Mutations in N-terminal Flanking Region of Blue Light-sensing Light-Oxygen and Voltage 2 (LOV2) Domain Disrupt Its Repressive Activity on Kinase Domain in the Chlamydomonas Phototropin*

Background: A plant photoreceptor “phototropin” is a light-dependent kinase containing the LOV photosensory domains. Results: Mutations in the N-terminal flanking region of LOV2 elevate kinase activity in darkness. Conclusion: The N-terminal flanking region is involved in intramolecular signaling from LOV2 to the kinase domain. Significance: This work provides insights into how the LOV domain can activate the kinase domain intramolecularly. Phototropin is a light-regulated kinase that mediates a variety of photoresponses such as phototropism, chloroplast positioning, and stomata opening in plants to increase the photosynthetic efficiency. Blue light stimulus first induces local conformational changes in the chromophore-bearing light-oxygen and voltage 2 (LOV2) domain of phototropin, which in turn activates the serine/threonine (Ser/Thr) kinase domain in the C terminus. To examine the kinase activity of full-length phototropin conventionally, we employed the budding yeast Saccharomyces cerevisiae. In this organism, Ser/Thr kinases (Fpk1p and Fpk2p) that show high sequence similarity to the kinase domain of phototropins exist. First, we demonstrated that the phototropin from Chlamydomonas reinhardtii (CrPHOT) could complement loss of Fpk1p and Fpk2p to allow cell growth in yeast. Furthermore, this reaction was blue light-dependent, indicating that CrPHOT was indeed light-activated in yeast cells. We applied this system to a large scale screening for amino acid substitutions in CrPHOT that elevated the kinase activity in darkness. Consequently, we identified a cluster of mutations located in the N-terminal flanking region of LOV2 (R199C, L202L, D203N/G/V, L204P, T207I, and R210H). An in vitro phosphorylation assay confirmed that these mutations substantially reduced the repressive activity of LOV2 on the kinase domain in darkness. Furthermore, biochemical analyses of the representative T207I mutant demonstrated that the mutation affected neither spectral nor multimerization properties of CrPHOT. Hence, the N-terminal flanking region of LOV2, as is the case with the C-terminal flanking Jα region, appears to play a crucial role in the regulation of kinase activity in phototropin.


Phototropin is a light-regulated kinase that mediates a variety of photoresponses such as phototropism, chloroplast positioning, and stomata opening in plants to increase the photosynthetic efficiency. Blue light stimulus first induces local conformational changes in the chromophore-bearing light-oxygen and voltage 2 (LOV2) domain of phototropin, which in turn activates the serine/threonine (Ser/Thr) kinase domain in the C terminus.
To examine the kinase activity of full-length phototropin conventionally, we employed the budding yeast Saccharomyces cerevisiae. In this organism, Ser/Thr kinases (Fpk1p and Fpk2p) that show high sequence similarity to the kinase domain of phototropins exist. First, we demonstrated that the phototropin from Chlamydomonas reinhardtii (CrPHOT) could complement loss of Fpk1p and Fpk2p to allow cell growth in yeast. Furthermore, this reaction was blue light-dependent, indicating that CrPHOT was indeed light-activated in yeast cells. We applied this system to a large scale screening for amino acid substitutions in CrPHOT that elevated the kinase activity in darkness. Consequently, we identified a cluster of mutations located in the N-terminal flanking region of LOV2 (R199C, L202L, D203N/G/V, L204P, T207I, and R210H). An in vitro phosphorylation assay confirmed that these mutations substantially reduced the repressive activity of LOV2 on the kinase domain in darkness. Furthermore, biochemical analyses of the representative T207I mutant demonstrated that the mutation affected neither spectral nor multimerization properties of CrPHOT. Hence, the N-terminal flanking region of LOV2, as is the case with the C-terminal flanking J␣ region, appears to play a crucial role in the regulation of kinase activity in phototropin.
To sense and respond to a fluctuating light environment, plants have evolved several classes of photoreceptor molecules that convert visible light stimuli into biological signals. Phototropin (PHOT) 2 is a blue light (BL) receptor, which is widely conserved in the plant kingdom (1). Most terrestrial plants possess two isoforms of phototropin designated PHOT1 and PHOT2. In Arabidopsis thaliana (At), PHOT1 and PHOT2 have redundant and distinct functions to mediate responses, such as phototropism (2,3), chloroplast movement (4,5), stomatal opening (6), and leaf photomorphogenesis (7,8) to optimize photosynthesis. On the other hand, the unicellular green alga Chlamydomonas reinhardtii (Cr) possesses a single phototropin homologue (CrPHOT), which has been proposed to regulate sexual differentiation (9,10) and the expression several photosynthetic genes. Interestingly, CrPHOT is fully functional in Arabidopsis, suggesting that the basic mechanism of phototropin signal transduction is highly conserved (11). Phototropins are BL-dependent Ser/Thr protein kinases that share well conserved structural properties. The kinase domain belongs to the AGC VIII subfamily (protein kinases A, G, and C) (12). Autophosphorylation of a conserved serine residue at the activation loop of the kinase domain is prerequisite for its physiological activity (13,14). To date, little is known about authentic phosphorylation substrates for phototropins. Recently, however, auxin efflux transporter ATP-binding cassette B19 (ABCB19) has been reported as a putative endogenous substrate in Arabidopsis (15).
In addition to the kinase domain, phototropin contains two N-terminal light-oxygen and voltage (LOV) domains, designated LOV1 and LOV2. The LOV domains, which are members of the Per-ARNT-Sim (PAS) domain superfamily, noncovalently bind flavin mononucleotide (FMN) as chromophores (16). Because the kinase fragment lacking LOV domains displays constitutive activity both in planta (17) and in vitro (18), LOV domains appear to suppress kinase activity in darkness intrinsically.
Upon BL absorption, a covalent adduct is formed between the FMN and a conserved cysteine residue within each LOV domain (19,20). Consequently, reorganization of the hydrogen bond network takes place in the FMN binding pocket (21,22). Inactivation of LOV domains by site-directed mutagenesis indicates that LOV2 plays a predominant role in regulating the kinase activity (18,23,24), whereas the role of LOV1 is limited to modulate photosensitivity (18). An NMR analysis of Avena PHOT1 has revealed an amphipathic ␣ helix, J␣, which is attached to the C terminus of the LOV2 core domain (25). The J␣ helix changes its structure in response to a light stimulus. Namely, it detaches from the LOV2 core (25) and unfolds (26,27). The prominent role of the J␣ helix in the regulation of kinase activity has been evidenced by site-directed mutagenesis in Arabidopsis PHOT1. An amino acid substitution that disrupts the LOV2-J␣ interaction leads to constitutive activation of the kinase in vitro (28). More recently, the crystallographic analysis of Avena PHOT1 LOV2 has suggested that not only the C-terminal J␣ but also the N-terminal flanking region display a light-induced structural change (22). These structural changes are presumed to disrupt the inhibitory interaction of LOV2 with the kinase domains. Accordingly, photoreversible contact/ separation between the LOV2 and the kinase domains has been suggested by small angle x-ray scattering in a LOV2 linker kinase fragment of Arabidopsis PHOT2 (29).
Besides the site-directed mutagenesis approach based on structural information, a random-mutagenesis approach utilizing Escherichia coli as a host has been applied successfully to uncover amino acid residues involved in the photochemistry of LOV2 in Avena PHOT1 (30). Although this method is quite effective for identifying spectral mutants in the LOV domains, it is not applicable for screening mutants that are altered in the photoregulation of the kinase activity. This is because no convenient way to assess the kinase activity of phototropin exists in E. coli.
In this study, we developed a new yeast experimental system in which CrPHOT triggered cell growth depending on its kinase activity. We applied this method to screen for mutations that disrupt repression of the kinase domain. Consequently, we revealed that the N-terminal flanking region of LOV2 is important for the suppression of kinase activity in darkness.

EXPERIMENTAL PROCEDURES
Media and Growth Conditions for Yeast-Yeast strains were cultured in YPDA-rich medium (1% yeast extract, 2% bactopeptone, 2% glucose, and 0.01% adenine). Strains carrying plasmids were selected in synthetic medium (SD) containing the required nutritional supplements (31). When appropriate, 0.5% casamino acids were added to SD medium without uracil (SDA-Ura). For induction of the GAL1 promoter, 3% galactose and 0.2% sucrose were added as carbon sources instead of glucose (YPGA and SGA-Ura). Growth sensitivity to duramycin was examined on YPDA plates containing 10 M duramycin (Sigma). Blue (peak at 470 nm) and red (peak at 660 nm) lightemitting diode panels (ISL-150X150 series; CCS, Tokyo, Japan) were used as light sources.
For yeast assays, pRS416-AtPHOT1, pRS416-AtPHOT2, and pRS416-CrPHOT were constructed as follows. The coding regions of PHOT1 and PHOT2 from Arabidopsis and PHOT from Chlamydomonas were amplified by PCR and cloned into the BamI/SalI site of pRS416 (33) to allow constitutive expression of phototropins tagged with N-terminal two tandem repeats of the influenza virus hemagglutinin epitope (2HA) under the control of the TPI1 promoter. Site-directed mutations were introduced by using the QuikChange site-directed mutagenesis kit (Stratagene) in accordance with the instructions of the supplier. All amino acid changes were verified by DNA sequencing. For overexpression of CrPHOT protein in E. coli, pET28a-CrPHOT was constructed as follows. Wildtype (WT) and mutant PHOT were amplified from pRS416-CrPHOT and cloned into NdeI/SalI sites of pET28a plasmid vector (Novagen) to allow expression of CrPHOT with an N-terminal histidine tag, which was composed of 20 residues including six histidine residues.
Immunoblot Analysis-Protein extraction from yeast was performed as described (34). Immunoblot analysis was performed essentially as described previously (11). The same amounts of protein extracts were run on 10.0% SDS-PAGE and blotted onto nitrocellulose membrane. Rat anti-HA monoclonal antibody (Roche Applied Science) and alkaline phosphatase-conjugated anti-rat IgG antibody (Promega) were used.
Yeast Growth Assay-The cells were grown in liquid SGA-Ura medium to A 600 of 0.6 -0.8, harvested by centrifugation, and resuspended into sterile water, resulting in an A 600 of 0.1. Serial 5-time dilutions were then prepared for loading 10 l each of the diluted cell suspension onto a plate with SGA-Ura or SDA-Ura medium. The plates were then placed under different light conditions.

PCR-based Random Mutagenesis and Mutant Screening for Pseudolit Phenotype-Error-prone PCR libraries of the N-terminal (1-1399) and C-terminal (1179 -2136) moieties of
CrPHOT flanked by sequences from the pRS416 vector (ϳ150 bp) were generated from the pRS416-CrPHOT using the Diversity PCR random mutagenesis kit (Clontech). The primer sequences are as follows: pRS416 forward for the N-terminal library, 5Ј-GTGAACTTGCAACATTTACTATTTTCCC; CrPHOT-1399 reverse for the N-terminal library, 5Ј-CGGTC-TGGATGGTGCAGTACAG; CrPHOT-1179 forward for the C-terminal library, 5Ј-CTACCAGGCGCTGCTGCAGCTG; and pRS416 reverse for the C-terminal library, 5Ј-AAAGCGG-GCAGTGAGCGCAACGC. Final concentrations of 320 M MnSO 4 and 16 M dGTP were chosen for the PCR to adjust the mutagenesis frequency to 1.5 mutations/kb. The pRS416-CrPHOT plasmid was digested with BamHI/ NruI and SalI/NruI to remove the sequence encoding the N-terminal and C-terminal moieties, respectively. The N-or C-terminal PCR fragments and a pRS416-CrPHOT-derived vector lacking the corresponding part of CrPHOT were mixed and transformed into the P GAL1 -CDC50 fpk1⌬ fpk2⌬ yeast strain, so that homologous recombination generated a vector containing mutagenized PHOT. The yeast strain was then plated on SDA-Ura agar and incubated in the dark at 28°C for 60 h. The growing colonies were isolated. Among isolated colonies, candidates were randomly selected, and plasmids were rescued using Wizard prep (Promega) according to the instructions in the DUALhybrid kit (Dualsystems Biotech). The candidate plasmids were sequenced in the mutagenized moieties. The second growth assay was performed by repeating the assays using the purified plasmids.
Expression and Purification of Recombinant Proteins-His 6tagged full-length CrPHOT and N-terminal fragment of Arabidopsis PHOT1 (AtP1-Nt) were prepared from E. coli according to the protocols used for LOV2 linker kinase fragment preparation (35) with minor modifications. The E. coli cells were lysed with a French pressure cell (Thermo Fisher Scientific) in a buffer containing 20 mM HEPES-NaOH (pH 7.8), 500 mM NaCl, and 1 mM PMSF. The purification was performed at 0 -4°C under a dim red light. The centrifugation supernatant of the lysate was subjected to Ni 2ϩ -ion affinity chromatography (Qiagen). The AtP1-Nt were then desalted by using HiTrap Desalting (GE Healthcare) and stored at Ϫ80°C.
The CrPHOT protein was further purified via size-exclusion column chromatography (Sephacryl HR 200; GE Healthcare) in a buffer containing 100 mM NaCl, 1 mM EGTA, and 20 mM Tris-HCl (pH 7.8). Glycerol was added to the eluted CrPHOT to a final concentration of 10% (v/v). The phosphorylation levels of the products after purification were examined by Pro-Q Diamond phosphoprotein gel stain (Invitrogen) according to the manufacturer's instructions if necessary.
For the spectral analysis, the above sample was further purified using an anion-exchange Resource Q column (GE Healthcare) equilibrated with a buffer containing 10% (v/v) glycerol, 1 mM EGTA, and 20 mM Tris-HCl (pH 7.8). The CrPHOT fraction that did not bind to the Resource Q column was collected and concentrated by ultrafiltration using the Amicon Ultra K15 (Millipore). The purity of CrPHOT in each step was examined by SDS-PAGE.
In Vitro Phosphorylation Assays-The phosphorylation assay was performed under dim green light (18,35). CrPHOT and AtP1-Nt polypeptides were incubated at 30°C in a kinase reaction mixture containing 20 mM Tris-HCl (pH 7.8), 100 mM NaCl, 10% glycerol, 10 mM MgCl, 500 M ATP, and 60 kBq of [␥-32 P]ATP. The reaction was terminated by the addition of concentrated SDS-PAGE sample buffer followed by boiling for 3 min. Then, the samples were run on SDS-PAGE to visualize phosphorylation bands using an imaging plate and a bioimaging analyzer (FLA2000; Fuji, Tokyo, Japan). The molecular masses of phosphorylated proteins were estimated by Coomassie Brilliant Blue (CBB) staining with reference to molecular mass standard samples (Sigma).
Spectral Analyses and Size-exclusion Chromatography-The absorption spectra of CrPHOT under BL illumination and the time course of absorption changes during thermal decay from the cysteinyl adduct state to the dark-adapted state were measured essentially as described previously (35) at 20°C and pH 7.8. The molecular mass of CrPHOT was estimated by size-exclusion chromatography using a Superdex 200 pg Hi-Load column (GE Healthcare) and the kinase reaction buffer as previously described (35).

RESULTS
Complementation of fpk1⌬ fpk2⌬ and cdc50⌬ Synthetic Lethality in S. cerevisiae by Chlamydomonas PHOT in a BL-dependent Manner-To establish an experimental system that can easily and quickly assess the light-dependent kinase activity of phototropin, we employed S. cerevisiae, in which the protein kinases named flippase-kinase 1 (Fpk1p) and Fpk2p were recently identified (32). Among the yeast kinases, Fpk1p and Fpk2p exhibit the highest sequence homology to the AGCVIII kinase domain of phototropins (32) (Fig. 1A and supplemental Fig. S1). Hence, we reasoned that phototropins might complement the loss of FPKs in yeast in a light-dependent manner.
CrPHOT Acts as a Light-dependent Flippase-Kinase in Yeast-We genetically examined whether the kinase activity of CrPHOT was indeed required for the complementation. The kinase domain fragments of phototropins exhibit constitutive kinase activity without light stimulus (17,18). As expected, the kinase domain fragment of CrPHOT restored yeast growth regardless of the light condition (Fig. 1B). Conversely, the D549N mutant of CrPHOT, which corresponded to the kinasedead mutation in AtPHOT1 and AtPHOT2 (17,23), failed to restore growth. The substitution of conserved Cys 39 to alanine in the LOV2 domain is known to decrease the extent of light activation in phototropins (18,23). The corresponding C250A mutant of CrPHOT restored growth but very weakly (Fig. 1B). Taken together, CrPHOT complemented the loss of Fpks in a kinase-like manner.
We then examined whether the complementation by CrPHOT indeed depended on its light activation. First, we confirmed that the level of CrPHOT in yeast was not altered under BL with an anti-HA tag antibody (Fig. 1C). This was because phototropin is known to be degraded more rapidly under BL in some cases (7,39). It should be noted that electric mobility shift due to BL-induced autophosphorylation was not observed. This was probably because such activity is low in CrPHOT (11).
We determined the dependence of complementation by CrPHOT on BL intensity (Fig. 1D). The degree of complementation increased as BL was increased from 0.01 to 10 mol m Ϫ2 s Ϫ1 . This result matched well with those from a previous analysis on the biological activity of CrPHOT expressed in Arabidopsis (11).

BL-dependent Regulation of Phospholipid Uptake by
CrPHOT-We further examined whether CrPHOT activated flippases in yeast as Fpks do. Phosphatidylethanolamine is enriched in the outer leaflet of the membrane in fpk1⌬ fpk2⌬ mutant due to reduced flippase activity. This leads to growth defect in the presence of duramycin, a cytotoxic small tetracyclic peptide that binds phosphatidylethanolamine on the outer leaflet. As expected, CrPHOT rescued the duramycin-sensitive growth defect in a BL-dependent manner (Fig. 2). In addition, the kinase fragment rescued the phenotype regardless of the light condition, whereas the putative kinase-dead mutant (D549N) of CrPHOT failed to restore the growth. Hence, CrPHOT was suggested to regulate the flippase activity in yeast.
Search for Novel CrPHOT Variants Showing Pseudolit Activity-The above experimental system should be potentially useful in identifying amino acid residues important for phototropin functions. To explore this possibility, mutations were introduced separately into the N-and C-terminal moieties of CrPHOT by a PCR-based method at the rate of 1.5 mutations/kb (Fig. 3A). Pools of transformed cells corresponding to 5.0 ϫ 10 5 each of successful recombination events were screened for mutations that allowed growth in darkness on SDA-Ura plates. Consequently, 1504 and 107 growth-positive colonies were obtained for the N-and C-terminal mutant pools, respectively.
We then sequenced 96 randomly selected clones for the N-terminal mutations and all 107 clones for the C-terminal mutations. Consequently, we found that many of them carried multiple amino acid substitutions. In addition, a large deletion, probably due to irregular homologous recombination, was found in some clones. Nonetheless, eight clones with single amino acid substitutions were identified (Fig. 3A). Additionally, four substitutions were shared in more than three independent clones with multiple mutations. Reintroduction of the latter mutations into the wild-type CrPHOT confirmed that they indeed conferred the pseudolit phenotype ( Fig. 3C and supplemental Fig. S3). We also confirmed that the expression levels of these mutant proteins were not elevated in yeast cells (supplemental Fig. S4).
Kinase Activities of Full-length CrPHOT Carrying Mutations in the N-terminal Flanking Region of LOV2-Interestingly, 8 of the above 12 mutations were clustered in a well conserved N-terminal LOV2-flanking region (residues Arg 199 -Arg 210 ) (Fig. 3, A and B). Hence, their kinase activity was examined in  vitro. For this purpose, we attempted to express the recombinant full-length CrPHOT protein tagged with an N-terminal His 6 tag in E. coli. Full-length CrPHOT protein was recovered as a yellow-colored, soluble protein from the E. coli extract. For the kinase assay, the CrPHOT protein was purified from the extract by Ni 2ϩ -affinity chromatography and size-exclusion chromatography. SDS-PAGE revealed that the resulting sample consisted of a main band of 90 kDa, which corresponded to the theoretical mass of the protein, 85.7 kDa, with several minor bands of lower molecular masses. The purity was then estimated to be ϳ75% (Fig. 4A), which was comparable with those of samples used for the previous kinase assays of LOV kinase fragments (18,35).
In vitro phosphorylation assay for the purified CrPHOT was performed using AtP1-Nt as a substrate (35). As expected, CrPHOT phosphorylated AtP1-Nt under BL, whereas phosphorylation was substantially reduced in darkness (Fig. 4B). The phosphorylation level increased during the incubation period of 60 min. The effects of BL were saturated at ϳ10 mol m Ϫ2 s Ϫ1 (Fig. 4C). It should be noted that the background activity in darkness was relatively high, which has been reported for other phototropin samples as well (40,41).
We then examined the kinase activities in 6 of 8 mutants carrying the mutation in the N-terminal flanking region of LOV2 (R199C, D203N, L204P, T207I, and R210H). The fulllength proteins were expressed in E. coli and purified as described above. Consequently, protein samples in similar qualities were obtained with the exception of R199C in which a larger amount of impurity at around 36.5 kDa was detected (Fig.  5A).
We examined the kinase activity using AtP1-Nt as a substrate. Consistent with the yeast phenotype (Fig. 3), the above mutants exhibited increased kinase activities in darkness compared with the WT with the exception of D203N (Fig. 5, B and  C). Hence, most of the CrPHOT mutants disrupted repression of the kinase activity in vitro.
Spectroscopic Analysis of Mutation in N-terminal Flanking Region of LOV2-We examined whether the present mutations affected the spectral nature of CrPHOT. For this purpose, the T207I mutant was selected as a representative mutant and expressed in E. coli. The protein sample conventionally prepared according to the procedure listed for the kinase assay was further purified by negative ion exchange chromatography. Consequently, a sample with Ͼ95% purity was obtained (Fig.  6A). The absorption spectroscopy revealed that both WT and T207I mutant exhibited the typical fine structured absorption spectra of protein-bound FMN with absorption maxima at 425, 447, and 472 nm (Fig. 6B). Under continuous BL irradiation, both WT and T207I CrPHOT proteins were converted to cysteinyl adduct state with absorption maxima at 390 nm (Fig. 6B). Those absorption spectra were nearly identical to those of the LOV1-hinge-LOV2 fragment of CrPHOT, which lacked the kinase domain (42).
At a fluence rate of 270 mol m Ϫ2 s Ϫ1 , adduct formation was saturated within 24 s, whereas the dark reversion was completed within 400 s after the illumination was turned off (data not shown). By exponential fitting analysis, the half-lives for dark decay of two photoproduct components in WT CrPHOT were calculated to be 15 and 50 s, whereas those for the T207I mutant were 15 and 56 s. Hence, the T207I mutant was almost indistinguishable from the WT. Taken together, the T207I mutation in the N-terminal flanking region of LOV2 activated the kinase domain without affecting the spectral nature of CrPHOT.
Estimation of Molecular Mass by Size-exclusion Chromatography-LOV1 domains in phototropins have been suggested to form a dimer (43)(44)(45). Hence, we examined the possibility that the T207I mutation affected kinase activity by altering the multimerization state of full-length CrPHOT. To test this possibility, we determined the molecular masses of full-length WT and T207I CrPHOT by size-exclusion chromatography. For this purpose, the samples were purified as for the spectral analysis.
The size-exclusion elution profile demonstrated that both WT and T207I CrPHOT proteins exhibited a major distinct peak at 71.14 ml, corresponding to a molecular mass of 123.7 kDa (supplemental Fig. S5). Hence, the T207I mutation did not alter the multimerization state of full-length CrPHOT. Another minor peak at 78 ml (corresponding to 68 kDa) probably represented a degradation product.

CrPHOT Complements Loss of Fpk1p and Fpk2p Kinases in
Yeast-A bacterial system has been employed successfully to identify amino acid residues important for the spectral integrity of the LOV domains of phototropins (30). The present study further extended this strategy to assess the signaling activity of phototropin derivatives in a microorganism.
Indeed, CrPHOT successfully rescued the fpk1⌬ fpk2⌬ phenotype to allow cell growth in a BL-dependent manner in yeast (Fig. 1). We further demonstrated that both photoexcitation of the LOV domain and the kinase activity of CrPHOT are necessary to promote yeast cell growth (Fig. 1B). Hence, we have developed an experimental system suitable for a large scale mutant screening for CrPHOT.
It remains unclear why the Arabiodopsis PHOT1 and PHOT2 did not function properly in yeast. To gain further insights, the complementation activities of respective kinase fragments were examined. Consequently, the PHOT1 kinase fragment failed to promote the cell growth, whereas the PHOT2 fragment did (supplemental Fig. S2). Hence, intrinsic kinase activities of Arabiodopsis PHOT1 and PHOT2 appeared to be quite different in yeast for an unknown reason.
Kinase Activity of Full-length CrPHOT-It had long been believed that expression of recombinant full-length phototropins in E. coli for biochemical analyses was very difficult. Recently, however, Pfeifer et al. successfully purified full-length CrPHOT in E. coli to conduct FTIR spectroscopy (46). Similarly, we were able to prepare full-length WT and mutated CrPHOT proteins pure enough for the in vitro kinase assay, photochemical kinetic analysis, and multimerization analysis.
The BL-dependent kinase activity of the purified CrPHOT was successfully shown in vitro using AtP1-Nt as a substrate (Fig. 4B). As has been shown already for CrPHOT expressed in insect cells (11), autophosphorylation activity was relatively low. This was not because the purified CrPHOT was already phosphorylated in E. coli (supplemental Fig. S6). We suppose that fewer phosphorylation sites exist in CrPHOT (11).
Spectral and Multimerization Properties of Full-length CrPHOT-Fast and slow components observed in the dark decay kinetics (Fig. 6) presumably represented LOV2 and LOV1 (t1 ⁄ 2 ϭ 15 and 50 s, respectively). Compared with the CrPHOT LOV1ϩLOV2 fragment (t1 ⁄ 2 ϭ 17 and 200 s) (47), the slower component was four times faster in the full-length CrPHOT. Hence, the presence of the C-terminal kinase domain might alter the photochemical kinetics of LOV1 as previously reported for Arabidopsis PHOT1 and PHOT2 (47).
The LOV1 domains of Arabidopsis PHOT1 and PHOT2 (43)(44)(45) and CrPHOT (48) have been suggested to form a dimer and higher oligomers. However, size-exclusion chromatography in this study hinted that full-length CrPHOT may exist as a monomer (apparent molecular mass of 123.7 kDa versus theoretical mass of 85.7 kDa) (supplemental Fig. S5). This possibility should be investigated further by other methods such as small angle x-ray scattering or crystallization analysis.
Mutations Outside LOV2 N-terminal Flanking Region-Prior to the present work, several amino acid substitutions in the N-terminal region of phototropins had been known to cause the pseudolit activation. Those include the ones in the J␣ helix (V601E, A605E, and I608E in Arabidopsis PHOT1) (28) and the LOV2 core region (G513N in Avena PHOT1) (49). By contrast, such a mutation had not been known within the C-terminal region. In the present study, four novel mutations were found in addition to those clustered in the N-terminal LOV2 flanking region (see below).
Among them, L354P resides in the vicinity of J␣ (Fig. 3A  and supplemental Fig. S3). This is not surprising because several mutations are already known in this region to cause the pseudolit phenotype (see above). By contrast, Y39C in the LOV1 core region (Fig. 3A and supplemental Fig. S3) was somewhat unexpected because LOV1 is less important than LOV2 to regulate the kinase activity (18,23,24). However, LOV1 modifies the light sensitivity of PHOT2 (18). Hence, this mutation may affect the presumed LOV1/LOV2 or LOV1/kinase interaction to substantially increase the light sensitivity.
The N610S/T mutations in the activation loop of the kinase domain conferred the pseudolit phenotype ( Fig. 3A and supplemental Fig. S3). Interestingly, the Asn 610 residue is adjacent to Ser 611 , which corresponds to the autophosphorylation site required for the activation of both PHOT1 and PHOT2 in Arabidopsis (13,14) (supplemental Fig. S1). We constructed the N610A CrPHOT to show that Ser 610 and Thr 610 but not Ala 610 caused the pseudolit activation (supplemental Fig. S3). Hence, the newly introduced Ser/Thr 610 might have acted as a secondary phosphorylation site to increase the kinase activity.
Involvement of N-terminal Flanking Region of LOV2 in Lightinduced Signal Transduction-In our screening, we identified a novel mutational hot spot in the N-terminal flanking region of LOV2 (Arg 199 -Arg 210 in Fig. 3A). These mutations substantially reduced repression of the kinase activity by the N-terminal moiety in darkness (Figs. 3 and 5) without affecting the photochemical (Fig. 6) or multimerization properties of CrPHOT (supplemental Fig. S5).
The crystal structure of LOV2 with N-and C-terminal flanking regions from Avena PHOT1 (residues Leu 404 -Ala 559 in Avena PHOT1) indicates that the N-terminal region (residues Leu 404 -Arg 414 in Avena PHOT1) is in the turn-helix-turn structure, in which the helix has been denoted as AЈ␣ (residues Thr 407 -Arg 410 in Avena PHOT1) (22). In this structure, AЈ␣ is packed against the surface of the ␤ sheet of the LOV2 core domain to interact with the J␣ helix (residues Asp 522 -Ala 543 in Avena PHOT1) in the C-terminal flanking region. Importantly, the light-induced structural rearrangements within the LOV2 core are associated with displacement of residues surrounding AЈ␣ (22). Hence, not only the C-terminal J␣ helix but also the N-terminal flanking region are proposed to act as interfaces relaying the signal from the LOV2 core to the outside domain.
Our mutational analysis sheds new light on the biochemical function of the N-terminal flaking region of LOV2. The sequences in this region (Arg 199 -Arg 210 ) are strikingly conserved among different phototropins (Fig. 3B). The alignment suggests that T207I and R210H reside within AЈ␣ whereas L204P is in the loop upstream of AЈ␣. Although the other three mutational sites, R199C, L202P, and D203N/ G/V, are excluded from the above structural study, they are in close proximity on the primary structure. It should be noted here that a high degree of homology does not warrant similar phenotype in corresponding mutants of other phototropins. Indeed, Arabidopsis PHOT1 carrying T469I (corresponding to T207I in CrPHOT) failed to show pseudolit activation (data not shown).
Hypothetical Extended AЈ␣ Helix May Play an Important Role for Intramolecular Signaling-Interestingly, a secondary structure prediction suggests that the AЈ␣ helix could be extended toward the N terminus to form a longer ␣ helix (Arg 199 -Arg 210 ), within which all the six mutational sites are mapped (supplemental Fig. S7A). Furthermore, the presumed ␣ helix shows amphipathic character (supplemental Fig. S7B). The hydrophobic surface of this helix might then be packed against the hydrophobic LOV2 core. If this prediction is correct, the mutated residues residing in the polar surface of the helix (Arg 199 , Asp 203 , Thr 207 , and Arg 210 ) might be directly involved in the physical interaction with other functional domains, such as the J␣ helix and the C-terminal kinase. The existence of and a structural role for the predicted helix must be investigated in future studies.