Issue topic: Blue-Light Response PCP-2012-E-00464 (revised)

Blue Light-Induced Conformational Changes in aLight-Regulated Transcription Factor, Aureochrome-1Osamu Hisatomi1,*, Ken Takeuchi1, Kazunori Zikihara2, Yuki Ookubo2, Yoichi Nakatani1,Fumio Takahashi3, Satoru Tokutomi2 and Hironao Kataoka4

1Earth and Space Science, Graduate School of Science, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka, 560-0043 Japan2Department of Biological Science, Graduate School of Science, Osaka Prefecture University, Gakuen-cho 1-1, Naka-ku Sakai, Osaka,599-8531 Japan3PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama, 332-0012 Japan4Botanical Gardens, Tohoku University, 12-2 Kawauchi, Aoba-ku, Sendai, 980-0862 Japan*Corresponding author: E-mail, [email protected]; Fax, +81-6-6850-5500.(Received November 1, 2012; Accepted November 21, 2012)

Aureochrome-1 (AUREO1) is a blue light (BL) receptor thatmediates the branching response in the stramenopile alga,Vaucheria frigida. AUREO1 harbors a basic leucine zipper(bZIP) domain at the N-terminus and a light–oxygen–voltage-sensing (LOV) domain within the C-terminalregion, and has been suggested to function as a light-regulated transcription factor. To understand the molecularmechanism of AUREO1, we have prepared three recombin-ant proteins: a full-length AUREO1 (FL), an N-terminaltruncated construct containing bZIP and LOV (ZL) and aLOV-only (LOV) construct. The constructs showed thesame absorption and fluorescent spectra in the dark stateand underwent the characteristic cyclic reaction as previ-ously observed in LOV domains upon BL excitation. FL andZL bound to DNA in a sequence-specific manner. BL ap-peared to induce a shift of the a-helical structure of theLOV domain to a b-sheet structure, but did not alter thehydrodynamic radius (RH) of this domain. ZL formed a dimerpossibly through disulfide linkages in the bZIP and the linkerregion between bZIP and LOV. BL induced an approximately5% increase in the RH of ZL, although its secondary structurewas unchanged. These results support a schema whereBL-induced changes in the LOV domain may cause conform-ational changes in the bZIP and/or the linker of a dimeric ZLmolecule. Since a 5% increase of the RH was also observedwith the FL construct, BL may induce global conformationalchanges similar to those observed for ZL, and formation ofthe FL dimer may facilitate DNA binding.

Keywords: Aureochrome � Blue light � bZIP �

Conformational change � LOV � Vaucheria frigida.

Abbreviations: AUREO, aureochrome; AUREO1,aureochrome-1; AUREO2, aureochrome-2; BL, blue light;bZIP, basic region/leucine zipper; CBB, Coomassie BrilliantBlue; CD, circular dichroism; D450, LOV in the dark state;DLS, dynamic light scattering; DTT, dithiothreitol; EMSA,

electrophoretic mobility shift assay; HK, histidine kinase;LED, light-emitting diode; LOV, light–oxygen–voltage-sensing;PBS, phosphate-buffered saline; phot, phototropin; phot1,phototropin1; phot2, phototropin2; RH, hydrodynamicradius; RH-app, apparent hydrodynamic radius; SEC, size exclu-sion chromatography; SAXS, small angle X-ray scattering;S390, LOV in the adduct state; �max, maximum absorptionwavelength; t1/2, half-lifetimes.

Introduction

Light plays an essential role for living organisms. For plants, lightplays an important role in the regulation of a variety of devel-opmental and physiological processes such as seed germin-ation, greening, flowering or tropic responses. Plants haveevolved two major types of photoreceptor proteins that con-vert the light stimuli into biological signals. They are phyto-chrome and blue light (BL) receptors. Phototropin (phot) isone of the most well studied BL receptor proteins in plants(Christie 2007). Phot mediates phototropism (Christie et al.1998), chloroplast movement (Kagawa et al. 2001), stomatalopening (Kinoshita et al. 2001) and leaf photomorphogenesis(Kozuka et al. 2011), and acts to maximize the efficiency ofphotosynthesis. Phot has two light–oxygen–voltage-sensing(LOV) domains (Taylor and Zhulin 1999), a subset of the PAS(Per-ARNT-Sim) superfamily, that act as the light sensorydomain and are located in the N-terminal region, and a Ser/Thr protein kinase domain at the C-terminus. Phot is thoughtto be a light-regulated protein kinase that transduces signaldownstream through autophosphorylation events (Christieet al. 2002) and/or substrate phosphorylation (Christie et al.2011, Okajima et al. 2011). Besides the phot-type receptors withtwo LOV domains and one kinase domain, there are manyproteins that harbor one LOV domain and a variety of signaloutput modules (Crosson et al. 2003). Such proteins include the

Plant Cell Physiol. 54(1): 93–106 (2013) doi:10.1093/pcp/pcs160, available online at www.pcp.oxfordjournals.org! The Author 2012. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists.All rights reserved. For permissions, please email: [email protected]

93Plant Cell Physiol. 54(1): 93–106 (2013) doi:10.1093/pcp/pcs160 ! The Author 2012.

Special

Focu

sIssu

e–

Regu

larPap

er

Downloaded from https://academic.oup.com/pcp/article-abstract/54/1/93/1903926by gueston 12 April 2018

LOV-F box-kelch repeat protein families such as ZTL, LKP2 andFKF1 (Imaizumi et al. 2003), YtvA, a LOV-STAS (sulfate trans-porter and anti-sigma antagonist) protein, of Bacillus subtilis(Avila-Perez et al. 2009) or LOV-histidine kinase (HK) proteinsfrom Brucella melitensis, Brucella abortus, Pseudomonas syrin-gae (Swartz et al. 2007), Caulobacter crescentus (Purcell et al.2007) and picoalga Ostreococcus (Djouani-Tahr et al. 2011).Furthermore, LOV proteins with a DNA-binding domain havebeen found in Vaucheria frigida (Takahashi et al. 2007),Neurospora (Malzahn et al. 2010) and the marine bacteriumErythrobacter litoralis (Nash et al. 2011).

The crystal structures of LOV domains resolved so far arecomposed of an assembly of short a-helices, an a-helix (helicalconnector) and a five-stranded antiparallel b-sheet (b-scaffold)(Crosson and Moffat 2001, Crosson and Moffat 2002, Fedorovet al. 2003, Harper et al. 2004, Halavaty and Moffat 2007,Moglich and Moffat 2007, Nash et al. 2011, Mitra et al. 2012,Rinaldi et al. 2012). All the LOV domains bind an FMNnon-covalently as a chromophore. The FMN in the LOV do-mains shows a unique cyclic photoreaction. Upon BL excitation,FMN in the dark state (D450) forms a cysteinyl adduct with ahighly conserved cysteine residue (adduct state, S390) via inter-system crossing to the triplet excited state (Salomon et al. 2000,Swartz et al. 2001, Iwata et al. 2002). S390 reverts to D450 withcharacteristic time constants, and the types of LOV domainsreflect their functions (Kasahara et al. 2002, Zikihara et al. 2006).The photochemical reactions induce conformational changesin the protein moiety that are thought to regulate the activitiesof the various signal output domains.

Among the reported LOV proteins with a DNA-bindingmodule, aureochrome (AUREO) is the only one known inplants. AUREO was first identified in a stramenopile alga,V. frigida (Takahashi et al. 2007). AUREO has two homologsnamed aureochrome-1 (AUREO1) and aureochrome-2(AUREO2). RNA interference experiments suggested thatAUREO1 and AUREO2 play roles in BL-induced branch forma-tion and the development of a sex organ, respectively(Takahashi et al. 2007). Both AUREO1 and AUREO2 have aLOV domain located in the C-terminal region and a basicregion/leucine zipper (bZIP) domain on the N-terminal side

of the LOV domain (Fig. 1). bZIP is an a-helical DNA-bindingmotif found commonly among the eukaryotic transcriptionfactors (Ellenberger et al. 1992). Recombinant AUREO1 showsthe characteristic triplet absorption peaks with the LOV do-mains at around 450 nm and was reported to bind the targetDNA sequence TGACGT in a light-dependent manner. It wastherefore suggested that AUREOs function as BL-regulatedtranscription factors (Takahashi et al. 2007). AUREO orthologshave been found only in photosynthetic stramenopiles, e.g.brown algae, some diatoms, Ochromonas, Chattonella andAureococcus (Ishikawa et al. 2009).

The LOV domain of AUREO1 also shows a similar photo-cycle when compared with other LOV proteins (Takahashi et al.2007). Interestingly, however, the effector domain of AUREO1and AUREO2 resides at the N-terminus and contrasts with thecommon arrangement of the other LOV proteins with theireffector domains located on the C-terminal side. This suggeststhat the AUREOs have a different molecular mechanism fromthat of other LOV-containing proteins with respect to the regu-lation of the effector domain. To gain insight into the molecularbasis for this different domain arrangement, we previously stu-died the global molecular changes of truncated AUREO1 by thetransient grating technique, which detects a change in the dif-fusion constant of molecules (Toyooka et al. 2011). TruncatedAUREO1 proteins used in the study were ZL and LOV (Fig. 1).ZL contained the bZIP and LOV domains, whereas LOV onlycontained the LOV domain. Upon BL excitation, both the ZLand the LOV constructs showed S390 formation with a timeconstant of 2.8 ms and subsequent changes in the diffusionconstant with time constants of 140 and 160 ms, respectively.However, the suggested origins for these diffusion constantchanges differ between the two molecules. We concludedthat the ZL construct existed in a dimeric form in both thedark and light states, and the diffusion constant change wasattributed to BL-induced conformational changes in the di-meric molecule. In contrast, for the LOV domain construct,an equilibrium between the monomer and the dimer speciesexisted and this was dependent on the concentration, and thediffusion constant change was suggested to come from aBL-induced transition from a monomer to a dimer. Thus, the

Fig. 1 Schematic drawing for the constructs of AUREO1 recombinant proteins. See the Results for details about the constructs.

94 Plant Cell Physiol. 54(1): 93–106 (2013) doi:10.1093/pcp/pcs160 ! The Author 2012.

O. Hisatomi et al.


role of the LOV domain in the conformational change of ZLremains unresolved.

In the present study, we have prepared a full-lengthAUREO1 (Fig. 1, FL) as well as the ZL and the LOV constructs,and investigated the global conformational changes in FL andthe role of the LOV in these changes using dynamic light scat-tering (DLS) analysis, size exclusion chromatography (SEC) andcircular dichroism (CD) spectropolarimetry. Furthermore, theinvolvement of the disulfide bond in the conformationalchange was studied by SDS–PAGE of the samples preparedunder non-reducing conditions, because this bond was sug-gested to be involved in the dimerization of the LOV1 inArabidopsis phototropin1 (phot1) (Nakasako et al. 2008). Thepresent results provide useful information concerning the con-formational changes induced in each domain of AUREO1 andtheir organization in the full-length protein. A schema for thesechanges is presented and discussed in connection with the lightregulation of DNA binding.

Results

Six AUREO1 recombinant proteins and theirbinding to DNA

We prepared six different recombinant AUREO1 proteins(Fig. 1): the full-length protein with a single deletion of Asn2(FL) and truncated proteins composed of the bZIP domain–linker region–LOV domain (G113–K348, ZL) and the LOVdomain only (P204–K348, LOV). All the recombinant proteinshave a tag containing a histidine hexamer LEHHHHHH orMHHHHHHS at either the C- (denoted as FLh, ZLh andLOVh) or the N-terminus (hFL, hZL and hLOV).

Fig. 2A shows the SDS–polyacrylamide gel patterns of pur-ified AUREO1 recombinant proteins denatured under reducing

conditions, i.e. 20 mM dithiothreitol (DTT). The recombinantproteins showed major bands with mol. wts of 43 kDa (FLh andhFL), 30 and 29 kDa (ZLh and hZL), and 16 and 17 kDa (hLOVand hLOV). These molecular weights match the monomericmolecular weights calculated from the amino acid sequences:39.6 and 39.5 kDa (FLh and hFL), 27.9 kDa (ZLh and hZL) and17.6 kDa (hLOV and hLOV). Fig. 2B shows the result of theelectrophoretic mobility shift assay (EMSA) to examine thebinding of the recombinant proteins to the AUREO1-specificoligonucleotide, Ap-oligo (Takahashi et al. 2007) and a controloligonucleotide, Cp-oliogo, under room light conditions.Band-shifts were found with the recombinant proteins harbor-ing the bZIP domain, i.e. FLh, hFL, ZLh and hZL (lanes 1, 3, 5 and7, respectively), that were not observed with the Cp-oligo (lanes2, 4, 6 and 8, respectively), indicating that they bound to DNA ina sequence-specific manner. On the other hands, LOVh andhLOV without the bZIP domain did not bind either theAp-oligo or the Cp-oligo (lanes 9 and 10). Thus, the bZIPdomain in the AUREO1 recombinant proteins functions to rec-ognize and bind to a specific DNA sequence.

Absorption and fluorescent emission spectra

AUREO1 recombinant proteins showed typical absorptionspectra of flavoproteins binding flavin in an oxidized form.They have absorption maxima (�max) at 447 ± 1 nm and anobvious triplet vibrational structure (Fig. 3A). The shapes ofthe minor doublet peaks at around 370 nm resemble thoseof LOV2 (Kasahara et al. 2002). The absorbance ratio A280/A447 of ZL and LOV was 3–4, whereas that of the FL constructwas larger and appeared to be approximately 5. This is probablybecause of the increase in the number of aromatic residues inFL, since both the LOV and the ZL constructs contain onetryptophan and six tyrosine residues, whereas the FL constructhas an additional tyrosine residue within the N-terminal region

A B

Fig. 2 SDS–PAGE and EMSA of AUREO1 recombinant proteins. (A) CBB-stained SDS–polyacrylamide gel of FLh (lane 1), hFL (lane 2), ZLh(lane 3), hZL (lane 4), LOVh (lane 5) and hLOV (lane 6) samples prepared under reducing (see Materials and Methods) conditions. (B) EMSA ofFLh, hFL, ZLh and hZL in the presence of Ap-oligo (lanes 1, 3, 5 and 7) or Cp-oligo (lanes 2, 4, 6 and 8), and LOVh and hLOV in the presence ofAp-oligo (lanes 9 and 10), respectively.


Light-induced structural change in aureochrome-1


(N-domain, corresponding to M1–S112 of AUREO1 in Fig. 1).Fig. 3B shows the fluorescence emission spectra measured withan excitation at 450 nm. All the spectra showed an emissionmaximum at 495 ± 1 nm with a shoulder at around 535 nm, andwere almost superimposable. Since the AUREO1 recombinantproteins have near identical absorption and fluorescence emis-sion spectra, the histidine tag, the N- and the bZIP domainshave a negligible effect on the p electron conjugation system ofthe flavin chromophore in the recombinant proteins.

BL-induced UV-visible spectral changes

Fig. 4A shows the absorption spectral change of ZLh inducedby BL illumination. The triplet absorption peak at 447 nm dis-appeared and a new peak appeared at approximately 390 nmthat was reverted to the initial absorption in the dark within48 min at 25�C (Fig. 4B). These spectral changes are character-istic of a BL-induced transient formation of a cysteinyl-flavinadduct (S390) of the oxidized flavin in the dark state (D450) inthe LOV proteins (Salomon et al. 2000, Swartz et al. 2001, Iwataet al. 2002). Similar spectral changes were observed with theother AUREO1 recombinant proteins (Supplementary Fig.S1), indicating that all the recombinant proteins used in thisstudy undergo a characteristic cyclic photoreaction with that ofthe known LOV proteins.

Dark regeneration of D450 from S390

Dark regeneration of D450 from S390 was measured with ZLh(Fig. 4) and the other AUREO1 recombinant proteins(Supplementary Fig. S1) at 25�C, pH 7.0. The time course ofthe regeneration reaction of ZLh was represented by an absorp-tion change at 450 nm and was well fitted by a single exponen-tial curve, indicating that the reaction was a first-order reactionof a single component (Fig. 4C). The measurement was re-peated several times, and the average calculated half-lifetime(t1/2) was 7.2 ± 0.4 min. This value is slightly longer than thepreviously reported value of 4.9 min at 25�C, pH 8.0 (Takahashiet al. 2007), probably due to the pH effect on the dark regen-eration (Guo et al. 2005). The dark regeneration was measured

at five different temperatures, and the t1/2 and the rate con-stant were calculated. From the Arrhenius plot of the rate con-stants (Fig. 4D), the activation energy was calculated as110 kJ mol�1. Similar analyses were performed with the datarecorded on the other recombinant proteins (t1/2 of theAUREO1 recombinant proteins ranged from 6.5 to 7.4 min at25�C), and the lifetime, the calculated activation energies andthe logarithm of frequency factors are summarized in Table 1.

SDS–PAGE of the samples prepared undernon-reducing condition

The LOV1 domain has been reported to act as a dimerizationsite in oat phot1 (Salomon et al. 2004), and Arabidopsis phot1and phototropin2 (phot2) (Katsura et al. 2009). Furthermore,the LOV1 of Arabidopsis phot1 has been shown to dimerize viathe formation of a disulfide bond (Nakasako et al. 2008).Therefore, the involvement of a disulfide bond in the oligomericstructure of the recombinant proteins was studied by SDS–PAGE of the samples prepared under non-reducing conditions(see the Materials and Methods) (Williams et al. 1991) (Fig. 5A).Bands of LOVh and hLOV were detected at a mol. wt of 17 kDa(lanes 5 and 6, respectively). This molecular weight is essentiallythe same as the molecular weight of the same samples preparedunder reducing conditions, suggesting no involvement of disul-fide linkage in the oligomeric structure of the LOV domain. Incontrast, the main bands of FLh and hFL were observed at mol.wts >97 kDa. ZLh and hZL gave rise to bands in the gel withmol. wts of about 58 and 62 kDa (lanes 3 and 4, respectively).The molecular weights are about twice as large as that of thesample prepared under reducing conditions. These results sug-gest that FL and ZL form a homodimer through an intermo-lecular disulfide linkage(s).

AUREO1 has six cysteine residues, two cysteines (C254 andC283) of which reside in the LOV domain, and C162 and C182locate in the bZIP and linker region, respectively (Fig. 1 upper).Since involvement of disulfide linkage in the dimerization ofLOV was negated, the contribution of the two cysteines in ZLhto the dimerization was studied by substituting them with

A B

Fig. 3 UV-visible absorption (A) and fluorescence emission (B) spectra of the recombinant AUREO1 protein over the concentration range of5–15 mM in 400 mM NaCl, 1 mM DTT, 20 mM Tris–HCl (pH 7.0) at 25�C. Spectra of FL, ZL and LOV are shown by solid, dashed and dotted lines,respectively. Pre-h and post-h are discriminated by black and gray shading, respectively. The UV/Vis absorption spectra and the fluorescentemission spectra are shown by normalizing both the maximal absorbance at 447 nm and the maximal emission at 495 nm as 1.0. Inset indicatesthe enlarged absorption spectra from 350 to 550 nm.


O. Hisatomi et al.


http://pcp.oxfordjournals.org/lookup/suppl/doi:10.1093/pcp/pcs160/-/DC1



serine (Fig. 5B). All the intact and cysteine/serine-substitutedZLh appeared as a single band with a mol. wt of about 30 kDaunder reducing conditions that can be assigned as the mono-mer band judging from the molecular weight. Undernon-reducing conditions, in contrast, dimer bands with a

mol. wt of about 58–65 kDa were detected as well as the mono-mer band with C162S and C182S single substitutes (lane 3 and5), whereas no dimer band was observed with the C162S/C182Sdouble substitute (lane 7). These results suggest that bothCys162 and Cys182 contribute redundantly to form a dimerthrough disulfide linkage.

Size exclusion chromatography

The oligomeric structure of LOVh was studied by SEC at 5�Cunder dark and light conditions. The elution profiles are shownin Fig. 6A. The molecular weight of LOVh in the dark state wasestimated as 24.6 kDa (filled triangle in Fig. 6B), which is largerthan the 17.6 kDa calculated for a monomeric LOVh. Under BLillumination, the elution peak shifted slightly to a higher mo-lecular weight (open triangle in Fig. 6B, 28.6 Da), indicatingBL-induced conformational changes to the LOVh construct.The eluted LOVh solution kept the absorption spectrum ofS390 and showed little dark regeneration of D450 due to aprolonged t1/2 of 94 min at the low temperature.

A

C D

B

Fig. 4 Dark regeneration of D450 from S390 of ZLh in a 400 mM NaCl, 1 mM DTT, 20 mM Tris–HCl (pH 7.0) solution at 25�C. (A) UV-visibleabsorption spectra in the dark state (D450; solid line), and immediately (S390; dashed line) and 48 min (dotted line) after the termination of BLillumination for 1 min. (B) Spectral changes during the dark regeneration in the wavelength region from 350 to 550 nm. The spectra weremeasured at the initial dark state (thin solid line), and immediately (thick dashed line), 2, 4, 8, 16, 32 (thin solid lines) and 48 min (thick dottedline) after the termination of the BL illumination. (C) Time-course of the dark regeneration of D450 from S390 at 15, 20, 25, 30 and 35�C. Graycircles are the absorption differences at 450 nm normalized by maximum absorption differences, �Amax. The solid lines are fitting curves basedon a single-exponential approximation. (D) The Arrhenius plot for the rate constants of the regeneration reaction in ZLh measured at 15, 20, 25,30 and 35�C.

Table 1 Half-lifetimes and Arrhenius parameters for the darkregeneration reaction from S390 to D450

q1/2 (min) lnA Ea (kJ mol�1)

FLh 7.4 ± 0.2 38 ± 1 110 ± 1

hFL 7.1 ± 0.2 38 ± 1 109 ± 1

ZLh 7.2 ± 0.4 38 ± 1 109 ± 2

hZL 6.9 ± 0.2 37 ± 2 107 ± 4

LOVh 6.6 ± 0.2 41 ± 1 116 ± 2

hLOV 6.5 ± 0.4 40 ± 1 113 ± 2

The average half-lifetimes (t1/2) were calculated from the rate constants (k) ofthe dark regeneration reaction from S390 to D450, and frequency factors (A)and activation energy (Ea) were estimated from the Arrhenius plots with stand-ard deviations (n� 3), an example of which is shown in Fig. 4C and D.




Dynamic light scattering

DLS of the AUREO1 recombinant protein solutions wasmeasured to obtain their Stokes hydrodynamic radius, RH,from the apparent hydrodynamic radius, RH-app, in dark, lightand light–dark conditions (see the Materials and Methods).Supplementary Fig. S2 shows examples of the size distributioncurves indicating good polydispersities. RH-app of the recombin-ant proteins depends quasi-linearly on the concentration(Fig. 7), as expected from the relationship in dilute solutions,RH-app� RH (1 + yC) where C is the concentration and the y is aconstant. RH determined by extrapolating the RH-app to a pro-tein concentration of zero is summarized in Table 2. FLh has anRH of 4.2 nm in the dark state that increased slightly to 4.4 nmafter the BL illumination. The increase was reversible in thedark. This observation corresponds to an increase of 15% inthe hydrodynamic volume of a spherical molecule. Likewise,the BL induced an increase in RH of ZLh from 3.6 to 3.8 nm.This observation corresponds to an increase in hydrodynamic

volume by 18%. In contrast, LOVh showed no RH change uponBL excitation. The molecular weights of the LOVh, ZLh and FLhwere calculated from their RH using the calibration curve ob-tained from spherical protein molecules (data not shown), andwere 35, 81 and 124 kDa, respectively.

CD spectra

The BL-induced secondary structural change of ZLh and LOVh

was studied by CD spectra measurements (Fig. 8) and their

component analyses (Table 3). ZLh in the dark state exhibited

a CD spectrum indicating that the protein is rich in a-helices

with only a very small change after BL illumination (Fig. 8A).

This is consistent with previous results (Toyooka et al. 2011).

On the other hand, LOVh showed a significant change between

the dark and the light states (Fig. 8B), in which the a-helix

content decreased by 7% and the b-sheet content increased

by the same percentage (Table 3).

A B

Fig. 5 SDS–PAGE of the recombinant AUREO1 protein samples including cysteine/serine substitutes. (A) Lanes 1–6 indicate the CBB-stainedbands of FLh, hFL, ZLh, hZL, LOVh and hLOV, respectively, prepared under non-reducing conditions (see the Materials and Methods).(B) CBB-stained bands of ZLh (lane 1 and 2), ZLh-C162S (lane 3 and 4) and ZLh-C182S (lane 5 and 6) substitutes, and the ZLh-C162S/C182Sdouble substitute (lane 7 and 8) prepared under non-reducing (lane 1, 3, 5 and 7) and reducing (lane 2, 4, 6 and 8) conditions, respectively.

1.2

1

0.8

0.6

0.4

0.2

040 60 80 100 120 140 160

A B

Fig. 6 (A) Elution profiles of SEC monitored by the absorption at 280 nm. LOVh in the dark state (black line) and under BL illumination(gray line). (B) Molecular weight estimation of LOVh in the dark state (filled triangle) and the light-activated state (open triangle), respectively,based on the SEC data. Open circles indicate the molecular weights of the proteins used for column calibration. The concentrations of theapplied and the eluted samples were 50 and 5 mM, respectively.


O. Hisatomi et al.



Discussion

Photoreaction of the AUREO1 recombinantproteins

All the AUREO1 recombinant proteins showed similar cyclicphotoreactions between the oxidized FMN, D450 and thecysteinyl-flavin adduct, S390. Although the t1/2 (2.8 ms) of theadduct formation reaction of LOVh and ZLh (Toyooka et al.2011) is comparable with the 4ms observed for oat phot1-LOV2(Swartz et al. 2001), the present t1/2 of their dark regenerationreaction from S390 to D450 (6.5–7.4 min) is >10-fold longerthan that of the oat phot1-LOV2 (Corchnoy et al. 2003, Harperet al. 2004) or of the plant phot-LOV (Kasahara et al. 2002). Onthe other hand, the t1/2 values of the LOV proteins with oneLOV domain vary and are dependent on the types of C-terminalfunctional modules (Losi et al. 2003, Swartz et al., 2007,Raffelberg et al. 2011, Zoltowsky et al. 2011). Thus, the involve-ment of the additional domains in the regeneration rate ofthe LOV domains is highly influential. Moreover, the activa-tion energy (Ea) of the dark regeneration reaction variesamong the LOV proteins from 55 to 100 kJ mol�1 (Corchnoyet al. 2003, Losi et al. 2003, Harper et al. 2004, Raffelberget al. 2011). Furthermore, this Ea value was found to varybecause of the type and position of an amino acidsubstitution (Raffelberg et al. 2011). These results suggest thatthe main factors regulating the dark regeneration may bemolecular structures in a LOV domain, and that the typeof function of the other domains may modulate the activityof the LOV domains. The addition of bZIP or bZIP plusthe N-terminal domains only slightly affected the t1/2 andthe Ea values, indicating that these domains may not inter-act strongly with the LOV domain to modify its darkregeneration.

Oligomeric structure of the AUREO1 recombinantproteins

SEC showed an apparent molecular weight of the AUREO1LOVh as 24.6 kDa. This value is 1.4 times larger than the calcu-lated molecular weight (17.5 kDa) based on the amino acidsequence. We conclude that the LOV domain at these

A

B

C

Fig. 7 Dependence of RH-app on the concentrations of FLh (A), ZLh(B) and LOVh (C) in a 400 mM NaCl and 20 mM Tris–HCl (pH 7.0)solution at 25�C. The regression lines for the ‘dark’, ‘light’ and ‘dark–light’ states (see the Materials and Methods) are shown with thinblack, gray and thick dashed lines, respectively, with SDs that wereobtained from >6 independent measurements.

15

10

5

0

-5

20

10

0

-10

190 200 210 220 230 240 190 200 210 220 230 240

A B

Fig. 8 CD spectra of ZLh (A) and LOVh (B) in the dark state (filled circles) and the light-activated state (open circles). Thin solid lines indicate thefitting curves calculated by the CDSSTR method.

Table 2 RH values estimated from DLS analyses

RH (nm)

Dark Light Light–Dark

FLh 4.2 4.4 4.2

ZLh 3.6 3.8 3.6

LOVh 2.7 2.7 2.7

RH values were estimated from the data in Fig. 7. See the Materials andMethods for details describing the estimation method and the light conditionsused.




concentrations is in a monomeric form and that the largermolecular weight determined from the SEC data arises from adeviation of the molecular shape from a sphere. Similar obser-vations have been made for Arabidopsis phot2-LOV2 (calcu-lated mol. wt 16.9 kDa), which was revealed to exist as amonomer by small angle X-ray scattering (SAXS; Nakasakoet al. 2004) and by chemical cross-linking (Katsura et al. 2009)and gave a similar increase in molecular weight by SEC (appar-ent mol. wt 25.5 kDa) (Nakasako et al. 2008). Our SEC results areconsistent with previous SEC results that showed that theLOVh construct exists as a monomer at 175 mM and partly asa dimer at 350 mM in the dark (Toyooka et al. 2011). Thus, theAUREO1 LOV domain appears to exist in a monomeric form atphysiological concentrations.

On the other hand, the molecular weight of LOVh wasestimated as 35 kDa using a RH calibration curve for sphericalproteins of known mass and oligomeric state. This estimatedmolecular weight is twice as large as the calculated molecularweight. DLS values of all the recombinant protein solutionswere monodisperse and the RH-app increased linearly with anincrease in the concentration from 0 to 100 mM. Therefore, themolecular weight of LOVh estimated at 0 mM should reflectthat of the monomer. The large apparent molecular weightsuggests that the LOVh domain adopts a significantly elongatedand fragile molecular structure in solution that influences thecalculated RH (Patel et al. 2012). The discrepancy between themolecular weight determined by SEC and DLS may be ascribedto the different physical processes that these techniques detect(Wyatt et al. 2009).

In contrast, the molecular weight of ZLh determined by SECin the dark at 175 and 250 mM was 65.1 kDa (Toyooka et al.2011), which is 1.2 times larger than the calculated molecularweight for a dimer (55.8 kDa), indicating that ZLh forms a dimerat these concentrations. DLS data gave a molecular weight of81 kDa, which is 1.5 times larger than the dimer molecularweight. This observation suggests a significant deviation froma spherical molecular shape for the ZLh construct. Furthermore,the molecular weight of the FLh domain estimated by DLS was124 kDa, suggesting that FLh also formed a dimer with an elon-gated molecular shape. To elucidate this further, DLS studies onnon-spherical proteins are required.

Involvement of the disulfide linkage in dimerformation

LOV1 of Arabidopsis phot1 and phot2 has been shown to forma dimer. In contrast, LOV2 of phot1 and phot2 has been re-ported to exist in equilibrium between a monomer and dimer,and this equilibrium is dependent on the concentration of theprotein (Nakasako et al., 2008, Katsura et al. 2009). AUREO1LOV has comparatively higher amino acid sequence homologyand identity with LOV2 than LOV1 of plant phot proteins, andthis sequence similarity is therefore consistent with the resultsrepresented herein showing that AUREO1 LOV is a monomerat physiological concentrations. In Arabidopsis phot1-LOV1, adisulfide linkage at the second cysteine plays a major role indimerization; this cysteine residue is not present in LOV2.AUREO1 LOV has the second cysteine (Cys283 in Fig. 1); how-ever, the position in the amino acid sequence is not homolo-gous with those of the LOV1 proteins. Recently, a crystalstructure of AUREO1 LOV with N- and C-terminal extensions(Arg199–Tyr336) has been presented (Mitra et al. 2012). Here,the subunits are arranged as a trimer of antiparallel dimers. Nodisulfide linkage was observed between the two neighboringsubunits in the crystal structure.

In contrast, ZL and FL were revealed to form a dimerthrough disulfide linkages (Fig. 5). There are two possibilitiesthat a cysteine(s) in the ZL region, Cys162 and Cys182, act(s) asthe linkage site(s) and/or that the addition of the bZIP and thelinker region produces an ‘active’ disulfide linkage in the LOVdomain through an induced conformational change. Theformer was proved by the cysteine/serine substitution experi-ments that revealed the redundant contribution of bothCys162 and Cys182 to the dimerization of ZL through the di-sulfide linkage. This may further contribute to the dimerizationof FL. The latter is unlikely because the bZIP domain and thelinker region may not interact strongly with the LOV domain, asdiscussed in the previous section.

In addition, the N-terminal domain harbors two cysteineresidues, Cys11 and Cys54, (Fig. 1 upper). Additional contribu-tions to the linkage by these two cysteine residues are possible.Currently, seven AUREO sequences including AUREO2 havebeen deposited in the DDBJ/EMBL/GenBank databases, butno amino acid sequence homology was observed amongthese sequences. Furthermore, a BLAST search detected nosignificant homology of the N-terminal domain with anyknown motifs, suggesting that the N-terminal domains havesignificantly diverged functions, or, conversely, no function.By considering the molecular structure and function of thebZIP domain (Fujii et al. 2000), the possibility that these regionsfunction for the dimerization may be excluded.

BL-induced conformational changes

SEC detected a 16% increase of the apparent molecular weightin an AUREO1 LOV monomer upon BL excitation. CD revealeda BL-induced 7% change of the a-helical structure to a b-strandsecondary structure of the LOV monomer in the same buffer,

Table 3 Secondary structure composition estimated from CDspectra

a-Helix b-Strand Turns Unordered

ZLh (dark) 17 34 11 38

ZLh (light) 18 32 11 39

LOVh (dark) 16 33 12 39

LOVh (light) 9 40 9 42

The CD spectra in Fig. 8 were decomposed and the percentage of each com-ponent is indicated. The compositions of a-helix and b-strand represent thesum of helix1 (regular a-helix)/helix2 (distorted a-helix) and strand1 (regularb-strand)/strand2 (distorted b-strand), respectively.


O. Hisatomi et al.


suggesting that a secondary structure change may be the originof the observed molecular weight increase. The increase in themolecular weight did not alter the RH. A similar situation wasobserved with Arabidopsis phot2-LOV2. Here, a similar CDspectral change to that observed for AUREO1 LOV wasfound, and a conformational change detected by SDS–PAGEinduced by BL illumination was also determined (Katsura et al.2009). However, SAXS analysis showed that the radius of gyr-ation remained unchanged (Nakasako et al. 2004). The origin ofthese paradoxical observations remains unresolved. AUREOLOV with N- and C-terminal extensions also showed a-helicalunfolding by 6% (Mitra et al. 2012). BL-induced a-helical un-folding appears to be a common structural feature among thephot-LOV proteins (Corchnoy et al. 2003) and the LOV domainof BL receptors with a C-terminal effector (Losi et al. 2005,Moglich and Moffat 2007).

In contrast, the CD spectrum of ZL showed negligible changeupon BL illumination. This can be explained in two ways. First,the CD spectral changes in the LOV domain were canceled bythe secondary structural changes induced in other regionsof the protein, e.g. the linker region. Secondly, the addition ofthese effector domains restricts changes in the LOV domain. Itis possible that both contribute to the CD result of the ZL;however, the latter contribution may be negligible becausethe bZIP and the linker did not noticeably influence the spec-troscopic properties and the photoreaction kinetics of the LOVdomain. Accordingly, it is probable that BL induces a-helicalunfolding in the LOV domain and folding in the other regions ofthe protein. These changes may contribute to the observed 5%increase of RH. BL may induce similar global conformationalchanges to those of ZL in a dimeric FL molecule since a 5%increase in RH was also observed with FL. Together with the

information regarding the disulfide linkage, a schematic modelfor the BL-induced conformational change of AUREO1 isdepicted in Fig. 9.

The molecular processes underlying the observed secondarystructural changes in AUREO1 are important to understand.Currently, the molecular properties of phot-LOV have been themost extensively studied. In plant phot-LOV, an a-helix namedJa exists downstream of the C-terminus of the LOV2 domain inthe linker region to the effector (Ser/Thr kinase) domain(Crosson and Moffat 2001). BL-induced changes in thisJa-helix are suggested to trigger the conformational changesinvolved in the signal transduction process. BL receptors withone LOV domain and a C-terminal effector(s), such as YtvA(Moglich and Moffat 2007), EL222 (Nash et al. 2011) or BrucellaLOV-HK (Rinaldi et al. 2012), also have an a-helix correspond-ing to Ja, but their topological orientations to the LOV cores aredivergent. The crystal structure of AUREO1 LOV (Mitra et al.2012) also revealed the presence of a Ja-like helix connected tothe C-terminus. However, the helix has almost no amino acidsequence homology with that of the plant LOV, and its orien-tation to the b-scaffold differs. Since AUREO1 has the effectorbZIP domain on the N-terminal side of the LOV domain andshows a unique domain order compared with the other LOVproteins, it is of importance to know if this helix unfolds uponBL excitation. Unfortunately, the observed light-induced con-formational changes reported with the AUREO1 LOV crystalwere limited to the vicinity of the isoalloxazine ring of the FMNchromophore, and unfolding of the helix was not detected(Mitra et al. 2012). This result is probably because crystal pack-ing restricted such changes, as has been seen with the crystals ofother LOV domains (Crosson and Moffat 2002, Fedorov et al.2003, Halavaty and Moffat 2007). The details of the BL-induced

A B

Fig. 9 A schematic model for the conformational change of AUREO1-ZL. (A) In the dark state, AUREO1-ZL forms a dimer in which a disulfidelinkage at Cys162 (bZIP domain) and/or Cys182 (linker region) is involved. The LOV domain does not act as a dimer site. (B) In the light-activatedstate. BL induces a-helical unfolding and folding in the LOV domain and a second (bZIP and/or linker region) region, respectively. The dimericstructure is unchanged. The apparent molecular weight of the LOV domain alone increases by 16%, whereas the RH of the whole moleculeincreased by 5%. This indicates a global molecular elongation or expansion. These changes may be involved in DNA binding. LOV, LOV domain;bZIP, bZIP domain; SS, disulfide bond.




conformational changes will be determined by solving the crys-tal and the solution structures of AUREO1 ZL or FL using X-raycrystallography, SAXS and NMR.

Binding to DNA

The conformational changes described above may be involvedin the regulation of DNA binding of the ZL and the FL con-structs. In the present study, BL-induced enhancement of DNAbinding was not detected by EMSA. This non-radioactive tech-nique has lower sensitivity (0.1 pmol for each band ofCy3-labeled DNA) when compared with autoradiographythat we used previously (Takahashi et al. 2007). Since the bind-ing of the DNA in the dark was considerably high, the presentEMSA may not have been able to distinguish the dark and thelight DNA binding state.

The present results indicate that BL illumination induceda-helical folding in the bZIP and the linker region. Yang andSchepartz (2005) have reported that an increase in a-helicity inthe DNA-recognizing polypeptide of the bZIP protein yeastgeneral control protein 4 (GCN4) enhances DNA binding.Such an observation supports the concept that the observeda-helical increase promotes DNA binding of AUREO1.Furthermore, Goren et al. (2001) have reported that the forma-tion of two disulfide linkages between a pair of bZIP domains ofthe cAMP-responsive element-binding protein (CREB) in-creases a-helicity and enhances binding to its cognate DNAsites. Ciaccio and Laurence (2009) also showed that the forma-tion of a single disulfide linkage stabilized the a-helicity in abZIP domain dimer of activating transcription factor 5 (ATF5).These reports reveal that disulfide linkages play an importantrole in the dimerization and stability of the dimer in bZIP pro-teins; although the BL-induced formation of the disulfide link-age was not observed in our present study. The mode of DNAbinding based on the present results is also illustrated schemat-ically in Fig. 9.

Concluding remarks

AUREO1 is the only one BL-regulated DNA-binding LOV pro-tein using a bZIP motif known so far. The present study re-vealed the involvement of a disulfide linkage in forming dimericstructure that is requisite for DNA binding, and a possible rolefor increased a-helices in the enhancement of DNA binding byBL. These results provide important clues to understand themolecular events that induce DNA binding of AUREO1 and alsothat regulate the effector domains uniquely located on theN-terminal side of the LOV domain.

Materials and Methods

Construction of expression vectors of AUREO1recombinant proteins

The expression plasmids were constructed as described previ-ously (Toyooka et al. 2011). In brief, an XhoI site (465–471) in

AUREO1 cDNA was mutated with aureo1-dXhoIF andaureo1-dXhoIR primers (Supplementary Fig. S3) using aPrimeSTAR mutagenesis kit (TAKARA BIO INC.). A truncatedDNA fragment of AUREO1 from Gly113 to Lys348 (for ZLh) wasamplified with aureo1-bZIPF and aureo1-CTR primers, and afragment from Pro204 to Lys348 (for LOVh) was amplified withaureo1-P204F and aureo1-CTR primers. The AUREO1 cDNAcontaining the complete coding sequence with a singleamino acid deletion of Asn2 (for FLh) was amplified withaureo1-NTF and aureo1-CTR primers. For preparations of re-combinant proteins bearing a histidine hexamer at theN-terminus (hFL, hZL and hLOV), the pET23-Nhis plasmidvector we constructed was used. This plasmid was derivedfrom the original histidine hexamer sequence in the pET23plasmid. Here, the His-tag was deleted with pET23-dHisF andpET23-dHisR primers, and a new histidine hexamer sequencewas inserted into the multicloning site with pET23-NHisF andpET23-NHisR primers. The aureo1-CTter primer was used in-stead of the aureo1-CTR primer for amplifications. AmplifiedDNA fragments were inserted between the NcoI and XhoI sitesof pET23d or pET23-Nhis plasmid vectors. In order to preparethe amino acid substitutions C162S, C182S and C162S/C182S,the pET23d plasmid containing ZLh cDNA was mutated withprimer sets AC1/C162S-F and -R (for C162S) or AC1/C182S-Fand -R (for C182S) (Supplementary Fig. S3), using aPrimeSTAR mutagenesis kit. Sequences of all constructs wereconfirmed using the Thermo Sequenase Primer CycleSequencing Kit (GE Healthcare) with a SQ-5500 DNA sequencer(Hitachi Hitech), and expression vectors of recombinantAUREO1s were introduced into BL21 (DE3) cells (Invitrogen).

Preparation of AUREO1 recombinant proteins

Cells were cultured in 2� YT containing ampicillin, and expres-sion of AUREO1 recombinant proteins was induced by theaddition of isopropyl-b-D-thiogalactopyranoside to a final con-centration of 1 mM. Cells were harvested by centrifugation anddisrupted by sonication in lysis buffer (400 mM NaCl, 2 mMMgCl2, 4 mM DTT, 0.4 mM phenylmethylsulfonyl fluoride,1 mg ml�1 aprotinin, 10 mg ml�1 leupeptin, 80 ng ml�1 DNase Iand 20 mM Tris–HCl, pH 7.0). After the cell debris was removedby centrifugation at 65,000�g for 10 min, the supernatantswere incubated with agitation for 10 min at 4�C by the additionof 0.25% (final concentration) polyethyleneimine (mol. wt50–100 kDa, MP Biomedicals) and DNA was pelleted by centri-fugation at 14,000�g for 10 min. Ammonium sulfate (final con-centration of 50% saturation) was added to the supernatantsand the precipitated proteins were centrifuged at 14,000�g for10 min. The pellets containing the precipitants werere-solubilized in the loading buffer (400 mM NaCl, 2 mMDTT, 0.2 mM phenylmethylsulfonyl fluoride, 20 mM Tris–HCl,pH 7.0) and centrifuged at 65,000�g for 15 min to remove anyinsoluble material. AUREO1 recombinant proteins were thenpurified twice with an Ni-NTA column (Ni SepharoseTM6 FastFlow, GE Healthcare) according to the manufacturer’s


O. Hisatomi et al.




instructions. For further purification, the recombinant proteinswere applied to a HiTrap Heparin HP column (GE Healthcare),washed with the loading buffer containing 0.2 M NaCl, andeluted with a stepwise gradient with the loading buffer con-taining 0.4–0.8 M NaCl at a flow rate of 1 ml min�1. Afterexchanging the solvents for the loading buffer with PDMidiTrap G-25 columns (GE Healthcare), the recombinant pro-teins were stored at 4�C until use. Concentrations of theAUREO1 recombinant proteins were determined from the ab-sorbance at 450 nm using an extinction coefficient of13,000 M�1 cm�1.

SDS–PAGE

AUREO1 recombinant proteins were run on SDS–PAGE afterdenaturation in an SDS–PAGE sample buffer containing 35%(final concentration) glycerol and 125 mM Tris–HCl (pH 6.8)either by heating at 99�C for 5 min in the presence of 20 mMDTT (reducing conditions) or without heating in the absence ofDTT (non-reducing conditions). Under non-reducing condi-tions, the final concentrations of DTT were <0.5 mM. Thegels were stained with Coomassie Brilliant Blue (CBB).

EMSA

Each AUREO1 recombinant protein (40 pmol) was incubatedwith 50 pmol of a Cy3-labeled AUREO1-specific oligonucleotide(Ap-oligo, Supplementary Fig. S3) or a negative control oligo-nucleotide (Cp-oligo, Supplementary Fig. S3) in the presenceof 0.6mg of dI–dC (GE Healthcare), in a 5ml reaction mixture(300 mM KCl, 5% glycerol, 1 mM EDTA, 2 mM DTT, 20 mMTris–HCl, pH 7.0) at 25�C for 20 min. After the addition of1.5ml of loading dye (70% glycerol, 0.1% bromophenol blue,0.1% xylene cyanol), reaction mixtures were loaded onto a 5%polyacrylamide gel. Electrophoresis was performed at 100 V for1 h with the 0.5� Tris-borate-EDTA buffer (45 mM Tris, 45 mMboric acid, 1.5 mM EDTA) in a cold chamber at 4�C. TheCy3-labeled oligonucleotides were visualized with a Fluor-STM

Max MultiImager (Bio-Rad Laboratories). All the procedureswere performed under room light supplied with fluorescenttubes, FLR40-SEX (NEC Co.).

Spectroscopic measurements

Stock solutions of the AUREO1 recombinant proteins werediluted to concentrations ranging between 5 and 15mM witha 400 mM NaCl, 1 mM DTT and 2 mM Tris–HCl (pH 7.0) buffersolution, and UV-visible absorption spectra were measuredusing a V530, V550 or V630 spectrophotometer (JASCO)equipped with a temperature-controlling system. To measureBL-induced absorption spectral changes of the recombinantproteins, their solutions in a spectrophotometer cuvette wereilluminated with BL supplied with a hand-made light-emittingdiode (LED) illuminator (�max at 470 nm) for 1 min (200 mmolphotons m�2 s�1 at the sample position). Spectral changesaccompanying the regeneration of D450 from S390 in ZLhwere monitored after termination of BL illumination for

1 min at 15, 20, 25, 30 or 35�C in the dark. Dark regenerationcurves of D450 were obtained from the absorbance at 450 nmof the absorption spectra. Before the measurements, the cu-vette with a sample was set in the cell holder of the spectro-photometers and left for >3 min in the dark to ensure that thesample had reached the set temperature. Each absorbance dif-ference curve was fit by a single exponential using the formula,�Amax(1� e–kt), where �Amax is a maximum absorption dif-ference and k is a rate constant of the reaction. Fluorescenceemission spectra were measured at 25�C with excitation at450 nm using an F-7000 fluorescence spectrophotometer(Hitach HiTech) with a temperature-controlling system.

SEC measurements

SEC was performed using an AKTA prime column chromatog-raphy system (GE Healthcare) with a Sephacryl S-100 HRcolumn (GE Healthcare) in a phosphate-buffered saline (PBS)solution in a cold room at 5�C. The flow rate was 1.3 ml min�1

and fraction sizes were 1.3 ml. The column volume was 320 mland the void volume was determined by the elution of bluedextran as 96.2 ml. To determine the molecular weight of therecombinant proteins, RNase A (13.7 kDa), chymotrypsinogenA (25.0 kDa), ovalbumin (43.0 kDa) and albumin (67.0 kDa)were used as size markers. To study the effect of BL, thesample solutions were illuminated for 1 min by BL and appliedto the column under the BL supplied with a BL LED bulb 100B0810 (SIGN). During the elution, the column was illuminated byBL supplied with a combination of four BL LED panels(ISL-150�150-BB, CCS) and a mirror.

DLS measurements

DLS of the AUREO1 recombinant protein solutions was mea-sured with a Zetasizer-mV system (Malvern Instruments) andthe data were analyzed according to the method described inthe handling manual. The stock protein solutions were dilutedto 25–100 mM with a 400 mM NaCl and 20 mM Tris–HCl (pH7.0) solution (viscosity, Z= 0.9166� 10�3 Pa s�1). A 60ml ali-quot of the protein solutions was incubated at 25�C for 20 minin the dark and then centrifuged at 36,000�g for 10 min at 25�Cto remove aggregates. A 50ml supernatant sample was trans-ferred to a quartz sub-micro fluorimeter cuvette (StarnaScientific) and DLS measurements were performed. The lightscattering was detected at 830 nm and collected in the auto-matic mode at 25�C using a refractive index of the solution of1.334. The cuvette containing a sample was placed in the cellholder of the DLS system and left for >3 min to equilibrate thesample temperature at 25�C. DLS of the recombinant proteinsin the dark state (dark) was measured by several cycles. DLS inthe light-excited state (light) was measured immediately afterthe termination of BL illumination of the sample in the cuvetteset in the cell holder for 1 min. This single measurement wasrepeated several times. To examine the light–dark reversibility(light–dark), the samples illuminated by BL for 1 min were leftfor 30 min in darkness and then DLS was measured several






times. DLS of the ‘dark’, ‘light’ and ‘light–dark’ samples wasmeasured with multiple preparations by varying the sampleconcentrations from 25 to 100 mM. SDs of all the RH-app

values obtained from the multiple measurements were<0.06 nm. BL was supplied with an epifluorescent illuminationunit LED470MS-EPI (�max at 470 nm, 200 mmol photons m�2 s�1

at the sample position, OptCode). There is a relationship be-tween the RH-app from the DLS data and the RH, whereRH-app� RH (1 + yC), where C is the sample concentrationand y is a constant under the condition that yC is sufficientlysmaller than 1. RH-app was plotted against the concentrationand RH was obtained from the extrapolated value at 0mM pro-tein concentration. In these analyses, DLS data giving a poly-dispersity index >0.2 were omitted. The z-average molecularsizes in terms of the RH-app in solution were determined usingthe Zetasizer Software (Version 6.20) (Malvern Instruments).

To estimate the molecular weight of the recombinant pro-teins, RH-app values of aprotinin (6.5 kDa), RNase A (13.7 kDa),carbonic anhydrase (29.0 kDa), ovalbumin (44.0 kDa) hemoglo-bin (64.5 kDa) and conalbumin (75.0 kDa) were measured, anda calibration curve for spherical proteins was made.

CD measurements

CD spectra of ZLh and LOVh in a PBS solution were measuredin the far UV (180–300 nm) region at 5�C with a J820 spectro-polarimeter (JASCO) equipped with an electrictemperature-controlling system under flowing N2 gas. The op-tical path length was 0.1 cm and the concentrations of the ZLhand LOVh solutions were set at 8.62 and 3.24 mM, respectively.For each measurement, a scan was repeated 20 times to averagethe data. This ensured a resolution of 1 nm. Spectra of therecombinant proteins were obtained by subtracting the spectraof the sample buffer solutions that were 1/400 and 1/200 PBSsolutions for LOVh and ZLh, respectively. Before the measure-ment, the samples were left in the spectropolarimeter for>10 min in the dark to adapt the samples to the N2 atmos-phere in the sample chamber. The setting of the cuvette in theinstrument and CD spectral measurements were carried out indarkness. The measuring light from a spectropolarimeter in thefar UV region has little actinic effect on the UV-visible absorp-tion spectra of ZLh and LOVh solutions. To obtain the CDspectra in the light-excited state, a scan was started immedi-ately after the termination of BL illumination for 3 min duringwhich the CD measuring light was shuttered. BL was suppliedwith a hand-made LED illuminator (�max at 467 nm) set besidethe sample cuvette in the spectropolarimeter (150mmol pho-tons m�2 s�1 at the sample position). To reduce the effect ofdark regeneration during the measurements, CD spectra weremeasured at 5�C to slow the dark regeneration of D450 fromthe photoproduct (half-lifetime was 94 min at 5�C), and the20 scans were completed within 8.5 min. The secondary struc-ture compositions of ZLh and LOVh were predicted basedon the CD spectra from 190 to 240 nm using the On-LineCircular Dichroism Analysis Internet Site, DICHROWEB

(http://dichroweb.cryst.bbk.ac.uk) (Whitmore and Wallace2004). Input and output units, and the wavelength step werey (mdeg) and 1.0 nm, respectively. The algorithm used wasCDSSTR (Sreerama and Woody 2000) and the reference data-base was SP175 (Lees et al. 2006).

Safety light

Experimental manipulations during the preparation of recom-binant proteins, EMSA, and DLS and spectroscopic measure-ments were performed under a dim yellow LED lamp (�max at590 nm).

Supplementary data

Supplementary data are available at PCP online.

Funding

This research was supported by the Ministry of Education,Culture, Sports, Science and Technology [a Grant-in-Aid forScientific Research on Innovative Areas (No. 23120517 toO.H. and No. 22120005 to S.T.) and a Grant-in-Aid forExploratory Research (No. 23657105 to S.T.]; the JapanSociety for the Promotion of Science [a Grant-in-Aid forScientific Research (C) (No. 22570162 to O.H.)].

Acknowledgments

The authors thank Dr. Norio Hamada and Dr. RyosukeNakamura (Osaka University) for helpful discussions.

References

Avila-Perez, M., Vreede, J., Tang, Y., Bende, O., Losi, A., Gartner, W. et al.(2009) In vivo mutational analysis of YtvA from Bacillus subtilis:mechanism of light activation of the general stress response.J. Biol. Chem. 284: 24958–24964.

Christie, J.M. (2007) Phototropin blue-light receptors. Annu. Rev. PlantBiol. 58: 21–45.

Christie, J.M., Reymond, P., Powell, G.K., Bernasconi, P., Raibekas, A.A.,Liscum, E. et al. (1998) Arabidopsis NPH1: a flavoprotein with theproperties of a photoreceptor for phototropism. Science 282:1698–1701.

Christie, J.M., Swartz, T.E., Bogomolni, R.A. and Briggs, W.R. (2002)Phototropin LOV domains exhibit distinct roles in regulatingphotoreceptor function. Plant J. 32: 205–219.

Christie, J.M., Yang, H., Richter, G.L., Sullivan, S., Thomson, C.E., Lin, J.et al. (2011) phot1 inhibition of ABCB19 primes lateral auxin fluxesin the shoot apex required for phototropism. PLoS Biol. 9: e1001076.

Ciaccio, N.A. and Laurence, J.A. (2009) Effects of disulfide bond for-mation and protein helicity on the aggregation of activating tran-scription factor 5. Mol. Pharm. 6: 1205–1215.

Corchnoy, S.B., Swartz, T.E., Lewis, J.W., Szundi, I., Briggs, W.R. andBogomolni, R.A. (2003) Intramolecular proton transfers and


O. Hisatomi et al.


http://dichroweb.cryst.bbk.ac.uk


structural changes during the photocycle of the LOV2 domain of

phototropin 1. J. Biol. Chem. 278: 724–731.Crosson, S. and Moffat, K. (2001) Structure of a flavin-binding plant

photoreceptor domain: insights into light-mediated signal trans-

duction. Proc. Natl Acad. Sci. USA 98: 2995–3000.Crosson, S. and Moffat, K. (2002) Photoexcited structure of a plant

photoreceptor domain reveals a light-driven molecular switch.

Plant Cell 14: 1067–1075.Crosson, S., Rajagopal, S. and Moffat, K. (2003) The LOV domain

family: photoresponsive signaling modules coupled to diverse

output domains. Biochemistry 42: 2–10.Djouani-Tahr, E.B., Christie, J.M., Sanchez-Ferandin, S., Sanchez, F.,

Bouget, F.Y. and Corellou, F. (2011) A eukaryotic LOV-histidine

kinase with circadian clock function in the picoalga Ostreococcus.

Plant J. 65: 578–588.Ellenberger, T.E., Brandl, C.J., Struhl, K. and Harrison, S.C. (1992) The

GCN4 basic region leucine zipper binds DNA as a dimer of uninter-

rupted alpha helices: crystal structure of the protein–DNA com-

plex. Cell 71: 1223–1237.Fedorov, R., Schlichting, I., Hartmann, E., Domratcheva, T.,

Fuhrmann, M. and Hegemann, P. (2003) Crystal structures and

molecular mechanism of a light-induced signaling switch: the

Phot-LOV1 domain from Chlamydomonas reinhardtii. Biophys. J.

84: 2474–2482.Fujii, Y., Shimizu, T., Toda, T., Yanagida, M. and Hakoshima, T. (2000)

Structural basis for the diversity of DNA recognition by bZIP tran-

scription factors. Nat. Struct. Biol. 7: 889–893.Goren, I., Tavor, E., Goldblum, A. and Honigman, A. (2001) Two cyst-

eine residues in the DNA-binding domain of CREB control binding

to CRE and CREB-mediated gene expression. J. Mol. Biol. 313:

695–709.Guo, H., Kottke, T., Hegemann, P. and Dick, B. (2005) The phot LOV2

domain and its interaction with LOV1. Biophys. J. 89: 402–412.Halavaty, A.S. and Moffat, K. (2007) N- and C-terminal flanking regions

modulate light-induced signal transduction in the LOV2 domain of

the blue light sensor phototropin 1 from Avena sativa. Biochemistry

46: 14001–14009.Harper, S.M., Neil, L.C., Day, I.J., Hore, P.J. and Gardner, K.H. (2004)

Conformational changes in a photosensory LOV domain monitored

by time-resolved NMR spectroscopy. J. Amer. Chem. Soc. 126:

3390–3391.Imaizumi, T., Tran, H.G., Swartz, T.E., Briggs, W.R. and Kay, S.A. (2003)

FKF1 is essential for photoperiodic-specific light signaling in

Arabidopsis. Nature 426: 302–306.Ishikawa, M., Takahashi, F., Nozaki, H., Nagasato., C., Motomura, T. and

Kataoka, H. (2009) Distribution and phylogeny of the blue light

receptors aureochromes in eukaryotes. Planta 230: 543–552.Iwata, T., Tokutomi, S. and Kandori, H. (2002) Photoreaction of the

cysteine S-H group in the LOV2 domain of Adiantum phyto-

chrome3. J. Amer. Chem. Soc. 124: 11840–11841.Kagawa, T., Sakai, T., Suetsugu, N., Oikawa, K., Ishiguro, S., Kato, T. et al.

(2001) Arabidopsis NPL1: a phototropin homolog controlling

the chloroplast high-light avoidance response. Science 291:

2138–2141.Kasahara, M., Swartz, T.E., Olney, M.A., Onodera, A., Mochizuki, N.,

Fukuzawa, H. et al. (2002) Photochemical properties of the flavin

mononucleotide-binding domains of the phototropins from

Arabidopsis, rice, and Chlamydomonas reinhardtii. Plant Physiol.

129: 762–773.

Katsura, H., Zikihara, K., Okajima, K., Yoshihara, S. and Tokutomi, S.

(2009) Oligomeric structure of LOV domains in Arabidopsis photo-

tropin. FEBS Lett. 583: 526–530.Kinoshita, T., Doi, M., Suetsugu, N., Kagawa, T., Wada, M. and

Shimazaki, K. (2001) Phot1 and phot2 mediate blue light regulation

of stomatal opening. Nature 414: 656–660.Kozuka, T., Kong, S.G., Doi, M., Shimazaki, K. and Nagatani, A. (2011)

Tissue-autonomous promotion of palisade cell development by

phototropin 2 in Arabidopsis. Plant Cell 23: 3684–3695.Lees, J.G., Miles, A.J., Wien, F. and Wallace, B.A. (2006) A reference

database for circular dichroism spectroscopy covering fold and sec-

ondary structure space. Bioinformatics 22: 1955–1962.Losi, A., Quest, B. and Gartner, W. (2003) Listening to the blue: the

time-resolved thermodynamics of the bacterial blue-light receptor

YtvA and its isolated LOV domain. Photochem. Photobiol. Sci. 2:

759–766.Losi, A., Ghiraldelli, E., Jansen, S. and Gartner, W. (2005) Mutational

effects on protein structural changes and interdomain interactions

in the blue-light sensing LOV protein YtvA. Photochem. Photobiol.

81: 1145–1152.Malzahn, E., Ciprianidis, S., Kaldi, K., Schafmeier, T. and Brunner, M.

(2010) Photoadaptation in Neurospora by competitive interaction

of activating and inhibitory LOV domains. Cell 142: 762–772.Mitra, D., Yang, X. and Moffat, K. (2012) Crystal structures of aureo-

chrome1 LOV suggest new design strategies for optogenetics.

Structure 20: 698–706.Moglich, A. and Moffat, K. (2007) Structural basis for light-dependent

signaling in the dimeric LOV domain of the photosensor YtvA.

J. Mol. Biol. 373: 112–126.Nakasako, M., Iwata, T., Matsuoka, D. and Tokutomi, S. (2004)

Light-induced structural changes of LOV domain-containing poly-

peptides from Arabidopsis phototropin 1 and 2 studied by

small-angle X-ray scattering. Biochemistry 43: 14881–14890.Nakasako, M., Zikihara, K., Matsuoka, D., Katsura, H. and Tokutomi, S.

(2008) Structural basis of the LOV1 dimerization of Arabidopsis

phototropins 1 and 2. J. Mol. Biol. 381: 718–733.Nash, A.I., McNulty, R., Shillito, M.E., Swartz, T.E., Bogomolni, R.A.,

Luecke, H. et al. (2011) Structural basis of photosensitivity in a

bacterial light–oxygen–voltage/helix–turn–helix (LOV-HTH)

DNA-binding protein. Proc. Natl Acad. Sci. USA 108: 9449–9454.Okajima, K., Matsuoka, D. and Tokutomi, S. (2011) LOV2-linker-kinase

phosphorylates LOV1-containing N-terminal polypeptide substrate

via photoreaction of LOV2 in Arabidopsis phototropin1. FEBS Lett.

585: 3391–3395.Patel, T.R., Reuten, R., Xiong, S., Meier, M., Winzor, D.J., Koch, M. et al.

(2012) Determination of a molecular shape for netrin-4 from hydro-

dynamic and small angle X-ray scattering measurements. Matrix

Biol. 31: 135–140.Purcell, E.B., Siegal-Gaskins, D., Rawling, D.C., Fiebig, A. and Crosson, S.

(2007) A photosensory two-component system regulates bacterial

cell attachment. Proc. Natl Acad. Sci. USA 104: 18241–18246.Raffelberg, S., Mansurova, M., Gartner, W. and Losi, A. (2011)

Modulation of the photocycle of a LOV domain photoreceptor

by the hydrogen-bonding network. J. Amer. Chem. Soc. 133:

5346–5356.Rinaldi, J., Gallo, M., Klinke, S., Paris, G., Bonomi, H.R., Bogomolni, R.A.

et al. (2012) The b-scaffold of the LOV domain of the Brucella

light-activated histidine kinase is a key element for signal transduc-

tion. J. Mol. Biol. 420: 112–127.




Salomon, M., Christie, J.M., Kneib, E., Lempert, U. and Briggs, W.R.(2000) Photochemical and mutational analysis of theFMN-binding domains of the plant blue light receptor, phototro-pin. Biochemistry 39: 9401–9410.

Salomon, M., Lempert, U. and Rudiger, W. (2004) Dimerization of theplant photoreceptor phototropin is probably mediated by theLOV1 domain. FEBS Lett. 572: 8–10.

Sreerama, N. and Woody, R.W. (2000) Estimation of protein secondarystructure from circular dichroism spectra: comparison of CONTIN,SELCON and CDSSTR methods with an expanded reference set.Anal. Biochem. 287: 252–260.

Swartz, T.E., Corchnoy, S.B., Christie, J.M., Lewis, J.W., Szundi, I.,Briggs, W.R. et al. (2001) The photocycle of a flavin-bindingdomain of the blue light photoreceptor phototropin. J. Biol.Chem. 276: 36493–36500.

Swartz, T.E., Tseng, T., Frederickson, M.A., Paris, G., Comerci, D.J.,Rajashekara, G. et al. (2007) Blue-light-activated histidine kinases:two-component sensors in bacteria. Science 317: 1090–1093.

Taylor, B.L. and Zhulin, I.B. (1999) PAS domains: internal sensors ofoxygen, redox potential and light. Microbiol. Mol. Biol. Rev. 63:479–506.

Takahashi, F., Yamagata, D., Ishikawa, M., Fukamatsu, Y., Ogura, Y.,Kasahara, M. et al. (2007) AUREOCHROME, a photoreceptorrequired for photomorphogenesis in stramenopiles. Proc. NatlAcad. Sci. USA 104: 19625–19630.

Toyooka, T., Hisatomi, O., Takahashi, F., Kataoka, H. and Terazima, M.(2011) Photoreactions of Aureochrome-1. Biophys. J. 100:2801–2809.

Wyatt, A.R., Yerbury, J.J. and Wilson, M.R. (2009) Structural character-ization of clusterin–chaperone client protein complexes. J. Biol.Chem. 284: 21920–21927.

Whitmore, L. and Wallace, B.A. (2004) DICHROWEB: an online serverfor protein secondary structure analyses from circular dichroismspectroscopic data. Nucleic Acids Res. 32: W668–W673.

Williams, S.C., Cantwell, C.A. and Johnson, P.F. (1991) A family of C/EBP-related proteins capable of forming covalently linked leucinezipper dimers in vitro. Genes Dev. 5: 1553–1567.

Yang, L. and Schepartz, A. (2005) Relationship between folding andfunction in a sequence-specific miniature DNA-binding protein.Biochemistry 44: 7469–7478.

Zikihara, K., Iwata, T., Matsuoka, D., Kandori, H., Todo, T. andTokutomi, S. (2006) Photoreaction cycle of the light, oxygen, andvoltage domain in FKF1 determined by low-temperature absorp-tion spectroscopy. Biochemistry 45: 10828–10837.

Zoltowski, B.D., Nash, A.I. and Gardner, K.H. (2011) Variations in pro-tein–flavin hydrogen bonding in a light, oxygen, voltage domainproduce non-Arrhenius kinetics of adduct decay. Biochemistry 50:8771–8779.


O. Hisatomi et al.


Issue topic: Blue-Light Response PCP-2012-E-00464 (revised)

Documents