A novel retinol-binding protein in the retina of the swallowtail butterfly, Papilio xuthus Motohiro Wakakuwa 1 , Kentaro Arikawa 1 and Koichi Ozaki 2 1 Graduate School of Integrated Science, Yokohama City University, Yokohama, Kanagawa; 2 Graduate School of Frontier Biosciences, Osaka University, Toyonaka, Osaka, Japan Retinoid-binding proteins are indispensable for visual cycles in both vertebrate and invertebrate retinas. These proteins stabilize and transport hydrophobic retinoids in the hydro- philic environment of plasma and cytoplasm, and allow regeneration of visual pigments. Here, we identified a novel retinol-binding protein in the eye of a butterfly, Papilio xuthus. The protein that we term Papilio retinol-binding protein (Papilio RBP) is a major component of retinal soluble proteins and exclusively binds 3-hydroxyretinol, and emits fluorescence peaking at 480 nm under ultraviolet (UV) illumination. The primary structure, deduced from the nucleotide sequence of the cDNA, shows no similarity to any other lipophilic ligand-binding proteins. The molecular mass and isoelectric point of the protein estimated from the amino-acid sequence are 26.4 kDa and 4.92, respectively. The absence of any signal sequence for secretion in the N-terminus suggests that the protein exists in the cytoplas- mic matrix. All-trans 3-hydroxyretinol is the major ligand of the Papilio RBP in dark-adapted eyes. Light illumination of the eyes increases the 11-cis isomer of the ligand and induces redistribution of the Papilio RBP from the proximal to the distal part of the photoreceptor layer. These results suggest that the Papilio RBP is involved in visual pigment turnover. Keywords: retinol-binding protein; rhodopsin; visual pig- ment; visual cycle. Retinalaldehyde (retinal) plays an essential role in animal vision as the chromophore of visual pigments that are generically called rhodopsins. In the rhodopsin molecule, retinal is bound to the protein, opsin, in the 11-cis configuration. Light energy first isomerizes the chromo- phore into its all-trans form that subsequently causes a conformational change of the opsin into an active form. The activated rhodopsin, usually called metarhodopsin, triggers the phototransduction cascade, that eventually controls the flow of ion currents through cation channels in the plasma membrane of the photoreceptor cell. Prolonged illumination will cause depletion of rhodopsin unless its chromophore is replenished. An important pathway for rhodopsin replen- ishment in all known photoreceptor cells is the recovery of all-trans retinal from opsin, its reverse isomerization to the 11-cis form, and subsequent recombination with opsin. Some processes in the pathway do not occur in the photoreceptive membrane, where rhodopsin molecules are embedded and function. Thus, the retinal has to be transported, when necessary, in hydrophilic matrices. As retinoids are highly hydrophobic and hardly soluble in water, hydrophilic retinoid-binding proteins are therefore required for stabilizing retinoids in the watery plasma as well as in the cytoplasm, and for transporting retinoids within and/or between cells [1]. In addition, recent studies have demonstrated that such protein is not simply a carrier of retinoid. Regulation of retinoid concentration and its delivery to various cells, protection of retinoid from degradation and protection of cells from the potentially toxic properties of free retinoid may also be biologically important functions of retinoid-binding proteins (reviewed in [2]). The rhodopsin recycling system, the visual cycle, is well characterized in vertebrates (reviewed in [3–5]). Briefly, all- trans retinol bound to serum retinol-binding protein (RBP) circulates in the blood and is targeted to the retinal pigment epithelial (RPE) cells. There it is possibly transferred to cellular retinol-binding protein (CRBP) and esterified to all- trans-retinyl ester. After hydrolysis and isomerization to the 11-cis form, it is transferred to cellular retinal-binding protein (CRALBP) and oxidized to 11-cis retinal. Several mechanisms for the isomerization from all-trans to 11-cis isomer have been proposed. These include coupling of the hydrolysis of all-trans-retinyl esters to isomerization gener- ating 11-cis-retinol [6], or the presence of an enzyme catalyzing the direct isomerization of all-trans-to11-cis- retinol through a carbocation intermediate [7]. In both cases, the isomerization requires the presence of CRALBP [6,7]. Another pathway for isomerization is mediated by RPE retinal G-protein-coupled receptor (RGR). RGR is a vertebrate homolog of squid retinochrome (see below), and catalyzes light-dependent isomerization of all-trans-to 11-cis-retinal [5,8]. The 11-cis-retinal formed in the RPE cells is then transported across the interphotoreceptor Correspondence to K. Ozaki, Graduate School of Frontier Biosciences, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043, Japan. Fax/Tel.: + 81 6 6850 5439, E-mail: ozaki@bio.sci.osaka-u.ac.jp Abbreviations: CRALBP, cellular retinal-binding protein; CRBP, cellular retinol-binding protein; IRBP, interphotoreceptor retinoid-binding protein; RBP, retinol-binding protein. Note: The nucleotide sequence reported in this paper has been deposited in the DDBJ/EMBL/GenBank under the accession number AB070628. (Received 12 February 2003, revised 4 April 2003, accepted 9 April 2003) Eur. J. Biochem. 270, 2436–2445 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03614.x matrix to the photoreceptor cells. Involvement of inter- photoreceptor retinoid-binding protein (IRBP) in this step has been advocated, but is, however, still in dispute (reviewed in [2]). In photoreceptor cells, retinal binds to opsin to form rhodopsin. All-trans-retinal, liberated from opsin after light absorption, is reduced into all-trans-retinol in the photoreceptor cells and then moved back to retinal pigment epithelial cells. Regeneration of rhodopsin in invertebrates is somewhat different from that of invertebrates, as studied intensively in cephalopods and insects. Metarhodopsins of these animals are usually thermostable, i.e. the opsin and the chromo- phore do not immediately separate as they do in vertebrates. Therefore, metarhodopsins can absorb light whose wave- length is different from the wavelength absorbed by rhodopsins. Upon light absorption by metarhodopsin, all- trans-retinal is reconverted to 11-cis form, and thus, rhodopsin is regenerated. This pathway is called photo- reconversion or photoregeneration. In addition to this photochemical reaction, there exists another pathway through which rhodopsin is metabolically regenerated (visual cycle). In squid, Todarodes pacificus, metarhodopsin, resulting from photoconversion of rhodopsin, transfers its all-trans-retinal to squid retinal-binding protein (squid RALBP) [9]. The protein transports the all-trans-retinal from the outer segment to the inner segment of the photoreceptor cell [10,11]. In the inner segment, all-trans retinal is transferred to retinochrome. Light absorption by the retinochrome-all-trans-retinal complex causes photo- isomerization of the all-trans-retinal to the 11-cis form, which is then transferred to the squid RALBP and sub- sequently transported back to the outer segment. The squid RALBP provides the attached 11-cis-retinal to metarho- dopsin and, in return, receives all-trans-retinal: the rhodop- sin is thus regenerated. In this system, squid RALBP functions as a shuttle carrying 11-cis- and all-trans-retinal back and forth between the inner and the outer segments [10,12]. A similar recycling system using retinochrome and RALBP is also found in gastropods [13,14]. Recently, Robles et al. suggested the direct interaction of rhodopsin with retinochrome, based on immunocytochemical obser- vations [15]. However, this finding does not completely rule out the involvement of RALBP in chromophore transport in the cephalopod visual cycle. The visual cycle in insect retina has been studied in several species. In the blowfly retina, metarhodopsin is degraded slowly into opsin and all-trans-3-hydroxyretinal [16]. HPLC analysis of retinoids suggested that the liberated all-trans-3- hydroxyretinal might be bound to a protein that mediates photoisomerization of the all-trans-3-hydroxyretinal to the 11-cis form [17,18]. A protein having required properties has been isolated from the honeybee retina [19,20], but not yet from fly. The 11-cis-3-hydroxyretinal is then reduced to alcohol (11-cis-3-hydroxyretinol) followed by slow re-oxi- dation to aldehyde (11-cis-3-hydroxyretinal). The aldehyde would be used as a chromophore to regenerate rhodopsin. Involvement of 11-cis-3-hydroxyretinol in this pathway was proposed based on the observation that the amount of 11-cis-3-hydroxyretinol was increased considerably by light- adaptation [17]. Also in the butterfly retina, it has been demonstrated that metarhodopsin is degraded rapidly [21], and abundant 3-hydroxyretinol is contained in the soluble fraction [22,23]. These findings suggest that a visual cycle similartothatintheflyalsoexistsinthebutterflyretina. In addition, it was demonstrated in the Japanese yellow swallowtail butterfly, Papilio xuthus, that the isomer com- position of the 3-hydroxyretinol changes between the light- and dark-adaptation, suggesting that the 3-hydroxyretinol is possibly involved in the visual cycle. Although these studies strongly suggest that some retinol-binding protein may be involved in the insect visual cycle, no such a protein has been identified. In addition to the above biochemical studies, we recently found that the Papilio compound eye consists of three distinct types of ommatidia, one of which emits strong fluorescence under ultraviolet light [24]. The microspectro- fluorometric study suggested that the fluorescence is due to 3-hydroxyretinol that can act as a UV absorbing spectral filter. These previous observations suggested strongly that some kind of retinol-binding protein possibly localized in the Papilio retina, and functions in the visual cycle and/or color vision. In this study, we therefore isolated a soluble retinol- binding protein from the Papilio retina, and performed molecular biological and biochemical analyses of the pro- tein. As the protein is a novel species of the hydrophobic- ligand-binding protein and solely binds 3-hydroxyretinol as an intrinsic ligand, we termed this protein the Papilio retinol-binding protein (Papilio RBP). Further analysis suggested that Papilio RBP is involved in the visual cycle rather than the ommatidial fluorescence. Materials and methods Animals We used both sexes of the Japanese yellow swallowtail butterfly, Papilio xuthus Linnaeus. The butterflies were reared on fresh citrus leaves at 25 °C under a light regime of 8-h light : 16-h dark. The pupae were stored at 4 °Cfor atleast3monthsandthenallowedtoemergeat25°C. When necessary, the butterflies were dark-adapted for 48 h in complete darkness, or light-adapted for 12 h by posi- tioning the animals 5-cm from a 15 W white fluorescent lamp. For light-adaptation, butterflies were immobilized by clipping their wings and fixed in appropriate positions. Column chromatography Papilio RBP was purified from a water-soluble fraction of the retina by two-step column chromatography. All of the following procedures were conducted under dim red light. Retinas of the dark-adapted butterflies were detached from the corneal cuticle of the compound eyes and homogenized in 63 m M Tris/HCl buffer (pH 6.8). The homogenate was centrifuged at 15 000 g for 15 min at 4 °C yielding a clear supernatant containing only soluble proteins. The proteins in the extract were first separated by anion-exchange chromatography using the SMART System (Amersham Pharmacia Biotech) equipped with a Mono Q column that was equilibrated with 20 m M bis/Tris/HCl buffer (pH 6.5) at room temperature. The proteins were eluted with a linear gradient from 0–0.4 M NaClinthesamebuffer.The fractions that emit bluish fluorescence under UV-irradiation Ó FEBS 2003 Retinol-binding protein in butterfly eye (Eur. J. Biochem. 270) 2437 were collected, concentrated by ultrafiltration, and subjected to further purification using size-exclusion chromatography. The chromatography was performed using the SMART System equipped with a Superdex 75 column. The column was equilibrated with 150 m M bis/Tris/HCl (pH 6.5) con- taining 0.15 M NaCl, and proteins were eluted with the same buffer at room temperature. The absorbance of the eluent was monitored at 280 nm and 330 nm. Gel electrophoresis Besides the column chromatography, native PAGE was also used for purification of Papilio RBP as follows. The compound eyes were homogenized in 63 m M Tris/HCl buffer (pH 6.8), and the homogenate was centrifuged at 15 000 g for 30 min at 4 °C. The supernatant was put on a 5% polyacrylamide concentrating gel (125 m M Tris/HCl, pH 6.8), and proteins in the supernatant were separated in a 10% polyacrylamide gel (375 m M Tris/HCl, pH 8.8) under electrophoresis using Tris/glycine (25/192 m M ) run- ning buffer. After electrophoresis, the gel was illuminated with UV light that visualizes a single band of Papilio RBP by strong whitish fluorescence. A piece of gel containing the fluorescing band was then cut out, and Papilio RBP was eluted electrophoretically out of the gel. Alternatively, the gel was placed in a whole gel elutor (Bio-Rad) immediately after electrophoresis, and fluorescing fractions were retrieved electrophoretically. Regular SDS/PAGE was also performed according to Laemmli (1970) by the use of 12% polyacrylamide gel [25]. The gel was then stained with Coomassie Brilliant Blue to visualize the proteins. Protein digestion and sequencing Papilio RBP was purified from 100 compound eyes as described above. The purified protein was digested with 10 pmol of lysyl-endopeptidase in 83 m M Tris/HCl buffer (pH 9.2) for 5 h at 37 °C. The reaction was stopped by adding trifluoroacetic acid to the reaction mixture at a final concentration of 0.04%. Peptides were separated and isolated by reverse-phase HPLC (SMART System) using a lRPC C 2 /C 18 column equilibrated with 0.1% trifluoro- acetic acid. Peptides were eluted with a 0–80% linear gradient of acetonitrile containing 0.1% trifluoroacetic acid. Elution was monitored at 215 nm and peaks were collected separately. Amino acid sequences of isolated peptides were determined using a protein sequencer (Model G1005A, Hewlett Packard). For nucleotide sequencing of Papi- lio RBP cDNA, Poly(A) + RNA was prepared from 40 compound eyes using a QuickPrep mRNA Purification Kit (Amersham Pharmacia Biotech), and used for synthesis of cDNA with oligo(dT) primer. We prepared three pairs of oligo nucleotide primers (ROLBP1-forward, 5¢-AARGAR GAYGTNTGG-3¢; ROLBP1-reverse, 5¢-CCANACRTC YTCYTT-3¢; ROLBP2-forward, 5¢-AARGCNGGNAT HYT-3¢; ROLBP2-reverse, 5¢-ARDATNCCNGCYTT-3¢; ROLBP3-forward, 5¢-AARGTNTGGWSNGA-3¢;ROLB P3-reverse, 5¢-TCNSWCCANACYTT-3¢)basedonthe amino acid sequences determined above (KEDVW, KAG IL, KVWSE). Using these primers, we amplified the Papilio retinal cDNA by PCR, and determined the nucleotide sequences of amplified cDNA products. The 3¢-and 5¢-RACE were employed to complete sequencing of the entire coding region of the Papilio RBP cDNA. For 3¢- RACE, the primer containing EcoRI–SacI–KpnIsites and poly(T) sequences (ROLBP-RT1, 5¢-GCCGAATT CGAGCTCGGTACCTTTTTTTTTTTTTTTTT-3¢)was prepared to synthesize the first strand cDNA from the Papilio retinal mRNA. Based on the nucleotide sequence of the above PCR products, specific forward primers (ROL- BP4-F, 5¢-TTGCTTCCTCACGGCACCAG-3¢; ROLBP5- F, 5¢-GACTAGTGGTGAACATGTGTATGCCGCAG- 3¢) were synthesized and used for PCR with the first strand cDNA (template) and the partial sequence of ROLBP-RT1 (T-RAP, 5¢-GCCGAATTCGAGCTCGGTACC) as a reverse primer. To synthesize the first strand cDNA for 5¢-RACE, a specific reverse primer (ROLBP-RT2, 5¢-TCTGCTCAATGATTGATGTC-3¢) was prepared. The poly(A) sequence was attached to the 5¢-end of the cDNA, which was then amplified by PCR, using a set of primers, ROLBP-RT1 and ROLBP7-R (5¢-GACTAG TATCGCTTCAGGGTCCTCCGCTG-3¢). The product was again amplified with the second set of primers, T-RAP and ROLBP7-R. Ligand analysis The ligand analysis experiments were carried out under dim red light. Ten compound eyes were used for each experi- ment. In order to analyze the geometric isomers of retinoids using HPLC, Papilio RBP was isolated by native PAGE from the crude extract of the light-adapted or dark-adapted retinas, and finally dissolved in 200 lL PAGE running buffer. Each sample (200 lL) was mixed with 60 lLof2 M hydroxylamine (NH 2 OH) and 400 lLofcold90%meth- anol to convert 3-hydroxyretinals, if any, to retinaloximes. Retinoids and retinaloximes were extracted with 500 lL dichloromethane and 6 mL n-hexane. The extract was then concentrated and separated with a Hitachi model 635 HPLC system equipped with a YMC A-012 column (5-lm silica gel, 6 · 150 mm, Yamamura Chemical Laboratory). Elution was carried out with n-hexane containing 25% ethyl acetate and 2% ethanol at a flow rate of 1.2 mLÆmin )1 ,and the eluent was monitored for absorbance at 340 nm. In this elution condition, isomers of 3-hydroxyretinaloximes and 3-hydroxyretinol were eluted between 10 and 35 min, while isomers of retinaloximes and retinol were eluted just after the solvent front without enough resolution. In the present study, we did not carry out further analysis of retinaloxims and retinol, as neither retinal nor retinol are contained in the Papilio retina [22]. Standard isomers of 3-hydroxyretinal were synthesized by M. Ito (Kobe College of Pharmacy, Japan) [22]. Isomers of 3-hydroxyretinol were prepared by reducing the corresponding isomers of 3-hydroxyretinal in ethanol with a trace amount of sodium borohydride. For routine analyses, isomers of 3-hydroxyretinal and 3-hydroxyretinol extracted from Drosophila heads were also used as a standard mixture. The molar ratio of retinol isomers was calculated by using their extinction coefficients at 340 nm in the eluent (all-trans, 39 100; 11-cis, 22 700; 13-cis, 42 500). In order to measure absorption and fluorescence spectra of Papilio RBP, the fluorescing protein was collected from dark-adapted compound eyes using a 2438 M. Wakakuwa et al. (Eur. J. Biochem. 270) Ó FEBS 2003 whole gel elutor as described above, dialyzed to remove acrylamide contamination, concentrated with a Centricon YM-10 (Millipore), and re-dissolved in 10 m M Tris/HCl (pH 8.0) buffer. Absorption and fluorescence spectra were measured with a Hitachi model U-3300 spectropho- tometer and a Hitachi model F-4500 spectrofluorometer, respectively. Localization of Papilio RBP in light- and dark-adapted eyes Light- or dark-adapted Papilio retina was divided into distal and proximal portions by pulling out the retina from the corneal cuticle. This manipulation allows the eyes to be separated into the distal portion, which contains distal one- third of the photoreceptor layer in addition to the cornea and the crystalline cone, from the rest that we call the proximal portion. After performing native PAGE (see above) in the distal and proximal portions separately, we compared the fluorescence intensity between the portions on the gel. The fluorescence intensity was measured directly with a CCD camera, stored using an ATTO AE6905C Image Saver, and quantified with NIH IMAGE program. The gel was then stained with Coomassie Brilliant Blue, and the protein content was measured via the absorbance of the stained bands using a Sharp JX-350 image scanner. We further analyzed the isomer composition of the intrinsic ligands of Papilio RBP extracted separately from the distal and proximal portions of the retina. Papilio RBP was extracted from each portion of the dark-adapted or light- adapted retina, and purified by native PAGE. The ligand was then analyzed by HPLC as described above. Results Purification of Papilio RBP Figure 1A shows the results of native PAGE of crude extract from Papilio compound eyes. We identified a single band emitting whitish fluorescence under UV illumination. Coomassie Brilliant Blue staining of the gel indicates that the fluorescing protein is one of the major components of soluble proteins in the crude extract. The surface of the fluorescing protein carries negative charge in total, because the protein expresses high mobility in the native gel. We purified the fluorescing protein from the gel by two-step column chromatography. We first separated the crude extract with an anion-exchange (Mono Q) column and then with a size-exclusion (Superdex 75) column (Fig. 1B). With this purification procedure, we isolated the protein from other soluble proteins, shown as a single band in a SDS/PAGE gel (Fig. 1C). The apparent molecular mass of this protein was 31 kDa on the SDS/PAGE gel, which was close to 34 kDa estimated from the size-exclusion chromatography in the native state (Fig. 1B, inset). Fig. 1. Purification of Papilio RBP. (A)NativePAGEofthecrude extractofthePapilio retina. Fluorescence under UV (left) and Coo- massie Brilliant Blue (CBB) staining (right). (B) Anion exchange (Mono Q, top) and size-exclusion (Superdex 75, bottom) chromato- graphs of Papilio RBP. The fluorescent fraction (arrow) in the anion exchange chromatography was collected, and re-chromatographed with Superdex 75 column. A well-separated fluorescent peak of Papi- lio RBP (arrow), whose molecular mass is estimated to be approxi- mately 34 kDa (open circle in inset) was isolated. During chromatography, eluents were continuously monitored via light absorption at 280 nm (solid lines) and 330 nm (dotted lines). (C) SDS/ PAGE analysis of the crude extract and the purified Papilio RBP. Ó FEBS 2003 Retinol-binding protein in butterfly eye (Eur. J. Biochem. 270) 2439 This suggests that the protein exists in a monomeric state in vivo. Biochemistry of Papilio RBP To determine the native ligand of the fluorescing protein, we extracted retinoids from the purified fluorescing protein collected from dark-adapted Papilio eyes, and analyzed the composition of retinoids with HPLC (Fig. 2). It appeared that the protein exclusively binds 3-hydroxyretinol. We therefore call the protein Papilio retinol-binding protein, or Papilio RBP. HPLC analysis demonstrated that protein prepared from dark-adapted animals contained the all-trans isomer as the major ligand, but significant amounts of 11-cis and 13-cis isomers were also detected. The UV-induced fluorescence of the Papilio RBP disap- peared after the intense UV-irradiation, probably because the ligand was degraded. To investigate whether Papi- lio RBP has the ability to bind exogenous retinol, we supplied, after irradiating the soluble fraction of the Papilio retina with UV light, all-trans-or13-cis-retinol to the fraction, and analyzed the fluorescence of the proteins with native PAGE. As shown in Fig. 3, both all-trans-and13-cis- retinols restored the fluorescence of Papilio RBP. This result indicated that the protein could bind the exogenously added retinols in vitro, irrespective of their isomeric form. We next performed spectrophotometry and spectro- fluorometry of the Papilio RBP. Besides the principal peak at 280 nm, corresponding to the absorption of the apopro- tein, the absorbance spectrum (Fig. 4A) of the Papilio RBP has a secondary peak at 330 nm, corresponding to the absorption of 3-hydroxyretinol. The rather broad emission spectrum elicited by 330-nm light (Fig. 4B), peaks at 480 nm, and is very similar to that of free 3-hydroxyretinol [24]. This indicates that the binding of the apoprotein has little influence on the fluorescence profile of 3-hydroxy- retinol. The excitation spectrum (Fig. 4B), measured at an emission wavelength of 480 nm, shows two maxima at 332 nm and at 280 nm. The principal peak at 332 nm corresponds to the absorbance spectrum of the ligand, 3-hydroxyretinol. The distinct secondary peak at 280 nm indicates energy transfer from the apoprotein to the ligand. Primary structure of Papilio RBP To determine the primary structure of the identified Papilio RBP, we first analyzed the amino acid sequences of lysyl- endopeptidase-digested fragments of purified protein. Based on the sequence results, we designed oligonucleotide primers and carried out RT-PCR to amplify fragments of cDNA encoding the protein, and determined its nucleotide sequence. Subsequently, we performed 3¢-and5¢-RACE protocols, and obtained the complete nucleotide sequence of the full-length cDNA encoding the protein (Fig. 5). The cDNA is approximately 1 kb in length, and contains an open reading frame of 708 bases encoding 235 amino acid residues. A stop codon (TAA at nucleotides )9to)7) precedes the ATG at nucleotides 1–3, suggesting that the coding region begins at this ATG. A polyadenylation signal, AATAAA, exists 16 bases upstream from the start of the poly(A) + tail. Fig. 2. HPLC analysis of the intrinsic ligand of Papi lio RBP. The lig- ands were extracted from Papilio RBP purified from the soluble fraction of the dark-adapted retina. Extraction and analysis were carried out under dim red light as follows. Purified protein was first mixedwith2 M hydroxylamine and cold 90% methanol to convert aldehydes, if any, to oximes. Retinoids and oximes were then extracted with dichloromethane and n-hexane, and separated by normal phase HPLC. Eluent was monitored for absorbance at 340 nm. Each isomer of 3-hyroxyretinaloxime and 3-hydroxyretinol was identified by its retention time compared to that of the standard compound. Purified Papilio RBP exclusively binds 3-hydroxyretinol. No isomers of 3-hydroxyretinal (detectable as 3-hydroxyretinaloxime, if present) were detected. AT, all-trans 3-hydroxyretinol; 13, 13-cis 3-hydroxyretinol; 11, 11-cis 3-hydroxyretinol. Fig. 3. Binding of exogenous ligands to Papilio RBP. Fluorescence of native PAGE (left) and Coomassie Brilliant Blue (CBB) stained gel (right). Soluble proteins from the dark-adapted Papilio retina (lanes 1 and 1¢) were irradiated with intense UV-light to degrade intrinsic ligand (lanes 2 and 2¢). To the UV-irradiated samples, all-trans (lanes 3 and 3¢)or13-cis (lanes 4 and 4¢) retinol was added, followed by 20-min incubation on ice. In this experiment, retinol was used for 3-hydroxyretinol. 2440 M. Wakakuwa et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Figure 5 shows the amino acid sequence of the Papi- lio RBP deduced from the cDNA sequence. The relative molecular mass of the protein calculated from the sequence was 26 412. This value is somewhat smaller than that of purified Papilio RBP as estimated by SDS/PAGE (31 kDa, Fig. 1). However, this difference is in the range of experi- mental variation: we have often encountered such an overestimation of molecular mass with SDS/PAGE when the proteins are negatively charged (squid RALBP [11]; lipophilic ligand-binding protein in the fly chemosensory hair [26]). The calculated pI value of the Papilio RBP is 4.92; the protein is highly acidic. This explains the high mobility of the protein in the native PAGE (Fig. 1). In order to determine the N-terminal sequence of the Papilio RBP in vivo, we sequenced intact Papilio RBP without lysyl-endopeptidase digestion. We acquired a sequence, XSRIYPKVWS, although the recovery rate was extremely low. This result indicates that the Papilio RBP undergoes post-translational modification: the N-terminal methionine is removed and the second residue, serine, carries some blocking residue. In addition, we could not identify any N-terminal signal sequence for secretion. Therefore, the Papilio RBP is most likely located in the cytoplasm. Based on the deduced amino acid sequence, we searched for homologous proteins in databases. Two partial sequences of the Papilio RBP, each consisting of less than 60 residues, showed low (<30%) identity to those of the chlorophyll-a/b-binding proteins and N-acetyltransferases. No protein was found that has significant similarity to the full length of Papilio RBP. Therefore, we conclude that the Papilio RBP is a member of a novel protein family. Papilio RBP in dark- and light-adapted eyes To address the question of whether Papilio RBP is involved in the visual cycle, we investigated the isomer composition of the ligand and the distribution of the protein in dark- and light-adapted eyes. In dark-adapted eyes, the molar ratio of all-trans,11-cis and 13-cis isomer was 48 : 39 : 13 (Fig. 6). When light-adapted, the ratio changed to 28 : 60 : 12, i.e. the fraction of 11-cis isomer significantly increased, whereas that of all-trans isomer decreased (Fig. 6). As the illumin- ation of purified RBP in vitro did not isomerize all-trans ligandtothe11-cis form (data not shown), it is clear that the light-induced change in isomer composition in vivo is not due to the direct isomerization of the all-trans ligand by Papilio RBP. Instead, it is possibly interpreted as a result of replacement of all-trans ligand with 11-cis isomer newly produced in the light-adapted eyes. In order to investigate the distribution of Papilio RBP in the compound eye, we divided the eye into two portions, distal and proximal, by gently pulling the retina off from the corneal cuticle with fine forceps. Figure 7A (right) shows a plastic section of the tissue layer containing the cornea, i.e. the distal layer. This layer contains about one-third of the photoreceptor layer as follows from a comparison with a section of the intact eye (Fig. 7A, left). We then prepared crude protein extracts separately from the distal and the proximal portions of the dark-adapted eye, and analyzed them with native PAGE. As shown in CBB-stained gel (Fig. 7B, left), both portions appeared to contain similar amounts of RBP. Also, the fluorescence intensity (Fig. 7B, right) is similar in both portions, indicating that the RBP binds the ligand ubiquitously. When the eyes were light- adapted, however, the amount of RBP, as well as the fluorescence of the ligand, increased in the distal portion and decreased in the proximal portion. This strongly suggests that the RBP together with its ligand migrates distally upon light adaptation. We next analyzed the isomer composition of the native ligands of Papilio RBP extracted separately from the distal and the proximal portions of the retina (Fig. 8). As expected from the fluorescence image analysis of the native PAGE (Fig. 7B), total amount of 3-hydroxyretinol was increased in the distal portion, and decreased in the proximal portion, by light adaptation of the eye. Light adaptation also induced the decrease of all-trans isomer both in the distal and proximal portions of the retina. In contrast, the increase of 11-cis ligand was observed in the distal but not in the proximal portion of the retina. Together with the results from the native PAGE (Fig. 7B), these findings strongly suggest that Papilio RBP exchanges its ligand from all- trans- to 11-cis-3-hydroxyretinol by light adaptation, and migrates from the proximal to the distal region within the retina. Fig. 4. Spectrophotometric and spectrofluorometric characteristics of Papilio RBP. Absorbance (A) and fluorescence excitation and emission (B) spectra of Papilio RBP were measured on the protein purified by native PAGE. The excitation spectrum was measured via emission at 480 nm, and emission spectrum was measured using excitation light at 330 nm. Ó FEBS 2003 Retinol-binding protein in butterfly eye (Eur. J. Biochem. 270) 2441 Discussion A novel retinol-binding protein, Papilio RBP We identified a novel type of protein that binds retinol, whichwetermedthePapilio retinol binding protein (Papilio RBP). The native ligand of this protein is 3-hydroxyretinol, whose isomer composition varies between dark- and light- adaptation. The deduced amino acid sequence of Papi- lio RBP shows no overall similarity to any other proteins so far described. However, part of the sequence has some similarity to that of chlorophyll-a/b-binding protein, whose ligand is also hydrophobic [27]. Furthermore, the hydro- phobicity profile of the C-terminal half of Papilio RBP resembles that of the human CRBP [28]. These results suggest that the binding proteins share some three-dimen- sional structure that is crucial for binding hydrophobic ligands. Biochemical studies suggest that the butterfly visual cycle may share a common system with the fly visual cycle (see Introduction). Nevertheless, no isoform of Papilio RBP could be found through the database search of the Drosophila genome. We demonstrated previously that a lipophilic ligand-binding protein in the fly chemosensory organs probably had conformational and functional simi- larities to the general odorant-binding protein in the moths, while the similarity between their amino acid sequences was very low [26]. This suggests that, in these ligand-binding proteins, amino-acids may be highly variable unless their protein conformation required for ligand binding is disrup- ted. Further knowledge on the protein structure of the Papilio RBP would be essential to address above question. Possible function of the Papilio RBP What is the biological function of the Papilio RBP? First, it is important to realize that free retinoids are highly labile and Fig. 5. cDNA and deduced amino acid sequence of Papilio RBP. The cDNA (923 bp) encodes an open reading frame for full-length Papilio RBP (708 bp, 235 amino acid resi- dues). The calculated M r and pI values are 26 412 and 4.92, respectively. Amino acid sequences revealed by sequencing the peptides with lysyl-endopeptidase digestion are underlined. Dotted underline indicates the N-terminal sequence obtained by sequencing the intact Papilio RBP. Underline in the 3¢-noncoding region shows a possible polyadenylation signal. 2442 M. Wakakuwa et al. (Eur. J. Biochem. 270) Ó FEBS 2003 possess various biological activities [1]. When stored in tissues, these labile and bioactive molecules need to be stabilized and inactivated. One way to achieve this is to bind to hydrophilic proteins. Apparently, the Papilio RBP iso- lated from the soluble fraction of the eye is hydrophilic and contributes to stabilize and inactivate 3-hydroxyretinol, the native ligand of the protein. We suspect that the primary function of Papilio RBP is involved in the visual cycle. The chromophore of the Papilio rhodopsins is 11-cis 3-hydroxyretinal [22]. The chromo- phore is converted into the all-trans form upon light absorption by a rhodopsin molecule. This photoconversion of rhodopsin to metarhodopsin triggers the phototransduc- tion process, resulting in a change in the photoreceptor membrane potential (depolarization). To maintain light sensitivity, photoreceptors have to restore the rhodopsin content. Metarhodopsins of arthropods are usually thermo- stable; opposite to vertebrate metarhodopsin. Therefore, they have enough time to reabsorb light and change back to rhodopsin by photoregeneration. However, invertebrate retinas possess additional proces- ses for rhodopsin recycling that reuses all-trans retinal released from metarhodopsin. Indeed, it has been demon- strated that metarhodopsin is bleached more rapidly than rhodopsin in the fly and butterfly retinas [16,21]. In Papilio, Shimazaki and Eguchi have proposed a process of rhodop- sin regeneration based on the HPLC analysis of retinoid in the eye [23,29]. According to their hypothesis, the isomeri- zation process has taken place when the chromophore is separated from opsin as in vertebrates. All-trans-3-hydroxy- retinol is somehow stored in the distal portion of the eye. The stored 3-hydroxyretinol is oxidized into all-trans-3- hydroxyretinal that is subsequently isomerized to the 11-cis form by light, and finally binds to opsin to form rhodopsin. In addition, they suggested that all-trans-3-hydroxyretinol in the proximal portion of the photoreceptor cell is transported to the distal portion to facilitate biogenesis of rhodopsins. Note that the present findings, namely the distal–proximal localization of retinoids (Fig. 7) and isomer composition (Fig. 8), basically matches the hypothesis Fig. 7. Light-induced relocation of Papilio RBP in the retina. (A) Longitudinal sections of the intact (left) and the distal portion (right) of the Papilio eye. (B) Native PAGE indicating the distribution of Papi- lio RBP in the distal and proximal portions of the retina. Eyes from dark-adapted and light-adapted animals were divided into the distal and proximal portions. Soluble proteins were extracted from each portion, and separated by native PAGE. The fluorescence of Papi- lio RBP was recorded under UV-illumination (right), and the proteins in the gel were stained with Coomassie Brilliant Blue (CBB) (left panel). The relative contents of the ligand and apoprotein were esti- mated via the intensity of fluorescence and the density of Coomassie Brilliant Blue, respectively. The ligand or apoprotein content in the distal portion (D) was compared with that in the proximal portion (P). Mean ± SEM (n ¼ 4) of the D/P ratio are shown at the bottom of the corresponding electrophoresis records. Fig. 6. Light-induced change in isomer composition of the intrinsic ligand of Papilio RBP. Ligands were extracted from Papilio RBP purified from dark-adapted or light-adapted retinas, and analyzed by HPLC. The molar ratio of all-trans,11-cis and 13-cis 3-hydroxyretinol was then calculated based on the absorbance and the molar extinction coefficient at 340 nm of each isomer. Mean ±SEM of three separate experiments are presented. **P <0.01(one-way ANOVA – Tukey test). Ó FEBS 2003 Retinol-binding protein in butterfly eye (Eur. J. Biochem. 270) 2443 proposed by Shimazaki and Eguchi [23,29]. The majority of the Papilio RBP ligand exists in the all-trans form, in dark- adapted eyes, and is then transformed to the 11-cis form when eyes are light-adapted. Light adaptation causes relocalization of the Papilio RBP from the proximal to the distal part of the retina. Coincidence of the present data with the hypothesis strongly suggests that the Papilio RBP is involved in the visual cycle. In order to elucidate how Papilio RBP functions in the visual cycle, it would be important to clarify the precise localization of the protein in the distal and proximal portions of the light- or dark- adapted retinas. Immunohistochemical localization of Papi- lio RBP would be adequate to address this question. Previously, we reported that the Papilio eye contains the ommatidia that fluoresce under UV [30]. The fluorescing ommatidia have a concentration of fluorescing material, most probably 3-hydroxyretinol that acts as a UV- absorbing spectral filter for the underlying photoreceptors [24]. However, the fluorescence is restricted to the distal 70 lm of the photoreceptor layer, and not detectable in the proximal portion. In addition, the fluorescence is extremely labile: it disappears in seconds under epi-fluorescence microscopy [24]. These results of fluorescence microscopy are not explained by the present features of the Papilio RBP. On the other hand, HPLC analysis of retinoids demonstrated that the amount of all-trans isoform extrac- ted from the whole eye tissue was greater than that from the purified Papilio RBP (data not shown). In contrast, there are no significant differences in the amount of other isomers between the whole eye tissue and purified samples. This result suggests that there may be unknown storage of all-trans-3-hydroxyretinol that is not bound to the Papi- lio RBP and is related to the ommatidial fluorescence. Immunohistochemistry would provide further insight on this issue. Acknowledgements We thank D. G. Stavenga for critical reading of the manuscript. We also thank S. Kawamura for useful discussion and kind permission to use analytical instruments. This work was supported partly by the Sasagawa Research Grant to M. W., and the Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science to K. O and K. A. References 1. Noy, N. (2000) Retinoid-binding proteins: mediators of retinoid action. Biochem. J. 348, 481–495. 2. Gonzalez-Fernandez, F. (2002) Evolution of the visual cycle: the role of retinoid-binding proteins. J. Endocrinol. 175, 75–88. 3. Stavenga, D.G., Schwemer, J. & Hellingwerf, K.J. (1991) Photo- receptor Evolution and Function (Holmes, M.G., eds), Academic Press, London. 4. Saari, J.C. (1999) Retinoids: the Biochemical and Molecular Basis of Vitamin A and Retinoids Action (Nau, H. & Blaner, W.S., eds), pp. 563–588. Springer-Verlag, Berlin. 5.McBee,J.K.,Palczewski,K.,Baehr,W.&Pepperberg,D.R. (2001) Confronting complexity: the interlink of phototransduction andretinoidmetabolisminthevertebrateretina.Prog. Retin. Eye Res. 20, 469–529. 6. Deigner, P.S., Law, W.C., Canada, F.J. & Rando, R.R. (1989) Membranes as the energy source in the endergonic transformation of vitamin A to 11-cis-retinol. Science 244, 968–971. 7. McBee, J.K., Kuksa, V., Alvarez, R., de Lera, A.R., Prezhdo, O., Haeseleer, F., Sokal, I. & Palczewski, K. (2000) Isomerization of all-trans-retinol to cis-retinols in bovine retinal pigment epithelial cells: dependence on the specificity of retinoid-binding proteins. Biochemistry 39, 11370–11380. 8.Yang,M.&Fong,H.K.W.(2002)Synthesisoftheall-trans- retinal chromophore of retinal G-protein-coupled receptor opsin in cultured pigment epithelial cells. J. Biol. Chem. 277, 3318–3324. 9. Hara, T. & Hara, R. (1991) Progress in Retinal Research (Osborne, N. & Chader, G.J., eds), pp. 179–206. Pergamon Press, Oxford, New York. 10. Ozaki, K., Terakita, A., Hara, R. & Hara, T. (1987) Isolation and characterization of a retinal-binding protein from the squid retina. Vision Res. 27, 1057–1070. 11. Ozaki, K., Terakita, A., Ozaki, M., Hara, R., Hara, T., Hara- Nishimura, I., Mori, H. & Nishimura, M. (1994) Molecular characterization and functional expression of squid retinal-binding protein. A novel species of hydrophobic ligand-binding protein. J. Biol. Chem. 269, 3838–3845. 12. Terakita, A., Hara, R. & Hara, T. (1989) Retinal-binding protein as a shuttle for retinal in the rhodopsin-retinochrome system of the squid visual cells. Vision Res. 29, 639–652. 13. Ozaki, K., Terakita, A., Hara, R. & Hara, T. (1986) Rhodopsin and retinochrome in the retina of a marine gastropod, Conomulex luhuanus. Vision Res. 26, 691–705. 14. Katagiri, N., Terakita, A., Shichida, Y. & Katagiri, Y. (2001) Demonstration of a rhodopsin-retinochrome system in the stalk eye of a marine gastropod, Onchidium, by immunohistochemistry. J. Comp. Neurol. 433, 380–389. Fig. 8. Light-induced changes in isomer composition of the ligand of Papilio RBP in distal and proximal portions. Ligands were extracted from Papilio RBP purified separately from distal and proximal por- tions of the light- and dark-adapted retinas, and analyzed by HPLC. The molar ratio of all-trans,11-cis and 13-cis 3-hydroxyretinol was then calculated based on the absorbance and the molar extinction coefficient at 340 nm of each isomer. Mean ± SEM of three separate experiments were presented. 11-cis isomer significantly increased in the distal portion of the light-adapted retina. **P <0.01(one-way ANOVA –Tukeytest). 2444 M. Wakakuwa et al. (Eur. J. Biochem. 270) Ó FEBS 2003 15. Robles, L.J., Camacho, J.L., Torres, S.C., Flores, A., Fariss, R.N. & Matsumoto, B. (1995) Retinoid cycling proteins redistribute in light-/dark-adapted octopus retinas. J. Comp. Neurol. 358, 605–614. 16. Schwemer, J. (1984) Renewal of visual pigment in photoreceptors of the blowfly. J. Comp. Physiol. [A]. 154, 535–547. 17. Schwemer, J. (1988) Cycle of 3-hydroxy retinoids in an insect eye. In Molecular Physiology of Retinal Proteins (Hara, T., ed.), pp. 299–304. Yamada Science Foundation, Osaka. 18. Schwemer, J. (1989) Facets of Vision (Stavenga, D. G. & Hardie, R. C., eds), pp. 112–133. Springer-Verlag, Berlin, Heidelberg. 19. Pepe, I.M., Cugnoli, C., Peluso, M., Vergani, L. & Boero, A. (1987) Structure of a protein catalyzing the formation of 11 cis-retinal in the visual cycle of invertebrate eyes. Cell Biophys. 10, 15–22. 20. Pepe, I.M., Schwemer, J. & Paulsen, R. (1982) Characteristics of retinal-binding proteins from the honeybee retina. Vision Res. 22, 775–781. 21. Bernard, G.D. (1983) Bleaching of rhabdoms in eyes of intact butterflies. Science 219, 69–71. 22. Seki, T., Fujishita, S., Ito, M., Matsuoka, N. & Tsukida, K. (1987) Retinoid composition in the compound eyes of insects. Exp. Biol. 47, 95–103. 23. Shimazaki, Y. & Eguchi, E. (1993) Synthesis of 3-hydroxy- retinal in the cytosol of the butterfly compound eye. Vision Res. 33, 155–163. 24. Arikawa, K., Mizuno, S., Scholten, D.G., Kinoshita, M., Seki, T., Kitamoto, J. & Stavenga, D.G. (1999) An ultraviolet absorbing pigment causes a narrow-band violet receptor and a single-peaked green receptor in the eye of the butterfly Papilio. Vision Res. 39, 1–8. 25. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 26. Ozaki, M., Morisaki, K., Idei, W., Ozaki, K. & Tokunaga, F. (1995) A putative lipophilic stimulant carrier protein commonly found in the taste and olfactory systems. A unique member of the pheromone-binding protein superfamily. Eur. J. Biochem. 230, 298–308. 27. Larouche, L., Tremblay, C., Simard, C. & Bellemare, G. (1991) Characterization of a cDNA encoding a PSII-associated chloro- phyll a/b-binding protein (CAB) from Chlamydomonas moewusii fitting into neither type I nor type II. Curr. Genet. 19, 285–288. 28. Colantuoni, V., Cortese, R., Nilsson, M., Lundvall, J., Bavik, C.O., Eriksson, U., Peterson, P.A. & Sundelin, J. (1985) Cloning and sequencing of a full length cDNA corresponding to human cellular retinol-binding protein. Biochem. Biophys. Res. Commun. 130, 431–439. 29. Shimazaki, Y. & Eguchi, E. (1995) Light-dependent metabolic pathway of 3-hydroxyretinoids in the eye of a butterfly, Papilio xuthus. J. Comp Physiol. (A). 176, 661–671. 30. Arikawa, K. & Stavenga, D. (1997) Random array of colour filters in the eyes of butterflies. J. Exp. Biol. 200, 2501–2506. Ó FEBS 2003 Retinol-binding protein in butterfly eye (Eur. J. Biochem. 270) 2445 . A novel retinol-binding protein in the retina of the swallowtail butterfly, Papilio xuthus Motohiro Wakakuwa 1 , Kentaro Arikawa 1 and Koichi Ozaki 2 1 Graduate. a novel retinol-binding protein in the eye of a butterfly, Papilio xuthus. The protein that we term Papilio retinol-binding protein (Papilio RBP) is a major