Functional evaluation of autism-associated mutations in NHE9.

Structural Modeling of NHE9 and Nhx1(A) Alignment of the sequences of human NHE9, S. cerevisiae Nhx1, human NHE1 and E. coli NhaA. Transmembrane segments are underlined and numbered. The positions of four NHE9 variants are boxed. (B) Hydrophobicity analysis, using the blue to-yellow color code shown in the color bar of the NHE9 model-structure shows that the lipid facing amino acids are (overall) hydrophobic, as they should63. (C) Model structure of yeast Nhx1 showing shared protein fold common to the NHE family, and residues targeted for mutation in red (stick representation).

Autism-associated NHE9 variants fail to alkalinize endosomal pHCells were loaded with FITC- and Alexa Fluor-tagged Transferrin (Tf) for 55 minutes and internalized Tf was quantified using flow cytometry from at least 5,000 cells, in triplicate. pH was calibrated using the ratio of internalized Tf-FITC (pH-sensitive) and Tf-Alexa Fluor (pH-insensitive). Cells were exposed to nigericin (100µM) and pH defined medium (pH 5.0 to pH 8.0) for calibration of pH dependent fluorescence. (A) Expression of NHE9, but not autism-associated variants, in HEK293 cells results in alkalinization of endosomal pH. (B) NHE9 and autism-associated variants were expressed in primary mouse astrocytes, with similar results as in (A). Error bars represent SD.

Modeling of autism-associated NHE9 variants(A) Top and side views of a model-structure of the membrane domain of NHE9 based on the structure of E. coli NhaA and colored according to degree of ConSurf conservation, with turquoise through maroon indicating variable through conserved amino acid positions. Three autism-associated variants (S438P, L236S, V176I) are shown in space-filled form. (B) Site-directed mutagenesis was used to introduce equivalent NHE9 mutations into yeast Nhx1 (A438P, I222S, and V167I) as well as ‘humanized’ variants A438S and I222L to mimic wild type NHE9. (C) Nhx1 constructs tagged with GFP were expressed in the nhx1Δ null strain and visualized (100× objective) as fluorescent punctae, characteristic of pre-vacuolar compartments. Scale bar: 20 µm (D) Immunoblot, with anti-HA, was used to detect similar expression levels of HA-tagged Nhx1 and variants. GAPDH was used as loading control.

Phenotype screening of autism-associated variants in yeast(A). Growth-sensitivity to Hygromycin B. Yeast nhx1Δ strains expressing the vector or indicated Nhx1 constructs were inoculated with equal numbers of cells in APG medium (pH 4.0) supplemented with hygromycin B. Growth (OD600) was measured after 17 h at 30°C and is expressed as percentage of growth in the absence of hygromycin. (B) Growth-sensitivity to KCl. Cultures, as in (A), were grown in a medium supplemented with KCl. (C) Growth-sensitivity to acidic pH. Cultures, as in (A) were grown in APG medium buffered to pH 4.0 or 2.7 for 21 h. Results shown for (A)–(C) are averages of triplicate determinations and are representative of at least three independent experiments. (D) Measurement of Vacuolar pH with BCECF. Cells were loaded with BCECF resulting in accumulation of the dye in yeast vacuoles, as seen in the fluorescent micrograph (100x objective). (D-inset: Scale bar: 20 µm). Fluorescence was normalized to cell number (NI485) and calibrated against vacuolar pH (E). Normalized, pH-sensitive fluorescence as in (D) for the yeast strains shown. Mean was plotted from at least three independent experiments for (D) and (E). All error bars represent SD. F) Sorting of carboxypeptidase Y (CPY). Extracellular CPY in culture supernatants (600 µl) was assessed by slot-blots. Samples were applied onto fixed slots by vacuum suction and the nitrocellulose filter treated as in a Western blot.

Developmental regulation of NHE9 in mouse and expression in primary brain cells(A) Raw expression levels of NHE9 in various regions of the mouse brain determined from in situ hybridization data obtained from Allen Brain Atlas [Available from: http://mouse.brain-map.org/]. ISOCTX: Isocortex, OLF: Olfactory areas, HPF: Hippocampal formation, CTXsp: Cortical subplate, STR: Striatum, PAL: Pallidum, CB: Cerebellum, TH: Thalamus, HY: Hypothalamus, MB: Midbrain, P: Pons, MY: Medulla. (B) NHE9 gene expression in developing mouse brain characterized by in situ hybridization (ISH) in sagittal plane across four embryonic and three early postnatal ages. Feulgen-HP yellow DNA counterstain, a nuclear stain, was used to add definition to the tissue. This counterstain is used in conjunction with ISH for all data shown except for P56, in order to provide tissue context to the ISH signal which is otherwise difficult to discern due to the very light tissue background for embryonic ISH. Images were obtained from Allen Institute for Brain Science, Allen Developing Mouse Brain Atlas [Available from: http://developingmouse.brain-map.org] (C) In situ hybridization data showing expression summary of NHE9 in the various regions of the mouse brain during development, obtained from Allen Developing Mouse Brain Atlas [Available from: http://developingmouse.brain-map.org]. RSP: Rostral secondary prosencephalon, Tel: Telencephalic vesicle, PHy: Peduncular (caudal) hypothalamus, p3: Prosomere 3, p2: Prosomere 2, p1: Prosomere 1, M: Midbrain, PPH: Prepontine hindbrain, PH: Pontine hindbrain, PMH: Pontomedullary hindbrain, MH: Medullary hindbrain. Scale bar: 3168 µm (D) qPCR analysis of NHE6 and NHE9 in primary murine neurons and astrocytes with mRNA normalized to two reference genes (GAPDH and 18S RNA) and expressed relative to NHE9 mRNA level. Error bars represent SD determined from triplicate measurements.

Subcellular Localization and Functional Analysis of NHE9(A) qPCR analysis of NHE6 and NHE9 mRNA in primary cortical astrocytes, normalized to two reference genes (GAPDH and 18S RNA) and expressed relative to NHE9 mRNA level. Error bars represent standard deviation determined from triplicate measurements. Baseline expression of NHE9 is significantly lower than NHE6 (note that the 8-fold difference corresponds to 3 cycles of PCR amplification on Log2 scale). (B) qPCR analysis showing the efficacy of overexpression of (NHE9) and shRNA knock-down (NHE9 and NHE6) in primary astrocyte culture. The data are plotted as average fold-change of mRNA levels relative to control levels, with standard deviations determined from triplicate measurements. (C) Subcellular localization of NHE9 in primary cultured cortical astrocytes determined by immunofluorescence confocal microscopy (63x objective) after fixation with 4% PFA. Top, NHE9-GFP (green) partly localizes with early endosome marker, EEA1 (red) as seen in the Merge. Middle, NHE9-GFP (green) partly localizes with recycling endosome marker, Rab11 (red) as seen in the Merge. Bottom, NHE9-GFP (green) does not localize with late endosome marker, LBPA (red). (D) Orthogonal views of subcellular localization of NHE9 from merged images in (C). (E) Overlapping subcellular localization of NHE6-GFP (green) and NHE9-DsRed (red) in primary cultured cortical astrocytes, as seen in Merge and (F) Orthogonal view. Scale bars for C and E: 50 µm

NHE9 regulates endosomal pH(A) Calibration of endosomal pH from fluorescence ratio of internalized Tf-FITC and Tf-Alexafluor. Cells were loaded with tagged Transferrin (Tf) for 1 hr, then exposed to nigericin (100µM) and pH defined medium (pH 5.0 to pH 8.0). Internalized Tf was quantified using flow cytometry. (B) NHE9 expression alkalinizes endosomal lumen. pH of Tfn-positive endosomes in primary cultured cortical astrocytes was determined in control, NHE9 overexpression, and NHE9 shRNA knock-down conditions. Results are averages of three biological replicates, each done in triplicate (* p < 0.05). (C) Knockdown of NHE9 acidifies Tfn-positive endosomes in primary cultured human glioma cells. Results are averages of three replicates (* p < 0.05). Statistical analysis was by Student’s t test; all error bars represent SD.

Expression and localization of NHE9 variants in primary astrocytes(A) Expression levels of NHE9 and autism-associated polymorphisms are similar in primary astrocytes. Immunoblot of total primary astrocyte cell lysate (100 µg) from Control (empty vector transfection) and cells expressing NHE9-GFP, L236S-GFP, S438P-GFP, and V176IGFP using anti-GFP antibody. (B) Localization of NHE9 and autism-associated variants to Transferrin-positive endosomes in primary astrocytes. Confocal fluorescence images (63× objective) of GFP tagged NHE9 and indicated patient polymorphisms (green) localize with Alexa Fluor tagged Transferrin after 55 minutes of uptake (red), as described in the Methods. Significant colocalization can be seen in merged images by the presence of yellow puncta. Scale bar: 50 µm.

Functional differences between NHE9 and variants revealed by Transferrin uptake(A) Maximum projection confocal images (63× objective) showing steady state Tfn-Alexafluor uptake in control (left) and NHE9-GFP expressing astrocytes (right). Scale bars: 50 µm (B) Immunobots, using anti-TfR antibody, showing the effects of 100µM cycloheximide (CHX) on TfR in control and NHE9-GFP expressing cells. TfR bands were normalized to Tubulin levels and expressed as percentage of controls lacking CHX. (C) Steady state Tfn-Alexafluor uptake was significantly elevated (** p < 0.005; Student's t-test, n= three biological replicates) upon NHE9-GFP expression and reversed upon subsequent knockdown in the same cells. (D) Steady state uptake of Tfn-Alexa Fluor in astrocytes expressing wild type NHE9-GFP or three autism-association variants. Variants (L236S, S438P, V176I) failed to elevate Tfn-Alexa Fluor uptake, showing a loss of function phenotype (n= three biological replicates). Error bars (C–D) represent SD.

Functional differences between NHE9 and variants revealed by glutamate uptake(A) Subcellular localization of GLAST and NHE9 in primary cultured mouse cortical astrocytes determined by immunofluorescence confocal microscopy (63× objective). Scale bars: 50 µm. GLAST (red) is distributed to vesicular compartments in untransfected astrocytes labeled with DAPI (blue), as seen in the Merge. (B) NHE9-GFP (green) partly localizes with GLAST (red) in transfected astrocytes, as seen in the Merge. (C) & (D) Orthogonal views of subcellular colocalization of NHE9 with GLAST from merged image in (B). (E) Glutamate uptake is elevated over control (empty vector) in astrocytes expressing NHE9-GFP but not the autism-associated variants. (F) Immunoblot (top) showing no significant change in total GLAST levels from astrocytes, after normalization to tubulin levels (graph), whereas (G) surface levels of GLAST, determined by biotinylation are elevated in cells expressing NHE9-GFP, but not autism-associated variants. Graphs represent average band intensity from densitometric scans of immunoblots from three biological replicates. GLAST levels were normalized to tubulin and shown relative to vector-transformed control. Statistical analysis was done using Student’s t-test (*p < 0.05). Error bars represent the average of three independent experiments with SD.