A novel phototaxis receptor hidden in the cyanobacterial genome.

Many microorganisms, plants and animals control their movements in response to light. The simplest motility response to light, bacterial phototaxis, was described more than a century ago (Engelmann, 1883). Although bacterial taxis is now the best-understood simple sensory system, no phototaxis receptors in bacteria have been described yet. More complex organisms are better studied in this respect: phototaxis in unicellular algae (Foster, 1984; Ebnet et al., 1999) and fungi (Saranak and Foster, 1997), as well as in Archaea (Spudich, 1998) was shown to be dependent on rhodopsin-like receptors. Complete genome sequencing of the photosynthetic cyanobacterium Synechocystis sp. PCC6803 (Kaneko et al., 1996) should have resulted in the identification of a phototaxis receptor. However, the bioinformatics methods used for the genome annotation in 1996 were not sensitive enough. An increased power of new algorithms for genomics, especially the introduction of the PSI-BLAST program (Altschul et al., 1997), provided new opportunities for discoveries in completely sequenced genomes. Therefore, I have revisited the Synechocystis genome in order to identify the first bacterial phototaxis receptor. BLAST searches against the Synechocystis genome database (available at www.kazusa.or.jp/cyano/cyano.html) by using a highly conserved signaling domain of bacterial chemotaxis receptors (Le Moual and Koshland, 1996) as queries have identified three genes coding for taxis receptors: sll0041, sll1294 and slr1044. Further in silico domain analysis including PSI-BLAST searches (Altschul et al., 1997) and multiple alignments revealed that the gene sll0041, which has been annotated as coding for a methylaccepted chemotaxis protein I or a serine chemotaxis receptor (Kaneko et al., 1996), coded in fact for a phototaxis receptor. In its C-terminal region, the sll0041 protein contains a conserved signaling module typical of bacterial taxis receptors (Figure 1). This sequence region is homologous to the signaling module from the serine chemoreceptor and other methyl-accepting chemotaxis proteins from Escherichia coli that caused the erroneous gene annotation in 1996. I have identified two GAF domains and two HAMP domains in the N-terminal region of the sll0041 protein (Figure 1). The GAF domains are sensor modules in various phototransducing proteins in both eukaryotes and prokaryotes (Aravind and Ponting, 1997). The best known examples of GAF-domain-containing photoreceptors are plant phytochromes (Reed, 1999) and ethylene response sensors (Chang and Shockey, 1999) and mammalian rod photoreceptor phosphodiesterases (Soderling and Beavo, 2000). HAMP domain is found in various signal transduction proteins. Aravind and Ponting have recently suggested that the HAMP domain is not just a linker between a transmembrane region and a signaling domain, as it was previously thought, but it is an important signaling domain itself (Aravind and Ponting, 1999). The presence of the second HAMP domain between the GAF2 domain and the signaling module in the cyanobacterial photoreceptor (Figure 1) is in a good agreement with their idea. All sensory and signaling domains of the sll0041 protein are predicted to be located in the cytoplasm (Figure 1).