Amyloid precursor proteins interact with the heterotrimeric G protein Go in the control of neuronal migration

Abstract

Amyloid precursor protein (APP) belongs to a family of evolutionarily conserved transmembrane glycoproteins that has been proposed to regulate multiple aspects of cell motility in the nervous system. Although APP is best known as the source of β-amyloid fragments (Aβ) that accumulate in Alzheimer's disease, perturbations affecting normal APP signaling events may also contribute to disease progression. Previous in vitro studies showed that interactions between APP and the heterotrimeric G protein Goα-regulated Goα activity and Go-dependent apoptotic responses, independent of Aβ. However, evidence for authentic APP-Go interactions within the healthy nervous system has been lacking. To address this issue, we have used a combination of in vitro and in vivo strategies to show that endogenously expressed APP family proteins colocalize with Goα in both insect and mammalian nervous systems, including human brain. Using biochemical, pharmacological, and Bimolecular Fluorescence Complementation assays, we have shown that insect APP (APPL) directly interacts with Goα in cell culture and at synaptic terminals within the insect brain, and that this interaction is regulated by Goα activity. We have also adapted a well characterized assay of neuronal migration in the hawkmoth Manduca to show that perturbations affecting APPL and Goα signaling induce the same unique pattern of ectopic, inappropriate growth and migration, analogous to defective migration patterns seen in mice lacking all APP family proteins. These results support the model that APP and its orthologs regulate conserved aspects of neuronal migration and outgrowth in the nervous system by functioning as unconventional Goα-coupled receptors.

Figure 1.

APPL is expressed by migrating EP cells and their motile processes. A–C, Schematic representation of the developmental sequence of EP cell migration (green) along the eight muscle band pathways (magenta) to form the enteric plexus of the ENS in Manduca (only the dorsal four muscle bands are shown). D–F, Montages of whole-mount preparations of filleted embryos that show the developing ENS from corresponding embryonic stages, immunostained for APPL (green) and Fas II (magenta; Fas II). (For these montaged images, Fas II staining was omitted from the EP cells to clearly show their alignment with the muscle bands; see Materials and Methods). A, D, By 55 hpf, the EP cells have spread bilaterally around the surface of the foregut, adjacent to the foregut/midgut (FG/MG) boundary; all of the neurons express APPL as they extend filopodia (arrows) preferentially onto the Fas II-positive muscle band pathways (b). B, E, By 58 hpf, subsets of EP cells have begun migrating along each muscle band pathway, while avoiding the adjacent interband regions (ib). APPL is strongly expressed throughout the motile cell bodies and within their leading processes that have extended along the bands (arrows). C, F, By 65 hpf, the EP cells have completed their migration (arrows) but will continue to grow axonal processes posteriorly along the band pathways for another 15 h, before eventually innervating the lateral visceral musculature. Throughout this developmental sequence, the EP cells maintain robust levels of APPL expression (particularly in their most motile regions) and remain confined to the muscle band pathways, although a small number occasionally extend processes from the foregut-midgut boundary onto the interband regions (asterisks). EN, esophageal nerve of the FG. White scale bar, 50 μm.

Figure 2.

Goα colocalizes with APP family proteins in the motile regions of developing insect and mammalian neurons. A, Schematic diagram of APPL and APP, indicating the conserved extracellular domains (E1 and E2), Aβ domains, and the cytoplasmic Go-binding domain (Go) and internalization domain (Y). Transmembrane domains (tm) are shown in gray. Epitopes targeted by the different anti-APPL and APP antibodies used in this study are indicated by the labeled black lines. Antibodies targeting APP and APPL extracellular domains are labeled in green and antibodies targeting cytoplasmic domains are labeled in magenta. B, Amino acid alignment of the conserved cytoplasmic domains in human (hu) APP, mouse (m) APP, Manduca (ms) APPL, and Drosophila (dm) APPL. Both the putative Go-binding domain (Go) and the internalization domain (Y) within the cytoplasmic (C-terminal) region of APP family proteins are highly conserved across species. Residues that are identical in at least three species are shaded in black; residues shared by two species are shaded in gray. The dotted black line indicates the putative Go-binding domain identified in APP₆₉₅ (Nishimoto et al., 1993; Okamoto et al., 1995). C, EP cells at 65 hpf immunostained with anti-nAPPL (green) and anti-cAPPL (magenta); the white box indicates the region highlighted in D–F. Individual channels in E (anti-nAPPL) and F (anti-cAPPL) are shown as monochrome images. Arrowheads indicate colocalized N- and C-terminal immunostaining (which appears white in D) at the plasma membrane, consistent with the presence of full-length transmembrane APPL. Images in C–F show compressed images of three optical sections acquired by confocal imaging. G, Leading processes of EP cells on the dorsal muscle band pathways, immunostained with anti-nAPPL and anti-cAPPL antibodies; the white box indicates the region highlighted in H–J, which includes three to four fasciculated growth cones. Arrowheads indicate colocalized N- and C-terminal immunostaining in the motile growth cones (white regions in H); individual channels in I (anti-nAPPL) and J (anti-cAPPL) are shown as monochrome images. Images in G–J show compressed images of 10 optical sections. K, EP cells coimmunostained with anti-nAPPL (green) and anti-Goα (magenta); white boxes indicate regions highlighted in L–N and O–Q. Arrowheads in L–N indicate the colocalization of APPL and Goα at the plasma membrane of a migrating neuron; arrowheads in O–Q indicate colocalization in the EP cell growth cone. Individual channels in M and P (anti-nAPPL) and N and Q (anti-Goα) are shown as monochrome images. The images in K–Q show single optical sections. R, Rat hippocampal neurons immunostained with antibodies against N-terminal APP (green, 22C11), C-terminal APP (red, pAPP), and Goα (blue); the white box indicates the highlighted region shown in S–T. S, Enlarged view of colocalized N-APP (green) and C-APP (magenta) immunostaining in the neuronal growth cones, indicating the presence of transmembrane APP. T, Corresponding image of anti-Goα immunostaining (shown in monochrome), consistent with the colocalization of Goα with full-length APP in mammalian growth cones. Images R–T show a compressed image of 10 optical sections. Scale bars: (in C,G, K) 7 μm; (in D–F, H–J, L–Q) 3 μm; (in R), 50 μm; (in S, T) 25 μm.

Figure 3.

Goα interacts with transmembrane APPL and APP in the nervous system. A, Schematic of the experimental protocol: tissue lysates were immunoprecipitated with antibodies specific for either the N- or C-terminal domains of APPL and APP, then immunoblotted with antibodies specific for different G protein subunits. B–E, Western blots of Manduca embryonic lysates immunoprecipitated with anti-nAPPL or anti-cAPPL and immunoblotted with antibodies against different Gα-subunits. B, Embryonic lysate immunoprecipitated with anti-nAPPL and immunoblotted with anti-Goα. Input lane shows endogenous Goα levels in the lysates before immunoprecipitation; IgG IP shows a matched negative control immunoprecipitation. C, Embryonic lysate immunoprecipitated with anti-cAPPL and immunoblotted with anti-Goα. Input lane shows endogenous Goα levels in the lysates; IgY IP shows a matched negative control immunoprecipitation. The size of Manduca Goα is 41 kDa. D–E, Western blots of Manduca embryonic lysates immunoprecipitated with anti-cAPPL and immunoblotted for either Giα (D) or Gsα (E). Input lanes show endogenous Giα and Gsα levels in the lysates before immunoprecipitation. Neither Giα nor Gsα coimmunoprecipitate with anti-cAPPL or with control IgY. The size of Giα is ∼41 kDa; Gsα is detected as a doublet at 48 and 52 kDa (Copenhaver et al., 1995). F–G, Western blots of immunoprecipitated Drosophila head lysates. F, Drosophila head lysates from wild-type and APPL^d flies (which lack APPL expression) immunoprecipitated with anti-cAPPL and immunoblotted with anti-Goα. Inputs show equivalent levels of endogenous Goα in both fly lines. Goα was coimmunoprecipitated with anti-cAPPL antibodies from wild-type lysates but not with control IgY, nor from APPL^d lysates. G, Western blot of lysates from Appl^d flies overexpressing mutant forms of APPL under the control of the GMR promoter. Appl^sd flies (“secretion-deficient”) express a transmembrane form of APPL lacking the juxtamembrane domain that is normally cleaved by secretases; Appl^sd_Δ^Cg flies express secretion-deficient transmembrane APPL, which also lacks the putative Go-binding domain. Fly head lysates were immunoprecipitated with anti-Gi/oα and immunoblotted with anti-cAPPL (135 kDa). Input lanes show abundant expression levels of APPL^sd and APPL^sdΔCg; no APPL-related proteins were detected in APPL^d flies. APPL^sd but not APPL^sdΔCg coimmunoprecipitated with Goα, indicating that the Go domain is necessary for APPL–Goα interactions in vivo. Appl^d flies served as the negative control. Probing for α-Goα verified that abundant levels of Goα were immunoprecipitated from each of the fly lines. Similar levels of Gβ also coimmunoprecipitated with Gi/o in all three fly lines, indicating equivalent expression of the heterotrimeric G protein complexes. The size of fly Gβ is ∼37 kDa. H–J, Western blot of mouse brain lysates immunoprecipitated with N-terminal- or C-terminal-specific APP antibodies and immunoblotted with antibodies targeting different G protein subunits. H, Immunoprecipitated mouse brain lysates labeled with anti-Goα. Input shows endogenous Goα levels in the lysate before immunoprecipitation. Lane 2: Goα coimmunoprecipitated with anti-nAPP. Lane 3: IgG = matched negative control immunoprecipitation. Lane 4: Goα also coimmunoprecipitated with anti-cAPP (8717; Fig. 2). Lane 5: Matched negative control immunoprecipitation with normal rabbit serum (NRS). I, Immunoprecipitated mouse brain lysates immunoblotted for Gsα. Input shows endogenous Gsα levels in the lysate before immunoprecipitation. Lanes 2–4 show that Gsα did not coimmunoprecipitate with anti-nAPP, control IgG, nor anti-cAPP. J, Mouse brain lysates immunoprecipitated with N-terminal- or C-terminal-specific APP antibodies and immunoblotted with α-Gβ. Input shows Gβ levels in the lysates before immunoprecipitation. Lanes 2–4: Gβ coimmunoprecipitated with both anti-nAPP and anti-cAPP antibodies (8717 and cAPP₆₆₈). Lane 5: IgG-negative control immunoprecipitation. K, Immunoprecipitated human brain lysates immunoblotted for Goα. Input shows endogenous Goα levels in the lysate before immunoprecipitation. Goα was coimmunoprecipitated with anti-nAPP but not with control IgG.

Figure 4.

APPL and Goα directly interact in cell culture. A–D, Representative images of COS7 cells at 18–24 h post-transfection, immunostained with anti-Goα (left columns) and anti-APPL (middle columns) to detect constructs of interest. Right-hand columns show green fluorescent BiFC signals that were produced by direct interactions between Vn1- and Vn2-tagged fusion proteins. A, B, Cell transfected with either APPL-Vn1 alone (A₃) or Vn2-Goα alone (B₃) did not emit detectable BiFC signals. C, Cell cotransfected with both APPL-Vn1 and Vn2-Goα exhibited BiFC signals at the plasma membrane and throughout growing processes (C₃, arrowheads). D, Cell cotransfected with Vn2-Goα and APPLΔGo-Vn1 (APPL lacking the putative Go-binding domain; amino acids 762–791) produced only minimal BiFC signals that were restricted to the Golgi/ER regions. No detectable BiFC signals were present at the plasma membrane or growing processes of the cell (D₃, arrowheads), despite the expression of both constructs in these regions (D₁, D₂). E–G, Lower magnification images of COS7 cells cotransfected with APPL-Vn1 plus equivalent levels of different Vn2-tagged Gα subunits. Forty-eight hours post-transfection, cells were coimmunostained with a Vn2-specific anti-GFP antibody and with anti-APPL (to detect APPL-Vn1). E, Cotransfection of APPL-Vn1 plus Vn2-Goα produced robust BiFC signals (C₃), indicating direct APPL-Goα binding (arrowheads). F, Cotransfection of APPL-Vn1 with Vn2-Giα produced only minimal BiFC signals that were primarily localized within the ER/Golgi regions (F₃). BiFC signals were not apparent in the plasma membrane or growing processes of the cells, despite the presence of both Vn-tagged constructs in these regions (F₁, F₂). G, Coexpression of APPL-Vn1 and Vn2-Gsα did not result in any detectable BiFC signals (G₃), although both proteins were expressed throughout the cells (G₁, G₂). Scale bars: 10 μm.

Figure 5.

APPL and Goα directly interact in Drosophila photoreceptors and developing synapses. Isolated eye discs from third instar larvae expressing UAS-APPL-Vn1, UAS-Vn2-Goα, or both constructs (controlled by GMR-GAL4). Discs were fixed and immunostained with α-GFP (Aves #GFP1020) to label both Vn1 and Vn2 epitopes (visualized with Alexa Fluor 568 secondary antibodies); BiFC signals produced by APPL–Goα interactions were imaged in the green channel. A₁, B₁, Eye discs expressing either APPL-Vn1 or Vn2-Goα alone did not exhibit detectable BiFC signals (A₂, B₂). C₁, Coexpression of APPL-Vn1 and Vn2-Goα resulted in robust BiFC signals throughout the fly photoreceptors (C₂) and their axonal projections extending into the optic stalk (arrowheads). Small white boxes in the low-magnification images indicate highlighted regions shown in the insets of each panel. D, E, Isolated eye disc-brain complexes from mid-third instar Drosophila larvae that expressed Vn2-Goα plus either wild-type APPL-Vn1 or APPLΔGo-Vn1. Magenta outlines demarcate the proximal brain lobe (br) in each preparation. D₁, E₁, Eye disc–brain complexes from mid-third instar larvae coexpressing Vn2-Goα plus mutated or wild-type forms of APPL-Vn1. Immunostaining for anti-GFP showed that the Venus-tagged constructs were expressed in the developing adult photoreceptors (adjacent to the morphogenetic furrow; arrowheads) and throughout Bolwig's nerve (bn), which carries the axons of 12 larval photoreceptors through the eye disc into the larval optic neuropil of the brain (lon; arrows). Larval photoreceptors are out of the field of view. D₂, The combined expression of APPL-Vn1 and Vn2-Goα produced robust BiFC signals in regions where both proteins were coexpressed (indicative of direct APPL–Goα interactions), including the synaptic projections of the larval photoreceptors in the lon. E₂, Coexpression of APPLΔGo-Vn1 plus Vn2-Goα produced no detectable BiFC signals, despite comparable expression levels for both APPL constructs in D and E (data not shown). D₃, Schematic representation of the late third instar larval eye disc–brain complex (equivalent to dissected preparations in D and E), illustrating the orientation of bn and the lon. Arrowheads indicate the morphogenetic furrow; dotted outline in the br indicates the adult optic anlage (oa), forming adjacent to the lon. F, Eye disc–brain complex from a white pupa expressing APPL-Vn1 plus Vn2-Goα. F₁, Immunostaining with anti-GFP revealed the expression of the Venus-tagged constructs in developing adult photoreceptors, their projecting axons in the optic stalk (os), and their differentiating synaptic terminals within the lamina cortex (la) of the brain. F₂, Magnified image of the boxed region in F₁ (compressed image of 16 optical sections) to highlight the presence of BiFC signals in the photoreceptor axons within the os and their synaptic terminals in the la. F₃, Schematic representation of the eye disc–brain complex at the white pupal stage (equivalent to preparation in F₁), illustrating the orientation of the os and la regions; small oblong circle (gray) indicates the position of the residual lon (not visible in F₁ and F₂). Scale bars: (in A–C) 25 μm; (in A–C inset boxes) 15 μm; D–F₁, 30 μm; (in F₂) 10 μm.

Figure 6.

APPL–Goα interactions are regulated by Goα activity. A, Schematic of the model that inactive Goα binds APPL as part of a heterotrimeric complex with Gβγ. Activation of Goα (normally due to the exchange of bound GDP for GTP) promotes its dissociation from APPL, whereas inhibiting Goα prevents this dissociation, resulting in an increase in basal APPL–Goα interactions. B, Replicate cultures of Manduca GV1 cells (which endogenously express APPL and Goα) were treated as indicated, then lysed, immunoprecipitated with α-APPL, and immunoblotted with anti-Goα. Relative levels of coimmunoprecipitated Goα were normalized to “no treatment” control run in parallel (lane 1). Stimulating G protein activity with GTPγS (a nonhydrolyzable activator of G proteins) led to a concentration-dependent reduction in the amount of Goα that coimmunoprecipitated with APPL (lanes 2–4). IgY = matched negative control immunoprecipitation (lane 5). Treatment with GDPβS (a nonhydrolyzable G protein inhibitor) caused a modest increase in APPL–Goα interactions (lane 6). C, Lysates from staged groups of embryos (60–65 hpf) were treated as indicated, immunoprecipitated with α-APPL, and immunoblotted with anti-Goα. Gel shows a representative Western blot of coimmunoprecipitated Goα (upper blot) and APPL (lower blot; labeled with anti-nAPPL). Double arrow indicates the immature (smaller) and mature (larger) forms of full-length APPL. Goα levels were calculated as a ratio of immunoprecipitated full-length APPL in the same sample, and ratios were normalized to untreated controls (lane 1). Treatment with 1.0 μm GTPγS decreased APPL–Goα interactions (lane 2). Treatment with Mas 7 (to activate Giα/Goα) decreased APPL–Goα interactions in a concentration-dependent manner (lanes 3–4). In contrast, inhibiting Goα with PTX enhanced APPL–Goα interactions (lanes 5–6). Treatment with 20 μm GDPβS caused only a minor increase in coimmunoprecipitated Goα levels (lane 7). IgY = negative control immunoprecipitation (lane 8). D, Combined analysis of multiple experiments in which Manduca embryonic lysates were treated with G-protein-specific reagents, then immunoprecipitated with anti-cAPPL and immunoblotted with anti-Goα. Each manipulation was repeated in at least three independent experiments and normalized to their respective untreated controls in the same assay. Treatment with both GTPγS and Mas 7 significantly reduced APPL–Goα interactions. Mas 17, an inactive mastoparan analog, had no significant effect, compared with controls. PTX induced a concentration-dependent increase in APPL–Goα interactions. Although treatment with GDPβS caused a slight increase in APPL–Goα interactions in some experiments, overall this effect was not significant at the concentrations tested. IgY represents the species-matched negative control immunoprecipitations performed in each assay. Pairwise statistical analyses were performed between the control and each experimental group using Student's two-tailed t tests, *p < 0.05; **p < 0.01. Error bars indicate SEM.

Figure 7.

Inhibiting Goα activity in the EP cells induces ectopic migration and outgrowth. A–D, Representative camera lucida drawings of Manduca embryos in which the EP cells were treated before migration onset with reagents targeting Goα, grown in culture for another 24 h, then fixed and immunostained with anti-Fas II to reveal the full extent of migration and outgrowth in the developing ENS. A, Control embryo immunostained at the onset of an experiment (52 hpf) to show the initial positions of the premigratory EP cells adjacent to the FG/MG boundary. By this stage, subsets of neurons had begun to extend exploratory filopodia onto their future muscle band pathways (b) but had not commenced their migratory dispersal. B, Embryo opened at 52 hpf and treated with control culture medium throughout 24 h of development showed the normal pattern of migration and outgrowth along the muscle bands, with only a few processes growing onto the interband musculature (ib). C, Treatment with 12.5 μm AlF₄₋ (an activator of heterotrimeric G proteins) caused almost complete inhibition of migration and outgrowth, stalling the neurons at the FG/MG boundary. D, Treatment with 100 ng/ml PTX (a specific inhibitor of Goα in insects) induced a distinctive pattern of ectopic migration (open arrows) and outgrowth onto the interband regions (black arrowheads). E, Average distances of neuronal migration and axon outgrowth along band pathways for each treatment group (normalized to matched sets of control embryos in each experiment). F, Quantification of the average number of neurons per embryo that exhibited ectopic migration into each interband region for each treatment condition. G, Quantification of the average extent of ectopic axon outgrowth in the interband regions of embryos treated with various reagents. In E–G, pairwise statistical analyses were performed between each experimental group and their corresponding controls, using Student's two-tailed t tests, *p < 0.05; ^#p < 0.02; **p < 0.01; ***p < 0.001. N = at least 10 per condition. Error bars indicate SEM.

Figure 8.

Disrupting APPL signaling in the EP cells induces ectopic migration and outgrowth. A, Western blot of Manduca GV1 cells treated with APPL-specific morpholinos (MOs) for 48 h, then lysed and immunoblotted with anti-APPL to demonstrate the selective knock down of APPL expression. Immunoblotting the same gel with anti-tubulin served as a control for nonspecific effects of the MOs. Lanes 1–2: Replicate GV1 cell lysates treated with 20 μm standard control MOs (plus 0.6% Endo-Porter). Lanes 3–8: Replicate lysates treated with 0.6% Endo-Porter plus increasing concentrations of APPL MOs. B, Quantification of replicate experiments demonstrating the effectiveness of the APPL MOs at different concentrations (normalized to tubulin levels); treatment with 5 and 20 μm APPL significantly inhibited APPL protein levels. Quantification of Goα levels in the same samples (also normalized to tubulin) showed that Goα was unaffected by the MO treatments. C–H, Inhibiting APPL signaling in the EP cells phenocopies the effect of inhibiting Goα activity on ectopic migration and outgrowth. C–E, Camera lucida drawings of Manduca embryos in which the EP cells were treated before migration onset with reagents targeting APPL signaling, allowed to develop in culture for another 24–48 h, then fixed and immunostained with anti-Fas II to reveal the full extent of migration and outgrowth. C, Control embryo treated with culture medium exhibited a normal pattern of migration and outgrowth (same preparation shown in Fig. 7B). D, Treatment with APPL-specific MOs (50 μg/ml) induced the same overall pattern of ectopic migration and outgrowth caused by inhibiting Goα activity. E, Treatment with anti-nAPPL (1 μg/ml) also induced a similar phenotype. Open arrows in D and E indicate ectopic neurons in the interband regions; black arrowheads indicate ectopic processes. F, Average distances of neuronal migration and axon outgrowth along the band pathways for each treatment group (normalized to matched sets of control embryos in each experiment). G, Quantification of the average number of neurons that exhibited ectopic migration onto the interband regions for each treatment condition. H, Quantification of the average extent of ectopic axon outgrowth per interband region for each treatment condition. *p < 0.05; ^#p < 0.02; **p < 0.01; ***p < 0.001; Student's two-tailed t tests. N = at least 10 per condition. Error bars indicate SEM.

Figure 9.

A model for the role of APPL-Goα signaling in the control of neuronal migration within the developing ENS. Transmembrane APPL is expressed in the motile processes of migrating EP cells, where it colocalizes with Goα. When exploratory filopodia extend off the band pathway (b) onto the adjacent interband regions (ib), they encounter ligands (as yet unidentified; black octagons) that induce APPL-dependent activation of Goα. In turn, local activation of Goα within the leading process induces Ca²⁺ influx via voltage-independent channels (Horgan and Copenhaver, 1998), resulting in filopodial retraction, thereby restricting inappropriate outgrowth and migration into these regions.