Fig 1: ATG9A is mislocalised in NDR1/2 KO brains.(A) Immunofluorescence staining of ATG9A in brain slices from 12-wk-old NDR1/2 knockout and control mice. White arrows indicate areas with increased endogenous ATG9A. Scale bars: 10 µm. (B) Western blot analyses of ATG9A in cortical lysates from 6-wk-old NDR1/2 KO and control mice. GAPDH was used as a loading control. The bar graph shows quantification of ATG9A normalised against GAPDH, and the data were analysed with an unpaired t test. n = 4–5 mice/group. (C) Immunofluorescence staining of ATG9A and the Golgi marker GM130 in the CA1 area of 12-wk-old mice. White arrows indicate areas where ATG9A co-localises with GM130. Yellow arrows indicate areas where ATG9A does not co-localise with GM130 in NDR1/2 knockout mice. Scale bars: 10 µm. The scatter plot shows quantification of ATG9A and GM130 co-localisation, expressed as a Pearson correlation coefficient, and the data were analysed with a Mann–Whitney test. n = 28 measurements from three mice/genotype. (D) Immunofluorescence staining of ATG9A and p62 at 12 wk. White dotted lines delineate the stratum oriens (s.o.) area and the CA1 cell body area (CA1 c.b.). Scale bars: 50 µm. (E) Immunofluorescence stainings of ATG9A and the presynaptic marker synaptophysin 1 in stratum oriens at 12 wk. White arrows show ATG9A co-localising with synaptophysin 1, and yellow arrows show ATG9A puncta that do not co-localise with synaptophysin 1. Scale bars: 5 µm. (F, G) Western blot analyses of surface biotinylation experiments in DIV12 rat cortical neurons infected with scramble (SCRM) shRNA and NDR1,2 shRNA lentivirus (F) or Raph1 shRNA (G). Surface levels of ATG9A were normalised against input. Bar graphs show surface ATG9A levels that are expressed as a percentage of the corresponding SCRM shRNA control level. The data were obtained in the same experiment shown in Fig 3B and E; therefore, the same tubulin control is used. The data were analysed using a one-sample t test. n = 6 samples/condition from three independent experiments. (H, I) Representative images (H) and kymographs (I) of ATG9A vesicle movement in the axons of DIV11 cultured rat hippocampal neurons, previously infected with a scramble shRNA lentiviral vector as a control or NDR1 and NDR2 shRNA–expressing lentiviruses. The total number of moving particles were quantified from kymographs. The data were analysed using Mann–Whitney tests. n = 138 mobile particles from 33 cells for scramble shRNA, and n = 37 mobile particles from 24 cells for NDR1/2 shRNA.Source data are available online for this figure.
Fig 2: Dual loss of NDR1/2 in adults is sufficient to induce p62 accumulation, ATG9A mislocalisation and neuropil loss.(A) Immunofluorescence staining of GFAP in the brains of NDR1/2 inducible knockout (NDR1/2 iKO) and control mice. White arrows show the distribution of neurons within the CA1 cell body layer. Arrowed lines show the length of stratum radiatum. Scale bars: 200 µm. (B) Graphs showing quantifications of cortex and stratum radiatum thickness. The data were analysed using paired t tests. n = 18 measurements from three mice/genotype. (C) Immunofluorescence staining of p62 in NDR1/2 iKO and control mice in the CA1 area. White arrows show Ai14-TdTomato–negative cells, which have not been recombined and lack p62 accumulations, and scale bars represent 50 µm. (D, E) Immunofluorescence staining of ATG9A (D) and transferrin receptor (TfR) (E) in the brains of NDR1/2 iKO and control mice in the CA1 area. White arrows show Ai14-TdTomato–negative cells, which have a perinuclear distribution of ATG9A (D) and lack TfR accumulations (E). Yellow arrows show Ai14-TdTomato–positive cells, which exhibit a punctate distribution of ATG9A (D) and a high number of TfR puncta (E). Scale bars: 10 µm. (F) Immunofluorescence staining of p62 in areas of the cortex with high Ai14-TdTomato expression. White circles highlight areas with p62 accumulations in NDR1/2 iKO brains. Scale bars: 20 µm. (G) Schematic diagram depicting membrane trafficking events affected by NDR1/2 in neurons. Loss of NDR1/2 kinases down-regulates endocytosis, likely via a reduction in phosphorylation of Raph1/Lpd. As a result, surface levels of ATG9A and TfR are increased. Altered endocytosis of ATG9A causes reduced axonal retrograde ATG9A transport and reduced cell body/Golgi localisation of ATG9A. ATG9A mislocalisation interferes with autophagosome formation, leading to progressive accumulation of p62-ubiquitinated proteins.
Fig 3: NDR1/2 kinases and Raph1 regulate transferrin endocytosis.(A) Western blot analyses of NDR1 and NDR2 protein levels in lysates from DIV13 rat cortex neurons infected with lentiviral vectors expressing the indicated shRNAs for ∼5 d. The black arrow shows the correct NDR2 band. The bar graphs show quantifications of the NDR1 and NDR2 bands normalised against the levels of the loading controls. The data were analysed using unpaired t tests. n = 6 samples/group from three independent experiments. (B) Western blot analyses of TfR in lysates from the cortex of 6-wk-old NDR1/2 knockout and control mice. GAPDH was used as a loading control. The bar graph shows quantifications of the TfR bands normalised against the GAPDH levels, and the data were analysed using an unpaired t test. n = 4–5 mice/group. (C) Immunofluorescence stainings of transferrin receptor (TfR) and the retromer component VPS35 in the CA1 area in brain slices from 12-wk-old mice. White arrows show co-localisation between TfR and VPS35 and the yellow arrows show TfR puncta that do not co-localise with VPS35. Scale bars: 10 μm. The scatterplot shows quantification of the overlap between TfR puncta and VPS35 puncta as assessed by measuring the adjusted Rand index. The data were analysed using a Mann–Whitney test. n = 29 measurements in controls and 39 in NDR1/2 KOs from three animals/genotype. (D) Western blot analyses of Raph1 levels in lysates from DIV13 rat cortex neurons infected with lentiviral vectors expressing scramble shRNA or Raph1 shRNA and treated with either DMSO or 100 nM Bafilomycin A1 for 4 h. Data are obtained from the same experiment as shown in Fig 4G; therefore, the same tubulin loading control is shown. The bar graph shows quantifications of the Raph1 bands normalised against the tubulin levels, and the data were analysed using ordinary one-way ANOVA with Tukey’s post hoc test. n = 6 samples/group from three independent experiments. (E) Scatterplot representing the amount of transferrin (Tf)-568 present in neurons at the end of a 20 min Tf-568 uptake step (pulse). The neurons were infected with the indicated shRNA-expressing lentiviruses and Raph1 was transfected in NDR1/2 shRNA-infected cells. The data were analysed using a Kruskal–Wallis test followed by Dunn’s test for multiple comparisons. (F) Representative images of DIV11 rat hippocampal primary neurons transfected with WT-Raph1-3×FLAG or S192A-Raph1-3×FLAG and stained with both a FLAG antibody and a Raph1 antibody. White arrows show areas with high Raph1 expression at the level of the cell membrane. Scale bars: 10 μm.Source data are available online for this figure.
Fig 4: Autophagy is impaired in NDR1/2 KO mice.(A) Immunofluorescence staining of the autophagy receptor p62 in brain slices of NDR1/2 KO and control mice at P20 and 12 wk of age. White arrows indicate areas of increased p62 signal. Scale bars: 200 µm. (B) Western blot analyses of p62 levels in lysates from the cortex of P20 and 6-wk-old NDR1/2 KO and control mice. GAPDH is used as a loading control. The bar graphs show quantifications of the p62 bands normalised against the GAPDH levels, and the data were analysed using unpaired t tests. n = 3–5 mice/group, two technical replicates. (C, E) Immunofluorescence staining of p62 (C) and LC3 (E) in the CA1 area of the hippocampus in brain slices from 12-wk-old Thy1-YFP–expressing mice. Scale bars: 50 µm. (C, E) Bar graphs show the number of p62 (C) or LC3 (E) puncta in neurons, normalised against the cell body area corresponding to the YFP signal. The data were analysed using a Mann–Whitney test for p62 and an unpaired t test for LC3. n = 38 control and 27 NDR1/2 KO neurons from two mice per genotype for p62, and n = 29 control and 24 NDR1/2 KO neurons from two mice per genotype for LC3. (D) Immunofluorescence staining of p62 and ubiquitin in the CA1 area at 12 wk. White arrows indicate co-localisation between p62 and ubiquitin puncta. Scale bars: 10 µm. The scatter plot shows quantification of p62 and ubiquitin co-localisation, expressed as a Pearson correlation coefficient. The data were analysed with a Mann–Whitney test. n = 30 measurements from three mice/genotype. (F, G) Western blot analyses of LC3 levels in lysates from DIV13 rat primary cortical neurons infected with lentiviruses expressing a scramble (SCRM) shRNA and shRNAs targeting NDR1 and NDR2 (F) or Raph1(G). The cells were treated with DMSO or 100 nM of Bafilomycin A1 (Baf.) for 4 h before lysis. Tubulin was used as a loading control. The bar graphs show quantification of the LC3II bands normalised against the tubulin levels, and the data were analysed using ordinary one-way ANOVAs with Tukey’s post hoc test. n = 6 samples/group from three independent experiments, two technical replicates. The scatter plots show quantifications of the absolute increase in LC3II between the DMSO and the bafilomycin A1 conditions. n = 6 measurements/group from three independent experiments, two technical replicates.Source data are available online for this figure.
Fig 5: NDR1/2 knockout mice have autophagy deficits.(A) Western blot analyses of NBR1 in lysates from the cortex of 6-wk-old NDR1/2 knockout and control mice. Tubulin was used as a loading control. The bar graph shows quantifications of the NBR1 bands normalised against the tubulin levels, and the data were analysed using an unpaired t test. n = 4–5 mice/group. (B, C) Immunofluorescence staining of p62 together with GFAP (B) or Iba1 (C) in the CA1 area of the hippocampus at 12 wk. White dashed lines show the outline of the GFAP-positive or Iba1-positive cells. Scale bars: 10 μm. (D) Western blot analyses of ubiquitin in lysates from the cortex of 6-wk-old NDR1/2 KO and control mice. GAPDH was used as a loading control. The bar graph shows quantifications of the ubiquitin signal within each lane, normalised against the GAPDH levels. The data were analysed using an ordinary one-way ANOVA with Tukey’s post hoc test. n = 4–5 mice/group. (E) Immunofluorescence staining of p62, LC3 and the lysosomal marker LAMP2 in the CA1 area of the hippocampus at 12 wk. White arrows show LC3 puncta, all of which co-localise with both p62 and LAMP2 in control and in NDR1/2 KO mice. The white squares highlight the zoomed areas. Yellow circles show p62 puncta in NDR1/2 KO mice that do not co-localise with either LC3 or LAMP2. Scale bars: 10 μm. (F, H) Quantification of LC3I levels normalised against tubulin levels in lysates from DIV13 rat cortex neurons, infected with a scramble shRNA and NDR1 and NDR2 shRNAs (F) or Raph1 shRNA (H). Before lysis, the neurons were treated with DMSO or with 100 nM Bafilomycin A1 (Baf.) for 4 h. The data were analysed using ordinary one-way ANOVAs with Tukey’s post hoc test. n = 6 samples from three independent experiments, two technical replicates. (G) Western blot analyses of LC3II levels in lysates from HEK293T cells transfected with a constitutively active NDR1-HA construct, a kinase-dead NDR1-HA construct or a control HA-expressing plasmid (Ctrl), and treated with either DMSO or 100 nM of Bafilomycin A1 (Baf.) for 4 h before lysis. GAPDH was used as a loading control. The graphs show quantifications of the LC3II bands normalised against the tubulin levels. The data were analysed using ordinary one-way ANOVA tests with Tukey’s post hoc test. n = 6 samples/group from three independent experiments.Source data are available online for this figure.
Supplier Page from Abcam for Anti-RAPH1 antibody