PMC

Proteomic analysis of native GBR complexes. a Protein abundance ratios for GBR proteome constituents (n = 3 measurements, data are presented as mean ± s.e.m.) in GBR IPs from membrane fractions of WT, GB1a^−/−, APP^−/−, AJAP^−/−, APP/AJAP^−/−, and PIANP^−/− brains solubilized with the intermediate stringency detergent CL91 (normalized to GB1/2). APP, AJAP-1, and PIANP directly bind to GB1a while APLP2 binds to GB1a via APP. Common changes in the GBR proteomes of GB1a^−/− and APP^−/− mice identify APP as a putative linker between GB1a and the trafficking machinery. b Immunoblot analysis of GBR constituents in brain membrane preparations from knock-out mice. Calnexin serves as a loading control. Lack of APP upregulates PIANP. c APP, AJAP-1, and PIANP assemble into independent GBR complexes. The table summarizes results of co-immunoprecipitations with anti-APP, anti-AJAP-1, and anti-PIANP antibodies from mouse brain membranes solubilized with mild (CL47) and intermediate stringency (CL91) detergents. Knock-out brains were used in control immunoprecipitations. Source data are provided as a Source Data file

APP localizes exogenous GB1a protein to axons in cultured hippocampal neurons. a Representative images of hippocampal neurons co-expressing Myc-GB1a and GFP in APP^−/− and control littermate (WT) mice. Neurons were transfected at DIV5, fixed at DIV10, permeabilized and then stained with anti-Myc antibodies. Note that APP^−/− neurons exhibit significantly reduced axonal Myc-GB1a expression. Dendrites were distinguished from axons using morphological criteria. Scale bar 10 μm. b Higher magnification images of distal axons and dendrites from APP^−/− and WT neurons expressing exogenous Myc-GB1a, GFP and mCherry. Scale bar 10 μm. c Images of distal axons and dendrites from APP^−/− neurons expressing exogenous Myc-GB1a, GFP and APPmCherry or APLP-2mCherry. Note that APPmCherry but not APLP-2mCherry rescues axonal localization of Myc-GB1a. Scale bar 10 μm. d Exogenous Myc-GB1a levels in axons and dendrites of transfected APP^−/− or WT neurons. Normalized fluorescence refers to the Myc-GB1a immunofluorescence intensity normalized to the GFP fluorescence intensity. ****P < 0.0001, unpaired Student’s t-test. e Axon:dendrite (A:D) ratio of Myc-GB1a in APP^−/− and WT neurons transfected with Myc-GB1a in the presence of mCherry, APPmCherry or APLP-2mCherry (DIV10). The n number of neurons analyzed is indicated. ***P < 0.001, ****P < 0.0001, one-way ANOVA. Data are presented as mean ± s.e.m. Source data are provided as a Source Data file

Reduced axonal GBR expression and presynaptic inhibition in APP^−/− mice. a Representative traces of evoked EPSC recordings in CA1 neurons of acute hippocampal slices of APP^−/− mice under baseline conditions (control, black) and during 50 μM baclofen application (red). Bar graphs show reduced baclofen-mediated EPSC amplitude inhibition in APP^−/− mice (**P < 0.01, unpaired Student’s t-test) while EPSC amplitude inhibition in AJAP-1^−/−, PIANP^−/− mice and WT littermate mice did not differ (P > 0.05). The n number of neurons is indicated. b Bar graphs showing reduced baclofen-mediated inhibition of the mEPSC frequency in CA1 pyramidal neurons of APP^−/− mice (***P < 0.001, unpaired Student’s t-test). Baclofen had no effect on the mEPSC amplitude in APP^−/− and WT littermate mice. Baseline mEPSC frequency and amplitude were similar in APP^−/− and WT littermate mice (P > 0.05, unpaired t-test). c Top: Immunofluorescence of endogenous GB1 protein in axons of hippocampal WT (left) and APP^−/− (right) neurons. Neurons expressing GFP were fixed at DIV10, permeabilized, and immunostained for endogenous GB1 (green) and the presynaptic marker piccolo (magenta). GFP served as a volume marker. Merged images show GB1 and piccolo co-localization. Scale bar 5 μm. Bottom: Intensity gray value profile graphs of GB1 (green) and piccolo (magenta). Source data are provided as a Source Data file. Data are presented as mean ± s.e.m

Interacting epitopes of APP, AJAP-1 and PIANP with GB1a. a The acidic domain (AcD) of APP comprising amino acids 191–294 is necessary for GB1a binding. Immunoprecipitations using anti-Myc-antibodies from HEK293 cells expressing Myc-GB1a together with APP or APP deletion mutants. Abbreviations: GFLD, growth factor-like domain; CuBD, copper-binding domain; CAPPD, central APP domain; TMD, transmembrane domain; AICD, APP intracellular domain. b The N-terminal SD1 is necessary for binding of APP. Immunoprecipitations from HEK293 cells expressing Flag-APP together with GB1a deletion mutants lacking SD1, SD2 or both SD1/2. c The AcD of APP interacts with recombinant SD1/2 protein via amino acid residues 202–219 containing a WG sequence motif. Top: Two-dimensional ¹H–¹⁵N heteronuclear single quantum coherence (HSQC) spectra of ¹³C/¹⁵N labeled APP (AcD residues 191–294), alone (blue) or in complex (red) with unlabeled recombinant SD1/2. In complex with SD1/2 several APP residues exhibit chemical shift changes or disappear. These residues participate in protein-protein interaction (red in the sequence alignment). Amino acid assignment of APP was performed from a standard set of 3D experiments using ¹³C–¹⁵N labeled AcD. Bottom: Alignment of the APP epitope with related sequences in AJAP-1 and PIANP (red). APLP-2 exhibits no sequence similarity with the binding epitope of APP. d Deletion or mutation to alanine of the binding epitopes in APP and AJAP-1 prevents binding to GB1a, as shown in co-immunoprecipitation experiments. Source data are provided as a Source Data file

GB1a stabilizes APP at the cell surface. a Cell surface biotinylation of APP in cultured hippocampal neurons of GB1a^−/−, GB1b^−/−, and control WT littermate mice. Bar graph summarizes the densitometric quantification of APP surface levels: WT 100.0 ± 4.1%, GB1a^−/− 69.6 ± 4.9%, **P < 0.01, unpaired Student’s t-test; WT 100.0 ± 2.9%, GB1b^−/−, 89.6 ± 11.2%, P > 0.05, Mann-Whitney. b Cell surface biotinylation of APP in HEK293 cells in the presence or absence of GB1a or GB1b. Bar graphs summarizes the densitometric quantification of APP surface levels. APP: 100 ± 0.9%; APP + GB1a/2: 129.7 ± 5.3%; APP + GB1b/2: 83.9 ± 17.5%; *P < 0.05, one-way ANOVA; n = 3 independent experiments. c To study APP internalization the α-BTX binding site (BBS) was fused to the extracellular N-terminus of APPmCherry (BBS-APPmCherry). BTX-488 and mCherry cell surface fluorescence of HEK293 expressing BBS-APPmCherry with or without GB1a/2 or GB1b/2 before (time 0’) and after BBS-APPmCherry internalization for 15 min at 37 °C (15’). Bar graphs show the mean surface BTX-488 and mCherry fluorescence intensity after 15 min of BBS-APPmCherry internalization. ***P < 0.001, one-way ANOVA, BBS-APPmCherry n = 11, BBS-APPmCherry + GB1a/2 n = 13, BBS-APPmCherry + GB1b/2 n = 13 independent experiments. Scale bar 20 μm. d Representative confocal images of the BTX-488 fluorescence in HEK293 cells expressing BBS-APPmCherry with or without GB1a/2 or GB1b/2 before (0’) and after BBS-APPmCherry internalization for 10, 20, and 30 min. Scale bar 10 μm. e Decrease of BTX-488 surface fluorescence in c over time. n = 14 cells per group, 3 independent transfections per group. ***P < 0.001, one-way ANOVA. Data are presented as mean ± s.e.m. Source data are provided as a Source Data file

GB1a inhibits BACE1-mediated APP proteolysis and Aβ40 generation. a Scheme indicating proteolytic cleavage sites in APP for α-secretase, β-secretase (BACE1) and γ-secretase. APP-FL, APP full-length; sAPP, soluble APP; APP-βCTF, β carboxy-terminal fragment of APP. b Immunoblot of APP cleavage products in HEK293 cells expressing Myc-BACE1, APP and GB2-YFP together with Flag-GB1a or Flag-GB1b. For sAPP analysis, the cell culture medium was filtered and concentrated 32 h post-transfection. Glyceraldehyde 3-phosphate dehydrogenase (GADPH) served as loading control. A significant reduction in the APP-βCTF/APP-FL and the sAPP/APP-FL ratio is observed in the presence of GB1a vs. GB1b. One-way ANOVA, n = 3–4 independent experiments. c Bar graphs of Aβ40 secretion into the culture medium of HEK293 cells expressing APP with or without Myc-BACE1 or Myc-ADAM10 in the presence of GB1a/2 or GB1b/2 (32 h post-transfection). Note that selectively GB1a/2 significantly prevents Aβ40 secretion. One-way ANOVA, n = 6 independent experiments. d GBR activity does not influence Aβ40 production. Bar graphs of the amount of Aβ40 secreted into the culture medium of WT hippocampal neurons after treatment for 7 days with baclofen (10 μM) or CGP54626 (CGP, 10 nM) or both. Values normalized to untreated (100%). One-way ANOVA, n = 3 independent neuronal cultures. e Bar graphs of the amount of Aβ40 secreted within 10 days into the culture medium of hippocampal neurons of GB1a^−/−, GB1b^−/− and control WT littermate mice. Neurons from GB1a^−/− but not GB1b^−/− mice exhibit increased Aβ40 secretion. Unpaired Student’s t-test, n = 4–7 independent neuron preparations. f Lentiviral expression of GB1a but not GB1b decreases Aβ40 secretion in neuronal cultures from GB1a^−/− and WT mice. Bar graphs of Aβ40 secreted within 10 days into the culture medium of hippocampal neurons infected with purified lentiviral particles expressing GFP, GFP-GB1a or GFP-GB1b. GB1a^−/− cultures; unpaired Student’s t-test, n = 4 independent neuronal cultures. WT cultures normalized to uninfected (100%); one-way ANOVA, n = 3 independent neuronal cultures. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as mean ± s.e.m. Source data are provided as a Source Data file

The APP/GB1a complex co-localizes with CSTN and JIP proteins in axons. a Scheme depicting the BiFC principle. Complex formation of APP-VN with GB1a-VC reconstitutes Venus fluorescence and leads to BiFC. GB1b-VC serves as a negative control. Representative confocal images show hippocampal neurons (DIV10) expressing APP-VN together with GB1a-VC or GB1b-VC. BiFC is observed in axons and dendrites for GB1a-VC. Microtubule-associated protein Map2 identifies dendrites; mCherry served as a volume marker. Neurons were imaged 7 h post-transfection. Scale bar 10 μm. b Higher magnification of axons and dendrites of hippocampal neurons transfected with APP-VN and GB1a-VC. The BiFC complex (Venus) partly co-localizes with piccolo (magenta) in axons. The BiFC complex is also present along dendritic shafts but excluded from spines (PSD-95, magenta). Scale bar 5 μm. c Partial co-localization (white, arrowheads) of the BiFC complex (green) with FLAG-CSTN-1, FLAG-CSTN-3, FLAG-JIP-1b, and FLAG-JIP-3 (magenta) and endogenous kinesin light-chain 1 (KLC1) (blue) in the axons of neurons. Scale bar 5 μm. d Quantification of the co-localization of the BiFC complex with FLAG-CSTN-1, FLAG-CSTN-3, FLAG-JIP-1b, and FLAG-JIP-3. The n numbers of neurons analyzed are indicated. e Scheme illustrating that APP together with interacting JIP and CSTN proteins link the GB1a/APP complex in cargo vesicles to axonal kinesin-1 motors. The neural adaptor protein X11-like (X11L) connects APP to CSTN-1²². f Time-lapse images of a well-separated APP-VN/GB1a-VC complex trafficking anterogradely in axons (acquisition times in seconds). White arrowheads mark a fluorescent APP-VN/GB1a-VC complex. A kymograph shows the entire time-lapse recording (right). Scale bars 25 μm. Data are presented as mean ± s.e.m. g Top: Analysis showing the percentage of mobile and non-mobile vesicles per axon within 5 min in hippocampal neurons expressing APPmCherry or the BiFC complex. Bottom: Number of vesicles moving antero-gradely and retrogradely per axons within 5 min. Data are presented in a min to max-box and whisker plot, with whiskers representing the smallest and largest values, the boxes representing the 25–75% percentile and the middle line representing the median. P > 0.05, one-way ANOVA. Source data are provided as a Source Data file