A Drosophila melanogaster G protein alpha subunit gene is expressed primarily in embryos and pupae.

A G protein alpha subunit gene has been isolated from a Drosophila melanogaster genomic library using a combination of bovine rod and cone transducin alpha subunit cDNAs as a probe under reduced stringency conditions. The gene, DG alpha 1, encodes a protein with an amino acid sequence 78% identical to bovine Gi alpha 1. However, unlike all reported Gi alpha subunit the DG alpha 1-encoded protein is not expected to be a pertussis toxin substrate, because it lacks a cysteine at the appropriate site. The protein coding region of the gene is split by four introns. The sequence of a head tissue cDNA clone, as well as amino acid similarities to mammalian G proteins, confirms this exon/intron structure. Northern blots of total cellular RNA reveal a major 2.3-kilobase transcript and a less abundant 1.7-kilobase transcript. These transcripts are most abundant in RNA from embryos and pupae. The DG alpha 1 gene is located on band 65C on the left arm of the third chromosome, on the basis of in situ hybridizations to Drosophila salivary gland polytene chromosomes.

A G protein a subunit gene has been isolated from a Drosophila melanogaster genomic library using a combination of bovine rod and cone transducin a subunit cDNAs as a probe under reduced stringency conditions. The gene, DGal, encodes a protein with an amino acid sequence 78% identical to bovine Gial. However, unlike all reported Gia subunits the DGal-encoded protein is not expected to be a pertussis toxin substrate, because it lacks a cysteine at the appropriate site. The protein coding region of the gene is split by four introns. The sequence of a head tissue cDNA clone, as well as amino acid similarities to mammalian G proteins, confirms this exonlintron structure. Northern blots of total cellular RNA reveal a major 2.3-kilobase transcript and a less abundant 1.7-kilobase transcript. These transcripts are most abundant in RNA from embryos and pupae. The DGal gene is located on band 65C on the left arm of the third chromosome, on the basis of in situ hybridizations to Drosophila salivary gland polytene chromosomes.
Guanine nucleotide binding proteins (G proteins) are a family of membrane-associated intracellular proteins which relay signals between activated membrane receptors and intracellular effector enzymes (Gilman, 1987;Stryer and Bourne, 1986). Many G proteins are now well characterized. G, interacts with hormone-bound @-adrenergic receptor and stimulates intracellular adenylyl cyclase. Transducins, G proteins found only in photoreceptors, carry the light-stimulated flow of information from bleached rhodopsin to a cGMP phosphodiesterase. The binding of a variety of hormones to their specific receptors activates Gi and causes inhibition of adenylyl cyclase. G proteins are also implicated in phosphoinositide breakdown and turnover (Kikuchi et al., 1986), ion channel gating (Pfaffinger et al., 1985;Logothetis et al., 1987;Yatani et al., 1987), and olfactory signal transduction (Pace and Lancet, 1986).
G proteins are heterotrimers which consist of an a subunit associated with a tight complex of @ and y subunits. They exist in two activity states, determined by the type of guanine nucleotide bound to the a subunit. In its active state the a * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession numbeds) 503903.
$ Supported by Public Health Service National Research Service Award 5T32 GM07270.
Presently in the Division of Molecular and Physiological Plant Biology, University of California, Berkeley, CA 94720. subunit binds GTP, dissociates from the @y complex, and interacts directly with an effector enzyme. A slow GTPase activity intrinsic to the a subunit hydrolyzes the bound GTP to GDP, and promotes the reassociation of a with @r, converting the G protein to its inactive state. Cholera and pertussis toxins catalyze covalent ADP-ribosylation of G protein a subunits, thereby altering G protein inactivation and activation kinetics, respectively.
Much of the biochemical characterization of G proteins has focused on vertebrate signal transduction systems. While these systems are useful for protein characterization studies, they are in general limited by the difficulty of generating and characterizing G protein mutants. In addition, the behavioral and neuronal impact of G protein function is difficult to study in uitro. For these reasons we have chosen to embark on a study of G proteins in Drosophila melanogaster, a species amenable both to genetic manipulation and neurobiochemical study.
We describe here the isolation and characterization of a gene encoding a D. melanogaster G protein a subunit. Isolated from a D. melanogaster genomic library by virtue of its crosshybridization to a bovine transducin a subunit cDNA probe, this Drosophila G protein gene encodes an amino acid sequence 78% identical to that of mammalian Gia. We also present data demonstrating that transcripts from this gene are most abundant in RNA from early embryos and pupae.

EXPERIMENTAL PROCEDURES
Fly Culture-Wild type Canton S strain D. melanogaster flies were maintained at room temperature on cornmeal/molasses/agar medium (Roberts, 1986), collected at various developmental stages, and frozen in liquid nitrogen prior to use for DNA or RNA preparation.

1.2070
Kauvar, and T. Kornberg, University of California, San Francisco) were screened using a nick-translated 0.35-kb EcoRI genomic fragment from the DGa1 gene ( Fig. 1). Hybridization conditions were identical to those described for the genomic library except that 30% formamide was used in the hybridization solution, and the hybridization was carried out at 30 "C. Library lifts were washed in 2 X SSC, 1% SDS at 48 "C.
Isolation of RNA and Northern Blot Analysis-Total RNA was isolated by the method of Cathala et al. (1983), with the following modifications: pelleted RNA was solubilized in 3 M LiC1, 1 M guanidinium isothiocyanate and the solution sheared by several passages through an 18-gauge needle; solubilization buffer contained 1% SDS; the crude RNA solution was extracted three times with 1 volume of phenol, two times with 1 volume of ch1oroform:l-butanol (4:1, v/v); RNA was precipitated with 2 volumes of ethanol and 0.1 volume of 3 M sodium acetate.
Total RNA (15 pg/lane) was electrophoresed in a 1.0% agarose, 2.2 M formaldehyde gel as described in Maniatis et al. (1982). 32P endlabeled DNA molecular weight standards (X HindIII digest, Bethesda Research Laboratories) and RNA molecular weight standards (RNA ladder, Bethesda Research Laboratories) were treated in the same manner as RNA samples and run on adjacent lanes of the agarose gel. Electrophoresed RNA was transferred to Nytran (Schleicher & Schuell) in 20 X SSC. Filters were prehybridized as described above for library lifts and were hybridized at 63 "C in the same hybridization Genomic DNA Isolation and Southern Blot Analysis-Isolation of Drosophila genomic DNA from frozen fly tissues was performed as described in Maniatis et al. (1982) except that RNase digestion preceded dialysis of the genomic DNA. Genomic DNA (5 pg/lane) was digestion with BamHI and electrophoresed on a 1% agarose gel. DNA was transferred to Nytran as described (Maniatis et al., 1982). Blots were prehybridized as described for library lifts and hybridized at 45 "C in the same hybridization solution as that used for screening the cDNA library. High stringency washes were performed in 0.1 X SSC, 0.05% SDS at 65 "C for 1-2 b. Low stringency wash conditions were 0.5 X SSC, 0.5% SDS at 45 "C for 1-2 h.
In Situ Hybridizations to Polytene Chromosomes-A nick-translated biotinylated 0.35-kb EcoRI genomic restriction fragment ( Fig.  1) was hybridized to polytene chromosomes in situ according to the method of Engels et al. (1986).

RESULTS
solution as that used to screen the head cDNA library. A 0.35-kb EcoRI genomic fragment ( Fig. 1) was used as a probe for gene A D. nelanogaster genomic library was screened with and Rosbash, 1984), was used as a probe for RNA integrity. After teen positively hybridizing plaques were isolated from the hybridization, filters were washed in 2 X SSC, 1% SDS at room initial screening. Two clones were purified and their inserts at 65 "c for 1 h. Autoradiography was performed with x-ray film and detail. ~i~. 1~ shows the restriction pattern of this clone, an intensifying screen at -70 "C.
clones were determined using the chain termination method of Sanger were using a 0*35-kb EcoR1 Of x4' et al. (1977) and Hattori and Sakaki (1986). Restriction fragments ( Fig. 1) to Screen a Drosophila Canton s adult head cDNA were subcloned into double-stranded plasmid vectors pUC 19 (Be-library and a Drosophila Oregon R embryo cDNA library.         1B shows the restriction pattern of one such clone derived from the adult head cDNA library, XSD7110.

A f A A C C G A I l l G G l T G C C A G l A G T C G C G~C~l A A C C T U C C A t C r G C M I l T A G I l A A C C G T G l C C A~~M A~l G t A C C C A T M G C T M C l M T T l A M~T I A G C C l A A C U U C M A l C l~G T T C G A C M A G~C A G C C G A T G C G l T T T G T A C
DNA restriction fragments of X4' and XSD7110 were subcloned into plasmids for sequence analysis (Fig. 1, A and B ) .
The sequences of the protein coding regions of the gene and the adult head cDNA (Fig. 2) were identical save for one base difference. Base 8, a guanosine in the genomic clone, is changed to an adenosine residue in the head cDNA, causing the amino acid sequence to change from a cysteine (TGT) to tyrosine (TAT) at codon 3. There were also minor differences in the sequences near the extreme 5' end of the cDNA which we attributed to cloning artifacts (data not shown). The one base difference in the protein coding region may be either a cloning artifact or a strain polymorphism, since a different cDNA, isolated from a 0-3 h Drosophila Oregon R strain embryo cDNA library, is identical to the gene at base 8 but differs at bases 177 and 751 (data not shown). The gene appears to code for a G protein a subunit, and therefore we have designated it DGa1 (for Drosophila G protein a subunit 1).
The DGa1 conceptual translation product was compared with different bovine G protein a subunit sequences (Fig. 3A). Several peptide sequences common to all G protein a subunits are present in DGal. These conserved amino acids are thought to be necessary for GTP binding and hydrolysis (Masters et al., 1986). DGal appears to be most similar to bovine Gia, the a subunit of the G protein which inhibits adenylyl cyclase. A comparison of the proteins encoded by DGal and Gia cDNAs derived from three different mammalian genes (Fig. 3B) reveals the highest degree of amino acid identity between DGa1 and bovine Gial. Seventy-eight percent of the corresponding amino acids in DGal and Gial are identical. Despite this high degree of similarity, there is a major difference between the DGal gene product and mammalian Gias. All mammalian Gias contain a cysteine 4 residues from the carboxyl terminus which serves as a substrate DGal :

Drosophila G Protein
In order to determine if there are other G protein a subunit genes within the Drosophila genome, a probe from the DGal gene was tested for cross-hybridization with other Drosophila genes. Fig. 4 shows a Southern blot of BamHI-digested Drosophila genomic DNA probed with the 0.35-kb EcoRI fragment from clone X4'. The DNA within this probe encodes a polypeptide sequence (amino acids 34-58 in Fig. 3B) which is conserved among all G proteins. Under high stringency conditions only the 6-kb BamHI fragment predicted from the DGal gene structure is detected. However, a low stringency wash of the Southern blot reveals several additional bands (designated by arrows) which do not correspond to the DGal gene, indicating cross-hybridization. The faint high molecular  (Roberts et al., 1986), larvae were predominantly at the third instar stage, pupae were not staged, and adults were collected 1-2 days after eclosion. Frozen adult bodies and heads were separated by sifting through wire mesh of defined sizes. B, same blot analyzed for RNA degradation by probing with a ribosomal protein gene (RP49, O'Connell and Rosbash, 1984). weight bands present on the high stringency genomic blot may represent partially digested genomic DNA.
Developmental expression of DGal was examined by Northern blot analysis of total cellular RNA using the 0.35kb EcoRI genomic fragment as a probe (Fig. 5 A ) . The predom-inant transcript, 2.3 kb long, is present in RNA from embryos, pupae, and (less abundantly) in adult heads. A 1.9-kb transcript was also detected in RNA from 0-3-h embryos. Because the DGal probe hybridized poorly with adult body RNA, we used a probe for the ribosomal protein 49 (RP49) transcript (O'Connell and Rosbash, 1984) to test for RNA integrity. Fig.  5B shows the RP49 probe hybridized with RNA in all the lanes, indicating intact RNA was present.
Drosophila salivary gland polytene chromosomes were probed in situ with the 0.35-kb EcoRI genomic fragment from DGal. Fig. 6 shows the gene-specific probe hybridized to a region on the left arm of chromosome 3 near region 65C. This was the only band which hybridized to the DGal probe.
The isolation and characterization of DGa1 also provides further evidence that the nervous system of D. melanogaster utilizes molecular signal transduction mechanisms similar to those of higher eukaryotes. The identification of ion channels (Salkoff et al., 1987;Papazian et al., 1987), visual pigments (O'Tousa et al., 1985;Zuker et al., 1985), adenylyl cyclase (Chen et al., 1986), and G protein , R subunit (Yarfitz et al., 1988) and a subunit (this work) reaffirms the choice of D. rnelanogaster as a model neurobiological system for higher eukaryotes.
The amino acid sequence of the DGa1 gene product is most similar to mammalian Gia1 (78% identical), and slightly less similar to Gia2 and Gia3 (77% identical). A value for the estimated evolutionary distance (KJ between DGal and each of the three mammalian Gias (based on the number of third base substitutions at synonymous codons) was calculated by the method of Kimura (1981). The K , value for DGal versus human Gia2 was 0.61 * 0.09, as compared to values much greater than 1 for DGal versus bovine Gial or human Gia3. Thus, despite the fact that the DGa1 amino acid sequence is most similar to bovine Gial, these calculations indicate that DGal is likely to have evolved from the gene that became Pertussis toxin catalyzes the covalent attachment of an ADP-ribose moiety to a cysteine residue near the carboxyl terminus of many G proteins. In the DGal gene product an isoleucine has been substituted for this conserved cysteine. Since the biological function of endogenous ADP-ribosylation is unknown, the significance of a missing pertussis toxin recognition site in the DGal gene product is unclear. Nonetheless, other pertussis toxin substrates are present in Drosophila tissues (Hopkins et al., in press). Perhaps a regulatory mechanism reflected by the ADP-ribosylation phenomenon has been evaded by the DGal gene product.
Although the function of the DGal gene product remains unknown, it is probably not a phototransduction enzyme. The northern blots in Fig. 6 show that the most abundant transcripts are present in cellular RNAs isolated from embryos Gia2. and pupae, and very little signal was found in head RNA. If this gene product were important in signal transduction primarily within Drosophila photoreceptors, a much stronger hybridization signal from head RNA might be expected. It also seems unlikely that a phototransduction-specific G protein would be found in early embryos, since photoreceptors are lacking at this stage of development. The presence of cross-hybridizing bands on genomic Southern blots (Fig. 4), evidence for pertussis toxin and cholera toxin substrates in Drosophih heads (Hopkins et al., in press), and light-dependent GTPase activity in house flies (Blumenfeld et al., 1985) all suggest that Drosophila may utilize more than one type of G protein in signal transduction. D G d transcripts were detected in the earliest staged embryo RNA isolated. Since very few genes are zygotically transcribed in the first 3 h after fertilization (Edgar and Schubiger, 1986), this result suggests that the DGa1 gene may be maternally transcribed and packaged into the unfertilized egg. If this is so, DGa1 could join the growing list of genes which are important in early Drosophila development (Scott and O'Farrell, 1986). It is noteworthy that improper regulation of CAMP levels is associated with infertility, developmental defects, and learning defects in Drosophila (Bellen et al., 1987).
There are no reports of mutants at the cytological locus of DGa1 at region 65C on the third chromosome. Therefore the identification of the function of this new G protein awaits localization, tissue-specific expression, and mutagenesis experiments.