitor changes in light, redox potential or overall energy level of a cell (Taylor and Zhulin 1999).

Sll0041 but not Sll0042 has two hydrophobic segments at the N-terminus, suggestive of anchoring in the thylakoid or cytoplasmic membrane (Fig. 4, underlined). As a result, the putative chromophore-binding region and the signaling domain of Sll0041 are assumed to be located in the cytoplasm. On the analogy of the homodimeric structure of chemoreceptor proteins (Milligan and Koshland 1988), we may postulate a model that Sll0041 and Sll0042 proteins form a heterodimeric photoreceptor.

Synechocystis cells displayed the positive phototactic movement and orientation to the light between 560 and 730 nm (Choi et al. 1999). A similar action spectrum was obtained for the positive phototaxis of a gliding filamentous cyanobacterium Anabaena variabilis, in addition to a slight effect around 440 nm (Nultsch et al. 1979). Furthermore, inhibitors of photosynthetic electron transport did not affect the phototactic orientation in both organisms. These results can be interpreted as a specific photoreceptor with phycocyanobilin or phytochromobilin-related open tetrapyrrole pigment. Our results appear to agree with this interpretation. Isolation of Sll0041 protein from Synechocystis cells will help elucidate the pigment identity. The bacterial phytochrome-like proteins including the well-characterized Cph1 have been reconstituted with phycocyanobilin or phytochromobilin in vitro but has not been isolated as holocomplex in vivo (Hughes et al. 1997, Davis et al. 1999).

sll0038, sll0039, sll0040 and sll0043 are weakly but significantly homologous to cheY, cheY, cheW and cheA genes respectively (Fig. 1) (Kaneko et al. 1996). These che genes are involved in the chemotactic regulation of the flagellar motility in bacteria (Eisenbach 1996). CheA, the histidine kinase of the two-component regulatory system in bacteria, transmits signals from Tsr or Tar to CheY, the cognate response regulator, by transferring the phosphoryl group from the histidine residue of CheA to the aspartate residue of CheY. The critical histidine and aspartate residues were conserved in Sll0043 (CheA homolog) and Sll0038/Sll0039 (CheY homologs). CheW is required for the interaction between Tsr/Tar receptors and CheA (Borkovich et al. 1989). Although we do not have any biochemical evidence for these products in Synechocystis, it is tempting to speculate that Sll0043 is activated by Sll0041 in response to light and phosphorylates Sll0038 and Sll0039, which in turn activate a motility machinery for the positive photoresponse. In contrast with the essential role of CheW in the signal transduction in E. coli, Sll0040 (CheW homolog) did not appear to be essential for but to stimulate the positive phototaxis (Fig. 1).

The arrangement of the gene cluster from sll0038 to sll0043 in Synechocystis sp. PCC 6803 is quite different from the arrangement of che genes in E. coli (Blattner et al. 1997). Namely, tar is located in the gene cluster of cheA/cheW/tar/tap/cheR/cheB/cheY/cheZ, while tsr is located far from this cluster. By contrast, the arrangement in Synechocystis remarkably resembles the pilG cluster in Pseudomonas aeruginosa (Darzins 1993, Darzins 1994, Darzins and Russell 1997), as shown in Fig. 1. In addition, the motility of both Synechocystis and P. aeruginosa has been shown to depend on a similar pilus structure, although cyanobacteria and γ-proteobacteria (P. aeruginosa) are distantly related to each other in the phylogeny of the bacteria. Based on these, we propose to designate sll0038, sll0039, sll0040, sll0041, sll0042, and sll0043 as pis (phototaxis) G, pisH, pisI, pisJ1, pisJ2, and pisL, respectively. However, there are critical differences between the two gene clusters. The pilK (or cheR) homolog was not present in the sll0038 cluster or even in the entire genome as mentioned above. Both Sll0039 (Synechocystis sp. PCC 6803), and PilG and PilH (P. aeruginosa) are simple CheY-type response regulators, whereas Sll0038 has unique N-terminal domain in addition to the C-terminal CheY domain. The overall sequence of Sll0038 was highly homologous to PatA, which is responsible in an unknown manner for correct control of the heterocyst pattern formation under nitrogen limitation in Anabaena sp. PCC 7120 (Liang et al. 1992). The mutant phenotype of the pilG gene cluster also differed from that of the sll0038 cluster. Namely, disruption of pilG, pilI or pilJ resulted in a loss of the type IV pilus structure together with the complete lack in the twitching motility, while disruption of the other genes caused a partial loss in both pilus structure and motility in P. aeruginosa (Darzins 1993, Darzins 1994). At present, it is unknown how the pilus structure generates torque and how it contributes to the directional movement of cells. Nevertheless, the twitching motility of Synechocystis cells strictly depends upon a pilus structure, which is supported by a number of pil genes as in P. aeruginosa (Bhaya et al. 1999, Bhaya et al. 2000, Okamoto and Ohmori 1999, Yoshihara et al. 2001). There are two possible explanations for the mutant phenotype in Synechocystis. Firstly, the Synechocystis cells might have two types of pili for motility, each responsible for positive or negative phototactic movement. Only the pilus structure responsible for the positive phototaxis may have been lost in our mutants. Secondly, a single type of pilus structure in Synechocystis may work for both positive and negative movement, for example, by alternate rotation of the pilus filaments as in the flagellar response. In either case, the phototactic motility of Synechocystis cells would be a novel model system for elucidation of the pilus-mediated directional movements.

Negative phototaxis is another important response to avoid light of excess fluence. Wavelength dependence of this photoresponse has also been studied in Synechocystis (Choi et al. 1999) or A. variabilis (Nultsch et al. 1979) but the results were not so clear due to coincidence of the positive phototactic response. Our phototactic mutants, which are genetically devoid of the positive phototaxis, seem to be a better material than PCC-N or other wild type strains for studies on the negative phototaxis. They are suitable for taking the action spectrum of the negative phototactic response and also could be a starting material suitable for isolation of non-phototactic mutants by random tagging or directed gene disruption.

Acknowledgements

This work was supported by Grants-in-Aid for Scientific Research on Priority Areas C“ Genome Biology” (12206002) (to M.I.) and for Scientific Research C (08836002) and B (11554035) (to M.I.) from the Ministry of Education, Science and Culture, Japan and by a grant for Scientific Research from the Human Frontier Science Program (to M.I.).

3

Corresponding author: E-mail, [e-mail]; Fax, [phone].

Fig. 1 Gene organization of the Synechocystis sll0038 cluster and the P. aeruginosa pilG cluster. Homologous regulatory components for the chemotactic flagellar switching in enteric bacteria are shown below the corresponding genes in P. aeruginosa. Designation of pis genes is described in the text. Genes homologous to each other are indicated with vertical lines. The direction of arrows indicates relative transcriptional orientation.

Fig. 2 Phototactic movements of colonies of the two wild-type clones (P: PCC-P, N: PCC-N) grown for 3 d under lateral illumination (arrowhead). Dotted line shows the initial position before the illumination. Bar indicates 5 mm.

Fig. 3 Phototactic movements of colonies of the gene-disrupted mutants and their parent strain PCC-P, which were grown for 4 d under lateral illumination (arrowhead). Note that two independent mutant clones for each gene are presented. P, wild type of positive phototaxis (PCC-P); 1 and 2, Msll0038; 3 and 4, Msll0039; 5 and 6, Msll0040; 7 and 8, Msll0041; 9 and 10, Msll0042; 11 and 12, Msll0043. Dotted line shows the initial position before the illumination. Bar indicates 10 mm.

Fig. 4 Sequence alignment of Sll0041 (PisJ1) and Sll0042 (PisJ2) with phytochrome-like RcaE protein from Fremyella diplosiphon and methyl-accepting chemotaxis proteins, Tsr and Tar from Escherichia coli. Underlined residues represent the predicted transmembrane region based on the hydropathy profiles. Note that only a part of the sequences for the chromophore binding in RcaE and for the signaling ofTsr and Tar are shown, while full-length sequences are presented for Sll0041 and Sll0042. Amino acid residues identical to RcaE and to Tsr or Tar are represented as white characters on a black background. # indicates the cysteine that is predicted to bind the chromophore in RcaE. The highly conserved domains of Tsr and Tar are shown with thick underline. Asterisks under the Tar sequence indicate those amino acids which are methylated during adaptation to signals. Two GAF domains are located at 230–374 and 400–546 in Sll0041 but not shown in this figure.

Table 1

Primers and positions of the insertional disruption for pis genes

Disrupted gene (Mutant strain)	Forward primer	Reverse primer	Position of insertional disruption/amplified DNA (bp)
sll0038 (Msll0038)	5-ATCATATGACAGCTCCCACCCCC-3	5-TTTCCGGGGCCTCCTTGG-3	653/1,233
sll0039 (Msll0039)	5-AACTTGACCCAAGGAGGC-3	5-AAGTTAACTGGAGGCAGC-3	401/1,218
sll0040 (Msll0040)	5-AACTTGACCCAAGGAGGC-3	5-AAGTTAACTGGAGGCAGC-3	679/1,218
sll0041 (Msll0041)	5-CTGAGTTCATGGCAGAGG-3	5-CCTCTTGCAATGACTGGG-3	1,466/2,658
sll0042 (Msll0042)	5-CAGACCATGGCTCTGGTG-3	5-ACCTTTCTTCGTCATCCC-3	302/1,229
sll0043 (Msll0043)	5-CAGACCATGGCTCTGGTG-3	5-ACCTTTCTTCGTCATCCC-3	813/1,229

Note that the same sets of primers were used for disruption of sll0039/sll0040 and sll0042/sll0043. This was because the neighboring ORFs in the same DNA fragment were insertionally interrupted at different positions.