Ras GTPase activating protein 1: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References

Gene name - Ras GTPase activating protein 1

Synonyms - GTPase-activating protein 1, Gap1

Cytological map position - 67D2--3

Function - Ras GTPase-activating protein

Keywords - eye, Ras pathway

Symbol - RasGAP1

FlyBase ID: FBgn0004390

Genetic map position - 3-[28]

Classification - PH domain, C2-domain and Ras GTPase-activating protein

Cellular location - cytoplasmic



NCBI links: Precomputed BLAST | Entrez Gene
BIOLOGICAL OVERVIEW

During normal eye development cone cell precursors are competent to become R7 photoreceptors and will adopt this cell fate if their Ras1 activity level is sufficiently high. Gap1, a guanine nucleotide exchange factor, functions in cone cell precursors to down-regulate Ras1 (Buckles, 1992; Gaul, 1992, and Rogge, 1992). Ras activity in these cells is likely elevated to a moderate degree by EGF receptor activity. Loss of Gap1 function causes the cone cell precursors to aberrantly adopt the fate of R7 cells. The number of supernumerary R7 cells produced within each ommatidium depends on the severity of the Gap1 allele (Buckles, 1992) indicating a sensitivity to different levels of Gap1 activity that can be exploited for the purposes of a structure-function analysis (Powe, 1999).

Ras signaling can be thought of in simple terms as occuring downstream of receptor tyrosine kinases, triggered by association of adaptor proteins with the phosphorylated cytoplasmic tails of these receptors. The analysis of Gap1 belies this simple picture, pointing to the involvement of Ca2+ and lipid signaling in moderating the strength and duration of the Ras signal. The Gap1 catalytic domain alone is insufficient for in vivo activity of Gap1, indicating a requirement for the additional domains. An inositol-1,3,4,5-tetrakisphosphate (IP4)-sensitive extended PH domain is essential for Gap1 activity, while Ca2+-sensitive C2 domains and a glutamine-rich region contribute equally to full activity in vivo. Furthermore, a strong positive genetic interaction occurs between Gap1 and phospholipase Cgamma (PLCgamma), an enzyme that generates inositol-1,4,5-trisphosphate, a precursor for IP4 and a second messenger for intracellular Ca2+ release. These results suggest that Gap1 activity in vivo is stimulated under conditions of elevated intracellular Ca2+and IP4 . Since receptor tyrosine kinases (RTKs) trigger an increase in intracellular Ca2 + and IP4 concentration through stimulation of PLCgamma, RTKs may stimulate not only activation of Ras but also its deactivation by Gap1 (Powe, 1999),

Several GAPs have been identified in both mammals and Drosophila, including neurofibromin, p120Gap and Gap1, which all share a catalytic domain required for stimulating Ras GTPase activity. However, GAPs possess additional non-catalytic domains that distinguish them from one another and that could be involved in their differential regulation or activity. Mammalian neurofibromin appears to constitutively down-regulate Ras, but the Drosophila homolog of this Gap protein might function upstream of Protein kinase A or in a parallel pathway, failing to interact with the Ras pathway (The, 1997). Nf1 and PKA appear to interact in a pathway that controls the overall growth of Drosophila (The, 1997). Mammalian p120Gap is required for moderating the extent and the duration of Ras activation after growth factor stimulation. Drosophila p120Gap has been implicated in the function of Breathless and Heartless, two Drosophila receptor tyrosine kinases of the fibroblast growth factor receptor family that are involved in wing morphogenesis, but no effect of RasGAP overexpression on photoreceptor development in the eye can be detected (Feldmann, 1999). Both Drosophila FGF receptors have exact matches to the optimal binding site (phospho YxxPxD, where x is any amino acid) for the SH2 domains of mammalian p120 Ras-GAP. The in vivo effects of RasGAP overexpression require intact SH2 domains, indicating that intracellular localization of RasGAP through SH2-phosphotyrosine interactions is important for RasGAP activity (Feldmann, 1999). Activation of mammalian p120Gap in response to growth factor stimulation may be achieved by both membrane translocation and stimulation of its catalytic activity: p120Gap contains SH2 domains that can be induced to bind to the activated RTK, thereby leading to the translocation of the cytoplasmic protein to the membrane. In addition, the N-terminal region of p120Gap appears to be required for full in vitro activity of the C-terminal catalytic domain (Powe, 1999 and references).

Gap1, which was first isolated in Drosophila as a negative regulator of RTK signaling, has two homologs in mammals (Gaul, 1992; Maekawa, 1994; Cullen, 1995). Apart from the catalytic domain in the middle of the molecule, all Gap1 family members share two N-terminal C2 domains and a C-terminal extended PH domain. Biochemical studies with the two mammalian homologs, Gap1m and Gap1IP4BP, have shown that the PH domain, which is required for membrane localization of an N-terminally truncated Gap1IP4BP, binds IP4 or phosphatidylserine (PS) in a mutually exclusive manner (Fukuda, 1996; Cullen, 1997), suggesting that an increase in intracellular IP4 levels leads to the release of the PH domain from the membrane. IP4 binding to the PH domain also appears to stimulate GAP activity in vitro (Cullen, 1995; Fukuda, 1996), implying that release of the PH domain from the membrane is accompanied by activation of the catalytic domain. The N-terminal C2a domain has been shown to bind phospholipids in a Ca2+-dependent manner (Fukuda, 1996 and Fukuda, 1997), while the role of the C2b domain remains unclear. However, the physiological relevance of any of these domains for Gap1 in vivo function has not been established. Resolution of these problems has been approached in Drosophila by carrying out an in vivo functional analysis of Drosophila Gap1 in the context of cone cell determination in the developing Drosophila eye. Under basal conditions Drosophila Gap1 is localized to the membrane through its extended PH domain in a less active form; upon elevation of Ca2+ and IP4 concentrations, the PH domain binds to IP4 thereby stimulating catalytic activity, while at the same time the C2a domain takes over the role of membrane tether. This model is supported by the observation that Gap1 appears to be constitutively associated with the plasma membrane (Powe, 1999).

To introduce altered Drosophila Gap1 transgenes into flies, the sevenless (sev) enhancer/promoter cassette, an eye-specific promoter, was used to drive expression in cone cell precursors at appropriate levels. Expression of one copy of wild type Gap1 under control of the sev enhancer/ promoter cassette (sevGap1) rescues the complete loss-of-function eye phenotype, while creating no phenotypic defects when placed in a wild type background containing two copies of the endogenous Gap1 gene. By contrast, high levels of Gap1 expression under sev enhancer/promoter control using amplification by the UAS/GAL4 system in an otherwise wild type background results in a rough eye phenotype due to loss of photoreceptors (Rørth, 1996). Therefore, fusing altered Gap1 transgenes directly to the sev enhancer/promoter cassette provides sufficient levels of expression without interfering with development. To assess their rescue ability, transgenes were placed into two different Gap1 mutant backgrounds: the partial loss-of-function allele Gap1S1-11 and the complete loss-of-function allele Gap1B2 (Gaul, 1992). Gap1 transgenes containing only the catalytic domain (sevGapM) fail to rescue either the partial or complete loss-of-function phenotype of Gap1. These findings clearly indicate, in contrast to other studies, that the catalytic domain of Drosophila Gap1 requires additional parts of the protein for in vivo activity (Powe, 1999).

Since the catalytic domain alone could not rescue the loss-of-function phenotype, a test was performed to see whether the C- or N-terminal portions are required along with the catalytic domain for in vivo function. In the C-terminus of the molecule two regions can be distinguished. Immediately adjacent to the catalytic domain lies a large stretch of sequence similarity that is shared by all Gap1 family members. This conserved region consists of a PH domain, followed by a Cys3His1 zinc finger, referred to as the BTK motif. The PH and BTK sequences are followed by a stretch of 123 amino acids with no known homology to other proteins and poor homology within the Gap1 family; however, since a sequence of eight amino acids (KYGSxxxPIGD) at the C-terminal part of this region is conserved among Gap1 family members, this 123 aa stretch is designated the KYG region. In mammalian Gap1, the BTK/KYG region is required for specific high affinity binding of IP4 by the PH domain (Fukuda, 1996). C-terminal to the KYG domain is a glutamine-rich stretch found only in the Drosophila homolog (Powe, 1999).

Transgenes containing the catalytic domain and the entire C-terminus (sevGapdeltaN) show a 50% rescue of the Gap1 partial loss-of-function phenotype. The expression level of the sevGapdeltaN transgene providing the most complete rescue is higher than that of the best transgene expressing full length Gap1, indicating that the inability to rescue completely is due to an alteration of intrinsic activity. Thus, in conjunction with the C-terminus, the catalytic domain is sufficient for partial rescue activity. However, the N-terminus is necessary for full in vivo activity, since transgenes lacking the N-terminus do not completely rescue loss-of-function mutants. The N-terminal region of Drosophila Gap1 possesses two domains with known function, namely C2a and C2b domains. Various transgenes containing just the N-terminus and the catalytic domain all essentially fail to rescue the mutant phenotype. These results demonstrate that, unlike the C-terminus, the N-terminus does not confer the ability to rescue onto the minimal catalytic domain, but is necessary for full rescue activity (Powe, 1999).

Attempts were made to determine which part of the C-terminus is responsible for endowing the catalytic domain with the ability to rescue. The importance of the glutamine-rich region was tested. Transgenes without the glutamine-rich region provide 50% rescue. Hence, the glutamine-rich region is required for full rescue, but not essential for in vivo activity. Next, the importance of the PH/BTK/KYG region was tested, beginning with BTK and KYG motifs. The sevGa-pdeltaBTK-Q transgenes lack the entire KYG domain and the last cysteine of the BTK zinc finger as well as the glutamine-rich region, and show virtually no rescue. This finding clearly demonstrates that the BTK and KYG regions are critical for rescue activity. To examine the role of the PH domain, a point mutation (sevGap1R786C), that severely reduces IP4 binding in mammalian Gap1 (Fukuda, 1996), was introduced and rescue ability was evaluated. Like the full length sevGap1 transgenes, sevGap1R786C transgenes are able to fully rescue the partial loss-of-function situation. However, unlike sevGap1, when placed in the complete loss-of-function background, sevGap1R786C only partially rescues (~70%) the phenotype. In the complete loss-of-function background, the average number of R7s per ommatidium for sevGap1R786C is 1.6, compared to 1.0 for sevGap1 and 2.9 in the absence of a Gap1 transgene. Thus, the R786C mutation does compromise Gap1 in vivo activity, even if only moderately. The effect of this point mutation is masked by residual activity in the partial loss-of-function situation and is only uncovered in the complete loss-of-function situation. In general, the rescuing ability of Gap1 transgenes is lower in the complete loss-of-function background than in the partial loss-of-function background (with the exception of the sevGapdeltaC2a transgene), suggesting that the residual activity in Gap1S1-11 partially compensates for the loss of activity in mutant Gap1 constructs (Powe, 1999).

These results demonstrate that the N-terminus of Gap1 is necessary for full in vivo activity, but only in the presence of critical C-terminal regions, suggesting a modulatory role for the N-terminus. It was therefore necessary to determine which of the two domains in the N-terminal region, C2a or C2b, is involved in the modulation of activity. This was done by separately deleting the entire C2a domain or half of the C2b domain. The C2a domain is required in the N-terminal region for full in vivo activity of Gap1. The presence of the C2b domain in the absence of the C2a domain confers a dominant negative activity on the protein. This finding in turn suggests that Gap1 molecules can physically interact either with one another or with other molecules. Deletion of the C2b domain leads to a complete loss of rescue activity in both the partial and the complete loss-of-function background, suggesting that the presence of the C2b domain is crucial for Gap1 in vivo activity. While it is possible that the presence of the C2a domain without the C2b domain inhibits the activity of the rest of the molecule, it seems more likely that the disruption of the C2b domain has a deleterious effect on the remainder of the protein, perhaps through misfolding (Powe, 1999).

Since the C2 and PH domains, which constitute potential regions of interaction with Ca2 + and IP4, are required for Gap1 in vivo activity, it became of interest to see whether genes involved in Ca2+ and phosphoinositide signaling regulate Gap1 function in vivo. One such candidate is the Drosophila gene small wing (sl), which encodes a phospholipaseCgamma (PLCgamma) homolog. PLCgamma comprises a catalytic core, two PH, two SH2, and one SH3 domains and a conserved tyrosine phosphorylation site, all of which act as signal-dependent regulators of enzymatic activity. Subsequent to RTK stimulation, PLCgamma is recruited to the receptor through its SH2 domains and then itself activated by tyrosine phosphorylation. Upon activation, PLCgamma converts phosphatidylinositol-4,5-bisphosphate into diacylglycerol and IP3; IP3, in turn, serves as a precursor for IP4, both of which act as signals for mobilization of intracellular Ca2+. Loss-of-function mutations in the sl locus display a phenotype that is consistent with PLCgamma playing a role as a negative regulator of Egfr signaling, similar to Gap1 (Thackeray, 1998). In the eye, sl complete loss-of-function mutations cause a supernumerary R7 phenotype in approximately 30% of ommatidia; this is qualitatively similar to the mutation caused by Gap1, but much less severe. In hemizygous sl1 mutants, the average number of R7s is 1.3. Thus, the sl complete loss-of-function phenotype is extremely mild compared to that of Gap1 partial or complete loss-of-function mutations. To probe the relationship between PLCgamma and Gap1, a test was performed to determine whether they genetically interact. In a wildtype background, Gap1/+ flies are phenotypically wildtype, i.e. all ommatidia have only one R7 cell (Gaul, 1992; Lai, 1992), while flies hemizygous for sl1 and heterozygous for Gap1 have a supernumerary R7 phenotype greatly enhanced over that seen in hemizygous sl1 flies; greater than 80% of ommatidia have multiple R7 cells. Thus, the phenotype of mutants hemizygous for sl and heterozygous for Gap1 closely resembles that of the partial loss-of-function Gap1S1-11 allele, demonstrating a synergistic positive genetic interaction between sl and Gap1. Given PLCgamma’s role in Ca2+ and phosphoinositide signaling and Gap1’s requirement for its Ca 2+ and IP4-sensitive domains, this result suggests that Gap1 activity is positively regulated by PLCgamma (Powe, 1999).

In order to examine the distribution of Gap1 protein in vivo, full length Gap1 was expressed under the control of the GMR promoter, which drives high levels of expression in all cells posterior to the morphogenetic furrow. Confocal microscopy of immunohistochemically stained eye discs reveals that full length Gap1 protein is concentrated at the plasma membrane in all cells, both inside and outside of developing ommatidial clusters. Examination of the surface of the eye disc shows Gap1 immunoreactivity marking the apical profiles of all cells posterior to the morphogenetic furrow. Using antibodies against HRP to mark the membrane of neuronal cells, Gap1 is found to colocalize with the HRP antigen in the apical microvillar protrusions of photoreceptor cells. These protrusions have been shown to be enriched for membrane-bound signaling molecules, such as Sevenless. Gap1 also appears to colocalize with the HRP antigen throughout the rest of the cell membrane, including axons. These findings suggest that Gap1 is constitutively localized to the plasma membrane, apparently independent of RTK signaling. In addition to its requirement for in vivo activity, the C-terminus, which includes the extended PH domain, is shown to be essential for membrane localization (Powe, 1999).

These findings argue for a link between RTK/Ras signaling and the second messengers Ca2+ and inositol phosphates. The in vivo activity of Gap1 is positively regulated by Ca2+-sensitive and IP4-sensitive domains that are adjacent to the Gap1 catalytic domain, thereby increasing the intrinsic GTPase activity of Ras and hence negatively regulating Ras activity. These results strongly suggest a negative regulation of RTK/Ras signaling by Ca2+ and IP4. Intriguingly, one of the signaling events responsible for increases in cytosolic Ca2+ and IP4 in non-excitable cells is the activation of RTKs themselves. Studies in vertebrates show that growth factor-stimulated RTKs, including Egfr, activate PLCgamma through association with the activated receptor and subsequent phosphorylation. Given the structural conservation of Egfr and PLCgamma between vertebrates and Drosophila, as well as the functional requirement for both in cone cell precursor development, this regulatory relationship is likely preserved. The observed interaction between Gap1 and PLCgamma, in conjunction with these findings, suggests that Egfr through PLCgamma activation could positively influence Gap1 activity. The RTK Torso, whose activity, together with Gap1 and Ras, determines terminal cell fates in the Drosophila embryo, may provide another example for a positive regulatory relationship between RTK, PLCgamma and Gap1. Torso possesses a conserved PLCgamma binding site that binds mammalian PLCgamma and whose removal leads to overactivity of in vivo signaling similar to removal of Gap1, suggesting that activation of the Torso RTK stimulates PLCgamma, which, by stimulation of Gap1, could result in the suppression of Ras1 activity. Thus, if RTKs trigger both activation and deactivation of Ras, then the duration and strength of the Ras signal in response to growth factor stimulation will depend on the kinetics of the pathways involved. Activation of Ras is thought to be achieved largely by membrane localization (and activation) of Sos, which is likely to be fast, whereas the deactivation of Ras by Gap1 may require a more complex series of events, including recruitment and stimulation of PLCgamma and the IP3-mediated increase in intracellular Ca2+, which has been shown to be slow and sustained in response to RTK stimulation (Powe, 1999 and references).


GENE STRUCTURE

cDNA clone length - 4690

Bases in 5' UTR - 434

Bases in 3' UTR - 768


PROTEIN STRUCTURE

Amino Acids - 1165

Structural Domains

Gap1 was first isolated in Drosophila as a negative regulator of RTK signaling (Gaul, 1992). Apart from the catalytic domain in the middle of the molecule, all Gap1 family members share two N-terminal C2 domains and a C-terminal extended PH domain. Biochemical studies with the two mammalian homologs, Gap1m and Gap1IP4BP, have shown that the PH domain, which is required for membrane localization of Gap1, binds IP4 or phosphatidylserine (Fukuda, 1996; Cullen, 1997), suggesting that an increase in intracellular IP4 levels leads to the release of the PH domain from the membrane. IP4 binding to the PH domain also appears to stimulate GAP activity in vitro (Cullen, 1995; Fukuda, 1996), implying that release of the PH domain from the membrane is accompanied by activation of the catalytic domain. The N-terminal C2a domain has been shown to bind phospholipids in a Ca2+-dependent manner (Fukuda, 1996 and Fukuda, 1997), while deletion of the C2b domain, whose function is still not understood, leads to a complete loss of rescue activity in both partial and complete loss-of-function backgrounds (Powe, 1999).


GTPase-activating protein 1: Evolutionary Homologs | Regulation | Developmental Biology | Effects of Mutation | References

date revised: 24 October 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.