INTRODUCTION

In the photoreceptor cells of animals' eyes, visual pigments trigger a G-protein-mediated phototransduction cascade, which eventually generates an electrical response of the cells. Two kinds of the phototransduction systems have been reported. One is the G_t¹-mediated system of vertebrate hyperpolarizing photoreceptor cells in which the visual pigment activates a cGMP-specific phosphodiesterase via a heterotrimeric G-protein, transducin (G_t) (1-3). The other is the G_q-mediated system of invertebrate depolarizing cells, such as cephalopod's and arthropod's, where phospholipase C is activated via a G_q-type G-protein (4-7). The visual pigments of these two systems show sequence homology, but clearly split into two subtypes (G_t- and G_q-coupled ones) in a molecular phylogenetic tree (8).

In addition to the depolarizing rhabdomeric photoreceptor cells present in invertebrates, scallop retinas contain ciliary photoreceptor cells that hyperpolarize in response to light (9). After the first electrophysiological recordings of Hartline (10), the mechanism of the hyperpolarizing response as well as its evolution have been discussed in comparison with vertebrate hyperpolarizing ciliary photoreceptor cells, rods and cones (11-14). It has been reported, however, that the hyperpolarizing response in the scallop cell is due to opening of a cGMP-sensitive potassium channel (11-14), which is different from that of the vertebrate cells (closing of a cGMP-sensitive cationic channel) (15). Our immunohistochemical experiments showed that an antibody to frog G_t did not cross-react to the scallop hyperpolarizing cell.² Therefore we expected the presence of an unknown G-protein-mediated phototransduction cascade other than a G_t-mediated one in the scallop hyperpolarizing cells, while the depolarizing photoreceptor cells contain a G_q-mediated cascade. Here, we show evidence that the phototransduction system in the invertebrate hyperpolarizing photoreceptor cells is not mediated by G_t or G_q, but rather uses a novel, G_o-mediated cascade.

EXPERIMENTAL PROCEDURES

Two gene fragments encoding a partial peptide of visual pigment homologues (SCOP1 and SCOP2) were obtained by PCR of scallop (Patinopecten yessoensis) genomic DNA with degenerated oligonucleotide primers: 5-TTYHTIHTIGCITRIACICCITAYRC-3 and 5-AYISCRTAIAYIADIGGRTT-3, where "I" stands for inosine, "R" for A/G, "H" for T/C/A, "D" for T/G/A, "Y" for T/C, and "S" for G/C. The primers were designed on the basis of comparison of the known amino acid sequences of visual pigments. As for each of the two genes, 5- and 3-rapid amplification of cDNA ends with the gene-specific oligonucleotide primers were accomplished with RNA prepared from scallop mantle eyes, to obtain the whole sequence of the protein cDNA. To confirm that the sequence thus obtained comes from a single species of mRNA, scallop eye cDNA was subjected to PCR to amplify the entire coding region of the proteins (1497 base pair (bp) for SCOP1 and 1197 bp for SCOP2). The cDNA sequence was confirmed at least twice for each gene.

G cDNA of Scallop Eyes

To obtain cDNA fragments encoding a region containing "helical domain" of G where the primary sequence is characteristic of each G subtype, PCR of scallop eye cDNA was carried out with a set of the degenerated oligonucleotide primers: 5-GGIAARWSIACIWTHRTIAARCARATG-3 and 5-AARCAITGDAWCCAYTT-3, where "W" stands for T/A. The primers were designed on the basis of comparison of the known amino acid sequences of G. The entire coding region for either G_q (1058 bp) or G_o (1071 bp), which localized in the photoreceptor cells (see "Results and Discussion"), was sequenced as described above for the visual pigment genes.

The ~0.2-kilobase pair DNA fragments were subcloned from the cDNA of SCOP2, G_o and G_q, using PCR. Digoxigenin-labeled antisense RNA probes were synthesized from these subclones, using the Dig RNA labeling kit (Boehringer Mannheim). The RNA probe for SCOP2 was a mixture of antisense RNAs from six separate regions of the cDNA (55-256, 259-458, 456-656, 659-858, 859-1058, and 1068-1268, which are base positions of the sense strand and numbered from the initiation ATG). The RNA probe for G_o was from the 838-1039 region of the cDNA, and the probe for G_q was from the 516-742 region of the cDNA. Mantle eyes dissected from the scallops were immersion-fixed in 4% paraformaldehyde, frozen with O.C.T. Compound (Miles Inc.), and sectioned at 10 µm. The eye sections were pretreated with proteinase K and hybridized with the antisense RNA probe. The probe on the sections was detected using alkaline phosphatase-conjugated anti-digoxigenin (Boehringer Mannheim) by a blue 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium color reaction.

Each of the peptides of helical domain encoded by G cDNA fragments was expressed in Escherichia coli and purified, using the QIAexpress kit (Qiagen), and used as antigen. BALB/c female mice were immunized four times with 100 µg of the peptides.

The eye slices prepared as described above were treated in order with diluted antisera (see figure legends), with a biotinylated goat anti-mouse IgG and with an avidin-biotinylated horseradish complex (Vector Laboratories), and visualized using diaminobenzidine (Sigma).

The multiple alignment of amino acid sequences of visual pigments was done with the aid of the Clustal W 1.4 program (16). Sequences were obtained from GenBank^TM and PIR data bases. The aligned sequences (235 amino acid residues for each sequence) included all the residues except for the N- and C-terminal fragments and some of the loop domains (loops IV-V, V-VI, and VI-VII as shown in Fig. 1). A tentative phylogenetic tree for the alignment was calculated with the Neighbor-Joining method by use of the PHYLIP 3.572 software package (17). The tree topology thus obtained was subjected to Maximum Likelihood analysis (ProtML) using the MOLPHY 2.3 software package (18) with the local rearrangement option and with the JTT-F option of amino acid substitution model. The bootstrap probabilities (percent) of local topologies were also estimated for the final tree with the maximum likelihood.

RESULTS AND DISCUSSION

Since phylogenetic analysis of visual pigments can roughly suggest the specific subclass of G-protein to which they couple (8), we examined into which subtypes the scallop visual pigments are classified. From the scallop eyes' cDNA, two kinds of cDNAs of visual pigments were obtained by several steps of PCR, and tentatively designated as SCOP1 and SCOP2. The amino acid sequence of SCOP1 was highly similar to those of squid and octopus G_q-coupled rhodopsins (hereafter referred to as G_q-rhodopsins). Since an antibody against the squid G_q-rhodopsin cross-reacted to the scallop depolarizing photoreceptor cells but not to the hyperpolarizing cells (data not shown), SCOP1 could be the cDNA of G_q-rhodopsin in the depolarizing cells. On the other hand, the amino acid sequence of SCOP2 (Fig. 1) did not show significant similarities to those of any other visual pigments, although SCOP2 had the functional amino acid residues conserved among all the known visual pigments. The molecular phylogenetic tree (Fig. 2) clearly indicated that SCOP2 was not a member of the G_t- nor G_q-coupled group of visual pigments. Furthermore, in the loop region between helices V and VI (Fig. 1), which is one of the G-protein interaction sites in bovine rhodopsin (19), SCOP2 was quite different in amino acid sequence from either G_t- or G_q-coupled visual pigments. These data strongly suggest that SCOP2 is a new subtype of visual pigment that couples with a G-protein other than G_t and G_q.

To test this prediction, we next investigated which subtype of G-protein was colocalized with SCOP2 in the hyperpolarizing cells. The cDNA fragments of five kinds of G encoding their helical domains were obtained by PCR against the scallop eyes' cDNA. On the basis of sequence similarities, four of them were classified as G_q, G_s, G_i, and G_o, respectively, while the other one was a new subtype (Table I). Then we prepared the specific antibody against each of the peptides encoded by these cDNA fragments. Among these antibodies, anti-G_q and anti-G_o antibodies strongly reacted with the scallop photoreceptor cells by immunohistochemical experiments (Fig. 3). The scallop retina has two layers of photoreceptor cells: the hyperpolarizing cell layer with its photoreceptive region adjacent to the lens and the depolarizing cell layer distant from the lens (20). As expected, anti-G_q antibody clearly stained the rhabdomeric depolarizing cells. Interestingly, anti-G_o antibody specifically stained the hyperpolarizing cells. The specificity of anti-G_q and G_o antibodies were confirmed by immunoblot analysis; they specifically reacted with 42- and 41-kDa peptides in ocular proteins, respectively (Fig. 4). The molecular weights were consistent with those calculated based on the whole amino acid sequences deduced from the cDNAs of the scallop G_q and G_o. The scallop G_o shows 83% identity in amino acid sequence to mammalian G_o and 90% to Drosophila G_o, and they show complete identical sequences in C terminus region, which is characteristic of each G subtype.

Table I. Percentage of amino acid identity between scallop and other G Amino acid identity, scallopa Gs Gq Go Gi G? % Drosophila   Gs 58 39 37 32 30   Gq 36 69 45 44 34   Go 42 41 87 54 36   Gi 35 45 55 58 41   Gf 36 34 34 34 34 Human   Gs 52 30 33 37 29   Gq 40 66 44 46 36   Go 40 46 80 60 39   Gi 39 45 61 63 39   Gt 36 44 53 53 41   Gz 35 38 50 50 37   G12 30 37 36 36 31 a Five kinds of cDNA fragments of G encoding their helical domains were obtained by PCR against the scallop eye cDNA. Based on the deduced amino acid sequence identities, they were classified as Gq, Gs, Gi, and Go, respectively, while the other one was a new subtype (G?).

Immunohistochemical localization of scallop G.

o

q

o

q

o

q

To further confirm the colocalization of SCOP2 with the G_o subtype of G-protein, we performed in situ hybridization experiments (Fig. 5). The results show that SCOP2 coexpresses with G_o, but not with G_q, in the hyperpolarizing cells. It should be noted that the G_o is localized only in the outer segment (photoreceptive region) of the cells in the immunohistochemical experiments (Fig. 3). Taken together, these results indicate that SCOP2 triggers a G_o-mediated phototransduction cascade in the hyperpolarizing cell. Therefore, we propose SCOP2 be named scallop G_o-rhodopsin.

o

q

Our results indicated that the phototransduction system leading to the scallop hyperpolarizing response is different from that in the vertebrate hyperpolarizing cells (G_t-mediated system). This indication is consistent with recent electrophysiological findings on their channels; the scallop hyperpolarizing response originates from opening of the potassium channel (11-13) by increase of cytosolic cGMP (14), whereas vertebrate cells respond with closing of the cationic channel resulting from decrease of cGMP (15). Thus, it is most likely that the scallop G_o-mediated phototransduction cascade couples with an effector enzyme, probably a guanylyl cyclase, to elevate cytosolic cGMP concentration.

The scallop G_o shows high similarity in amino acid sequence to mammalian G_o, which is localized mainly to brain and nervous cells (21, 22). The sequence similarities suggest that the G_o phototransduction cascade is similar to a G_o-cascade in the nervous system. However, little is known about a specific effector enzyme(s) directly coupling with G_o, although it has been reported that mammalian G_o, especially its subunit, is involved in regulating voltage-sensitive calcium channels in the synaptic region of neuronal cells (23-26). The phototransduction in the scallop hyperpolarizing cells can be an alternative system to identify the G_o-coupled effector enzyme because of the highly specific expression of G_o with photoactivable receptor proteins (G_o-rhodopsin) in the cells.

The molecular phylogenetic tree of visual pigments (Fig. 2) strongly suggests that G_o-rhodopsin diverged from an ancestral visual pigment before the divergence of animals (700 million years ago) into higher invertebrates (deuterostomia) and vertebrates (protostomia). This was also supported by the fact that visual pigment-like proteins (RGR), recently found in mammals (27), clustered with G_o-rhodopsin (at the point of the closed triangle in Fig. 2) with relatively high bootstrap probability (89%) when these extra members were added to the tree. The deepest root of the tree (closed circle in Fig. 2) corresponds to the generation of three different genes for visual pigments and not to the divergence of animal species. Thus the multiple phototransduction systems of vision may have emerged before the divergence of animals into vertebrates and invertebrates in the course of evolution. It is likely that for some time following this divergence, both vertebrates and invertebrates kept each of the multiple phototransduction systems. The G_o-mediated phototransduction system might spread over a wide variety of animal species, since the mechanism of photoresponse in the scallop hyperpolarizing cells appears to be shared by some other species; the ciliary photoreceptor cells of a lizard "parietal eye" respond to light by increasing cytosolic cGMP and opening channels (28), suggesting the presence of a similar G_o-mediated phototransduction system.