The Lnk/SH2B adaptor provides a fail-safe mechanism to establish the Insulin receptor-Chico interaction

Abstract

Background: Insulin/insulin-like growth factor signalling (IIS) has been described as one of the major pathways involved in growth control and homeostasis in multicellular organisms. Whereas its core components are well established, less is known about the molecular functions of IIS regulators. The adaptor molecule Lnk/SH2B has been implicated in IIS but the mechanism by which it promotes IIS activity has remained enigmatic.

Results: In this study, we analyse genetic and physical interactions among InR, Chico and Lnk in Drosophila tissues. FRET analysis reveals in vivo binding between all three molecules. Genetically, Lnk acts upstream of Chico. We demonstrate that Chico's plasma membrane localisation is ensured by both its PH domain and by the interaction with Lnk. Furthermore, Lnk is able to recruit an intracellular InR fragment to the membrane.

Conclusions: Thus, by acting as a scaffolding molecule that ensures InR and Chico enrichment at the membrane, Lnk provides a fail-safe mechanism for IIS activation.

Figure 1

In vivo FRET analysis reveals physical interactions among Lnk, Chico and InR. (A-A”’) Chico-RFP (A’) and InR-CFP (A) strongly co-localise in salivary glands (A”). FRET^c shows regions with high energy transfer between CFP and RFP (colour code: black - low intensities, red - high intensities). (B-B”’) Lnk-CFP (B) and Chico-RFP (B’) co-localise (B”) showing positive FRET^c (B”’). (C-C”’) InR-CFP (C) and Lnk-RFP (C’) localisation (C”) and FRET (C”’) at the cortical membrane. (D-D”’) InR-CFP (D) and Dl-RFP (D’) do not exhibit energy transfer (D”’). (E-E”’) Lnk-CFP (E) and Dl-RFP (E’) do not show positive FRET^c (E”’). (F) Table showing data from FRET analyses. Number of samples analysed: (A) n = 21 (+ins) and n = 10 (−ins), (B) n = 20 (+ins) and n = 10 (−ins), (C) n = 17 (+ins) and n = 11 (−ins), (D) n = 15, (E) n = 14. Scale bars represent 50 μm.

Figure 2

Lnk acts upstream of Chico ensuring its proper localisation. (A-D) Low phospho-PKB levels in chico^−/− MARCM clones are not rescued by overexpression of lnk-CFP. Clones are marked by GFP expression (’ panels) and outlined by a dashed line in the detail pictures (”” panels). Phospho-PKB staining is shown in ” panels. (A-A””) control clones. (B-B””) Levels of phospho-PKB (B”) are decreased in chico^−/− clones. (C-C””) Overexpression of lnk-CFP within control clones promotes high levels of phospho-PKB (C”). (D-D””) lnk-CFP overexpression within chico^−/− clones does not rescue low levels of phospho-PKB (D”). (E-G) Chico overexpression restores localisation of the tGPH reporter at the membrane in lnk^−/− salivary glands. (E) tGPH reporter localises at the membrane in wild-type salivary glands. (F) lnk^4Q3 salivary glands show diffuse tGPH reporter signal in the cytoplasm. (G) tGPH reporter is recovered to the membrane of lnk^4Q3 tissue by overexpressing chico. (H-K) Subcellular localisation of RFP-tagged Chico. The RFP signal is shown in ’ panels. Actin staining marks the cellular cortex in ” panels. Merged pictures are shown in ”’ panels. Signal intensities along the lines indicated in the ”’ panels are displayed in ”” panels. Quantifications in Additional file 1: Figure S1A. (H-H””) Cortical localisation of Chico-RFP in a wild-type situation. (I-I””) Chico-RFP in a lnk mutant background. (J-J””) A PH domain mutant form of Chico localises at the cortical membrane. (K-K””) Chico-PH*-RFP in a lnk mutant background. Number of samples analysed: (A) n=9, (B) n=9, (C) n = 10, (D) n = 12, (E) n=5, (F) n=5, (G) n=6, (H) n=6, (I) n=9, (J) n=6, (K) n = 6. Scale bars represent 50 μm.

Figure 3

Lnk recruits the intracellular part of InR to the membrane. CFP-tagged InR (fragments) are shown in ’ panels. Actin staining marks the cellular cortex in ” panels except for E and F where Lnk-RFP is shown in the ” panels. Merged pictures are shown in ”’ panels. Signal intensities measured along the indicated lines in ”’ panels are displayed in ”” panels. Quantifications in Additional file 1: Figure S1B. (A-B) InR-CFP is substantially reduced at the cellular cortex in lnk mutant tissue (B-B””) in comparison to wild-type salivary glands (A-A””). (C-C””) Localisation of InR^INTRA-CFP (C’) in wild-type tissue. (D-D””) InR^INTRA-CFP in lnk^−/− tissue. (E-E””) Co-localisation of InR^INTRA-CFP (E’) and Lnk-RFP (E”) at the cellular membrane. (F-F””) Co-localisation of InR^INTRA-CFP (F’) and Lnk-PH*-RFP (F”) in the cytoplasm. Number of samples analysed: (A) n=5, (B) n=6, (C) n=6, (D) n=7, (E) n=6, (F) n = 5. Scale bars represent 50 μm.

Figure 4

Lnk reinforces membrane localisation of Chico-RFP and InR^INTRA-CFP. CFP-tagged InR^INTRA is shown in ’ panels, RFP-tagged Chico versions in ” panels. Merged pictures are shown in ”’ panels. Signal intensities measured along the indicated lines in ”’ panels are displayed in ”” panels. Quantifications in Additional file 1: Figure S1C. (A-A””) InR^INTRA-CFP (A’) and Chico-RFP (A”) in a wild-type background. (B-B””) InR^INTRA-CFP (B’) and Chico-RFP (B”) in lnk^4Q3. (C-C””) InR^INTRA-CFP (C’) and Chico-PH*-RFP (C”) in wild-type tissue. (D-D””) InR^INTRA-CFP (D’) and Chico-PH*-RFP (D”) in lnk^−/−. (E-E”’) Model of Lnk function in IIS. In wild-type tissue, Lnk ensures enrichment of InR and Chico at the cortical membrane (E). In the absence of Chico, the p60 subunit of PI3K is able to directly bind to InR (E’). In lnk mutants, despite the reduction of InR at the cellular membrane, Chico is able to bind to residual InR; hence the IIS pathway is partially active (E”). By contrast, when Lnk and Chico are lacking, p60 localisation by the residual InR at the membrane is not sufficient to promote pathway activity (E”’). Number of samples analysed: (A) n=8, (B) n=7, (C) n=7, (D) n = 6. Scale bars represent 50 μm.