BIOLOGICAL OVERVIEW

The G proteins, which have been conserved from yeast to humans, are heterotrimeric GTPases that behave as signal transduction proteins. Their three subunits (alpha, beta, and gamma) form a complex that relay signals arising from a receptor (serpentine receptor) composed of seven transmembrane segments: this complex may be activated by a variety of ligands, including hormones or neuropeptides. In the inactive complex, GDP is bound to the alpha subunit. Activation of the serpentine receptor results in an exchange between GDP and GTP and in the dissociation of the heterotrimer into a monomer (alpha subunit) and a dimer (beta and gamma). Both entities can target the activity of various enzymes and of ionic channels, which in turn control the level of intracellular second messengers.

The gene encoding the alpha subunit of the Drosophila G_o protein is expressed early in embryogenesis in the precursor cells of the heart tube, visceral muscles, and nervous system. This early expression coincides with the onset of a mesenchymal-epithelial transition, which takes place within two cell populations: the cardial cells and the precursor cells of the visceral musculature. A detailed analysis of G-oalpha47A expression suggests that the cardioblasts originate from two subpopulations of cells in each parasegment of the dorsal mesoderm. These cardioblasts depend on the wingless and hedgehog signaling pathways (which originate from the overlying ectoderm) for both their determination and specification. In the nervous system, the expression of G_oalpha shortly precedes the beginning of axonogenesis. Mutants produced in the G-oalpha47A gene harbor abnormalities in the three tissues in which the gene is expressed. In particular, the heart does not form properly and interruptions in the heart epithelium are repeatedly observed. Furthermore, in G-oalpha47A mutant embryos, in a sporadic manner, the epithelial polarity of cardial cells is not acquired (or maintained) in the cardiac tube. In the nervous system of homozygous Df(2R)47A embryos, which lack the G-oalpha47A gene, longitudinal axons are often missing and important modifications are observed in the guidance and axonal growth of motoneurons. It is thought that G_oalpha might be involved in vesicular traffic of membrane proteins that are responsible for the acquisition of epithelial polarity (FrÈmion, 1999 and references).

In Drosophila, genes coding for proteins that belong to the three major families of alpha subunits (alpha_s, alpha_i, and alpha_o) have been cloned and characterized. Two G_salpha proteins differing by four amino acids are produced by alternative splicing of a unique gene (Quan, 1990). The unique gene for G_ialpha produces two transcripts that encode the same protein (Provost, 1988). Finally, two G_oalpha proteins have been deduced from cDNA sequence analyses and are issued from a unique gene transcribed in three different transcripts (de Sousa, 1989, Schmidt, 1989, Thambi, 1989 and Yoon, 1989). Two other alpha subunits have been identified as homologs of vertebrate G_falpha (Quan, 1993) and G_qalpha-3 (Talluri, 1995), respectively. All these proteins are expressed throughout embryogenesis in various tissues in a dynamic pattern (Wolfgang, 1991) and, consequently, may play major roles in development although no precise developmental function has been ascribed to any of them. In contrast, it is known that the activity of concertina, which encodes the alpha subunit of a G₁₂ or G₁₃ protein, is necessary for the coordination of cell shape changes in the ventral furrow at gastrulation (Parks, 1991). In addition, concertina is a likely candidate to act downstream of a putative secreted protein, Folded gastrulation, whose receptor is as yet unidentified (Costa, 1994).

Among the different families of G proteins, the G_o protein is perhaps less well characterized in terms of the nature of its associated signal transduction pathway. On the plasma membrane, G_o proteins may regulate Ca²⁺ and K⁺ channels (Van Dongen, 1988; Kleuss, 1991; Lledo, 1992), distinct types of phospholipase C (Moriarty, 1990), and the mitogen-activated protein kinase cascade (Van Biesen, 1996). In addition to interacting with serpentine receptors such as opiate, alpha₂-adrenergic, D₂ dopaminergic, and somatostatin receptors, the G_o protein has been described recently as a regulator of vesicular traffic (Lagriffoul, 1996, Gasman, 1998).

The role of Drosophila G_o in heart development will be documented in this overview. In stage 11 embryos, the zygotic G-oalpha47A mRNA is first detected in the precursors of the cardioblasts in a repeated pattern of 11 clusters in the dorsal mesoderm. These clusters, constituted initially of two to four cells, are located in the anterior compartment of each mesodermal parasegment in a position anterior to the domain of Engrailed ectodermal stripes and closely neighboring the Wingless-expressing cells in parasegments 2-12 of the overlying ectoderm. These cardioblast precursors belong to the functional domain of, but are distinct from, Ladybird (Lbd) and Evenskipped (Eve) expressing cells in the mesoderm which are the precursors of a fraction of the pericardial cells and of a dorsal muscle. Within the cluster, the expression of the G-oalpha47A mRNA is initially very intense in only two of the four cells cells and then it becomes detectable in all four cells with an almost uniform intensity. Due to their position relative to the parasegment boundary, these cells are called A (for anterior) cardioblasts. Soon after the onset of G-oalpha47A expression in the first cluster, a second cluster of stained cells appears posterior to the first one and not in continuity with it. This group of cells called P (for posterior) cardioblasts is located in the posterior domain of the next parasegment and it never contains more than two G-oalpha47A-expressing cells. Their position coincides with that of an area posterior to En stripes in the ectoderm (FrÈmion, 1999).

During germ-band retraction (late stage 11 and stage 12), the cardioblast progenitors, which at this stage are mesenchymal in nature, begin to establish contacts among themselves, to extend filopods, and to reorganize their shapes to form a continuous layer on each side of the dorsal opening. This reorganization process coincides in time with the encounter of the A and P groups of G-oalpha47A-expressing cells in each parasegmental domain. Later in development and due to the shortening of the segment during germ-band retraction, the segmentally repeated six cell clusters spread along the anterior-posterior axis to finally join as a monolayer of polarized epithelial cells. Still further along in the process, the two rows of cardial cells, together with the pericardial cells that are attached to cardial cells' basal membrane, fuse at the dorsal midline to form the heart tube. These observations have led to the conclusion that a mature heart at the end of embryogenesis is made up of 52 cardial cells per hemiembryo, consisting of six cells in every segment from T3 to A7 and in only four cells in the T2 segment. Each segment is made up of two distinct populations of cardial cells: two groups; one composed of four cardial cells and one of two cardial cells (together consitituiting the six cells per hemisegment) that derive from mesodermal cells and that are located in domains corresponding, respectively, to the anterior and posterior compartment of each parasegment. None of the cardioblasts present in parasegments 2 and 3, and only a part of those found in PS4, participate in the mature dorsal vessel: however, it is difficult to follow their fates, since the expression of G-oalpha47A vanishes rapidly from these cells during development. Similarly, by using the EC11 antibody specific for the pericardial cells, 36 such cells have been found distributed 4 cells per segment, from T2 to A7. Eve-expressing cells from the labial segment and the segment T1 do not contribute to the pool of pericardial cells and they probably give rise to other structures that could not be recognized (FrÈmion, 1999).

The cardial cells, which are also muscle cells, are considered to be epithelial, based on the expression of several markers for polarity. For example, antibodies directed against alpha-spectrin specifically label the basal-lateral membrane of the epithelium; staining by anti-Neurotactin antibodies (Nrt) is restricted to the lateral and apical domains of the cell surface. In contrast, phosphotyrosyl proteins and Armadillo are weakly expressed in the heart cell membranes and have no polarized localization, in good agreement with the lack of authentic Zonulae adherens in Drosophila secondary epithelia. Finally, the basal membrane of the epithelium was visualized with the mAb EC11. This antibody recognizes an antigen that is most probably secreted by the pericardial cells and which is detected on the basal membrane of the cardial cells and around the pericardial cells. In the visceral mesoderm, which also undergoes a mesenchymal-epithelial transition before its differentiation into the visceral musculature, G-oalpha47A is expressed in a repeated pattern of clusters of cells in the same 11 parasegments as the precursors of the cardial cells and anterior to them. These precursor cells in parasegments 4-12 are ultimately transformed into a polarized epithelial ribbon, one cell wide (FrÈmion, 1999).

Homozygous Df(2R)47A embryos, lacking G-oalpha47A gene, harbor defects in the three tissues in which G-oalpha47A is expressed. All homozygous Df(2R)47A embryos have defects in the heart, the visceral mesoderm, and the nervous system. When probed with anti-EC11 antibodies, the mutant embryos show interruptions in their dorsal vessel that are also visible when DMef 2 expression is examined. The cardial cells are present and they migrate properly in the dorsal mesodermal position, but, in some places, they are no longer arranged as a continuous layer; rather, they form as unorganized clusters of cells. Fas III expression reveals the same types of defects in the visceral mesoderm. Similar alterations are detected in G-oalpha47A⁰⁰⁷ mutants with the same total penetrance (FrÈmion, 1999).

The heart and visceral mesoderm phenotypes of Df(2R)47A embryos can be rescued at least partially by expressing the class II G-oalpha47A cDNA under the control of 24B-GAL4 that drives the expression of UAS-cDNAs in the myogenic lineage, including the cardioblasts and the visceral mesoderm. In >65% (n = 250 observed embryos) of homozygous Df(2R)47A embryos, a normal development of these tissues has been restored. In contrast, the phenotype in the nervous system is not abolished as expected, first because the deficiency uncovers the lola gene, and second because of the restricted specificity of expression of the 24B-GAL4 line. The G-oalpha47A⁰⁰⁷ mutation that affects only the G-oalpha47A gene can also be rescued by a G-oalpha47A transgene (FrÈmion, 1999).

The loss of G-oalpha47A function produces heart and visceral mesoderm phenotypes that suggest a participation of the G_o protein in epithelia formation. During the mesenchymal-epithelial transition, subsets of membrane and cytoskeletal proteins localize to distinct regions of the cell surface to create the apical and basal-lateral membrane domains that confer to the cells their epithelial polarity. The G-oalpha47A mutant phenotype is more accurately described with the aid of markers localized specifically to these different domains. Analysis was focused on regions of the mutant heart in which the cells form a continuous uninterrupted layer but which display abnormal marker expression. The localization of two polarity markers is affected in several places in the heart epithelium of G-oalpha47A mutant embryos. alpha-Spectrin, specific for the basal-lateral membrane of epithelial cells, is abnormally expressed on the entire surface of the mutant cells, however with a somewhat lower intensity than in wild-type cells. In addition, these mutant cells remain round, have no signs of shape remodeling, and fail to express the EC11 antigen in their basal membrane. The pericardial cells that are normally attached to the basal membrane of the cardioblasts are absent from these same regions and are rather often associated in clusters in the domains of high EC11 antigen expression. Similarly, Neurotactin, which is localized to the apical and lateral membranes of wild-type cardial cells, in G-oalpha47A embryos is scattered in several locations on the entire surface of those cells that are round and in which the epithelial array is disorganized. Finally, Fas III expression also reveals phenotypes in the visceral mesoderm is characterized in some regions by a uniform and lower Fas III staining on the surface of still round cells with a partially destroyed epithelial structure (FrÈmion, 1999).

The myoendothelial heart tube is considered as a secondary epithelium that forms by a mesenchymal-epithelial transition. The mesenchymal cardial precursor cells, after their migration from the ventral site of gastrulation in the direction of their final dorsal position, reorganize their plasma membrane to acquire their cellular polarity and establish cell junctions to build up the cardial epithelium. A signal must be received by the mesenchymal cells to create the first asymmetry on their surface. Interactions with localized extracellular matrix components are believed to be largely responsible for this first event. For example, Drosophila mutants in Laminin A or in an integrin subunit present severe disruptions in their dorsal vessels. The initiating signal is probably emitted in response to inductive interactions between the mesenchymal cells and the overlying dorsal ectoderm. The newly created asymmetry could then trigger a reorganization of the cytoskeleton. Finally, the different membrane domains (apical and basal-lateral) are established by the acquisition of different combinatorials of membrane proteins that have been specifically routed towards them; this sorting-out step relies on an absolute specificity of vesicular traffic. In the absence of G-oalpha47A function, the mesenchymal cardial cells do not remodel their shape and, in the case of the most extreme phenotype, they failed to form a continuous row of cells. In addition, whenever such a layer is eventually formed, in several places some (but not all) cells neither acquire nor maintain a proper polarity (FrÈmion, 1999 and references).

Based on the timetable of G-oalpha47A expression, its participation in the first step, creating the asymmetry of the cardial cells, is not very likely. Rather, it might be required for the subsequent steps that concern cell shape change and polarization, by addressing membrane proteins to their respective domains (FrÈmion, 1999).

It has been suggested that heterotrimeric G proteins could contribute to the vesicular protein traffic by regulating early steps in the secretory pathway. This hypothesis stems from the observation that AlF4^-, an activator of heterotrimeric but not of monomeric G proteins, inhibits ER to Golgi and intra-Golgi transport as well as vesicle budding from the trans-Golgi network. In particular, G_o proteins have been implicated in granule exocytosis from chromaffin cells (Vitale, 1993 and Gasman, 1998); insulin secretion (Lang, 1995), and transcytosis (Bomsel, 1992). It has been shown that the secretion of the protease Nexin-1 by glioma cells is under the control of G_o1 (Lagriffoul, 1996). The G_oalpha1 protein has been detected on the membrane of small intracellular vesicles and the secretion of Nexin-1 is stimulated by G_oalpha1 overexpression and by activators of G_o proteins, such as mastoparan. It has been further suggested that the GTPase activity of the G_oalpha1 protein can be stimulated in the absence of a classical serpentine receptor (Lagriffoul, 1996).

Thus, it is tempting to speculate that the Drosophila G_o protein has a function in a particular type of vesicular traffic responsible for the acquisition or maintenance of cell polarity in the cardial and visceral mesoderm cells. Preliminary observations on the subcellular localization of G_o in embryonic cells are consistent with this idea. In the totality of the cells examined, Drosophila G_o is located to the cytoplasm rather than associated to the cell membrane and the staining pattern reveals a typical granular appearance (FrÈmion, 1999).

Since the exportation and the localization to the plasma membrane of the protein markers that were used were all equally affected in G-oalpha47A mutants, it has been concluded that G-oalpha47A is involved in a general aspect of vesicular traffic rather than in the specific process of the sorting out of membrane proteins. G-oalpha47A might also be required for the reorganization of the cytoskeleton. The protein G encoded by the concertina gene participates in cell shape changes taking place at gastrulation (Parks, 1991), probably via a modulation of the invaginating blastoderm cell cytoskeleton resulting from the activation of RhoA (Hall, 1998). The early role of the G_o protein in the formation of the heart epithelium does not exclude a function in later events leading to the formation of the heart or to the acquisition of heart function. It has been shown recently that knocking-out the G_oalpha gene in the mouse results in heart dysfunction. G_o-deficient mice have lost the muscarinic inhibition of isoproterenol-stimulated cardiac L-type Ca²⁺ currents (Valenzuela, 1997 and Jiang, 1998). It will be interesting to investigate whether the Drosophila G_o protein could also be involved in such a process (FrÈmion, 1999).

GENE STRUCTURE

A Drosophila gene (G-oalpha47A) encoding a G protein alpha subunit has been isolated by screening genomic and adult head cDNA libraries using bovine transducin alpha subunit cDNA as probe. The gene, which maps to 47A on the second chromosome, encodes two proteins that are both 354 amino acids long but differ in seven amino acids in the amino-terminal region. The deduced amino acid sequences of the two proteins are 81% identical to the sequence of a rat Go alpha subunit. Analysis of genomic clones reveals that there are eight coding exons and that the putative transcripts for the two proteins differ in the 5'-noncoding regions and the first coding exons but share the remaining six coding exons. The arrangement of two different 5'-noncoding regions on the gene suggests that two different promoters regulate the expression of the transcripts encoding the two proteins. RNA blot analysis detected three transcripts: a 3.9-kilobase (kb) transcript found at all stages of development; a 5.4-kb transcript present predominantly in adult heads, and a 3.4-kb transcript present only in adult bodies. In situ hybridizations of a cDNA probe to adult tissue sections shows that the gene is expressed abundantly in neuronal cell bodies in the brain, optic lobe, and thoracic ganglia (Yoon, 1989).

Drosophila cDNA clones coding for the alpha subunit of a Go-like G protein have been isolated. The sequence of two cDNA clones shows there is alternative splicing in the 5'-coding region which, on conceptual translation, would give rise to two proteins with slightly different amino termini. A partial genomic clone indicates there are four introns in the carboxyl-terminal half of the clone. Two transcripts, 3.8 and 5.3 kilobases long, are expressed at a high level in the heads of adult flies and also in larvae, pupae, and embryos. Hybridization of the cDNA probes to sections of adult flies indicates the RNA is present predominantly in nervous tissue (in the cortex of the brain and the thoracic ganglion); it is also expressed in the ovaries. Transcripts corresponding to both cDNAs are present in the central nervous system, but only one of them is found in detectable levels in the ovaries. The gene maps to 47A on the Drosophila second chromosome (de Sousa, 1989).

Exons - Six common and two different 5' exons

PROTEIN STRUCTURE

Amino Acids - 354

Structural Domains

The sequence similarity is particulary striking between Drosophila G protein oalpha 47A proteins and the alpha subunit of rat Go (81% identity). The two Drosophila proteins also show 70% and 61% sequence identity to rat and Drosophila Gi alpha subunits, respectively (Yoon, 1989)

date revised: 1 September 99

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.