Edited by Alexander Namiot Glazer, University of California, Berkeley, CA, and approved May 23, 2011 (received for review March 16, 2011)
Positive phototaxis systems have been well studied in bacteria; however, the photoreceptor(s) and their downstream signaling components that are responsible for negative phototaxis are poorly understood. Negative phototaxis sensory systems are important for cyanobacteria, oxygenic photosynthetic organisms that must contend with reactive oxygen species generated by an abundance of pigment photosensitizers. The unicellular cyanobacterium Synechocystis sp. PCC6803 exhibits type IV pilus-dependent negative phototaxis in response to unidirectional UV-A illumination. Using a reverse genetic approach, together with biochemical, molecular genetic, and RNA expression profiling analyses, we show that the cyanobacteriochrome locus (slr1212/uirS) of Synechocystis and two adjacent response regulator loci (slr1213/uirR and the PatA-type regulator slr1214/lsiR) encode a UV-A–activated signaling system that is required for negative phototaxis. We propose that UirS, which is membrane-associated via its ETR1 domain, functions as a UV-A photosensor directing expression of lsiR via release of bound UirR, which targets the lsiR promoter. Constitutive expression of LsiR induces negative phototaxis under conditions that normally promote positive phototaxis. Also induced by other stresses, LsiR thus integrates light inputs from multiple photosensors to determine the direction of movement.
Cyanobacteria rely on light for photosynthesis, and therefore possess many photoreceptor systems for adaptation to limiting or excess light. Although this typically entails regulation of genes involved in light-regulated metabolism, critical signaling pathways regulate movement toward or away from weak or strong light, instead using light to regulate a behavioral cue. Phototactic motility of the cyanobacterium Synechocystis sp. PCC6803 (hereafter referred to as Synechocystis) is based on the synthesis and retraction of type IV pili (1). Light wavelengths ranging from the green (560 nm) to the red (720 nm) region mediate positive phototaxis of Synechocystis. Negative phototaxis is observed in response to UV-A light (360 nm; hereafter abbreviated as UV) as well as in response to blue (470 nm) and red (600–700 nm) light at high intensity (2, 3). Although light avoidance is regulated by sensory rhodopsins in Archaea and in a few proteobacterial and algal species (4), no such proteins are encoded within the Synechocystis genome. Known photosensors that play key roles in Synechocystis phototaxis include the phytochrome-related cyanobacteriochromes (CBCRs) PixJ1/TaxD1 (5, 6) and Cph2 (7), the BLUF photosensor PixD (8) and the CRY-related cyanopterin Sll1629 (9). Despite new insight into the regulatory roles of these proteins in UV-mediated phototaxis in Synechocystis (10), photoreceptors that directly sense unidirectional UV are unknown.
One potential candidate for a negative UV phototaxis sensor in Synechocystis is encoded by the slr1212 locus (hereafter designated uirS for UV intensity response Sensor based on this work). A member of the CBCR photoreceptor family (11), UirS possesses three N-terminal transmembrane (TM) helices similar to those of the ethylene-binding domain of the Arabidopsis ethylene receptor ETR1 (12), along with two PAS domains, a CBCR GAF domain, and an H-box containing histidine kinase (HiK) domain at the C-terminal end (13). UirS belongs to a CBCR subfamily with two conserved Cys residues within its phycocyanobilin (PCB)-binding GAF domain (14), with other notable examples being Synechocystis SyPixJ/TaxD1/Sll0041 (5, 15) and TePixJ/Tll0569 and Tlr0924 from Thermosynechococcus elongatus (14, 16, 17). Previous studies have shown that the region containing TM helices is responsible for ethylene binding in UirS (12), whereas its GAF domain can bind to the linear tetrapyrrole PCB to yield a violet/green light-switchable CBCR (18).
UirS is encoded in 1 of 16 two-component signaling gene clusters in Synechocystis (19). This cluster also harbors two response regulator genes, slr1213 (hereafter designated uirR for UV intensity response Regulator) and slr1214 (hereafter designated lsiR for light and stress integrating response Regulator) (Fig. 1A). An AraC family transcription regulator (20), UirR possesses an N-terminal receiver domain and a C-terminal DNA-binding domain (Fig. 1B). Because AraC subfamily members often act as activators of stress responses, such as MarA, SoxS, and Rob in Escherichia coli (21), we reasoned that UirR may function to transmit a UV light signal perceived by UirS. The present studies were undertaken to address the hypothesis that UirS functions as a negative UV phototaxis sensor. Our studies establish that UirR is indeed a transcriptional activator of the lsiR gene, which encodes one of six PatA-related regulators found in the Synechocystis genome (5, 22) previously shown to be induced by varied stresses, such as inorganic carbon and iron limitation (23, 24), exposure to H2O2 (25), and high light (26). Using a combination of biochemistry and forward/reverse genetics, we establish that the uirS, uirR, and lsiR loci encode key components of a UV photosensory signaling system that supports negative phototaxis.
uirS/uirR/lsiR locus and domain organization. (A) Arrangement of cobN (slr1211), uirS (slr1212), uirR (slr1213), lsiR (slr1214), and slr1215 on the chromosome of Synechocystis. (B) Predicted domain architecture of UirS (S), UirR (R1), and LsiR (R2) proteins. PVB is predicted to covalently bind to Cys533 (C533) and Cys561 (C561) of UirS. His658 (H658) of UirS, Asp54 (D54) of UirR, and Asp297 (D297) of LsiR are conserved residues of a putative His-Asp phosphorelay system. DB, DNA-binding domain; GAF, GAF domain; HiK, HiK domain; PAS, PAS domain; REC, receiver domain.
Expression of CBCR GAF domains in E. coli strains engineered for PCB production has been frequently used to examine the photosensory specificity of these phytochrome-related photoswitches (27⇓–29). In our hands, UirS-GAF adopted two photointerconvertible forms that maximally absorb in the UV-violet (Puv) and UV-green (Pg) spectral regions (Fig. 2A and Fig. S1A). Irradiation of Puv with UV (UV light source emission is shown in Fig. S2) yielded Pg, whose absorption maximum in the green region (534 nm) is accompanied by two smaller peaks in the UV region (325 nm and 382 nm) and a blue-absorbing (414 nm) shoulder (Fig. 2A). Irradiation of Pg with green light restored the Puv spectrum with a broad absorption maximum at 382 nm, results similar to those previously reported for recombinant UirS/Slr1212-GAF (18). Absorption and difference spectra of UirS-GAF expressed in Synechocystis cells were nearly identical to those of the E. coli-expressed version (Fig. 2B). We conclude that these spectral properties accurately reflect the Pg photosensory specificity of UirS.
Spectrophotometric properties of recombinant Synechocystis UirS-GAF holoprotein absorption (A and C) and difference (B and D) spectra of purified recombinant UirS-GAF isolated from PCB-producing E. coli (A–D) and from Synechocystis cells (B). (A) UV-absorbing form (Puv, blue trace) and green-absorbing form (Pg, green trace) formed after saturating green and UV illumination, respectively. Difference spectra of native UirS-GAF isolated from PCB-producing E. coli (green trace) and from Synechocystis (blue trace) cells for green minus UV light illumination (Puv-Pg) (B) and dark minus green illumination of denatured Puv (black trace) or green minus dark of denatured Pg (green trace) (D) are shown. (C) Recombinant UirS-GAF from PCB-producing E. coli was denatured by acidic urea (8 M, pH 2.0) in the dark (black) and then illuminated with green light (green).
Mutagenesis studies have established two conserved Cys residues to be required for the blue/green photocycles of related CBCRs (14, 18). To test the importance of both Cys residues for the light-sensing activity of UirS, we expressed single (Cys561Ala, C561A; Cys533Gly, C533G) and double (C533G/C561A, CGA) mutants of UirS-GAF in PCB-producing E. coli cells. Mutagenesis of Cys561, conserved in plant and cyanobacterial phytochromes and in all CBCRs, resulted in the loss of PCB binding, as shown by a lack of near-UV or visible absorbance (Fig. S1 D and E) and an absence of Zn2+-dependent fluorescence in SDS gels, a hallmark of covalently linked biliproteins (Fig. S1B). By comparison, substitution of the blue/green CBCR subfamily-specific Cys533 ablated normal photochemistry and reduced but did not abolish chromophore binding: the C533G mutant retained visible light absorption and measurable, albeit reduced, Zn2+-dependent fluorescence in SDS gels (Fig. S1 B and C). These studies corroborate studies on blue/green CBCRs that implicate thioether linkages between both Cys residues and the bilin chromophore.
Denaturation studies were performed to assess the structure of the bilin chromophore bound to recombinant UirS-GAF. When denatured with acidic urea [8 M urea/HCl (pH 2.0)] in the dark, Puv was transformed to a green-absorbing (Pg) species with a double peak at 564 and 603 nm (Fig. S1F, black trace). This dramatic spectral shift was similar to that seen on acid denaturation of the blue-absorbing (Pb) state of blue/green CBCRs (16, 17, 30). Denaturation of the Pg state of UirS-GAF in the dark yielded a distinct green-absorbing species with a single major peak at 550 nm (Fig. 2C, black trace). Irradiation of both denatured samples with green light revealed that the denatured Pg state was photoactive, whereas the denatured Puv was not (Fig. 2C and Fig. S1F, green traces). The green-minus-dark difference spectrum of the denatured Pg form (Fig. 2D) of recombinant UirS-GAF was indistinguishable from that of acid-denatured TePixJ-GAF (30), showing a pair of peaks corresponding to a majority population of phycoviolobilin (PVB) and a minority population of PCB. Taken together, these data indicate that UirS possesses a dual thioether linkage and a mixed population of 15Z PCB and PVB chromophores, whose photoisomerization to the respective 15E states initiates cleavage of the second Cys thioether linkage.
To examine the function of the Synechocystis uirS, uirR, and lsiR loci, uirS−, uirR−, and lsiR− lines were constructed. Mutant lines defective in the positive phototaxis sensor (pixJ1) and the adenylate/guanylate cyclase (cya1/2) genes were used as control lines; pixJ1− cells exhibit negative phototaxis (5, 6), whereas cya1− (and cya1−2−) cells are nonmotile under unidirectional blue and red light (31). The use of the cya1−2− mutant as a control allows us to distinguish between directional cell movement (taxis) and preferential growth, because our assay does not assess individual cell movement per se. We also examined cph2− mutants, previously reported to exhibit positive (10) and negative (32) phototaxis under UV light. Phototactic motility was compared on 1% soft agar with unidirectional illumination of different spectral quality (light spectra are shown in Fig. S2). As has been observed under considerably higher fluence rates of visible light (3), WT Synechocystis cells exhibited positive (<5 μmol·m−2·s−1), nonmotile (25 μmol·m−2·s−1), and negative (≥50 μmol·m−2·s−1) phototaxis as UV fluence rates are increased (Fig. 3A). This fluence rate dependency, as well as strain differences, likely accounts for distinct UV phototaxis behaviors reported for Synechocystis WT between laboratories (3, 7, 9, 10, 32).
Phototaxis and uirS/uirR/lsiR gene expression of WT and phototaxis mutant strains of Synechocystis. Phototactic motility of Synechocystis WT and uirS−, uirR−, lsiR−, cya1−2−, pixJ1−, and cph2− mutant exposed to unidirectional UV illumination (A, <5, 25, 50, 100, and 200 μmol·m−2·s−1; B, 85 μmol·m−2·s−1) or 16 μmol·m−2·s−1 of blue, green, or red light for 2 d. B, blue, G, green, R, red. (C) Transcript levels of uirS (black), uirR (red), and lsiR (blue) from Synechocystis WT and mutants. Total RNAs were isolated from cells incubated for 2 d under unidirectional UV (85 μmol·m−2·s−1) or 16 μmol·m−2·s−1 of blue, green, or red light. (A and B) Arrows indicate the direction of light, and the dotted lines represent the initial position of cells before illumination. (C) Data are mean ± SE (n = 3–5). Asterisks denote differences between WT and mutants or between UV and visible light (blue, green, and red), with a statistical significance set at P < 0.05 (t test). ΔΔCT, threshold cycle value.
By contrast with WT, uirS−, uirR−, and lsiR− mutants exhibited positive UV phototaxis at all fluence rates of UV except under the highest light intensity tested (200 μmol·m−2·s−1), where negative phototaxis was observed (Fig. 3A). Consistent with previous reports, the cya1−2− mutant proved nonmotile under all fluence rates of UV, whereas pixJ1− mutants showed nonmotile (≤50 μmol·m−2·s−1) and negative (≥100 μmol·m−2·s−1) phototaxis. Interestingly, cph2− mutants behaved similar to the uirS−, uirR−, and lsiR− mutants; however, in this case, reversal from positive to negative phototaxis occurred at lower fluence rates (∼50–100 μmol·m−2·s−1 vs. 100–200 μmol·m−2·s−1 for uirS−). These experiments establish that although UirS, UirR, LsiR, and Cph2 all influence negative UV phototaxis, another UV sensing system(s) is responsible for the very high-fluence UV response.
Phototactic behaviors of WT and the three mutants were next compared under unidirectional visible light. In agreement with previous studies (2, 3), WT moved toward blue (464 nm), green (527 nm), and red (650 nm) sources but away from UV light (355 nm) at the fluence rates specified (Fig. 3B). In contrast, the uirS−, uirR−, and lsiR− mutants all displayed wavelength-independent positive phototaxis. Control pixJ1− cells exhibited negative phototaxis under all light conditions, whereas cya1−2− cells proved nonmotile as expected (5, 10, 31). These studies indicate that only UV-specific phototactic behavior is altered by the loss of the uirSR two-component module and that this module and lsiR are both required for the negative UV phototaxis response.
To identify potential targets of the negative UV phototaxis signaling pathway, transcript profiles of uirS−, uirR−, and lsiR− mutants exposed to unidirectional UV were compared under fluence rates that trigger negative phototaxis in WT. High-density Synechocystis DNA microarray analyses detected only two significantly misregulated genes common to both uirS− and uirR− mutant lines: the two adjacent genes lsiR/slr1214, whose transcript level was <10% that of the WT, and the conserved hypothetical protein slr1215 (Table S1). Because the response regulator lsiR might function as the bona fide target of the UirSR two-component system, we chose the former for further study. The transcription profiling results were confirmed by quantitative RT-PCR analysis, which revealed that lsiR transcript levels were 14- to 25-fold lower than WT for all three mutants (Fig. 3C). lsiR transcription was also specific for UV because its induction was not seen in blue-, green-, or red-grown WT (Fig. 3C).
No other gene implicated in phototactic motility was jointly misregulated in the three mutants (Dataset S1), including genes for the photoreceptors pixJ1, cph2, pixD, and sll1629; the adenylate cyclase cya1; pili biosynthesis/assembly pilA1-A4; the pili motor pilT1; and the twitching motility protein pilT2. Neither was the UV-dependent expression of lsiR altered in cya1/2−, pixJ−, or cph2− mutant backgrounds (Fig. 3C). We also observed that uirS and uirR transcript abundance was unaffected by light quality or by the loss of functional cya1/2, pixJ, or cph2 genes (Fig. 3C). Taken together, these results implicate LsiR as the only transcriptional target of the UirSR two-component signaling system.
Complementation and overexpression studies were next undertaken. Expression of uirS or uirR in the respective mutants fully complemented negative UV phototaxis (Fig. 4A, UirS and UirR), while also restoring lsiR transcript levels to those seen in UV-treated WT (Fig. 4B). Similarly, WT lsiR (Fig. 4A, LsiR) fully complemented the lsiR− mutant. To test whether the conserved Asp residues in the receiver domains of UirR and LsiR were required for complementation, site-directed mutant alleles of UirR (D54G) or LsiR (D297A) were expressed in the respective mutants. Both mutant alleles failed to rescue negative UV phototaxis (Fig. 4A), revealing that the Asp residues of both response regulators are required for signal transduction. Unexpectedly, WT lsiR complemented both uirS− and uirR− mutants (Fig. 4C, labeled LsiRs and LsiRr). These results indicate that negative phototaxis under unidirectional UV can be achieved by constitutive expression of LsiR even in the absence of UirS and UirR proteins. Constitutive lsiR expression was also sufficient to reverse the positive phototactic response to red light (Fig. 4D). Thus, despite their close proximity in the Synechocystis genome, the uirSR module and lsiR are functionally distinct: uirSR is a UV-specific activator of lsiR expression, whereas lsiR is instead able to leverage UirS and an unknown red light photoreceptor(s) for avoidance behavior.
UV phototactic motility and transcript levels of uirS/uirR/lsiR genes in WT, uirS locus mutants, and complemented cell lines of Synechocystis. (A) UV-dependent phototaxis of Synechocystis WT; uirS−, uirR−, and lsiR− mutants; sigE promoter-driven uirS overexpressor (UirS) in uirS− mutant backgrounds; trc promoter-driven overexpressors of uirR (UirR) or site-directed mutant D54G (D54G) in uirR− mutant backgrounds; lsiR promoter-driven complementary strains of lsiR (LsiR) or site-directed mutant D297A (D297A) in lsiR− mutant backgrounds; pILA vector controls expressed in uirS− (pIL1) or lsiR− (pIL2) backgrounds; and pSL1211 in uirR− (pSL1) backgrounds. (B) Transcript levels of uirS (black), uirR (red), and lsiR (blue) from Synechocystis WT, overexpressors, and complementary strains in respective mutant backgrounds. (C and D) UV-dependent (85 μmol·m−2·s−1) and red-dependent (16 μmol·m−2·s−1) phototaxis of WT, uirS−, and uirR− mutants and psbA2 promoter-driven lsiR overexpressors in uirS− (LsiRs) and uirR− (LsiRr) mutant backgrounds and pILA vector controls (pIL3 and pIL4). R, red. (A, C, and D) Arrows indicate the direction of light, and the dotted lines represent the initial position of cells before illumination. (B) Total RNAs were isolated from cells incubated for 2 d under unidirectional UV (85 μmol·m−2·s−1). Data are mean ± SE (n = 3–5). Asterisks denote differences between WT and overexpressors or complementary strains, with a statistical significance set at P < 0.05 (t test).
To address the hypothesis that UirS regulates lsiR expression by modulating the binding of UirR to the lsiR promoter, we screened for UV-responsive cis-acting elements in the lsiR promoter region. A nested series of DNA fragments (Fig. 5A, P1–P5) were cloned upstream of a LUX reporter cassette (33). Only the reporter strain harboring the full 534-bp intergenic region from the uirR stop codon to the translation initiation start site of lsiR (L1) showed an approximately sixfold greater bioluminescence level relative to the control strain. Moreover, the entire intergenic region was also required to rescue negative phototaxis by lsiR expression under control of its own promoter (Fig. 5B). Thus, we conclude that the −534- to −439-bp region of the lsiR promoter unique to the P1 construct is vital for UV-activated transcription.
UirR binding to the lsiR promoter region. (A) Luminescence from lsiR promoter fragment (P1, −534- to −1-bp region; P2, −438- to −1-bp region; P3, −337- to −1-bp region; P4, −228- to −1-bp region; and P5, −112- to −1-bp region) and promoterless control-driven luxAB reporter strains (L1–L5, CO). Exponential phase cells grown under white light were exposed to unidirectional UV (85 μmol·m−2·s−1) for 16 h. Numbers refer to the nucleotide positions of lsiR promoter fragments relative to the translation +1 start site. CO, control; PL, promoterless. (B) Phototactic motility of Synechocystis WT, lsiR− mutant, and complemented lsiR− mutant lines (P1–P5, PL) expressing lsiR promoter fragment-driven lsiR (P1–P5) and promotorless lsiR (PL) under unidirectional UV illumination for 2 d. (C) Electrophoretic mobility shift assay of lsiR promoter fragments (E1–E2) coincubated with purified GST-UirR protein. Thirty femtomoles of the biotin-labeled DNA segments E1 (−534- to −452-bp region, lanes 1–8) and E2 (−462- to −347-bp region, lanes 9 and 10) was incubated for 30 min at 22 °C with increasing concentrations of purified GST-UirR (0 ng for lanes 1 and 9, 20 ng for lane 2, 50 ng for lane 3, 100 ng for lane 4, and 250 ng for lanes 5–7 and lane 10). Specific (200-fold molar excess of the nonlabeled E1 segment, lane 6) and nonspecific (200-fold molar excess of lsiR ORF fragment, +361- to +457-bp region, lane 7) inhibitors were included. In control (lane 8), GST (250 ng) was included instead of GST-UirR.
We performed electrophoretic mobility shift assays using recombinant GST-UirR protein and biotin-labeled lsiR promoter fragments to test whether UirR could recognize this region. A distinct DNA–protein complex was observed with DNA fragment E1, corresponding to the −534- to −462-bp region of the lsiR promoter (Fig. 5C, lanes 1–7). No such complex was observed with fragment E2 (−462- to −347-bp region, lanes 9 and 10) or with GST alone (lane 8). The amount of E1–UirR complex increased with increasing UirR, with depletion of the band corresponding to the free probe (lanes 1–5). Binding could be specifically inhibited by excess unlabeled E1 DNA (lanes 6 and 7). These results support the presence of a UirR recognition sequence within the −534- to −462-bp region upstream of the lsiR ORF.
Should UirS and UirR function as a canonical two-component signaling system, protein-protein interactions and/or phosphotransfer would be expected to occur between the UirS HiK domain and the receiver domain of UirR. In contrast to Arabidopsis ETR1 (34), His kinase autophosphorylation activity for TM-deleted UirS (UirSΔ) was not observed, and we could not detect UirSΔ-phosphotransfer to UirR with or without UV illumination. Because HiK activity is not required for ETR1 signaling (35), we examined whether UirS could physically interact with UirR using yeast split-ubiquitin and two-hybrid assays. Using full-length UirS and truncated UirSΔ (Fig. S3A), yeast split-ubiquitin assays showed that UirS (and UirSΔ) can self-interact based on histidine auxotrophy, red/white, and β-galactosidase analyses (Fig. S3B), a result consistent with HiK dimerization (36). Yeast two-hybrid assays showed that UirSΔ can also interact with full-length UirR polypeptide via the expected interaction between the UirS HiK domain and the UirR receiver domain (Fig. S3C). Interaction between UirS and UirR was further supported by in vitro pull-down assays (Fig. S3D), which showed copurification of GST-UirR and His-tagged UirSΔ.
Our studies establish that the membrane-bound CBCR HiK UirS and its cognate AraC-family response regulator UirR function as a UV-sensing two-component signaling system in Synechocystis. Interestingly, AraC subfamily members in E. coli often act as activators of stress responses (21). UirR binds to the promoter of the adjacent gene slr1214 that encodes the PatA-family response regulator LsiR. Among other PatA-related proteins encoded by the Synechocystis genome are TaxP1 (PixG/Sll0038) and PixE (Slr1693), both of which are located within phototaxis operons (5, 6), and thus may share the same regulatory target(s) as LsiR. Similar to the CBCRs SyPixJ1, TePixJ, and Tlr0924 (15⇓⇓–18), UirS possesses a dual-Cys–linked PVB chromophore whose photoactivation triggers conversion between UV (Puv) and green (Pg) light-absorbing states. Physiological analyses indicate that LsiR functions as a signal output regulator of avoidance movement of Synechocystis cells under unidirectional UV illumination.
Our studies also show that constitutive lsiR expression is sufficient for phototaxis reversal under unidirectional light. Because LsiR is induced by multiple cellular stresses (23⇓⇓–26), this response regulator would appear to integrate unidirectional light stress with general stress signaling systems. Based on our interaction data, we propose a model in which UV activation of UirS induces lsiR expression via Pg-dependent release of UirS-bound UirR (Fig. 6). This model is based on the observation that UirS and UirR are constitutively expressed, regardless of light quality, and the knowledge that UirS is an integral plasma membrane protein (37). UirR thus functions as a mobile signal carrier that ultimately must locate the lsiR promoter (plsiR). Although we have been unable to detect phosphotransfer from UirS to UirR, we show that the conserved Asp residue of UirR is required for function. It is thus possible that phosphorylation is required for UirR to function as an activator of lsiR transcription, possibly by promoting UirR homodimerization (20). The identity of the phosphorylating agent(s) responsible, if any, remains unknown at present. The regulatory activity of LsiR also requires the equivalent Asp; however, UirS is clearly not required for LsiR function.
Proposed model for the UirS/UirR/LsiR-based negative UV phototaxis signaling pathway in Synechocystis.
How might LsiR reverse the direction of Synechocystis cell movement? Phototactic motility of Synechocystis uses a type IV pilus motor to drive pilus extension, tethering to the surface or to neighboring cells and then retracting to move toward the direction of extension (1). To achieve negative phototaxis, differential LsiR activation on the irradiated “proximal” and shaded “distal” sides of the cell could be used either to activate the distal motor or to inactivate the proximal motor. This could be accomplished by differential activation of a kinase or inactivation of a phosphatase (or vice versa), which alters LsiR stability and/or localization. Alternatively, a second-messenger gradient could be used to alter the activity of LsiR differentially. Both hypotheses require that the light gradient sensed by a photoreceptor trigger a differential response on the proximal and distal sides of the cell. In a second-messenger model, it is interesting to note that cph2− mutants exhibit a phenotype very similar to loss-of-function mutants in the UirS signaling system shown here (Fig. 3A). Cph2 possesses GGDEF and EAL regulatory output domains, which could generate a UV-dependent gradient of di-cGMP to affect the activities of LsiR and/or the pilus motor differentially. However, expression of LsiR results in negative phototaxis under red light, indicating that multiple photoreceptor pathways support negative phototaxis when the LsiR protein is present.
It is well established that the blue/green CBCR SyPixJ/TaxD1/Sll0041, which possesses a methyl-accepting chemotaxis signaling domain, functions to promote positive phototaxis under visible light (1). For this reason, SyPixJ is probably antagonized by LsiR action, particularly under unidirectional blue or green light. Because SypixJ− mutants exhibit negative phototaxis under all wavelengths of visible light, Synechocystis cells must contain other visible light sensors responsible for perceiving visible light gradients, and these may also involve a role for LsiR. Like those present in other cyanobacteria, multiple photosensors that could serve this function are representatives of the cyanobacterial phytochrome, CBCR, BLUF, and CRY families. Of particular note, the blue-absorbing CRY-related cyanopterin Sll1629 (9), the BLUF protein SyPixD/Slr1694 (8), the red/green CBCR Slr1393 (38), and the red/far-red light sensor Cph1 (39) all could contribute to sustaining the signaling gradient necessary for LsiR-dependent (and potentially LsiR-independent) negative phototaxis at elevated fluence rates of visible light. This work also illuminates the exquisite complexity with which cyanobacteria tune not only their metabolism but their behavior in response to light and other environmental cues.
The motile Synechocystis sp. PCC 6803 WT was used to generate mutants and transformants that contained genes conferring spectinomycin, chloramphenicol, kanamycin, and gentamicin resistance. All strains were grown on 1% BG-11 agar plates buffered with 10 mM N-Tris (hydroxymethyl) methyl-2-aminoethanesulfonic acid (TES)-KOH (pH 8.0) at 28 °C under the white light (10 μmol·m−2·s−1) with or without 10–20 μg·mL−1 appropriate antibiotic(s).
Exponentially growing Synechocystis cells were streaked onto 1% BG-11 agar trays containing 10 mM glucose and then incubated under unidirectional illuminations of UV, blue, green, and red light sources (Fig. S2) for 2 d.
Gene expression profiling was conducted with the Synechocystis high-density 134K NimbleGen microarray designed to accommodate 3,673 transcripts (Roche NimbleGen Inc., Madison, WI).
Protein interactions among UirS and UirR were examined by the yeast mating-based split-ubiquitin system and the yeast Matchmaker GAL4 Two-Hybrid system (Clontech).
His6-UirSΔ and GST-tagged UirR proteins were purified by nickel affinity chromatography (Qiagen) and glutathione Sepharose 4B (GE). Purified His6-UirSΔ protein was immobilized by nickel affinity chromatography on nickel-nitrilotriacetic acid (Ni-NTA) metal-affinity resin (Qiagen) and then incubated with purified GST or GST-UirR. GST fusion proteins bound to His6-UirSΔ were detected by Western blotting on a PVDF membrane (Millipore) using a rabbit anti-GST antibody (Santa Cruz).
Biotin-labeled DNA fragments containing the −534- to −452-bp (E1) and −462- to −367-bp (E2) regions in the lsiR promoter region, specific (nonlabeled E1 fragment) and nonspecific (+361- to +457-bp region within lsiR ORF) competitors of E1, purified GST-UirR, and GST proteins were used in gel mobility shift assays (Chemiluminescent Light Shift EMSA kit; Pierce Biotechnology).
We thank Y. H. Chung, H. Pakrashi, M. Hageman, P. Wolk, and W. B. Frommer for their gifts of pixJ1− and cph2− mutants, pSL1211, pILA vectors, and E. coli cells for the conjugation system and mating-based split-ubiquitin system, respectively. We also thank N. C. Rockwell, G. Choi, and H. G. Nam for helpful discussions and review of the manuscript. This work was supported by Grant 2010-0029728 from the Advanced Biomass Research and Development Center (to Y.-I.P.), Grant PJ8205 from the Agricultural Genomics Center of Korea (to Y-I.P.), and Grants GM068552 and DOE DE-FG02-09ER16117 from the National Institutes of Health and the Department of Energy Office of Basic Energy Biosciences (to J.C.L.).
Author contributions: J.-Y.S., J.C.L., and Y.-I.P. designed research; J.-Y.S., H.S.C., J.-I.C., and J.-S.J. performed research; J.-Y.S., H.S.C., J.-I.C., J.-S.J., J.C.L., and Y.-I.P. analyzed data; and J.-Y.S., J.C.L., and Y.-I.P. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1104242108/-/DCSupplemental.
