Expression, subcellular localization, and mechanistic analysis of intellectual disability syndrome protein ABBA

A missense mutation in the MTSS2 gene, which encodes the I-BAR domain protein ABBA (Mtss1l/Mtss2), has been linked to an intellectual disability syndrome. To better understand the MTSS2 mutation-related effect in the brain, we elucidated the cells expressing ABBA and the localization of ABBA in these cells to get insights into which cells and processes might be dysfunctional in mutation-carrying patients. As a novel discovery, we found that ABBA was highly expressed in GABAergic inhibitory neurons, such as parvalbumin-positive interneurons in the hippocampus. At the subcellular level, ABBA localizes to the edges of membrane protrusions in various cells in the brain, suggesting a role in cell migration and spinogenesis. Overexpression of ABBA in pyramidal excitatory and inhibitory neurons increased dendritic spine density. Through live-cell imaging, we demonstrated that ABBA facilitates spine initiation by clustering on the plasma membrane before a new filopodium appears. However, our live cell imaging data also revealed that ABBA localized not only to small focal points, typical for filopodia formation, on the plasma membrane, but also more broadly on the edge of lamellipodial structures. Compared to its close homolog MIM, ABBA appears to be a more general facilitator of protrusion formation, from dendritic filopodia to lamellipodial structures. Altogether, our findings provide insights into ABBA expression, localization, and functional mechanisms, advancing our understanding of its role in neurodevelopmental processes and disorders.

Methods
Animals

In the present study, we have used wild-type and transgenic Tg(Thy1-EGFP)MJrs/J mice (Jackson Laboratory) on a C57Bl/6J background. Tg(Thy1-EGFP)MJrs/J mice express EGFP in sparse subsets of neurons within specific populations. All experimental procedures were carried out according to Finnish laws and ethics under the EU directive 2010/63/EU (licenses: ESAVI/7404/2021, ESAVI/30031/2024, and GMO 3/S/12). Animals were kept in cages as groups of 2-4 mice in a controlled environment (temperature 21±1°C, humidity 50±10%, light period 06:00 AM to 6:00 PM) and supplied with food and water ad libitum.

For tissue collection, animals were deeply anesthetized with pentobarbital solution (200 mg/kg) containing lidocaine and transcardially perfused using cold PBS. The brains were dissected and kept in 4% PFA for 24 hours and then cryoprotected in 20% sucrose solution for at least 48 hours before cryo-sectioning. The Thy1-GFP brains were cut sagittally into 50 µm sections with a cryostat (Leica) and stored in cryoprotectant solution (30% ethylene glycol, 20% glycerol in PBS) at -20°C until further processing.

The wild-type mice's left-brain hemisphere was fixed with 4% PFA for 3 days, followed by 70% ethanol for 3 days (daily refreshing solutions), and then paraffinized during a 10-hour adult mouse brain program. Then, 5 μm sections were cut for staining. The right brain hemisphere was dissected into the cortex, hippocampus, and cerebellum, which were fast frozen in liquid nitrogen and stored for Western blot at -80°C.

Organotypic slices

RccHan: WIST (Inotiv) rat pups were used for organotypic brain slices. At postnatal days 5–10, rat pups were first anesthetized by using isoflurane and then euthanized by decapitation. Both hippocampi were dissected and sliced transversely into 400 μm cross-sections using a tissue chopper. The slices were placed onto sections of nitrocellulose membrane ~5 mm2, which were placed on a Millicell cell culture insert with a pore size of 0.4 μm and a diameter of 30 mm on the 6-well plate containing 1.2 ml of culture medium. The culture medium was composed of 95.5 ml Opti-MEM (Gibco), 50 ml of HBSS (Gibco), 50 ml of Hof Horseshoe serum (Gibco), and 2 ml of D-glucose (50%), and the pH was set to 7.2. After plating, the hippocampal slices were kept at 37°C in 5% CO2. At DIV3, anti-mitotic (0.5 mM uridine, 0.5 mM ARA-C, 0.5 mM 5-Fluoro-deoxy-uridine) was added to inhibit the glial growth. At DIV4, the media was replaced with fresh media. Later, the culture media were refreshed twice a week. At DIV10, slices were treated either with 180 ng of BDNF for 24 hours or 30 μM Bicuculine for 1 hour, diluted in culture media. After treatments, the slices were put back into refreshed culture media and incubated for 4 h, 1 day, and 2 days from the start of the treatment. Slices were then snap-frozen and stored at -80°C until protein extraction.

Characterization of ABBA antibody

In western blotting and immunostaining, we used an earlier characterized rabbit ABBA antibody. This was originally generated in the Lappalainen laboratory against the central region of the mouse ABBA protein that is not conserved in other I-BAR-domain proteins, such as the loss homologue MIM or Irsp53. This study demonstrated with Western blot analysis that the antibody detected only a single band of expected mobility (around 100 kDa) from brain lysates and in NIH3T3 cells expressing a GFP-fusion protein of ABBA. Furthermore, the antibody did not recognize other I-BAR-domain proteins, IRSp53 and MIM, on a western blot, suggesting that the antibody is specific to ABBA. The antibody was later reproduced in Rivera´s laboratory, and in their recent study, this antibody was used to detect decreased ABBA expression in shRNA ABBA-treated C6 cells, indicating that it is indeed ABBA that the antibody recognizes. Uncropped western blots shown in Figure 1 are provided as Supplementary Figure 1.  

Immunostaining/immunohistochemistry

Both paraffin sections (5 µm, Fig. 1) and cryosections (50 µm, Fig. 2) were used for stainings. Briefly, paraffinized brain sections went through deparaffinization and rehydration. After antigen retrieval (1h boiling in 0.01 M citrate buffer, pH 6.0), blocking was done by using 2% BSA, 5% milk, 5% sucrose, and 1% Triton X-100 in PBS. Cryosections were first permeabilized with 0.1% Triton X-100 in PBS and then blocked for 1.5 hours in a blocking buffer containing 0.3% Triton X-100 and 5% goat serum in PBS.

The sections were then incubated in blocking solution with primary antibodies: rabbit anti-ABBA (1:500 [5], guinea pig polyclonal antiserum against parvalbumin (1:500, synaptic system, 195004), chicken polyclonal anti-GFAP (1:4000, Abcam, Ab4674), and chicken monoclonal recombinant IgY anti-Iba1 (1:1000, synaptic system, 234009) overnight at 4°C. The next day, sections were incubated in secondary antibodies: anti-rabbit Alexa-488, anti-rabbit Alexa-647 (1:1000, ThermoFisher, A21206, A31573), anti-guinea pig Alexa-568 (1:1000, ThermoFisher, A11075), and anti-chicken Alexa-647 (1:1000, ThermoFisher, A-21449) for 2h and mounted on glass coverslips using Immu-mount (Thermo Scientific,999041). The stained fixed brain slices were scanned using a 3DHISTECH Panoramic 250 FLASH III digital slide scanner using a 20X/0.8 NA objective at the FIMM digital microscopy and molecular pathology unit. For better resolution and 3D images, the brain slices were further imaged through a Zeiss LSM880 inverted confocal microscope using 20X/0.80 NA dry or 63X 1.4 NA oil immersion objectives.

Western blotting

The hippocampus, cortex, and cerebellum of wild-type mice were homogenized and lysed in RIPA buffer (50 mM Tris-HCl, pH 7.4, 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA) supplemented with 10% protease inhibitor, 10% phosphatase inhibitor cocktail (Roche), and 1% SDS, using a bead mill homogenizer with 1.4 mm ceramic beads for 15-20 seconds. Organotypic brain slices were similarly processed with sonication. Protein concentration was measured using the BCA Protein Assay (ThermoFisher Scientific, 23227). For analysis, 20 µg of protein was separated on 10% SDS-PAGE gels and transferred to PVDF membranes. Membranes were blocked for 5 minutes in Bio-Rad EveryBlot blocking buffer (Bio-Rad, 12010020) and incubated overnight at 4°C with primary antibodies diluted in blocking buffer: Rabbit anti-ABBA (1:1000 [5], rabbit anti-srGap3 (1:1000, Novus, NBP1-88831), and mouse anti-Gas7 (1:1000, Santa Cruz, SC-365385). Membranes were washed three times with TBS-T and incubated with StarBright secondary antibodies (anti-rabbit, 1:2500, Invitrogen, 1200Y161; anti-mouse, Invitrogen, 12005866) diluted in blocking buffer and 0.02% SDS for 1 hour at room temperature. Membranes were again washed three times with TBS-T and imaged using the ChemiDocTM MP Imaging System (Bio-Rad). The protein levels were quantified using Image Lab software (Bio-Rad) and normalized against total protein per lane. Precision Plus Protein Unstained (Bio-Rad, 161- 363) and All Blue Standards (Bio-Rad, 1610373EDU) were used as molecular weight references.

Hippocampal cultures and transfections

The plasmids pmCherry-C1 and pEGFP-C1 were purchased from Clontech Laboratories, Inc. GFP-ABBA and mCherry-ABBA were gifts from Juha Saarikangas. The RFP-LifeAct construct was a gift from Roland Wedlich-Söldner (University of Münster, Münster, Germany). GFP-LifeAct was a gift from Emmanuel Lemichez (an Institute Pasteur, Université Paris Cité, France). The murine N-WASP-mCherry construct was provided by Maria Vartiainen (Institute of Biotechnology, University of Helsinki). mCherry-actin plasmid was a gift from Martin Bähler (Westfälische Wilhelms-University, Münster, Germany). The Scar W and WA constructs [13] were a gift from Laura Machesky (Beatson Institute for Cancer Research, Glasgow, United Kingdom).  mCherry-Akt-PH was provided by Vesa Olkkonen (Minerva Foundation Institute for Medical Research). GFP-Tubby was given by Lawrence Shapiro (Columbia University, USA). GFP-Rac1 N17 and GFP-Rac1 V12 were gifts from Johan Peränen (University of Helsinki).

The hippocampal culture preparation was done by the previously proposed method. The dissection and dissociation of cells were carried out by the Neuroscience Center core facility (HiLIFE, University of Helsinki). Briefly, hippocampi were dissected from embryonic day 16–17 Wistar rat fetuses. Cells were dissociated in 0.05% papain and triturated in Ca2+ and Mg2+-free HBSS medium with 1mM sodium pyruvate and 10mM HEPES (pH 7.2). The obtained cells were then plated on 13mm glass coverslips (VWR) (100000 cells) or 13 mm high-precision glass coverslips for SIM (Marienfeld) (15000 cells) coated with poly-L-Lysine (0.01 mg/ml, Sigma Aldrich) in a 24-well plate. The culturing of the cells was done in a Neurobasal medium (Invitrogen) supplemented with L-glutamine (Invitrogen), B-27 (Invitrogen), and primocin (InvivoGen) in humidified incubators at 37°C and 5% carbon dioxide (CO2) with media refreshing twice a week. The transfection was performed at DIV14 by using Lipofectamine 2000 (Invitrogen) with 0.5 μg plasmid/well in 24-well plates, as described earlier [15]. Normally, the transfected cultures were fixed after 24 hours of the transfection. For testing effects of F-actin depolymerization or PI3-kinase activity, cells were treated with 5 mM Latrunculin B (Sigma-Aldrich) or 100 mM PI3-kinase inhibitor LY294002 (Sigma-Aldrich) before fixation.

Immunostaining

Hippocampal cultures were fixed with pre-warmed (37°C) 4% paraformaldehyde (PFA) for 13 minutes at room temperature (RT), and after this, they were permeabilized using 0.2% TritonX-100 in PBS. Blocking was done for 30 minutes using 3% normal donkey serum and 0.5% bovine serum albumin (BSA) in PBS. The fixed cells were incubated with primary antibodies in a blocking solution for 1 hour at RT, followed by incubation with secondary antibodies for 1 hour. The coverslips with cells were mounted on glass slides by using Shandon Immu-Mount (Thermo Fisher Scientific,9990402) or Prolong Gold (ThermoFisher, P10144). The primary antibodies used were rabbit anti-ABBA (1:400 [5], mouse anti-myc (9E10) (1:200, ThermoFisher, MA1-980), rabbit monoclonal anti-GAD65/67 (1:400, Abcam, AB 183999), guinea pig polyclonal antiserum against parvalbumin (1:500, Synaptic Systems, 195004), and chicken polyclonal anti-GFAP (1:4000, Abcam, Ab4674). The Secondary antibodies were anti-rabbit Alexa-488, anti-rabbit Alexa-647 (1:1000, ThermoFisher, A21206, A21446), anti-guinea pig Alexa-568 (1:1000, ThermoFisher, A11075), and anti-chicken Alexa-647 (1:1000, ThermoFisher, A-21449). F-actin was visualized using Alexa 488-conjugated phalloidin, Alexa 568-conjugated phalloidin, or Alexa 633-conjugated phalloidin (1:200, ThermoFisher, A12379, A12380, and A22284).

Imaging

Confocal images were obtained using either a Zeiss LSM780 or LSM880 inverted confocal microscope. The fixed samples were imaged at room temperature, and the live cell imaging was performed in a chamber where the temperature was maintained at 37°C and the CO2 levels were 5%.  The 63X 1.4 NA oil immersion objective was used for Z-stack imaging of each neuron. The step size was set to 0.2 μm. Cells or fields of view were selected randomly for imaging. The laser power and gain settings were adjusted to maximize the signal-to-noise ratio. The Fiji software was used to process the image files obtained from the imaging.

The used imaging system was the DeltaVision OMX SR system (GE Healthcare Life Sciences) with a 60X 1.42 NA PlanApo N oil immersion objective. AcquireSR software was used for acquisition, and SoftWoRx for image reconstruction and alignment.

Dendritic spine morphology and density analysis

Neuron Studio was used for the analysis of dendritic spine morphology and density. Neuron Studio is a software package designed for three-dimensional detection of dendritic spines from fluorescence microscopy images. It also allows the classification of spines into different types based on their morphology. Images were re-numbered and ordered randomly before analysis, so that the analysis was done blinded to the experimental condition. The Fiji software was used to convert Zeiss files with the z-stacks of 20–30 optical sections to the TIFF files. The images had a voxel size of 0.066 μm x 0.066 μm x 0.2 μm, and the EGFP or the mCherry channels were used for analyzing the dendritic spine morphology and density. After modeling the dendrite surface, protrusions with a minimum volume of 5 voxels (0.020 μm3), a length between 0.1 μm and 5 μm, and a maximal width of 3 μm were retained as spines. Following the default settings of the program and the empirical classification rule defined by spines with a minimum head diameter of 0.35 μm and a minimum head vs. neck ratio of 1.1, were classified as mushroom spines. Non-mushroom spines with a minimum volume of 10 voxels (0.040 μm3) were classified as stubby spines. All other spines were considered thin. The spines detected and modeled by the software were carefully verified and corrected manually. The wrongly labeled structures were removed or corrected, and the spines that were not detected by the software were added and classified. Measurements obtained by Neuron Studio were transferred to a spreadsheet application (MS Excel) for further analysis.

Co-localization Analysis

The co-localization analysis between GFP and mCherry-tagged constructs was done on the maximum intensity projections obtained from 3D Z-stacks. After making a maximum intensity projection, the GIMP software was used to outline the region of interest of the image for co-localization analysis. GIMP is an open-source drawing and annotation software (https://www.gimp.org/). The co-localization analysis was carried out with an EzColocalization Plugin in the Fiji software by following the procedure created, used, and reported in previous research. Pearson’s correlation coefficient (PCC) value was determined for each image to measure the co-occurrence of two-channel signals. The range of PCC values is from 1 to -1, indicating a strong co-localization and a strong anti-co-localization, respectively. PCC=0 reflects no correlation between the two signals.

Analyzing the membrane vs diffuse protein ratio for GFP-Tubby

The quantification of fluorescence intensity was done using single confocal focal planes in the middle (in the z direction) of the measured dendrites. We first draw lines along the sides presenting the plasma membrane, and in the middle of a dendrite, presenting diffuse protein using Fiji software. Measurements were averaged to get one average value for the membrane and one for diffuse protein, and the ratio was calculated using these average values.

Statistical analyses

Statistics were made in GraphPad Prism (GraphPad Software Inc.). We used either the two-sample t-test or two-way ANOVA for parametric data and the Kruskal-Wallis test or the Mann-Whitney U test for non-parametric data. The tests used are specified for each analysis in the Figure legends.

Figures

Microscope images were processed in Fiji, Imaris (Oxford Instruments), and Photoshop (Adobe). Graphs were done in Prism (GraphPad Software Inc.). The final layout of figures was done in Inkscape.