Blue light-induced autophosphorylation ofphototropin is a primary step for signalingShin-ichiro Inoue*, Toshinori Kinoshita*†, Masaki Matsumoto‡, Keiichi I. Nakayama‡, Michio Doi§, andKen-ichiro Shimazaki*¶

*Department of Biology, Faculty of Science, Kyushu University, Ropponmatsu, Fukuoka 810-8560, Japan; ‡Department of Molecular and Cellular Biology,Medical Institute of Bioregulation, Kyushu University, Maidashi, Fukuoka 812-8582, Japan; and §Center for Research and Advancement in HigherEducation, Kyushu University, Ropponmatsu, Fukuoka 810-8560, Japan

Edited by Winslow R. Briggs, Carnegie Institution of Washington, Stanford, CA, and approved February 4, 2008 (received for review September 27, 2007)

Phototropins are autophosphorylating protein kinases of plant-specific blue light receptors. They regulate various blue lightresponses, including phototropism, chloroplast movements, hypo-cotyl growth inhibition, leaf flattening, and stomatal opening.However, the physiological role of autophosphorylation remainsunknown. Here, we identified phosphorylation sites of Ser or Thrin the N terminus, Hinge1 region, kinase domain, and C terminusin Arabidopsis phototropin1 (phot1) by liquid chromatography–tandem mass spectrometry in vivo. We substituted these Ser or Thrresidues with Ala in phot1 and analyzed their functions by inspect-ing the phot1-mediated responses of stomatal opening, phototro-pism, chloroplast accumulation, and leaf flattening after the trans-formation of the phot1 phot2 double mutant. Among these sites,we found that autophosphorylation of Ser-851 in the activationloop of the kinase domain was required for the responses men-tioned above, whereas the phosphorylation of the other Ser andThr, except those in the activation loop, was not. Ser-849 in theloop may have an additional role in the responses. Immunologicalanalysis revealed that Ser-851 was phosphorylated rapidly by bluelight in a fluence-dependent manner and dephosphorylated grad-ually upon darkness. We conclude that autophosphorylation ofSer-851 is a primary step that mediates signaling between photo-chemical reaction and physiological events.

protein kinase � stomata � photoreceptor � light signaling

Phototropins (phot1 and phot2) mediate multiple blue lightresponses in Arabidopsis, including phototropism, chloro-

plast movements, leaf f lattening, leaf positioning, stomatalopening, and rapid inhibition of hypocotyl growth (1–5). Theseresponses enhance photosynthesis and optimize plant growth,particularly under weak light (6). Under strong light, phot2induces a chloroplast-avoidance response to prevent photodam-age to photosynthetic machinery (7). Phototropins, therefore,are essential proteins for the survival of higher plants and theextension of their living areas, particularly under ever-changingenvironments of light.

Phototropins are blue light receptor protein kinases with twolight, oxygen, voltage (LOV) domains in the N terminus and aSer/Thr protein kinase in the C terminus (8). The LOV domainspossess noncovalent binding sites for the chromophore flavinmononucleotide (FMN) and cysteine residues (2, 9–11) andproduce a covalent cysteinyl adduct with FMN when LOVs areilluminated with blue light (12, 13). The cysteinyl adduct for-mation is the primary photochemical process; the adduct for-mation induces conformational changes in the LOV2 domain(14, 15) and J�-helix (16) and leads to phototropin phosphory-lation and subsequent physiological processes (17, 18). Pho-totropin phosphorylation by blue light is demonstrated to beautophosphorylation through the use of recombinant proteins ofPHOT1 and PHOT2 expressed in insect cells (2, 9, 17).

Phototropin was initially found as a plasma membrane-associated phosphorylated protein in etiolated seedlings of pea(Pisum sativum) under blue light (19), and extensive investiga-

tions established correlations between phosphorylation of theprotein and phototropism (20, 21). A typical example is therelationship between blue light responsiveness in phototropismand levels of phosphorylation in Zea mays coleoptiles (22).Another is the asymmetric phosphorylation of the membrane-associated protein between the irradiated side and the shadedside of Avena sativa coleoptiles (23). These results suggest thatautophosphorylation plays essential roles in phototropism. How-ever, no direct evidence of the role of autophosphorylation inphototropism has been obtained. It is even suspected thatautophosphorylation is not the primary process, because itrequires higher light intensity than is needed for phototropin-mediated responses (4).

Mapping of autophosphorylation sites in vitro identified eightsites of phot1a from A. sativa (24). Two of the sites (Ser-27 andSer-30) were located in the N terminus, and the other six(Ser-274, Ser-300, Ser-317, Ser-325, Ser-332, and Ser-349) werein the hinge region between LOV1 and LOV2 (Hinge1 region).In guard cells of Vicia faba, phot1a and phot1b underwentautophosphorylation by blue light, and the sites were determinedto be Ser-358 for phot1a and Ser-344 for phot1b in the Hinge1regions as the sites of 14-3-3 protein binding (25). In Arabidopsis,however, any of the autophosphorylation sites of phototropinhave not been determined, and thus this information will becrucial for the functional analyses of phototropin using this plantspecies.

In this study, we prepared phot1 protein from blue light-irradiated etiolated seedlings of Arabidopsis and determined thephosphorylation sites in vivo. We then investigated the roles ofthese phosphorylation sites by transforming the phot1 phot2double mutant with the mutated phot1 constructs, in which theidentified Ser or Thr residues had been substituted with Ala, andpaid special attention to the stomatal responses. We demon-strated that autophosphorylation of the kinase domain is essen-tial for initiating signaling to downstream components in phot1-mediated responses.

ResultsPreparation of Phot1 from Etiolated Seedlings Stimulated by BlueLight. Etiolated seedlings of Arabidopsis labeled with [32P]-orthophosphate were irradiated with a pulse of blue light, afterwhich microsomal membranes were immediately prepared.

Author contributions: S.I., T.K., and K.S. designed research; S.I., M.M., and M.D. performedresearch; S.I., T.K., M.M., K.N., M.D., T.K. and K.S. analyzed data; and S.I. and K.S. wrote thepaper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

†Present address: Division of Biological Science, Graduate School of Science, NagoyaUniversity, Chikusa, Nagoya 464-8602, Japan.

¶To whom correspondence should be addressed. E-mail: kenrcb@mbox.nc.kyushu-u.ac.jp.

This article contains supporting information online at www.pnas.org/cgi/content/full/0709189105/DCSupplemental.

© 2008 by The National Academy of Sciences of the USA

5626–5631 � PNAS � April 8, 2008 � vol. 105 � no. 14 www.pnas.org�cgi�doi�10.1073�pnas.0709189105

Phot1 proteins were isolated by immunoprecipitation from themembranes. Autoradiogram showed that the radiation denselylabeled phot1 by 32P with less mobility, whereas there was littlelabeling without irradiation (Fig. 1A). This was blue light-induced autophosphorylation in planta, because the phot1 fromthe kinase-dead D806N-2 line (see Table 1) showed no increasein phosphorylation (Fig. 1 A). A 14-3-3 protein bound to phot1upon autophosphorylation (Fig. 1B). The immunoprecipi-tated phot1 proteins were stained by Coomassie brilliant blue(Fig. 1C).

Mapping of in Vivo Phosphorylation Sites of Phot1. The phot1proteins were digested with trypsin. The digested mixture ofpeptides was subjected to liquid chromatography–tandem massspectrometry (LC-MS/MS). Three separate experimentsmapped eight peptides derived from phot1 with phosphate

modification (80 Da) by a Mascot database search. A typicalcollision-induced dissociation spectrum of phosphopeptide408KSpSLSFMGIK417 is presented (Fig. 1D). The fragmentationpatterns of b- and y-ions allowed us to identify the phosphory-lation site as Ser-410. The ion peaks with the phosphate groupshowed a mass shift by 80 Da as a result of the phosphatasetreatment (data not shown), indicating that the peptide wasphosphorylated.

At least seven other phosphorylation sites were identified inthe same way. The phosphorylation sites included Ser-58, Ser-170, and Ser-185 in the N terminus; Ser-350 and Ser-376 in theHinge1 region; Ser-849 and/or Ser-851 in the kinase activationloop; and Thr-993 in the C terminus (Fig. 1E). Ser-58 andSer-170 were found in the dark, and Ser-350 was found in bothdark and light, but other sites were found in their phosphorylatedforms after irradiation. Because of a lack of informative frag-ment ions, we could not determine whether Ser-849, Ser-851, orboth were autophosphorylated.

To show whether blue light autophosphorylates Ser-851, weidentified phosphorylated-Ser-851 (pSer-851) by an immunolog-ical method. We found a little pSer-851 in the WT-11 line underdarkness, and the amount of pSer-851 increased when theseedlings were irradiated with blue light (Fig. 2A). In thekinase-dead line D806N-2, the amount of pSer-851 did notincrease with blue light. The pSer-851 was absent in phot1 fromthe S851A-3 line (Table 1). These results indicate that Ser-851is autophosphorylated by phot1 under blue light. The smallamount of pSer-851 in both WT-11 and the D806N-2 lines mightbe produced by other kinase(s) in the dark. Phosphorylation ofSer-851 by blue light was fluence-dependent (Fig. 2B). Thephosphorylation proceeded rapidly in response to blue light, andthe pSer-851 was completely dephosphorylated within 10 min inthe dark (Fig. 2C).

Functional Role of Phosphorylation in Phot1 in Stomatal Opening. Weconstructed single- and multiple-site-mutated phot1 genes bysite-directed mutagenesis via substitutions of Ser or Thr withAla. These included a kinase-dead construct (D806N; ref. 17)and constructs of single amino acid substitutions, of the removalof 14-3-3 protein-binding sites (S350A S376A; data not shown)of the activation loop substitutions (S849A, S851A, and S849A/DS851A/D), and of all of the identified phosphorylation sitesubstitutions except the activation loop (Others) (Table 1). Wetransformed the phot1 phot2 double mutant (phot1-5 phot2-1)with these constructs under the control of native PHOT1 pro-moter and selected the transgenic plants that expressed mutantphot1 proteins with levels similar to that of the control (gl1)(Fig. 3C).

We initially assessed the blue light-specific stomatal opening.Stomata in the epidermis of gl1 leaves opened slightly by red lightand widely by a weak blue light superimposed on the red light.The phot1 phot2 double mutant did not respond (Fig. 3A). Theblue light-specific stomatal opening was restored completely bythe wild-type construct but not by the kinase-dead constructD806N (WT-11 and D806N-2 lines).

We measured blue light-specific stomatal opening in trans-genic plants expressing mutant phot1 proteins with a singleamino acid substitution of S58A, S170A, S185A, S350A, S376A,S849A, S851A, or T993A. The openings were restored almostcompletely in all of these plants (data not shown) except for theS851A-3 line. The stomatal response was only slightly restored inthe S851A-3 line (Fig. 3A).

We determined the responses in the transgenic plants express-ing mutant phot1 proteins with multiple amino acid substitu-tions. The lines lacking 14-3-3 protein-binding sites of S350AS376A-4 and those of Others-4 with substitutions of the identi-fied Ser and Thr (Table 1) showed complete stomatal openingin response to blue light (Fig. 3A). However, the line of S849A

Fig. 1. Identification of the phosphorylation sites of Arabidopsis phot1. (A)Blue light-induced autophosphorylation of phot1 in vivo in Arabidopsis.Etiolated seedlings labeled by 32P were kept in the dark (Dk) or illuminatedwith blue light at 100 �mol m�2�s�1 for 1 min (BL). The phot1 protein wasisolated by immunoprecipitation from the microsomes of the seedlings, andseparated by SDS/PAGE. An autoradiogram is illustrated. (B) Phosphorylation-dependent binding of a 14-3-3 protein to the immunopurified phot1. Proteinblot of phot1 was done with glutathione-S-transferase (GST)-GF14�. (C )Coomassie brilliant blue stain of immunopurified phot1. (D) A typical case ofa phosphorylation site mapping in phot1 by LC-MS/MS. Phot1 protein wasdigested with trypsin in gel. A MS/MS spectrum of a phosphorylated peptide408KSpSLSFMGIK417 is represented. pS indicates phosphorylated Ser. (E) Allidentified phosphorylation sites in phot1. Individual phosphopeptide se-quences and positions of Ser and Thr are indicated. Asterisks indicate thepeptides identified in error-tolerant search against phot1. Double asterisksindicate that a precise phosphorylation site was not determined because ofthe lack of informative fragment ions.

Inoue et al. PNAS � April 8, 2008 � vol. 105 � no. 14 � 5627

PLA

NT

BIO

LOG

Y

S851A-6 lost stomatal opening in response to blue light (Fig. 3A).These results indicate that Ser-851 is essential for phot1-mediated stomatal opening, with Ser-849 having some additionalrole, and that other phosphorylation sites are not required forstomatal opening.

We further substituted both Ser-849 and Ser-851 with Asp,which was expected to mimic phosphorylation (26). Transfor-mants of the double mutant with this construct restored thestomatal response by blue light (Fig. 3A, S849D S851D-2 andS849D S851D-3 lines). However, these substitutions did notaffect the stomatal aperture in the dark. These results suggestthat the substitution of Ser by Asp mimicked phosphorylation,and that the phosphorylation of the activation loop is notsufficient in inducing the stomatal opening and blue light isrequired for the opening. The kinase-dead phot1 protein withAsp substitutions did not restore stomatal opening by blue light(Fig. 3A, S849D S851D D806N-15).

Functional Role of Phosphorylation in Phot1 in H� Pumping. Blue lightactivates the plasma membrane H�-ATPase via phototropins,induces H� pumping in guard cells, and initiates stomatalopening with concomitant inhibition of anion channels (27–32).We isolated guard cell protoplasts of Arabidopsis and deter-mined blue light-induced H� pumping (Fig. 3B). Guard cell

Fig. 2. Autophosphorylation of the Ser-851 in the activation loop of phot1in vivo by blue light. (A) The phosphorylation of Ser-851. Phot1 was immu-nopurified from 200 �g of proteins of microsomal membranes as shown in Fig.1. The obtained phot1 was visualized with anti-pSer-851 antibodies (Upper)and immunodetected (Lower). (B) Fluence dependencies in phosphorylationof the Ser-851 in response to blue light. Etiolated seedlings were irradiated bya pulse of blue light for 30 sec at the indicated fluence rates. The seedlingswere disrupted 1 min after the start of blue light, and microsomal membraneswere immediately isolated from the seedlings. The membranes of 40 and 20�g of protein were used for the immunoblots of pSer-851 (Upper) and phot1(Lower), respectively. (C) Time courses of phosphorylation and dephosphor-ylation of the Ser-851 in response to blue light. The seedlings were irradiatedby blue light at 100 �mol m�2�s�1 for 30 sec and were disrupted at indicatedtimes after the start of blue light. Microsomal membranes were immediatelyisolated, and used for immunoblots of pSer-851 (Upper) and phot1 (Lower).Experiments repeated on three occasions gave similar results.

Fig. 3. Blue light-induced stomatal responses in transgenic plants. (A)Stomatal opening in the epidermis. Epidermal peels were irradiated by redlight (60 �mol m�2�s�1; RL) or red (50 �mol m�2�s�1) and blue (10 �mol m�2�s�1;RL � BL) light for 3 h. Values are means of three independent experimentswith standard deviations, with 45 stomata measured in each experiment. (B)Blue light-dependent H� pumping in guard cell protoplasts, determined by pHdecrease. Guard cell protoplasts (100 �g of proteins) were irradiated with redlight at 600 �mol m�2�s�1 and superimposed with blue light at 100 �molm�2�s�1 for 30 sec. The protoplasts were added by 10 �M fusicoccin. Measure-ments were done at 24°C. BL, blue light; FC, fusicoccin. Experiments repeatedon three occasions gave similar results. (C) Expression of the phot1 proteins inthe transgenic Arabidopsis plants. Immunoblot of the phot1 protein in gl1, thephot1 phot2 double mutant, and all of the transgenic plants. Immunoblot wasperformed by using 20 �g of microsomal proteins prepared from etiolatedseedlings. Experiments repeated on two occasions gave similar results.

Table 1. List of transgenic plants with various phot1 constructs

Construct Description

WT No mutationD806N (ref. 17) Kinase-dead: binding site of Mg2�-ATP in phot1 kinase is mutatedS350A S376A Binding sites of a 14-3-3 protein in phot1 are mutatedOthers Simultaneous substitutions of Ser-58, Ser-170, Ser-185, Ser-350, Ser-376, Ser-410, and Thr-993

with AlaS849A Substitution of the Ser-849 in the activation loop with AlaS851A Substitution of the Ser-851 in the activation loop with AlaS849A S851A Simultaneous substitutions of the Ser-849 and Ser-851 in the activation loop with AlaS849D S851D Simultaneous substitutions of the Ser-849 and Ser-851 in the activation loop with AspS849D S851D D806N Simultaneous substitutions of the Ser-849 and Ser-851 in the activation loop with Asp in the

kinase-dead construct of D806N

5628 � www.pnas.org�cgi�doi�10.1073�pnas.0709189105 Inoue et al.

protoplasts from the gl1 plants showed H� pumping in responseto blue light, and the protoplasts from the phot1 phot2 doublemutant lost H� pumping (data not shown; ref. 32). The proto-plasts exhibited complete H� pumping when the double mutantwas transformed with the wild-type construct (WT-11 line), butthe protoplasts from the D806N-2 transgenic line lost theresponse. Guard cell protoplasts from the S849A S851A-6 lineshowed a little pumping. Interestingly, the protoplasts from theS849D S851D-2 line showed much larger pumping than thosefrom the gl1 and WT-11 lines. Such large H� pumping isprobably brought about by a longer pumping period after bluelight irradiation. The pumping durations after the pulse were 7.6min for the WT-11 line and 19.1 min for the S849D S851D-2 line,and the maximum rates of H� pumping showed similar values inthese lines [supporting information (SI) Table S1]. Guard cell

protoplasts from gl1 and transgenic plants exhibited similar ratesof H� pumping by fusicoccin, a H�-ATPased activator, suggest-ing that pump activities were not suppressed in these transgeniclines.

The above results of H� pumping in the protoplasts are inaccord with the stomatal responses in the epidermis and intactleaves. The little pumping in the S849A S851A-6 line resulted inthe loss of blue light-specific stomatal opening (Fig. 3A). Incontrast, the sustained nature of H� pumping in the S849DS851D-2 line might result in larger stomatal conductance in theintact leaf of this line (Fig. S1). In this line, the stomatalconductance continued to increase even after pulse stimulation.

Functional Role of Phosphorylation in Other Phot1-Mediated Re-sponses. We measured other phot1-mediated responses. Bluelight-induced hypocotyl bending was abolished in the doublemutant and restored completely in the WT-11 line but not in thekinase-dead D806N-2 line (Fig. 4A). The S849A-2 line showedpartial phototropic bending, and the S851A-3 and S849AS851A-6 lines showed only slight bending. The S350A S376A-4and Others-4 lines showed almost complete bending.

Chloroplast accumulation was measured by local irradiation ofleaves with blue light. The gl1 leaves showed a green band bychloroplast accumulation, but the double mutant leaves did notshow such a band. WT-11 leaves showed a green band, butD806N-2 leaves did not (Fig. 4B). S851A-3 leaves displayed apale-green band because of a low level of chloroplast accumu-lation, and S849A S851A-6 leaves did not show a band. Incontrast, S350A S376A-4 and Others-4 leaves showed typicalgreen bands.

We inspected the leaf shapes of these transgenic plants grownunder white light. The gl1 leaves were flattened and suitable forcapturing light (5, 6), but the double-mutant leaves curleddownward (Fig. 4 Band C). The WT-11 leaves were flattened, butthe D806N-2 leaves were curled. The S851A-3 leaves werepartially f lattened, and the S849A S851A-6 line had no flattenedleaves at all. However, the leaves of S350A S376A-4 andOthers-4 lines were flattened.

The S849D S851D-3 line completely restored chloroplastaccumulation and leaf flattening and partially restored photot-ropism (Fig. 4).

From these results, we conclude that phosphorylation of theSer-851 in the activation loop is required for stomatal opening,phototropism, chloroplast accumulation, and leaf flattening, andthat of Ser-849 seems to have some role in these responses. Bycontrast, other phosphorylation sites, including the 14-3-3 pro-tein-binding sites, are not essential for these responses.

Phot1 Kinase Activities in Transgenic Plants. We determined bluelight-dependent autophosphorylation activity of the phot1s invivo in transgenic plants through the binding of a 14-3-3 proteinto these phot1s. The phosphorylation-dependent binding sitesfor a 14-3-3 protein were identified at both Ser-350 and Ser-376in the Hinge1 region of Arabidopsis phot1 (Fig. 5A, S350AS376A-4 line). The binding of 14-3-3 protein depended onautophosphorylation, because the binding was shown in theWT-11 line but not in the kinase-dead D806N-2 line (Fig. 5A).We found that the binding was evident in the phot1 protein ofthe S849A S851A-6 line. This in vivo autophosphorylation ofphot1 from the S849A S851A-6 line was confirmed by bothstaining of phos-tag, a specific indicator of phosphorylation (33)(Fig. 5B Upper), and mobility shift of the protein (Fig. 5B Lower).The results suggest that these substitutions did not significantlyinhibit autophosphorylation activity of phot1 in vivo.

DiscussionIdentification and Functional Analyses of Phosphorylation Sites inPhot1. We identified phosphorylation sites in vivo using phot1protein isolated from etiolated Arabidopsis seedlings by LC-

Fig. 4. Phot1-mediated responses in transgenic plants. (A) Phototropism.Etiolated seedlings were irradiated with unilateral blue light at 0.5 �molm�2�s�1 for 14 h. Values are means of 49–70 hypocotyls with standard errors.(B) Slit assays of chloroplast accumulation. Detached leaves were irradiatedwith blue light at 2.5 �mol m�2�s�1 for 30 min through a slit of 1-mm width.Arrowheads indicate the irradiated areas. Experiments repeated on twooccasions gave similar results. (C) Leaf flattening. The plants were grownunder white light at 50 �mol m�2�s�1 for 3 weeks. (Scale bar, 1 cm.) Experi-ments repeated on two occasions gave similar results.

Inoue et al. PNAS � April 8, 2008 � vol. 105 � no. 14 � 5629

PLA

NT

BIO

LOG

Y

MS/MS (Fig. 1). The sites contained two 14-3-3-binding sites ofSer-350 and Ser-376 (Fig. 5A), which had been identified usingthe recombinant phot1 fragments (data not shown). This obser-vation verifies the reliability of the LC-MS/MS method. Unfor-tunately, the LC-MS/MS method could not determine whetherSer-849, Ser-851, or both are autophosphorylated by blue light(Fig. 1E). To investigate this further, we generated antibodiesagainst pSer-851 and demonstrated that Ser-851 was autophos-phorylated in a fluence-dependent manner (Fig. 2). Unfortu-nately, we could not raise antibodies against pSer-849. However,the phenotypic analyses of phot1-mediated responses indicatedthe S849A S851A construct restored the phot1-mediated re-sponses less than the S851A construct (Figs. 3 and 4), suggestingSer-849 is likely an autophosphorylation site.

Salomon et al. (24) determined eight phosphorylation sites invitro in Avena phot1a. However, those sites did not include anyof the sites in the kinase domain or in the C terminus shown inthis study. The difference in results is probably due to thephosphorylation condition of the phot1 protein. We phosphor-ylated Arabidopsis phot1 in planta by an endogenous phot1kinase, but Salomon et al. (24) phosphorylated the recombinantAvena phot1a by PKA in vitro. Furthermore, the techniques usedto identify phosphorylation site differed between the two ex-periments. We may have missed phosphorylation sites that existin too-long or -short peptides produced by tryptic digestion,because these peptides were beyond the detection range ofLC-MS/MS.

We hypothesized that divergent blue light responses via phot1might be generated by various combinations of the phosphory-lation sites in phot1, which might transmit distinct signals to thedownstream components; the distinct constructs would restorethe different phot1-mediated responses. In support of thishypothesis, there are at least eight phosphorylation sites in Avenaphot1a that undergo autophosphorylation in response to differ-ent intensities of blue light (4, 24). However, the transgenic lines,except for S849A-2, S851A-3, and S849A S851A-6, restored allof the responses (Figs. 3 and 4). The transgenic S849A-2 lineshowed less impairment in these responses than the S851A-3line. From these results, we conclude that the autophosphory-lation of two Ser residues in the activation loop is required for

these phot1-mediated responses, and Ser-851 is more importantthan Ser-849. The other autophosphorylation sites may havesome other regulatory roles in phot1 functions.

Role of Phototropin Autophosphorylation in the Activation Loop. Theactivation loop is defined as the region spanning the conservedsequences DFG/D from kinase subdomain VII and APE fromsubdomain VIII. In general, phosphorylation in the loop bringsabout kinase activation both by the creation of the binding sitefor the substrate and the enhancement of catalytic activity(34–36). In the present study, the simultaneous substitution ofSer-849 and Ser-851 with Ala almost completely eliminated thesignaling (Figs. 3 and 4). The same substitution did not signif-icantly affect autophosphorylation activity determined in vivo(Fig. 5). From these results, we conclude that the blue light-specific autophosphorylation of the activation loop in phot1 maybe the mechanism underlying endogenous substrate recognition.Further study will be needed to elucidate this.

Regulation of Phototropin Signaling State. The S849D S851D con-struct mimicked the phosphorylation of the Ser residues in theactivation loop. However, the construct did not induce anyresponses without blue light in planta (Fig. 3 A and B). This isexplained by a recent report that LOV2 acts as a kinase inhibitorby binding to the kinase domain and that the kinase becomesable to phosphorylate the substrate by dissociation of LOV2 viablue light (18). In accord with those results, the phot2 kinasedomain, which is devoid of LOV domains, induces phot2-dependent responses without blue light (37). Therefore, it ismost likely that LOV2 prevents both activation loop autophos-phorylation and substrate transphosphorylation in the dark; theblue light-induced dissociation of LOV2 from the kinase domainallows the phosphorylation of the activation loop and thencreates the binding cleft for the endogenous substrates throughthe release of the loop from the cleft (34).

We suspected that the signaling would be sustained if thephosphorylation levels of the activation loop were maintained.The stomatal responses of the S849D S851D-2 line may reflectthe case where the substituted Asp in the activation loop cannotbe dephosphorylated. The negative charges of Asps in the loopwould prevent its refolding to the catalytic cleft and hinder therebinding of LOV2 to the kinase domain. In the S849D S851D-2line, H� pumping lasted longer (Fig. 3B), and greater stomatalconductance was found in the intact leaf (Fig. S1). In light ofthese results, dephosphorylation of the activation loop (Fig. 2C)would stimulate the termination of signaling, and some proteinphosphatase(s) may regulate this process. In accord with thisidea, dephosphorylation and H� pumping were completed insimilar periods; pSer-851 was dephosphorylated within 10 min,and H� pumping was sustained for 7.6 min (Figs. 2 C and 3B;Table S1).

The alignment of the activation loop in phototropin kinasesindicates that Ser-851 is conserved in both phot1 and phot2 ofseed plants, ferns, mosses, and green algae (Fig. S2). Thissuggests that autophosphorylation of the Ser residue may be acommon event for phototropin-mediated responses in theseplants.

Materials and MethodsPlant Materials and Growth Conditions. Plants of Arabidopsis thaliana gl1 as acontrol, the phot1-5 phot2-1 mutant, and all transformants were grown for3–5 weeks under 14-h fluorescent light (50 �mol m�2�sec�1)/10-h dark cycle. Allplants were grown at 24°C under a relative humidity of 55–75% in growthrooms. Etiolated seedlings were grown according to a previous method (17).

Isolation of Phot1 by Immunoprecipitation. Microsomal membranes wereprepared from 12,000 etiolated seedlings for MS analyses (25). Phot1 proteins

Fig. 5. Autophosphorylation kinase activity in vivo. Autophosphorylationactivity by phot1 kinase in vivo was determined through protein blot (A),staining of phosphorylation (B), and mobility shift (B). Etiolated seedlingswere kept in the dark (Dk) or irradiated with a blue light pulse at 100 �molm�2�s�1 for 1 min (BL), and microsomal membranes were immediately isolatedfrom the seedlings. (A) Phot1 in the membranes was separated by SDS/PAGEand transferred to nitrocellulose membrane. The nitrocellulose membranewas incubated with GST-GF14� to measure the phosphorylation-dependentbinding of a 14-3-3 protein to phot1 (Upper). Immunoblot of phot1 (Lower).(B) Phot1 was immunoprecipitated from 200 �g of the microsomal mem-branes. The immunopurified phot1 proteins were separated by SDS/PAGE andstained by phos-tag (Upper). Immunoblot of phot1 (Lower). Experimentsrepeated on three occasions gave similar results.

5630 � www.pnas.org�cgi�doi�10.1073�pnas.0709189105 Inoue et al.

were isolated by immunoprecipitation from the membranes using antibodiesagainst Arabidopsis phot1 (29).

Determination of in Vivo Phosphorylation Sites by Electrospray Ionization–LC-MS/MS. The phot1 proteins were excised from the gel and subjected to in-geldigestion with trypsin (Promega). The resulting peptides were dried, dissolvedin 50 mM Tris�HCl, 0.1 mM ZnCl2, and 1 mM MgCl2 (pH 8.0), and divided intotwo fractions. Two microliters of alkaline phosphatase-conjugated F7m (Mo-BiTech) was added to one of them, and both were incubated for 3 h at 37°C.These peptides were desalted and analyzed by an ion-trap mass spectrometer(LCQ-Deca, Finnigan) equipped with an HPLC system (Magic 2002, MichromBioResources). The collision-induced dissociation spectra acquired were com-pared with the National Center for Biotechnology Information nonredundantprotein database of the Mascot algorithm.

Determination of Phot1 Phosphorylation Levels in Etiolated Seedlings. Fivehundred etiolated seedlings were immersed in 2 ml of Mes buffer (5 mMMes-NaOH, 10 mM KCl, pH 5.7) containing [32P] orthophosphate at 0.25 mCiml�1 and kept for 14 h in the dark at room temperature with gentle agitation.After the removal of excessive 32P, the seedlings were illuminated with a pulseof blue light at 100 �mol m�2�s�1 for 1 min. Microsomal membranes wereprepared, and 200 �g of microsomal proteins was used for immunoprecipi-tation as described above. Autoradiography was performed as described (29).Phosphorylation was also measured by the method of phos-tag (PerkinElmer)according to the manufacturer’s instructions.

Immunoblotting and Protein Blotting. Immunoblotting and protein blottingwere performed according to previous methods (25). Polyclonal antibodies

against phosphorylated Ser-851 of Arabidopsis phot1 were raised in rabbit bys.c. injection of the phosphorylated synthetic peptide (RASNpSFVGTEEY,where pS represents phosphorylated Ser) as an antigen (Peptide Institute).Protein blotting was done by using 14-3-3 protein of Arabidopsis GF14�.

Construction of Transformation Vector. We constructed gene transfer vectorsbearing the genomic PHOT1 gene under the control of the native PHOT1promoter, as described in ref. 3. Detailed information is provided in SI Text.

Site-Directed Mutagenesis of Gene Transfer Vector. Single amino acid substi-tutions were made by using the QuikChange II Site-Directed Mutagenesis Kit(Stratagene) according to the manufacturer’s instructions. Nucleotide substi-tutions were introduced into the genomic PHOT1 fragment in pBluescript II KS(�) as templates for PCRs. The PCRs were conducted by using the oligonucle-otide primers described in SI Text. All constructs were sequenced to verifyspecific mutations. These mutated fragments of the PHOT1 gene were recon-stituted into the gene transfer vector as described above.

Phenotypic Analyses. Measurements of stomatal aperture, stomatal conduc-tance, blue light-induced H� pumping, phototropic curvature, chloroplastaccumulation, and leaf flattening are described in SI Text.

ACKNOWLEDGMENTS. This work was supported by grants from the Ministryof Education, Science, Sports, and Culture of Japan [nos. 17084005 and16207003 (to K.S.) and 14704003 (to T.K.)] and by a grant from the JapanSociety for Promotion of Science for Young Research Fellows (S.I.).

1. Briggs WR, Christie JM (2002) Phototropins 1 and 2: versatile plant blue-light receptors.Trends Plant Sci 7:204–210.

2. Sakai T, et al. (2001) Arabidopsis nph1 and npl1: Blue light receptors that mediate bothphototropism and chloroplast relocation. Proc Natl Acad Sci USA 98:6969–6974.

3. Sakamoto K, Briggs WR (2002) Cellular and subcellular localization of phototropin 1.Plant Cell 14:1723–1735.

4. Christie zJM (2007) Phototropin blue-light receptors. Annu Rev Plant Biol 58:21–45.5. Inoue S, Kinoshita T, Takemiya A, Doi M, Shimazaki K (2008) Leaf positioning of

Arabidopsis in response to blue light. Mol Plant 1:15–26.6. Takemiya A, Inoue S, Doi M, Kinoshita T, Shimazaki K (2005) Phototropins promote

plant growth in response to blue light in low light environments. Plant Cell 17:1120–1127.

7. Kasahara M, et al. (2002) Chloroplast avoidance movement reduces photodamage inplants. Nature 420:829–832.

8. Huala E, et al. (1997) Arabidopsis NPH1: A protein kinase with a putative redox-sensingdomain. Science 278:2120–2123.

9. Christie JM, et al. (1998) Arabidopsis NPH1: A flavoprotein with the properties of aphotoreceptor for phototropism. Science 282:1698–1701.

10. Christie JM, Salomon M, Nozue K, Wada M, Briggs WR (1999) LOV (light, oxygen,voltage) domains of the blue-light photoreceptor phototropin (nph1): Binding sites forthe chromophore flavin mononucleotide. Proc Natl Acad Sci USA 96:8779–8783.

11. Salomon M, Christie JM, Knieb E, Lempert U, Briggs WR (2000) Photochemical andmutational analysis of the FMN binding domains of the plant blue light receptor,phototropin. Biochemistry 39:9401–9410.

12. Salomon M, et al. (2001) An optomechanical transducer in the blue light receptorphototropin from Avena sativa. Proc Natl Acad Sci USA 98:12357–12361.

13. Swartz TE, et al. (2001) The photocycle of a flavin-binding domain of the blue lightphotoreceptor phototropin. J Biol Chem 276:36493–36500.

14. Swartz TE, Wenzel PJ, Corchnoy SB, Briggs WR, Bogomolni RA (2002) Vibration spec-troscopy reveals light-induced chromophore and protein structural changes in theLOV2 domain of the plant blue-light receptor phototropin 1. Biochemistry 41:7183–7189.

15. Nozaki D, et al. (2004) Role of Gln1029 in the photoactivation processes of the LOV2domain in Adiantum phytochrome3. Biochemistry 43:8373–8379.

16. Harper SM, Christie JM, Gardner KH (2004) Disruption of the LOV-J� helix interactionactivates phototropin kinase activity. Biochemistry 43:16184–16192.

17. Christie JM, Swartz TE, Bogomolni RA, Briggs WR (2002) Phototropin LOV domainsexhibit distinct roles in regulating photoreceptor function. Plant J 32:205–219.

18. Matsuoka D, Tokutomi S (2005) Blue light-regulated molecular switch of Ser/Thr kinasein phototropin. Proc Natl Acad Sci USA 102:13337–13342.

19. Gallagher S, Short TW, Ray PM, Pratt LH, Briggs WR (1988) Light-mediated changes intwo proteins found associated with the plasma membrane fractions from pea stemsections. Proc Natl Acad Sci USA 85:8003–8007.

20. Short TW, Briggs WR (1990) Characterization of a rapid, blue light-mediated change indetectable phosphorylation of a plasma membrane protein from etiolated pea (Pisumsativum L.) seedlings. Plant Physiol 92:179–185.

21. Reymond P, Short TW, Briggs WR, Poff KL (1992) Light-induced phosphorylation of amembrane protein plays an early role in signal transduction for phototropism inArabidopsis thaliana. Proc Natl Acad Sci USA 89:4718–4721.

22. Palmer JM, Short TW, Briggs WR (1993) Correlation of blue light-induced phosphory-lation to phototropism in Zea mays L. Plant Physiol 102:1219–1225.

23. Salomon M, Zacherl M, Rudiger W (1997) Asymmetric, blue light-dependent phos-phorylation of a 116-kilodalton plasma membrane protein can be correlated with thefirst- and second-positive phototropic curvature of oat coleoptiles. Plant Physiol115:485–491

24. Salomon M, Knieb E, von Zeppelin T, Rudiger W (2003) Mapping of low- and high-fluence autophosphorylation sites in phototropin 1. Biochemistry 42:4217–4225.

25. Kinoshita T, et al. (2003) Blue-light- and phosphorylation-dependent binding of a14–3-3 protein to phototropins in stomatal guard cell of broad bean. Plant Physiol133:1453–1463.

26. Guo Y, Halfter U, Ishitani M, Zhu J-K (2001) Molecular characterization of functionaldomains in the protein kinase SOS2 that Is required for plant salt tolerance. Plant Cell13:1383–1399.

27. Assmann SM, Simoncini L, Schroeder JI (1985) Blue light activates electrogenic ionpumping in guard cell protoplasts of Vicia faba L. Nature 318:285–287.

28. Shimazaki K, Iino M, Zeiger E (1986) Blue light-dependent proton extrusion by guardcell protoplasts of Vicia faba. Nature 319:324–326.

29. Kinoshita T, Shimazaki K (1999) Blue light activates the plasma membrane H�-ATPaseby phosphorylation of the C-terminus in stomatal guard cells. EMBO J 18:5548–5558.

30. Kinoshita T, et al. (2001) phot1 and phot2 mediate blue light regulation of stomatalopening. Nature 414:656–660.

31. Marten H, Hedrich R, Roelfsema MRG (2007) Blue light inhibits guard cell plasmamembrane anion channels in a phototropin-dependent manner. Plant J 50:29–39.

32. Ueno K, Kinoshita T, Inoue S, Emi T, Shimazaki K (2005) Biochemical characterizationof plasma membrane H�-ATase activation in guard cell protoplasts of Arabidopsisthaliana in response to blue light. Plant Cell Physiol 46:955–963.

33. Kinoshita E, Kinoshita-Kikuta E, Takiyama K, Koike T (2006) Phosphate-binding tag, anew tool to visualize phosphorylated proteins. Mol Cell Proteom 5:749–757.

34. Hanks SK, Hunter T (1995) Protein kinases 6. The eukaryotic protein kinase superfamily:kinase (catalytic) domain structure and classification. FASEB J 9:576–596.

35. Huse M, Kuriyan J (2002) The conformational plasticity of protein kinases. Cell 109:275–282.

36. Adams JA (2003) Activation loop phosphorylation and catalysis in protein kinases: Isthere functional evidence for the autoinhibitor model? Biochemistry 38:601–607.

37. Kong S-G, et al. (2007) The C-terminal kinase fragment of Arabidopsis phototropin 2triggers constitutive phototropin responses. Plant J 51:862–873.

Inoue et al. PNAS � April 8, 2008 � vol. 105 � no. 14 � 5631

PLA

NT

BIO

LOG

Y