daughter of sevenless

Protein Interactions

The pleckstrin homology (PH) domain-containing protein Daughter of Sevenless (Dos) is an essential component of the Sevenless receptor tyrosine kinase (SEV) signaling cascade, which specifies R7 photoreceptor development in the Drosophila eye. Previous results have suggested that Dos becomes tyrosine phosphorylated during Sev signaling and collaborates with the protein tyrosine phosphatase Corkscrew (Csw). This possibility was investigated by identifying tyrosine residues 801 and 854 of Dos as the phosphorylated binding sites for the Csw SH2 domains. These sites become phosphorylated in response to Sev activation and phosphorylation of both sites is required to allow Csw to bind Dos. Mutant Dos proteins in which either Y801 or Y854 of Dos has been changed to phenylalanine are unable to function during signaling by Sev and other receptor tyrosine kinases. In contrast, a mutant Dos protein in which all tyrosine phosphorylation sites except Y801 and Y854 have been removed is able effectively to provide Dos function during Sev signaling and to rescue the lethality associated with dos loss-of-function mutations. These results indicate that a primary role for Dos during signaling by Sev and other receptor tyrosine kinases is to become phosphorylated at Y801 and Y854 and then recruit Csw (Herbst, 1999).

Structural and enzymatic studies of SHP-2, the mammalian homolog of Csw, have shown that binding of the N-terminal SH2 domain to a phosphorylated tyrosine-containing peptide leads to a marked increase in SHP-2 catalytic activity. Thus, an expected consequence of the binding of Csw to Dos is a substantial increase in Csw catalytic activity. Given the requirement for the Csw catalytic activity during Sev signaling, a key question is the identity of the key Csw target(s) whose dephosphorylation is required for R7 photoreceptor development. Dos itself might be a crucial target of Csw. In particular, Dos might contain sites of tyrosine phosphorylation that serve as docking sites for proteins inhibitory to Sev signaling and these sites might be dephosphorylated by Csw. This idea is based both on studies showing that Dos is a potential Csw substrate and on genetic studies indicating that reduced Csw or Dos function decreases the strength of Sev signaling. Given this model, an expectation is that mapping the sites of tyrosine phosphorylation in catalytically inactive Csw^CS-bound Dos might identify both the Csw SH2 domain-binding sites and additional Csw target sites. Instead, this analysis reveals phosphorylation of only the two Csw SH2 domain-binding sites, Y801 and Y854. Together, the (1) failure to find additional sites of Dos tyrosine phosphorylation in Csw^CS-bound Dos and (2) the ability of Dos^YY (in which all tyrosines outside of the Dos PH domain except Y801 and Y854 are either removed by deletion or changed to phenylalanine) to provide Dos function during Sev signaling, suggest that the primary role of Dos may be to activate the catalytic activity of Csw towards other, as yet unidentified, proteins. In this model, the role of Csw's dephosphorylation of Dos would be to down-regulate signaling by eliminating its own binding sites (Herbst, 1999).

Dos is a phosphorylation target of Sev, but a direct interaction between Dos and Sev has not been so far demonstrated. Previously it had been shown that Gab1 recruitment to RTKs can be accomplished through interaction with Grb2. Gab1 can form a constitutive complex with Grb2 which is primarily mediated via the C-terminal SH3 domain of Grb2 (Fixman, 1997; Nguyen, 1997). To determine whether the Drosophila Grb2 homolog Downstream of receptor kinase (Drk) can bind to Dos via its N- or C-terminal SH3 domain, precipitation experiments were performed. The N- and C-terminal Drk SH3 domains were expressed as glutathione S-transferase (GST)-fusion proteins in bacteria [Drk-SH3(N) and Drk-SH3(C)]. As a source of Dos protein, transfected HEK293 cells expressing the dos cDNA under the control of the CMV promotor were used. Protein lysates from transfected and non-transfected cells were prepared and analyzed for expression of Dos using an anti-Dos antiserum or incubated with purified Drk-SH3(N) or Drk-SH3(C) proteins. Western blot analysis of the purified protein complexes shows that Dos binds selectively to Drk-SH3(C) but not to Drk-SH3(N) in vitro. Very similar results were obtained with protein extracts from transgenic flies expressing the dos cDNA under the control of the heat shock (hsp70) promotor. As a further control for specificity the SH3 domain of PLCgamma, another putative interaction partner of Dos, was tested: no binding of the PLCgamma SH3 domain to Dos was detected (Feller, 2002).

Dos has several PxxP motifs, but none of them fits the PxxPxR motif identified in the guanine nucleotide exchange factor Sos as the binding site for the Drk SH3 domains (Raabe, 1995). More recently it has been shown that the Grb2 C-terminal SH3 domain binds to the sequence motif PxxxRxxKP. Dos contains two of these sites (pos 640: PPVNRKLKP and pos 692: PSVDRKCKP). To map the interaction sites between Dos and Drk, several Dos deletion constructs were expressed in HEK293 cells and tested for their ability to bind to the Drk SH3 (C) domain in vitro. Only deletion of Dos protein sequences encompassing both PxxxRxxKP sites (positions 477-878) completely abolishes the ability of Dos to associate with Drk-SH3(C) while deletion of the PxxP containing region (positions 211-507) does not affect binding (Feller, 2002).

To determine whether both PxxxRxxKP motifs in Dos indeed can function as binding sites for the C-terminal SH3 domain of Drk, the conserved arginine residues were mutated in either predicted binding site alone (R644K and R696K) or in combination (R644, 696K) and the ability of the corresponding proteins upon expression in HEK293 cells to bind to Drk-SH3(C) was tested. Mutation of both binding sites completely abolishes binding of Dos to Drk-SH3(C) in vitro while mutation of either binding site alone results in a severe reduction of binding. This result indicates that both PxxxRxxKP motifs in Dos are able to bind to Drk-SH3(C). In order to confirm this result, peptides corresponding to amino acids 638-650 (Dos-P1) and 690-702 (Dos-P2) were synthesized and used to compete for the Dos-Drk interaction. A peptide derived from the human SLP-76 protein (positions 231-243) was used as a positive control. This peptide contains a PxxxRxxKP motif and binds with high affinity to the Grb2-SH3(C) domain. As a control for specificity, a point-mutated SLP-76 peptide bearing an arginine to lysine substitution was included. For the peptide competition experiments, protein extracts from flies expressing the dos cDNA under the control of the hsp70 promoter were used. Flies were heat shocked for 1 h at 35°C to induce ubiquitous expression of the dos-transgene and allowed to recover at room temperature for 5 h. Protein extracts were directly analyzed for Dos expression or incubated with the Drk-SH3(C) protein in the presence or absence of the indicated peptide. Both Dos-derived peptides and the wild-type SLP-76 peptide efficiently block binding of the Drk-SH3(C) domain to Dos. In contrast, the mutant SLP-76 peptide bearing a substitution of the central arginine residue by lysine is impaired in its ability to compete for the Dos-Drk-SH3(C) interaction. Direct binding of the Dos-P1 and Dos-P2 peptides to the Drk-SH3(C) domain was also confirmed by tryptophan fluorescence measurements and isothermal titration calorimetry (Feller, 2002).

Because these are based on in vitro interactions of the isolated Drk-SH3(C) domain with Dos, coimmunoprecipitation experiments were performed to determine whether full length Drk can associate in vivo with the wild-type Dos protein or a Dos protein with both Drk binding sites mutated (Dos^R644,696K). A Myc-tagged Drk protein was coexpressed with Dos or Dos^R644,696K in HEK293 cells and the immunopurified protein complexes were analyzed by Western blot analysis. Only the wild-type Dos protein interacts with Drk thus confirming the requirement of the identified PxxxRxxKP sites for binding of Drk to Dos in vivo (Feller, 2002).

In summary, these experiments indicate that the two PxxxRxxKP motifs in Dos constitute the major sites for a direct interaction with the C-terminal Drk SH3 domain. Therefore, these binding sites were named Drk binding site 1 (DBS1, amino acids 638-650) and Drk binding site 2 (DBS2, amino acids 690-702) (Feller, 2002).

To determine whether the identified SH3 domain binding sites on Dos have an in vivo functional role in RTK signaling, transgenic flies were generated expressing the Dos, Dos^R644K, Dos^R696K or Dos^R644,696K protein (R644 and R696 are Drk binding sites) under the control of the sevenless (sev) enhancer and hsp70 promotor sequences (sE/hsp). This allowed cell type specific expression of the transgenic Dos proteins in all Sev expressing cells during eye development, or ubiquitous expression upon heat shock induction. Protein expression of the different transgenes was verified by Western blot analysis. Upon heat induction, similar levels of Dos protein accumulate in the transgenic strains expressing either the wild-type or the mutant proteins (Feller, 2002).

As a first functional test, the ability of the Dos, Dos^R644K, Dos^R696K and Dos^R644,696K proteins to compensate for the removal of the endogenous Dos protein function in the Sev pathway was tested. Expression of a constitutively activated version of the Sev RTK under the control of the sev enhancer (sE-sev^S11) causes a rough eye phenotype due to the transformation of cone cells into additional R7 cells. The number of supernumerary R7 cells is dependent on the expression level of the activated Sev protein and can be modulated by altering the gene dosage of downstream signaling molecules. In the heterozygous state, the dos^R31 mutation largely suppresses the rough eye phenotype (Raabe, 1996) such that most ommatidia have the wild-type number of photoreceptor cells. The dos^R31 allele carries a nonsense mutation leading to premature termination of translation at amino acid position 463 (Bausenwein, 2000). Thus, in dos^R31, both identified Drk SH3 domain binding sites are deleted. As previously shown (Bausenwein, 2000), one copy of the sE/hsp-dos construct driven by the sE-enhancer is sufficient to restore the multiple R7 phenotype in sE/hsp-dos/+; dos^R31/sE-sev^S11 flies. In contrast, expression of the sE/hsp-dos^R644,696K construct in a dos^R31/sE-sev^S11 background does not result in a multi-R7 cell phenotype. Compared to dos^R31/sE-sev^S11 flies, the average number of R7 cells per ommatidium is not increased, indicating that a Dos protein lacking DBS1 and DBS2 is unable to fulfil its normal function during Sev signaling. To determine whether both SH3 domain binding sites are of equal importance for Sev signaling, flies expressing either the Dos^R644K or Dos^R696K protein were tested in the same assay. Both single site mutant transgenes were largely non-functional although a slight increase in the number of ommatidia containing more than one R7 cell is consistently observed with different independent transgenic lines bearing the sE/hsp-dos^R644K construct. This indicates that the Dos^R644K protein still has some residual activity. In summary, these experiments provide evidence that both Drk SH3 domain binding sites identified in this study have an essential function to transduce the signal from the Sev RTK (Feller, 2002).

Dos also functions downstream of the DER and Torso RTKs. Removal of Dos function in the eye and the wing results in phenotypes comparable to the phenotypes observed with Egfr loss-of-function alleles or mutations in other genes of the Ras/MAPK signaling cassette, namely absence of photoreceptor cell development and lack of wing veins (Raabe, 1996). Heat shock-induced expression of the wild-type dos cDNA is sufficient to rescue the lethality of homozygous dos^R31 animals. Flies expressing the wild-type dos cDNA in a dos^R31 background display no abnormalities; in particular, no defects were observed in eye or wing vein morphology. Heat shock-induced expression of the sE/hsp-dos^R644K transgene rescues the lethality of homozygous dos^R31 animals, although these flies display a rough eye phenotype with many ommatidia lacking one or more photoreceptor cells. This phenotype becomes much more pronounced when the sE/hsp-dos^R696K transgene is expressed in a homozygous dos^R31 background. Only a few flies survive and the eyes of these flies are reduced in size and most ommatidia lack one or more photoreceptor cells. In addition, the lack of wing vein structures is frequently observed. In contrast to the Dos^R644K or Dos^R696K proteins, the Dos^R644,696K protein completely fails to rescue the lethality of dos^R31 animals. From these experiments it is concluded that Drk binding sites DBS1 and DBS2 in Dos are essential for Egfr mediated signaling in the eye. Mutant Dos proteins lacking either binding site alone are only impaired in their normal function whereas the Dos^R644,696K protein is non-functional in this assay (Feller, 2002).

As a further control, all transgenes were tested for their ability to rescue the lethality of transheterozygous dos^R31/dos^P115 and homozygous dos^P115 animals. Dos^P115 contains a P-element insertion within the first intron of the dos transcription unit (Raabe, 1996). Since the translation start site lies in the first exon, dos^P115 was considered to be a complete loss of function allele. The lethality of dos^R31/dos^P115 or homozygous dos^P115 animals is rescued by heat shocked induced expression of the Dos^R644K and Dos^R696K proteins, but the eye phenotypes of the rescued flies are less pronounced than those seen in a homozygous dos^R31 background. In contrast to the results obtained with the dos^R31 allele, expression of the sE/hsp-dos^R644,696K construct also results in a few rescued dos^R31/dos^P115 or homozygous dos^P115 flies having greatly reduced eyes with only few photoreceptor cells present in most ommatidia. Wing vein formation is also disrupted. The phenotypes are characteristic of reduced Egfr signaling. The different behavior of the dos^R644,696K transgene to rescue the lethality of the dos^R31 or dos^P115 animals could indicate that dos^P115 is a hypomorphic allele. Alternatively, the Dos^R644,696K protein has residual activity that is sufficient to rescue the lethality of dos^P115 animals. In this case, expression of a truncated Dos protein in dos^R31 animals could negatively interfere with the Dos^R644,696K protein masking its residual function (Feller, 2002).

Nevertheless, from the genetic studies it is evident that both Drk SH3 domain binding sites in Dos have an important function for signal transduction by least two different RTKs. In general, mutation of DBS2 causes more severe defects than mutation of DBS1 (Feller, 2002).

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

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