Biochemical Properties of CikA, an Unusual Phytochrome-like Histidine Protein Kinase That Resets the Circadian Clock in Synechococcus elongatus PCC 7942*

We recently described the cikA (circadian input kinase A) gene, whose product supplies environmental information to the circadian oscillator in the cyanobacterium Synechococcus elongatus PCC 7942. CikA pos-sesses three distinct domains: a GAF, a histidine protein kinase (HPK), and a receiver domain similar to those of the response regulator family. To determine how CikA functions in providing circadian input, we constructed modified alleles to tag and truncate the protein, allow-ing analysis of each domain individually. CikA covalently bound bilin chromophores in vitro , even though it lacks the expected ligand residues, and the GAF domain influenced but did not entirely account for this function. Full-length CikA and truncated variants that carry the HPK domain showed autophosphorylation activity. Deletion of the GAF domain or the N-terminal region adjacent to GAF dramatically reduced autophosphorylation, whereas elimination of the receiver domain increased activity 10-fold. Assays to test phosphorelay from the HPK to the cryptic receiver domain, which lacks the conserved aspartyl residue that serves as a phosphoryl acceptor in response regulators, were negative. We propose that the cryptic receiver is a regulatory domain that interacts with an unknown

The wide variety of organisms in which biological clocks have been described share the important feature that their endogenous clocks can be set to "local time" by external environmental signals such as light and temperature (1)(2)(3)(4). Whereas light perception to reset the circadian clock in mammals, mice, and Drosophila is likely mediated via opsins and cryptochromes (5)(6)(7), plants such as Arabidopsis thaliana receive clock-setting information through cryptochromes and phytochromes that work synergistically (8). Cryptochromes are blue light photoreceptors that most likely evolved from DNA photolyases. They bind two chromophores, flavin-adenine dinucleotide, and a pterin (methenyltetrahydrofolate) or a deazaflavin (7,8-didemethyl-8-hydroxy-5-deazariboflavine). The latter functions in light perception by transducing excitation energy to flavinadenine dinucleotide, which then catalyzes an electron transfer reaction in the DNA photolyase (5,9,10). Plant phytochromes are red/far-red photoreceptors that bear linear tetrapyrrole (bilin) chromophores (11,12) attached to an N-terminal sensory module known as a GAF domain (13,14). Plant phytochromes also have sequence similarity to histidine-protein kinases (HPK) 1 but instead possess serine/threonine (S/T) kinase activity (15,16). Members of the phytochrome family have also been identified in prokaryotes, including cyanobacteria (17)(18)(19), purple photosynthetic bacteria (20,21), and nonphotosynthetic eubacteria (22). Many of these proteins exhibit a similar domain structure to plant phytochromes with an N-terminal bilin-binding domain that is associated with an HPK module (23). The HPK domain of cyanobacterial Cph1 has been shown to have authentic HPK activity (19).
We recently described the cikA gene of the cyanobacterium Synechococcus elongatus PCC 7942, which encodes a phytochrome-related protein with similarity to HPKs of bacterial two component signal transduction systems (24 -26) and is involved in signal perception for resetting the circadian clock in response to environmental cues (27). CikA was first identified from a Tn5 mutant that has subtle alterations in light-responsive regulation of the photosystem II gene psbAII (28). The mutant has an expression level of a psbAII::luxAB reporter fusion strain that is 50 -80% of that in wild type (WT) under low light conditions and exhibits an exaggerated increase in expression upon exposure to higher light intensities. However, a more striking circadian phenotype was observed, because both the period length and amplitude of circadian rhythms of different luxAB reporter strains (kaiB::luxAB, purF::luxAB, and kaiA::luxAB) were reduced (27,29,30). In addition, the relative phasing of expression (timing of circadian peaks) was altered for some genes. The additive effect on period length reduction of cikA inactivation in both short and long period kai mutants of S. elongatus suggested that CikA and the Kai proteins have independent, nonoverlapping functions. A lack of phase resetting in the CikA-free mutant in response to a 5-h dark pulse that resets WT revealed that CikA functions as a sensor component on the input pathway of the circadian clock.
The deduced amino acid sequence of cikA reveals it to be a member of the extended phytochrome family (27). CikA exhibits a typical bacteriophytochrome domain structure with an N-terminal GAF domain (13) and a C-terminal histidine kinase motif possessing H-, N-, D/F-, and G-boxes, as defined previously by Stock et al. (31). By contrast, CikA lacks the PAS domains present in plant phytochromes (27,32). Relative to the plant phytochromes PhyA-E and the cyanobacterial bacteriophytochromes Cph1 of Synechocystis sp. PCC 6803 (18,33) and RcaE of Fremyella diplosiphon (17), the GAF domain of CikA has a deletion near the usually conserved cysteine ligand for bilin binding; this makes it difficult to evaluate which residues, if any, play a role in bilin association with the CikA apoprotein. Indeed, CikA appears to lack both the conserved Cys residue that serves as the site for bilin attachment in plant phytochrome photoreceptors and the adjacent His residue to which phycocyanobilin (PCB) or biliverdin (BV) binds for some BphPs from nonphotosynthetic bacteria (22,34).
The C-terminal region of CikA has similarities to receiver domains of two component response regulators (RR) such as PhoB. However, the mode of action of this RR domain may be different, because the conserved Asp residue necessary for phosphoryl transfer is missing (24 -27). In this regard, the CikA structure is similar to the family of so-called pseudoresponse regulators (APRR) that has been identified in A. thaliana (35)(36)(37)(38). APRRs also possess N-terminal CONSTANS "Myb-related" motifs, which regulate transcriptional activity in eukaryotes, in addition to their pseudo-receiver domains. In contrast to the authentic response regulator proteins of A. thaliana, APRRs lack two of three invariant residues of receiver domains, which apparently prevents them from accepting phosphoryl groups from A. thaliana histidine kinases in vitro (35). APRR mutants also exhibit significant circadian phenotypes, such as delayed flowering and shortened periods of several rhythmic markers, including expression of the CAB (chlorophyll a/b-binding protein) gene (37,38).
To determine whether CikA is an input photoreceptor to the cyanobacterial circadian clock and to clarify its mode of action, we made a variety of plasmid constructs for expression of tagged full-length and truncated CikA variants in Escherichia coli and in S. elongatus. After showing functionality of the His-tagged CikA variant in complementation experiments, we investigated the ability of affinity-purified CikA to covalently attach a bilin chromophore in vitro. We also tested the influence of other predicted domains on the autophosphorylation activity of the HPK. The results suggest regulatory roles for the N-terminal, GAF, and pseudo-receiver domains.
Construction of Mutant Strains and DNA Manipulations-Plasmid clone analysis, cleavage with restriction endonucleases, agarose gel electrophoresis, ligations, and transformation of E. coli strains were performed according to standard procedures (41). Plasmids, strains, and oligonucleotide primers for PCR amplification with Pwo polymerase (Roche Applied Science) are listed in Table I.
Expression and Purification of His 6 -and Trx-tagged CikA and Truncated Versions of CikA in E. coli-Expression of recombinant CikA 6HIS was achieved using the Qiaexpress overexpression system (Qiagen) for synthesis of N-terminal His 6 -tagged proteins. A PCR-derived fragment using primers AMO558 and F0748, encoding the entire cikA gene except the first methionine, was cloned into SacI/SalI-digested pQE30/ 80L expression vector and used to transform E. coli SG13009. The cells were grown to OD 600 nm ϭ 0.5, induced with 0.9 mM IPTG, and grown for an additional 3-5 h at 25°C to minimize the formation of insoluble protein. The cells were harvested and passed twice through a prechilled French press at 137.9 MPa. This extract was clarified by centrifugation for 30 min at 12,000 ϫ g to prepare a lysate for affinity purification on nickel-nitrilotriacetic acid (Ni-NTA) matrix (Qiagen). For expression of truncated versions of CikA, we used the pET overexpression system (Novagen), generating fusion proteins that carry an N-terminal thioredoxin domain (Trx) for enhanced solubility as well as N-and/or Cterminal His 6 tags for purification by a Ni-NTA matrix (Table II).
Expression and Purification of Cik 6HIS in S. elongatus-5 ml of stationary phase AMC1006 was added to 500 ml of fresh BG-11M medium, incubated in an aquarium at 30°C and a light intensity of ϳ200 microeinsteins m Ϫ2 s Ϫ1 , and bubbled with 1% CO 2 in air. When the cells were grown to an OD 750 nm of 1.5, production of CikA 6HIS was induced with 2 mM IPTG. After 48 h, the cells were harvested, washed, and resuspended in 20 ml of Buffer 1 (25 mM Tris-HCl, pH 7.5, 500 mM NaCl). After passage through a prechilled French press at 137.9 MPa, a cleared lysate was prepared by centrifugation at 100,000 ϫ g for 60 min. CikA 6HIS in the soluble protein fraction was diluted to 200 ml in Buffer 1. Protein was bound to Ni-NTA-agarose (Qiagen), washed, and eluted with Buffer 1 that contained 50 and 250 mM imidazole, respectively.
Assay of Bioluminescence in 96-well Microtiter Plates-Inocula of AMC913, AMC914, AMC1006, and the corresponding control strains were streaked from solid medium onto 300-l BG-11 M agar pads in 96-well plates. After applying a 12-h dark pulse to the cultures to synchronize the circadian clock, the psbAI::luxAB and kaiB::luxAB reporter strains were incubated in continuous light, and bioluminescence was measured using a Packard TopCount luminometer (Packard Instrument Company, Meriden, CT) (29). Period length and amplitude were calculated using I&A and FFT-NLLS software (www.scripps.edu/ cb/kay/shareware/) (30).
[␥-32 P]ATP Autophosphorylation Assays-CikA autophosphorylation assays were done as previously described (42,43), except that gels were dried for exposure to phosphorimaging plates in most cases. Briefly, the standard phosphorylation buffer stock to which purified protein was added was 50 mM Tris, pH 7.5, 100 mM KCl, 2 mM dithiothreitol, 0.1 mM ATP (unlabeled), and 0.15 M [␥-32 P]ATP. Chemical stability measurements were performed as described by McCleary and Zusman (42) with minor modifications. The assays to test for transphosphorylation from 32 P-labeled CikA 6HIS or Trx-⌬RR (Table II) to the receiver domain of CikA were done according to previously published protocols (44). The activities were quantitated using a Fujix BAS 2000 phosphorimaging system.
In Vitro Bilin Lyase Assays with Purified CikA-The assays were performed according to a previously published protocol with purified phycocyanobilin, phytochromobilin, and phycoerythrobilin (12). Biliverdin was obtained as described previously (45). After the reaction, the mixtures were concentrated by acetone precipitation and applied to 0.1% SDS-8% PAGE. The proteins were electrophoretically transferred to polyvinylidene difluoride (PVDF) membranes (Immun-Blot TM ; Bio-Rad) for 3 h at 45 V. After incubation of the PVDF membranes with 1.3 M zinc acetate, the bilin adducts were visualized by fluorescence with a Molecular Dynamics Storm 860 instrument with a setting of red fluorescence and Photomultiplier tube voltage ϭ 1000.

Complementation of a cikA Mutant with an Affinity Tagged
Allele-We constructed a 6-histidine tag-encoding full-length allele of cikA (cikA 6HIS ) in the IPTG-inducible expression vector pQE30/80 to express, purify, and characterize recombinant CikA in both E. coli and S. elongatus (Table II). To verify that the tagged allele has equivalent functions to WT cikA, the cikA 6HIS gene was introduced into the S. elongatus CikA-free  (27). These include alteration of period length (shortened by about 2 h) and reduction in amplitude of the oscillation of expression from reporter genes, e.g. kaiB::luxAB, purF::luxAB, psbAII::luxAB, and psbAI::luxAB, as well as a change in the relative phasing of circadian peaks for some genes, such as the light-regulated psbA family. A construct that bears the cikA 6HIS allele (pAM2477: P TRC -OP-OP-cikA 6HIS ) was introduced into strain AMC568 via a neutral site I-targeting overexpression vector pAM2428 (P trc and laqI regulation system) ( Table I). We monitored bioluminescence from the complemented strain, AMC913 (pAM2477), along with the parent strain AMC568 and the WT psbAI::luxAB reporter strain AMC669, to assess the circadian phenotypes. The cikA 6HIS allele complemented the cikA phenotype with respect to period length and amplitude (compare traces from of AMC568 and AMC913 without IPTG; Fig. 1 Induction of CikA 6HIS expression in AMC913 by adding 2 mM IPTG led to an arrhythmic phenotype of psbAI::luxAB reporter activity (Fig. 1). The effect of CikA 6HIS overexpression on the more robust, high amplitude rhythm of kaiB::luxAB in a CikAfree mutant background (AMC1006) also led to arrhythmic bioluminescence (data not shown).  Chromophore Ligation Assays with Various Substrates-We assayed in vitro ligation of bilin chromophores to full-length CikA 6HIS and a number of variants to determine whether the GAF-like domain of CikA can autocatalytically attach a bilin chromophore, indicating lyase activity. Table II and Fig. 2 describe the proteins used in this study. The proteins were overexpressed in E. coli and affinity-purified by His 6 tags. Truncated versions of CikA required, additionally, a Trx tag at the N terminus to improve solubility, even though the fulllength protein is soluble when expressed with only a His 6 tag. Zinc blot analysis, which detects tetrapyrroles, gave clear fluorescent signals for CikA 6HIS with phytochromobilin (P⌽B) and PCB chromophores (Fig. 3). The fluorescence with these substrates was well above the background obtained when the protein was incubated with Me 2 SO alone as a control, as well as above that detected with either BV or phycoerythrobilin, to which CikA 6HIS does not appear to attach covalently (Fig. 3).
Assaying GAF as a Chromophore Attachment Domain-CikA 6HIS , Thioredoxin-tagged CikA (Trx-CikA) and its variants were assayed for chromophore ligation with purified PCB in vitro. In this experiment, we decreased the PCB concentration to 2.0 M to reduce nonspecific binding. As shown in Fig. 4 (A  and B), recombinant proteins that included the GAF domain (CikA 6HIS , Trx-CikA, ⌬RR, ⌬ATP, ⌬HPK, NϩGAF, and ⌬N) emitted obvious zinc-induced fluorescence, indicating the presence of protein-chromophore interaction. However, the ⌬GAF variant had a reduced signal, only slightly higher than background negative controls (Fig. 4B). We concluded that GAF, but no other single motif, is important for bilin attachment. However, a Trx-GAF variant, composed of only GAF and the Trx tag, produced no signal above that obtained with a Trx-RR, in which the only the receiver-like domain of CikA is fused to Trx (Fig. 4C). These results suggest that the GAF domain we defined is not sufficient for bilin lyase activity or that the domain does not fold stably when removed from its natural protein context.
Although lyase assays showed that CikA can bind a bilin in vitro, CikA 6HIS isolated from S. elongatus gave no fluorescent signal on a zinc blot (Fig. 4D) and showed no spectral evidence of a chromophore (data not shown). This negative result was obtained whether the protein was isolated from cells under normal illumination or in darkness. We reasoned that CikA might associate noncovalently with a bilin in vivo and that the co-factor might be lost during purification. However, co-expression of the protein with either PCB or BV in E. coli, under conditions in which other bacteriophytochromes form photoac-tive adducts (46,47), showed no evidence of association between CikA and the chromophores (Fig. 5). We concluded from these various lines of data that CikA is unlikely to form a biliprotein complex naturally in S. elongatus.
Autophosphorylation Assays with Overexpressed and Partially Purified CikA 6HIS -We performed in vitro autophosphorylation assays with affinity-purified CikA 6HIS apoprotein under various conditions to test whether the protein has authentic HPK activity as predicted from its sequence. CikA 6HIS showed strong autophosphorylation, with a temperature optimum around 25°C and detectable activity at 0°C (Fig. 6A). Maximum phosphorylation was detected after 10 min of incubation at 25°C under these conditions (Fig. 6B). Autophosphorylation activity showed a strong dependence on substrate concentration (Fig. 6C), and cold ATP competed with [␥-32 P]ATP as expected for an enzymatic activity (Fig. 6D). The phosphoryl linkage remained stable under basic conditions (100% signal remaining; Fig. 6E, lane 1) and decreased after treatment with hydroxylamine (14% signal remaining; Fig. 6E, lane 2) and under acidic conditions (11% signal remaining; Fig.  6E, lane 3). This result is consistent with a phosphoryl linkage to the His 393 determined as an H-box by sequence alignment (27). The holoprotein generated in vitro by addition of 2 or 4 M PCB showed the same autophosphorylation activity as the apoprotein purified from E. coli, and CikA 6HIS purified directly from the cyanobacterium showed autokinase activity comparable with that observed for E. coli-derived protein (data not shown).
Influence of other CikA domains on HPK Activity-The assay conditions described for partially purified CikA did not produce robust activity when the protein was more highly purified, either as a His 6 or Trx fusion. The addition of 2 g/l (w/v) BSA to the reaction mixture allowed a linear increase of autophosphorylation of CikA variants up to 120 min at 30°C, whereas the mixture without BSA reached a plateau at 15 min (data not shown). Therefore, the following reactions were carried out in the presence of 2 g/l BSA at 30°C for 1 h.
To identify which regions of the protein influence autophosphorylation activity, we produced variants that are missing specific residues or motifs and performed in vitro autophosphorylation assays. Because truncated CikA variants were soluble only if fused to Trx, we first compared autophosphorylation of full-length CikA 6HIS and Trx-CikA (Fig. 7). The addition of the Trx tag dropped autophosphorylation to 20 -25%; this variant served as the control for the following mutant derivatives.
All of the recombinant proteins that have a critical deficiency of the HPK domain (⌬HPK and Trx-RR) or a point mutation at the expected phosphoryl acceptor His residue in the kinase domain (Trx-H393A) were completely defective in autophosphorylation activity (Fig. 7, B and C). ⌬GAF, which conserves an intact HPK domain (Fig. 2), also had no autophosphorylation signal (Fig. 7, B and C). Additionally, only an extremely weak signal was detected after 60 min from ⌬N (Fig. 7, B and  C), which retains all recognizable motifs of the protein (Fig. 2), and was lower than the detectable sensitivity at 30 min (data not shown). In sharp contrast, removal of the receiver domain enhanced autophosphorylation activity by about 2-fold over CikA 6HIS , and more than 10-fold over the more comparable Trx-CikA full-length control. We confirmed the linearity of these single time point experiments by performing time courses for the active CikA variants (CikA 6HIS , Trx-CikA, and ⌬RR; Fig. 8A), except for ⌬N, which needs longer than 60 min to visualize activity. The assays showed a linear increase of autophosphorylation as expected for kinase activity dependent on the HPK domain, as depicted for CikA 6HIS in Fig. 8B.
Absence of Internal Phosphoryl Transfer in CikA-We tested whether CikA transfers a phosphoryl group from the HPK domain to the atypical receiver domain in transphosphorylation assays. In the first strategy, we used autophosphorylated full-length CikA 6HIS or ⌬RR as a phosphodonor and Trx-RR as a potential acceptor. No phosphorylation of Trx-RR was observed (data not shown). Because HPKs tend to dimerize, we reasoned that if phosphoryl transfer occurs in CikA, it might involve the H-box of one monomer and the receiver domain of the other. To test this possibility, we used ⌬RR as a donor that is lacking its own receiver and Trx-H393A as a potential receiver that is incapable of autophosphorylation (Fig. 9A). No phosphorylation of Trx-H393A was observed (Fig. 9B). The same negative result was obtained when the Trx-H393A sequence was modified to substitute Asp for Ala at the site that most closely aligns with Asp of bona fide receivers of twocomponent phosphorelay systems (data not shown).

DISCUSSION
The identification of cikA as a locus that affects circadian phasing and induction of light-regulated genes and resetting of circadian rhythms in response to a dark pulse implicates the CikA protein as important for transducing light signals from the environment, especially as an entry point to the circadian clock (27). Its structure is similar to bacteriophytochromes, suggesting a direct role as a photoreceptor. However, if CikA is a phytochrome-like photoreceptor, its properties are very different from previously reported bacteriophytochromes. CikA can form adducts with bilins in vitro that survive conditions used for SDS-PAGE, despite the absence of residues expected to be necessary for covalent linkage of a bilin (Fig. 3) (27). Mutation of the GAF domain at residues highly conserved among bacteriophytochromes (E236A, E285A, or C299A) did not abolish in vitro bilin binding (data not shown). Although the GAF domain was important for bilin lyase activity, it was not sufficient (Fig. 4C), and its absence did not completely abolish binding (Fig. 4B). This suggests either that the covalent  Table  I. Open boxes, identifiable motifs of the CikA protein; black bars, protein coding regions for each recombinant protein variant; star, mutagenized histidyl to alanyl substitution. ⌬RR, NϩGAF, and Trx-GAF variants have a second His 6 tag at the C terminus (Table II) attachment site(s) lies outside of the GAF proper or that GAF alone does not assume the tertiary structure needed to support the physical structure of the bilin-binding pocket. Because we observed clear covalent binding with PCB from both ⌬N and NϩGAF (Fig. 4B), which lack either the N-or C-terminal region neighboring the GAF domain, respectively, neither of those regions is specifically critical for the autocatalytic attachment reaction of linear tetrapyrrole chromophores.
No spectral activity was detected from CikA in the presence of PCB or P⌽B, and recombinant CikA isolated from S. elongatus did not co-purify with a bound bilin. Recent studies with phytochromes and bacteriophytochromes that lack the conserved cysteine site of attachment have established that noncovalent adducts are photochromic and could serve as photoreceptors (48 -50). However, co-expression of CikA with BV or PCB in E. coli did not produce a pigmented holoprotein (Fig. 5). It is possible that CikA does not serve as a photoreceptor itself but oligomerizes with a partner that binds a photoactive co-factor. One of the potential CikA partners identified in a yeast two-hybrid screen is a protein that carries a canonical GAF. 2 Whether or not CikA naturally binds a co-factor that would serve as a chromophore, the GAF domain likely serves a regulatory role, as evidenced by the negative effect of GAF removal on HPK activity. We favor interaction with a partner as the regulatory role of the CikA GAF domain.
As is implicit in the name of the gene cikA (circadian input kinase A), the deduced amino acid sequence predicts a well conserved transmitter module typical of the sensor kinases of bacterial two-component signal transduction systems and is expected to supply a phosphoryl group to a specific, yet unidentified, response regulator(s) (27,31). A histidyl residue in the region designated as the H-box is the site of autophosphorylation in HPKs. In CikA, this residue is His 393 , and a single point mutation at that site completely abolished the autophosphorylation with [␥-32 P]ATP (H393A in Fig. 7). Additionally, the 32 P-labeled protein obtained by in vitro autophosphorylation was alkali-tolerant in 3 M NaOH on PVDF membrane but was not stable to hydroxylamine or acidic conditions. These results also support the expectation that the histidyl residue in the H-box works as a typical phosphoacceptor residue (42).
In vitro autophosphorylation assays with a series of truncated CikA variants identified regions that affect autokinase activity. Purified full-length CikA tagged by His 6 or Trx and some of the Trx-tagged truncated variants that carry the HPK domain showed clear autophosphorylation activity. However, deletion of the GAF domain or the N-terminal region adjacent to the GAF domain dramatically reduced the autophosphorylation despite the existence of an intact HPK domain (⌬N and ⌬GAF; Fig. 7, B and C). These variants were soluble and expected to be folded correctly. From this, we conclude that both the GAF domain and the undefined N-terminal region are important for activating the kinase (Fig. 7D). Interaction of each domain with a protein partner or small organic molecule (bilin or other) could provide a mechanism to modulate the activating functions. Fusion of Trx to the CikA N terminus, which was used to achieve soluble protein expressed in E. coli, reduced the activity as compared with CikA 6HIS . Because Trx is a larger tag than His 6 , the physical structure of the N-terminal region of CikA may have been modified. This is consistent with the finding that a functional domain that influences kinase activity resides in this portion of the protein.
The C terminus of CikA contains a similar sequence to the receiver domains of response regulators involved in two-component signal transduction systems. Although this cryptic receiver domain conserves other key residues of the motif, the invariable aspartyl residue that is phosphorylated by a cognate HPK in bona fide receivers is absent from the sequence (27). From the negative results of various in vitro trans-phosphorelay experiments using the HPK domain and the C-terminal domain of CikA (such as in Fig. 9), we conclude that the latter is a pseudo-receiver and does not accept a phosphoryl group 2 J.-S. Choi, S. Canales, and S. S. Golden, unpublished data.
FIG. 5. Co-expression of CikA and bilins in E. coli. Strains of E. coli engineered to produce PCB or BV were transformed with a plasmid that encodes CikA. As standards for holoprotein formation, the PCB-producing cells were transformed with the gene for Cph1 of Synechocystis sp. Strain PCC 6803 (46) and the BV cells with the gene for BphP1 of Agrobacterium tumefaciens (47). and 60 min at 25°C. C, concentration dependence of autophosphorylation activity with 1 g of affinity-purified Cik 6HIS and 10 min of incubation at 25°C with 0, 1, 10, 50, 100, 250, and 500 nM [␥-32 P]ATP. D, competition experiment with [␥-32 P]ATP and 1, 10, 50, 100, 500, or 1000 nM unlabeled ATP. Other components of the reaction mixture were standard, and the incubation was carried out for 10 min at 25°C. E, chemical stability of phosphoryl linkage after 1 h of treatment at 37°C with 2 M NaOH, 0.8 M NH 2 OH pH 6.8, 1 M HCl, 50 mM Tris-HCl, pH 7.5, and no treatment (lanes 1-5, respectively). Incubation was in standard stock buffer with 1 g of purified protein for 15 min at 25°C. The gel was electroblotted onto PVDF membranes and exposed to a phosphorimaging plate for 15 min prior to chemical treatment and re-exposed for 15 min after treatment.
from the kinase domain of CikA. The N-terminal domain of the circadian clock protein KaiA has been identified structurally as a pseudo-receiver that is unlikely to be involved in phosphorelay activity (51). Sequence also predicts a pseudo-receiver, unrelated to phosphoryl transfer, in the A. thaliana putative clock component TOC1 and its APRR family members (35)(36)(37)(38). We predict that each of these acts as a protein-protein interaction domain that induces conformational changes in another do-main to modulate its activity. Unlike the N-terminal region and GAF domain, elimination of the pseudo-receiver domain increased autokinase activity 10-fold over full-length Trx-CikA. These results suggest that the receiver domain functions as an autophosphorylation suppressor of the CikA HPK transmitter but not as a phosphoryl receiver. This is consistent with a role for the pseudo-receiver similar to that in AmiR, an antiterminator whose RNA-binding domain is regulated by interaction of A, phosphorimage of autophosphorylation time courses from CikA variants. The protein concentration was adjusted to achieve comparable signal intensity among variants (400 ng of Trx-CikA, 100 ng of CikA 6HIS , and 40 ng of ⌬RR). B, graphic representation of the time course from CikA 6HIS . Autophosphorylation was quantified, and the average of the maximum activity from four separate assays was set as 100% (n ϭ 4).
FIG. 9. In vitro assay to detect potential trans-phosphorylation of the receiver-like domain by the CikA HPK. A, scheme for experimental design. A donor CikA variant having kinase activity but lacking the receiver-like domain (⌬RR) was allowed to autophosphorylate and then mixed with a recipient that is full-length but defective for kinase activity (Trx-H393A). B, results of test for trans-phosphorylation between dissimilar CikA variants. The ⌬RR phosphodonor and Trx-H393A potential recipient (0.5 g each) were incubated in solution containing 25 mM Tris, pH 7.5, 100 mM KCl, 5 mM MgCl 2 , 3000 Ci/mmol [␥-32 P]ATP, 0.01 mM unlabeled ATP, 1 g/l BSA, and 2 mM dithiothreitol at 30°C. The proteins were separated by 10% SDS-PAGE, dried to filter paper, and exposed to a phosphorimaging plate. The migration positions of the two variants are indicated.
Bacteriophytochrome two-component systems are present in other cyanobacterial species and have been characterized best from Synechocystis sp. PCC 6803. Cph1 carries a typical GAF domain that includes the critical cysteine residue for bilin attachment and has an HPK domain; Rcp1 is its response regulator (19). Like plant phytochromes, Cph1 exists in photoconvertible Pr and Pfr forms that are generated by absorption of red or far-red light, respectively. Cph1 autophosphorylation is attenuated by red light and increased by far-red light; the phosphorylated Pr form transfers a phosphoryl group to Rcp1 (19). In the case of plant phytochromes, light affects a Ser/Thr protein kinase activity (15,16). CikA reconstituted in vitro with purified PCB had the same level of autophosphorylation activity as CikA apoprotein (data not shown) and showed no spectral activity. These differences between CikA and phytochromes/ bacteriophytochromes suggest a very distinct mechanism for modulating signal transduction, despite an expected shared reliance on phosphoryl transfer to a partner response regulator protein.
CikA plays a central role in transducing environmental information to the cyanobacterial circadian clock. In addition to acting as a mediator for light/dark signals (27), it is required for response to temperature and clock-generated signals as well. 3 Whether multiple signals converge on CikA as an integrator before transduction to the clock or CikA is part of a complex that can sense various inputs directly remains to determined. The basic function of the protein as a kinase has been established; further understanding its role depends on discovering the other molecules with which it directly interacts.