A Novel Gγ Isolated from Drosophila Constitutes a Visual G Protein γ Subunit of the Fly Compound Eye*

Visual transduction in the compound eye of flies is a well established model system for the study of G protein-coupled transduction pathways. To characterize key components of the phototransduction cascade we performed substractive hybridization screening. We cloned the cDNA coding for the visual Gγ (Gγe) subunit from Drosophila which had so far eluded identification at the molecular level. Northern blot analysis revealed the presence of a major, 1.4-kilobase(kb) Gγe transcript and two minor transcripts of 1.8 and 6 kb in size. The major 1.4-kb mRNA is expressed preferentially in the eye. The spatial expression pattern determined for Gγe as well as co-immunoprecipitation experiments demonstrated that Gγe dimerizes with Gβe to form the heterodimeric Gβγ subunit which functions in visual transduction in the Drosophila compound eye. Gγe shares common characteristics with the visual Gγ subunits of human rod and cone photoreceptors although different classes of Gα subunits are employed in vertebrate and invertebrate phototransduction. By the molecular cloning and characterization of the visual γ subunit ofDrosophila one of the few missing links in the well studiedDrosophila phototransduction cascade has been characterized to complete our knowledge about the Drosophila visual transduction pathway.

The visual cascade in the microvillar photoreceptors of the Drosophila compound eye is activated by light absorption of five rhodopsins which are differentially expressed in a distinct pattern within an ordered array of eight photoreceptor cells (1). The G protein couples light activation of rhodopsin to the activation of the main target enzyme of the phototransduction cascade, the phosphoinositide-specific phospholipase C␤ (norpA) 1 (2)(3)(4)(5). At this level the transduction pathways activated by either one of the five distinct rhodopsins converge to a unitary cascade. As a result of G protein activation phosphoinositide-specific phospholipase C␤ hydrolyzes the membrane lipid phosphatidyl inositol bisphosphate to form the second messengers diacylglycerol and 1,4,5-inositol trisphosphate (6,7). Phosphoinositide-specific phospholipase C␤ itself is assembled by the PDZ domain protein INAD into a signaling complex together with the major light-activated Ca 2ϩ channel transient receptor potential and the eye-specific protein kinase C (8 -14).
A detailed understanding of the activation mechanism transmitting rhodopsin activation to the INAD signaling complex is still missing. Part of the deficit in this knowledge is due to a lack of information on the structure of the visual G protein.
Photoreceptor-specifically expressed genes encoding the G␣ (G␣ q ) and G␤ (G␤ e ) subunits of the visual G protein have been isolated (15)(16)(17), whereas the visual G␥ (G␥ e ) subunit had eluded identification at the molecular level. The function of G␣ q and G␤ e in phototransduction has been studied to some extent by biochemical and genetic studies. Analysis of Drosophila mutants defective in G␣ q revealed a requirement of this G␣ subunit in the activation of the phototransduction cascade, demonstrating that G␣ q indeed is the visual G␣ subunit (18). While G␣ q is believed to activate phosphoinositide-specific phospholipase C␤, the main target of the visual ␤␥ subunit has not been identified yet. As has been shown by analyzing Drosophila G␤ e mutants, G␤ e is essential for phototransduction, especially for proper termination of the signaling cascade (19).
In the absence of any information on the existence of a distinct visual G␥ subunit, it has been suggested that Drosophila G␥ 1 , the only G␥ subunit cloned from Drosophila so far, might associate with diverse G proteins, including the visual G protein (20). G␥ 1 , however, is preferentially expressed in the brain. Here we report the isolation of a gene coding for a G␥ e subunit specifically expressed in the Drosophila eye. By immunoprecipitation experiments we show that G␥ e is associated with the visual G␤ subunit. Furthermore, the G␥ e subunit shares common characteristics with the vertebrate visual G␥ subunits of rod and cone photoreceptors.

EXPERIMENTAL PROCEDURES
Fly Stocks-Male Calliphora vicina Meig., chalky mutant, were raised at 25°C in a 12 h light/12 h dark cycle and were used for the experiments at an age of 8 -10 days after eclosion. The Drosophila fly strains were raised on a standard corn meal diet and were kept under a 12 h light/12 h dark cycle. Drosophila inaA P226 mutants were provided by W. L. Pak.
Differential Hybridization Screen-An oligo(dT)-primed cDNA library in vector ZAP II (Stratagene) was produced from poly(A ϩ ) RNA isolated from the compound eye of C. vicina. The differential hybridization screen of the library was carried out using cDNA probes derived from mRNA of retinal tissue, of muscle tissue, and with a mixture of cDNAs encoding previously cloned retina-specific genes (Rh1, Rh6, arr1, arr2, dgq, inaD, inaC, trp, trpl, and D19). The hybridizations were performed in 5 ϫ SSC (1 ϫ SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0), 0.1% laurylsarcosinate, 0.02% SDS, 1% blocking reagent (Roche Molecular Biochemicals) at 65°C according to standard protocols (21). For generating the retinal cDNA probe and the muscle cDNA probe, total RNA was isolated from retinas or flight muscles, respectively, using the TRIzol TM Reagent (Life Technologies). mRNA was isolated with the PolyATtract mRNA Isolation System (Promega) according to the manufacturer's instructions and 3 g of mRNA were transcribed into cDNA (Ready to Go Kit, Amersham Pharmacia Bio-* This work was supported by European Union Grant BMH4-CT97-2341. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ 250440 (DmG␥ e ) and AJ 250441 (CvG␥ e ).
Reverse Transcriptase-Polymerase Chain Reaction, cDNA Sequencing, and Sequence Analysis-Reverse transcriptase-polymerase chain reaction reactions were carried out using the Titan TM One Tube reverse transcriptase-polymerase chain reaction System (Roche Molecular Biochemicals). 1 g of total RNA isolated from heads of Drosophila inaA mutants was reverse transcribed into cDNA and the coding region of G␥ e was amplified using sequence-specific primers containing either an EcoRI or a XhoI restriction-site (5Ј-GGCTGAATTCCTCTTGTGTCT-GGGGTGGTAT-3Ј, 5Ј-TCGGCTCGAGGGCGGTATTTCTGTGGTTTA-CG-3Ј). The amplified product was cloned into the EcoRI/XhoI site of the pBluescript II SK vector (Stratagene) and sequenced. DNA sequencing was performed with an automated sequencer (Alfexpress, Amersham Pharmacia Biotech) by the dideoxy chain termination method (22) using either Cy5-labeled vector-primers (Thermosequenase Kit, Amersham Pharmacia Biotech) or appropriate sequence-specific primers (Alfexpress Autoread Sequencing Kit, Amersham Pharmacia Biotech). Data base searches for homologous proteins were performed using public internet resources of the National Center for Biotechnology Information (NCBI). Pairwise and multiple sequence alignments and phylogenetic tree calculations were performed using Vector NTI software (Informax) utilizing the ClustalW algorithm and the Neighbor Joining algorithm (23), respectively. Amino acid sequence identities were calculated from pairwise sequence alignments.
Chromosomal Localization of the DmG␥ e Gene and Northern Blot Analysis-The chromosomal localization of DmG␥ e was established by hybridization of squashed polytene chromosomes with a biotinylated 0.9-kb BamHI fragment of DmG␥ e cDNA according to Bloomquist et al. (2). The biotinylated probe was detected using a streptavidin-coupled peroxidase (Detek-Horseradish Peroxidase kit, Enzo Diagnostics).
Northern blot analysis with digoxigenin-labeled antisense cRNA probes transcribed from DmG␥ e cDNA was performed as has been described previously (24). Immunological detection of the probe was performed with the DIG Luminescent Detection Kit (Roche Molecular Biochemicals) using CDP-Star (Tropix, Bedford, MA) as luminescent dye.
Antibodies-Polyclonal antibodies directed against Calliphora G␤ e and G␥ e were generated by immunizing rabbits with recombinantly expressed polypeptides as described previously (8,9). The peptide used to generate anti-G␥ e antibodies corresponded to the entire coding region of CvG␥e, whereas the recombinantly expressed CvG␤ e -peptide comprised amino acids 65-346. The antibodies were affinity purified using the respective antigen. The anti-DmG␣ q antibody was kindly provided by C. Zuker (18).
Immunohistochemistry-For localization of G protein subunits by confocal laser scanning microscopy Drosophila wild type heads were fixed in 2% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.2) for 1 h at room temperature, followed by 3 washes in 10% sucrose for 15 min. The heads were infiltrated with 25% sucrose overnight at 4°C, embedded in boiled liver, covered with Tissue Tek, and cryofixed in melting isopentane. They were sectioned at 10-m thickness in a cryostat at Ϫ25°C and transferred to coverslips precoated with 0.01% aqueous poly-L-lysine. The cryosections were washed in 0.1 M sodium phosphate buffer, incubated in 0.01% Tween 20 for 20 min, followed by two washes in 0.1 M sodium phosphate buffer. After blocking the sections in 0.1% ovalbumin, 0.5% cold water fish gelatin in 0.1 M sodium phosphate buffer for 30 min at room temperature, the sections were incubated with the primary antibodies diluted in blocking solution (anti-DmG␣ q 1:50, anti-CvG␤e 1:20, anti-CvG␥e 1:25) overnight at 4°C. The sections were washed 3 times in 0.1 M sodium phosphate buffer (pH 7.2) and then incubated with a rhodamine-conjugated secondary goat anti-rabbit antibody for 3 h at room temperature. The sections were mounted in Mowiol 4.88 and examined with a confocal laser scanning microscope (LSM-SP, Leica).

SDS-PAGE and Western Blot Analysis-For
Western blot analysis of the visual G protein subunits, compound eyes or heads of Drosophila and retinas or rhabdomeral membranes of Calliphora were collected in 1ϫ SDS-PAGE buffer and homogenized with a plastic pestle. The proteins were extracted for 20 min at room temperature. Insoluble material was sedimented by centrifugation at 100,000 ϫ g for 10 min at 4°C. Protein concentrations were determined by the bicinchoninic acid procedure (29). Extracts of Calliphora tissues containing either 5 g of protein (equivalent of 0.13 retinas or rhabdomeral proteins out of 5 retinas) for G␣ q and G␤ e or 10 g of protein for G␥ e immunoblots and extracts of Drosophila tissues containing 17 g of protein (equivalent of 1.5 heads or 5 eye cups) were subjected to SDS-PAGE according to Laemmli (30) on 8 -20% gradient gels (Amersham Pharmacia Biotech, Midget System). For immunoblotting, proteins were electrophoretically transferred to polyvinylidene difluoride membranes (Bio-Rad) in a transfer buffer containing 50 mM Tris base, 0.1% SDS, 20% methanol for 1 h. The blot membranes were blocked in 10 mM sodium phosphate buffer, pH 7.0, 100 mM NaCl, 2% bovine serum albumin, 0.5% gelatin, 0.1% NaN 3 which greatly increased the sensitivity for detecting G␥ (31). Probing the blots with primary antibodies was performed in TBS-Nonidet P-40 (50 mM Tris-HCl, pH 7.3, 150 mM NaCl, 0.01% Nonidet P-40) overnight. Thereafter, the membranes were washed with TBS-Nonidet P-40 and incubated either with alkaline phosphatase-conjugated protein A or with 125 I-labeled secondary antibodies in TBS-Nonidet P-40 for 1 h. The protein bands were visualized through a chromogenic reaction with 5-bromo-4-chloro-3-indolyl phosphate/4-nitro-blue-tertazolium chloride or by exposing the blots to an X-Omat AR Film (Kodak).

RESULTS
Isolation and Sequencing of a Novel Drosophila G␥ Subunit-To isolate novel cDNA clones preferentially expressed in the photoreceptor cells of flies, we performed a differential hybridization screen. A high degree of specificity was obtained by hybridizing a Calliphora retinal cDNA library with probes derived from mRNA of retinal tissue, muscle tissue, and with a mixture of cDNAs encoding previously cloned retina-specific proteins such as Rh1 rhodopsin. The clones supposed to encode novel, retina specifically expressed proteins were sequenced. One cDNA clone of 1.5 kb in length contained a 216-base pair open reading frame encoding a polypeptide with homology to ␥ subunits of heterotrimeric G proteins. The coding region of this clone was used to probe a Drosophila head cDNA library. The Drosophila G␥ clones isolated by this homology screen were distinct from a previously reported Drosophila G␥ gene expressed in the brain (20). The novel genes for Calliphora and Drosophila G␥ subunits will hereafter be referred to as CvG␥ e and DmG␥ e , respectively, in analogy to the designation of G␤ e which encodes the eye-specifically expressed ␤ subunit of the Drosophila visual G protein (17).
The longest Drosophila G␥ e clone contained 246 base pairs of the 5Ј-untranslated region with an in-frame stop codon 15 nucleotides before the first AUG start codon, an open reading frame coding for 72 amino acids (M r ϭ 8398), and a 642-base pair 3Ј-untranslated region. For deducing the amino acid sequence, the translation initiation site was assigned to the first AUG codon (CCGCCAUGG) of the open reading frame which fits with the consensus sequence for translation initiation sites, CC(A/G)CCAUGG (32). The deduced amino acid sequences of Calliphora G␥ e and Drosophila G␥ e are identical except for a Ser 3 Asp and an Ile 3 Val substitution at positions 26 and 43, respectively (Fig. 1A). Both, CvG␥ e and DmG␥ e , exhibit a CAAX motif, CVIM (C, cysteine; A, aliphatic amino acid; X, any amino acid), at their C termini (Fig. 1A) suggesting that G␥ e may become post-translationally modified by isoprenylation. Comparison of the amino acid sequences of DmG␥ e with other G␥ subunits revealed the highest homology to a G␥ subunit (gpc-2) of Caenorhabditis elegans (41.9% amino acid identity) and an amino acid identity of about 30% to the visual G␥ subunits of vertebrate photoreceptors. The intra-specific homology of DmG␥ e to DmG␥ 1 (25.7% identity) is lower than the inter-specific homology to vertebrate visual G␥ subunits (Fig.  1B). An unusual stretch of highly charged amino acids located near the N terminus of visual G␥ from squid (33) has no parallel in DmG␥ e .
To obtain comprehensive information on the conservation of the G protein subunits expressed in the fly eye we also cloned G␣ q of Calliphora by a homology screen with DmG␣ q (15) and G␤ e of Calliphora by differential hybridization as described for isolating CvG␥ e . Both subunits proved to be highly homologous to the corresponding genes of Drosophila showing 95.5 and 89.3% amino acid identity for G␣ q and G␤ e , respectively. The extremely high homology of the Calliphora and Drosophila G protein subunits indicates the conservation between the phototransduction cascades of these flies which is also evident when the homology of other phototransduction proteins of Drosophila and Calliphora is compared, for example, rhodopsin (86%; 34), arrestin 2 (91%; 35), or transient receptor potential protein (77%; 9). It is therefore possible to combine results obtained with either of these visual systems into a common model for fly phototransduction.
Genomic Structure of the Novel Drosophila G␥ Subunit-In situ hybridization of a 0.9-kb BamHI fragment of the DmG␥ e cDNA clone to salivary gland chromosomes of Drosophila indicated a single locus for G␥ e at position 30A on chromosome 2L ( Fig. 2A). Since this part of the Drosophila genome is already sequenced, we screened the data base of the Drosophila genome project with the DmG␥ e cDNA sequence for matching sequences. This search revealed that the cDNA sequence of DmG␥ e is distributed on the contigs AC005889 and AC005125, the latter of which had to be inverted in order to correspond to FIG. 1. Comparison of the primary structure of selected G␥ subunits. A, amino acid alignment of selected G␥ subunits. The deduced amino acid sequences are shown in single-letter code. Identical amino acids are indicated by an asterisk. Amino acids which are identical in four of the six sequences aligned are marked with dots. The CAAX motif at the COOH terminus of the G␥ subunits is underlined. Black boxes depict amino acids which are characteristic for visual G␥ subunits. B, phylogenetic relationships between selected G␥ subunits. Visual ␥ subunits are marked with an asterisk. Accession numbers for sequences used: bovine G␥ 2 , P16874; bovine G␥ 3 , P29798; bovine G␥ 7 , P30671; bovine G␥ 12 , Q28024; bovine G␥ cone, P50154; bovine G␥ rod, P02698; C. elegans G␥ 1 (gpc-1), CAA91806; C. elegans G␥ 2 (gpc-2), AAC78236; calliphora G␥ e , AJ 250441, dog G␥ cone, AAC98924; Drosophila G␥ 1 , P38040; Drosphila G␥ e , AJ 250440, human G␥ 4 , P50150; human G␥ 5 , P30670; human G␥ 7 , AAC32595; human G␥ 10 , P50151; human G␥ 11 , P50152; human G␥ cone, O14610; human G␥ rod, Q08447; mouse G␥4, P50153; rat G␥ 9 , P43426; squid G␥, Q01821. DmG␥ e (Fig. 2B). The cytogenetical map position of both contigs is in agreement with the cytogenetical localization of DmG␥ e to position 30A. AC005889 maps to the chromosomal region 30A3-30A6 and AC005125 maps to 30A7-30A8. The genomic sequence of 34,810 kb allowed us to analyze the exonintron structure of DmG␥ e . The genomic organization of DmG␥ e reveals three exons which are separated by a 15,259-kb intron located in the 5Ј-untranslated region and a 18,447-kb intron that disrupts the coding region (Fig. 2C). Both introns are flanked by consensus GT/AT splicing sites.
Furthermore, we investigated whether there is evidence for a mutant in the DmG␥ e gene with defects in the phototransduction cascade. The only identified gene locus for a gene showing misfunctions in visual behavior at the cytogenetical map position of G␥ e is the inactivation no afterpotential A (inaA) mutant which is localized in the chromosomal interval 27F1-38E9 (36). However, Northern blot analysis of the DmG␥ e gene product of inaA flies revealed that G␥ e is normally expressed with no obvious difference in transcript size or transcript level as compared with the G␥ e mRNA isolated from wild type flies (Fig.  3A), indicating that inaA mutants are not null mutants or deletion mutants of G␥ e . Reverse transcriptase-polymerase chain reaction of the coding region of the G␥ e gene of inaA mutants and sequencing of the amplified product revealed that the G␥ e gene of the inaA mutant exhibits the same nucleotide sequence as the G␥ e gene of wild type flies. Consequently, inaA mutants contain a stable and fully functional G␥ subunit.
DmG␥ e Is Predominantly Expressed in Photoreceptor Cells and Forms the Visual G␤␥ Complex with G␤ e -Although the novel G␥ e subunit was isolated by a differential hybridization screen designed to isolate clones which are preferentially expressed in the eye, it is not certain that G␥ e codes for the ␥ subunit of the visual G protein. Therefore, we studied the cellular and subcellular localization of the G␥ e mRNA and the G␥ e protein and compared it with the spatial distribution of the visual G␣ q and G␤ e subunits. To determine the tissue distribution of the DmG␥ e transcripts, we performed Northern blot analysis (Fig. 3A). Hybridization with an antisense cRNA transcribed from the full-length DmG␥ e cDNA revealed three transcripts of approximately 1.4, 1.8, and 6 kb (Fig. 3A). The major 1.4-kb transcript was detected only in RNA isolated from heads of wild type flies (Fig. 3A, lane 1) but not in RNA obtained from eyes absent (eya) mutants, indicating that this transcript is predominantly expressed in the eye. The 1.8-and 6-kb RNA are present in wild type and eya heads, showing that the expression of these minor transcripts is not restricted to the eye.
For monitoring the expression pattern of the G protein subunits by Western blot analysis we generated specific antisera against recombinantly expressed peptides corresponding to the coding region of CvG␥ e and amino acids 65-346 of CvG␤ e . The anti-DmG␣ q antibody has been described previously (18). The anti-CvG␥ e antibody recognizes a 8-kDa protein extracted from retina and purified photoreceptor membranes of Calliphora, and from heads and eye cups of Drosophila wild type flies (Fig.  3B). The apparent molecular mass of 8 kDa deduced from the electrophoretic mobility after SDS-PAGE of DmG␥ e corresponds to the molecular mass of 8.398 kDa calculated from the amino acid sequence. When triplicate Western blots were probed with antibodies directed against G␣ q , G␤ e , and G␥ e a similar distribution of these G protein subunits was observed (Fig. 3B). A strong signal was obtained in lanes loaded with proteins of purified photoreceptor membranes with each of the three antibodies. This indicates that all three G protein subunits are associated with the light-absorbing photoreceptor membrane which harbors rhodopsin and the INAD signaling complex. G␣ q and G␤ e were not detected in protein extracts of Drosophila eya mutants lacking the compound eyes (Fig. 3B, lane 5), whereas a weak signal was observed when Western blots of eyes absent extracts probed with anti-G␥ e antibodies were developed for a prolonged time. The presence of minor amounts of G␥ e in heads lacking compound eyes may result from the translation of the 1.8-and 6-kb G␥ e transcripts which are not restricted to the eye.
Immunolocalization studies were carried out to elucidate the spatial distribution of visual G protein subunits. Laser scanning microscopy reveals that G␥ e is localized in the photoreceptor cells of the retina. Antibody reactivity was also detected in the lamina and the medulla presumably in the axons of the photoreceptor cells which project to these optic ganglia (Fig.  4C). Exactly the same labeling pattern is observed for G␣ q and G␤ e subunits (Fig. 4, A and B). This finding supports the hypothesis that G␥ e is part of an eye-specific heterotrimeric G protein. As labeling was not restricted to the rhabdomeres of the photoreceptor cell, we investigated the subcellular localization of G␤ e and G␥ e by electron microscopy. The labeling pattern shown in Fig. 5 reveals that both G␤ e and G␥ e signals are mainly associated with the rhabdomeres, but labeling is also found in the cytoplasm of the photoreceptor cell. The labeling density of G␤ e is higher than the labeling density of G␥ e , which may result from differences in the affinities of the anti-G␤ e and anti-G␥ e antibodies for their antigens or from a lower accessi- bility of the antigenic epitopes on G␥ e which is embedded in the 5-fold larger G␤ e protein.
It is possible that photoreceptor cells contain more than one G protein and, accordingly, more than one G protein ␥ subunit. Therefore, it is of crucial importance to identify the ␤ subunit to which G␥ e is attached. We immunoprecipitated proteins of rhabdomeral photoreceptor membranes of Calliphora with anti-CvG␤ e antibodies in order to concentrate our study on proteins of the photoreceptive membrane. G␤ e and G␥ e were detected by Western blot analysis in the immunoprecipitates obtained with anti-G␤ e antibodies, but not in control precipitations without antibody (Fig. 6). This indicates that G␥ e specifically co-precipitates with G␤ e and confirms that the newly isolated G␥ subunit is interacting with G␤ e to constitute a ␤␥ complex in the photoreceptive membrane.
Taken together, the eye-specific expression and the spatial distribution of G␥ e which coincides with the distribution of G␣ q and G␤ e , and the co-immunoprecipitation with G␤ e strongly suggest that this newly isolated G␥ subunit associates with G␤ e of the visual G protein of Drosophila photoreceptors. DISCUSSION Heterotrimeric G proteins are obligatory members of the phototransduction pathways activated by rhodopsins. Apart from the squid eye (33) there has been no indication for a distinct visual G␥ subunit in the majority of invertebrate eyes. Here we report the cloning and characterization of a novel G protein ␥ subunit (DmG␥ e ) expressed in the Drosophila eye. Western blot analysis and immunohistochemical localization of DmG␥ e at the cellular as well as subcellular level revealed that G␥ e is co-localized with the visual G␣ q and G␤ e subunits of Drosophila. Co-immunoprecipitation of G␥ e with G␤ e demonstrated a direct association of G␥ e and G␤ e into a functional heterodimeric G␤␥ subunit. Since G␤ e has been shown to function as a subunit of the visual G protein with an essential role in terminating the active state of the phototransduction pathway (19), the G␥ e cloned in the present study most likely represents the visual G␥ subunit.
It has already been noticed that vertebrate G␥ subunits can be classified on the basis of sequence information into subgroups that reflect functional similarities, for example, with respect to post-translational modifications and interaction with similar ␣ and ␤ subunits (37). Comparison of the amino acid sequences of the newly isolated G␥ e subunits from Drosophila and Calliphora with other G␥ subunits provides evidence that G␥ e is more closely related to visual G␥ homologs of vertebrates  than to most non-visual G␥ subunits (Fig. 1B). Thus, rhodopsins and G␥ subunits are members of phototransduction pathways that are conserved, irrespective of the photoreceptor cell type, rhabdomeric or ciliar, and the reaction sequence downstream of the relevant G protein. Common characteristics which distinguish the group of visual G␥ subunits from nonvisual G␥ subunits include a (Asp/Glu)-(Arg/Lys)-Asp motif near the amino terminus (position 10 -12 of DmG␥ e ), followed by Leu-Lys (position 14 -15), a basic amino acid (Asn, Lys) at position 54, a Asp-Lys sequence (position 57-58), and a glycine at position 67. These sites appear to allow the grouping into visual and non-visual G␥ subunits, despite a relatively low amino acid identity of about 30% between the most divergent members within the group of visual G␥ subunits. An exception exists with respect to G␥ 11 (30, 5% identity with DmG␥ e ), which shows distinct features of a visual G␥ subunit, but is expressed in non-retinal tissues (37). A high homology of G␥ e exists also to a G␥ subunit from C. elegans (gpc-2, 41.9% identity) which has been deduced from a cosmid sequence (F08B6) (38) but has not yet been functionally characterized. Drosophila G␥ 1 , the first G␥ subunit isolated from Drosophila, is more closely related to non-visual G␥ subunits which are expressed in a variety of tissues and may therefore form G␤␥ subunits capable of interacting with different types of G␣ subunits. Both, the phylogenetic relationship and demonstration of the existence of a novel, photoreceptor specific G␥ subunit indicate that G␥ 1 is not a visual G␥ as discussed previously (20). The conservation of the visual G␥ subunits is also reflected by the exon-intron structure of G␥ genes. In DmG␥ e and in the visual G␥ genes of vertebrates, the number and the location of the introns is conserved, with a first intron located in the 5Ј-untranslated region near to the coding region and a second intron splitting the coding region. Some non-visual G␥ subunits, notably Drosophila G␥ 1 , show a different gene structure in which the second intron is missing.
Considering that the G␣ subunits of transducins and of the fly visual G proteins belong to different G␣ classes which either activate a cyclic nucleotide phosphodiesterase or a phospholipase C, respectively, the homology of the corresponding G␥ subunits is striking. The question arises why G␥ subunits of G proteins which couple to rhodopsins are conserved, particularly as the associated G␤ subunits do not mirror this phylogenetic relationship. One reason might be that conserved interactions of visual G␥ subunits with rhodopsin exist at some stage of G protein activation. It has already been shown that the visual G␥ subunit from bovine photoreceptors is required for the interaction of rhodopsin with transducin as well as for transducin activation (39 -43). In this interaction the farnesylation of the COOH terminus of G␥ appears to be a crucial requirement. A COOH-terminal CAAX motif commonly found in other G␥ subunits modified by isoprenylation is also present in DmG␥ e (CVIM). Since G␥ subunits which exhibit a consensus sequence with a serine or methionine at the carboxyl-terminal position usually carry a farnesyl group (44 -46), DmG␥ e subunits are likely to become farnesylated rather than geranylgeranylated. Indeed, apart from the visual G␥ subunit of squid, all visual G␥s share a CAAX motif with a consensus sequence for farnesylation. Non-visual G␥ subunits, except for G␥ 11 (37), have a leucine or valine at the carboxyl-terminal position and are therefore expected to become geranylgeranylated. Thus, fly G␥ e subunits and G␥ subunits of vertebrate transducins are likely to share functions depending on post-translational fatty acid modification, for example, binding to the membrane and interaction of the G protein with rhodopsin.