SFPQ and Tau: critical factors contributing to rapid progression of Alzheimer’s disease

Dysfunctional RNA-binding proteins (RBPs) have been implicated in several neurodegenerative disorders. Recently, this paradigm of RBPs has been extended to pathophysiology of Alzheimer’s disease (AD). Here, we identified disease subtype specific variations in the RNA-binding proteome (RBPome) of sporadic AD (spAD), rapidly progressive AD (rpAD), and sporadic Creutzfeldt Jakob disease (sCJD), as well as control cases using RNA pull-down assay in combination with proteomics. We show that one of these identified proteins, splicing factor proline and glutamine rich (SFPQ), is downregulated in the post-mortem brains of rapidly progressive AD patients, sCJD patients and 3xTg mice brain at terminal stage of the disease. In contrast, the expression of SFPQ was elevated at early stage of the disease in the 3xTg mice, and in vitro after oxidative stress stimuli. Strikingly, in rpAD patients’ brains SFPQ showed a significant dislocation from the nucleus and cytoplasmic colocalization with TIA-1. Furthermore, in rpAD brain lesions, SFPQ and p-tau showed extranuclear colocalization. Of note, association between SFPQ and tau-oligomers in rpAD brains suggests a possible role of SFPQ in oligomerization and subsequent misfolding of tau protein. In line with the findings from the human brain, our in vitro study showed that SFPQ is recruited into TIA-1-positive stress granules (SGs) after oxidative stress induction, and colocalizes with tau/p-tau in these granules, providing a possible mechanism of SFPQ dislocation through pathological SGs. Furthermore, the expression of human tau in vitro induced significant downregulation of SFPQ, suggesting a causal role of tau in the downregulation of SFPQ. The findings from the current study indicate that the dysregulation and dislocation of SFPQ, the subsequent DNA-related anomalies and aberrant dynamics of SGs in association with pathological tau represents a critical pathway which contributes to rapid progression of AD. Electronic supplementary material The online version of this article (10.1007/s00401-020-02178-y) contains supplementary material, which is available to authorized users.


Immunocytochemistry
The cells were grown in T75 flasks with the culture medium (DMEM, 10% FBS, 1% PS). At confluency (70-90%), cells were trypsinized and seeded (5 x 10 4 ) on glass coverslips (13 mm) in 24 well plates. After undergoing the stress treatment as described above, the cells were fixed with 4% PFA for 20 min at RT, followed by 3x washes with ice-cold PBS for 5 min each. Permeabilization was achieved using 0.2% Triton X-100 in PBS for 10 min, followed by 3x washes with PBS for 5 min each. To avoid non-specific binding, cells were incubated with blocking buffer (1% BSA, 10% FBS in PBS) for 30 min at RT. Primary antibodies (in case of double labelling, both antibodies) were diluted in 1% BSA in PBS, followed by overnight incubation at 4°C. The cells were washed 3x with PBS (each wash was for 5 min) followed by incubation with secondary antibodies (AlexaFlour 488 and AlexaFlour 546) diluted in 1% BSA in PBS for two hrs at RT in the dark. The cells were washed 3x with PBS for 5 min each in the dark to remove the nonspecific binding. The cells were then counterstained with DAPI (one min) or a RedDot 2 Far red nuclear stain (20 min). After nuclear staining, the cells were washed 3x with PBS, followed by mounting with one drop of mounting medium (immuo-mount, Thermo Fisher Scientific). The slides were stored in the dark at -20°C or +4°C. Detail procedure for imaging the slides has been provided in the Coimmunoprecipitation protocol described in the main Manuscript. The average number of stress granules in each cell was calculated using FIJI software.
Lysates were sonicated using an ultrasonicator on ice, followed by incubation for one hr at 4°C with shaking. The lysates were centrifuged at 14000 rpm for 30 min at 4°C. The supernatants were transferred to new tubes. To harvest cells after transient transfections, cells were trypsinized and washed with 1x PBS followed by centrifugation at 400×g for 5 min at 4°C. The washed cell pellets were lysed, and protein quantification was performed with Bradford assay.

MTS assay
Cells were grown in T75 flasks with the culture medium (DMEM, 10% FBS, 1% PS). At confluency (70-90%), cells were trypsinized and then washed with 1x PBS, and seeded (10 x 10 3 ) in 96-well plates and incubated for 18-24 hours at 37°C. Transfections were performed with WT-tau and P301L-tau plasmids for variable periods of time (24 and 48 hours). To measure cell viability, MTS assay (ab197010, Abcam) was used according to the manufacturer's instructions. The culture media was replaced with a fresh medium before treatment with MTS [3- (4, 5-dimethylthiazol-2-yl)-5-(3-carboymethoxyphenyl)-2-(4-sulfophenyl)-2Htetrazolium, inner salt]. To estimate the effect of tau-expression on cell viability, reduced MTS tetrazolium complex (colored formazan product) was measured. This conversion is a property of metabolically active cells. For color development, cells were incubated at 37 o C for one hour and the absorbance measurement was taken at 490 nm using a Perkin Elmer Wallac 1420 Victor microplate reader (GMI, USA). Absorbance from the control wells (background) was subtracted from the experimental sample wells.

Trypan blue exclusion assay
Cell viability was also assessed through the trypan blue exclusion dye test. Briefly, the cell suspension in the culture medium (25 µL) after trypsinization was mixed with 25 µL of 0.4% trypan blue. After mixing, the 10 uL of cell suspension was loaded onto hemocytometer. Both viable (colorless) and dead (blue) cells were counted in each large square of the haemocytometer under (40x) objective for both untreated (control) and sodium arsenite treated (stress) cells.

Library preparation
Analytical-grade reagents were used for protein extraction and digestion. Ampuwa water (Ampuwa sterile water, Handels GmbH, Germany) was used to prepare all buffers and solutions. To prepare spectral peptide library, digested protein extracts (normalized for protein amounts) from each sample were pooled to a total amount of 220 µg and separated into fourteen staggered pooled fractions, using an Aekta pure (GE Healthcare) with a Hypersil Gold C18 column (Diam. 150x2,1 mm, Particle size: 3 µ). Digested proteins were analyzed on an Eksigent nanoLC425 nanoflow chromatography system associated with TripleTOF 5600 + , a hybrid triple quadrupole TOF mass spectrometer with a Nanospray III ionization source (Ionspray voltage 2400V, Interface heater temperature 150°C, Sheath gas setting 12). The peptides were dissolved in loading buffer (0.1% formic acid, 2% acetonitrile in optima water (Thermo Scientific) to a final concentration of 0.3 µg/µL and spiked with a synthetic peptide standard used for retention time alignment (iRT Standard, Schlieren, Schweiz). For every measurement, digested proteins (1.5 µg) were enriched on a precolumn (PharmFluidics µPAC Trapping Column) and separated on a PharmaFluidics µPAC micro Chip based seperations analytical column with 50 cm length) using a 120 min linear gradient of 5%-45% ACN, 0.1% FA) at a flow rate of 300 nL/min. Qualitative LC-MS/MS measurement was carried out using a Top 20 data-dependent acquisition (DDA) mode with a mass range of m/z 350-1250 for 250 milliseconds (ms). The resolution for data extraction was 30,000 full width at half maximum (FWHM). MS/MS scans of mass range m/z 180-1600 at a resolution of 17,500 FWHM for 85 ms and a precursor isolation width of 0.7 FWHM with a cycle time of 2.9 sec was used. For MS/MS, the criteria for precursor ions selection was: a threshold intensity of more than 125 cps with charge states of 2+, 3+, and 4+, and the dynamic exclusion of 30 sec. In every reversed-phase fraction, two technical replicates were analyzed for the construction of the spectral library.

Quantitative SWATH measurement
For the SWATH measurement, MS/MS data was obtained using windows of 65 variable sizes across a mass range of 400-900 m/z. For each biological sample, two technical replicates were performed. Protein identifications were obtained using ProteinPilot Software version 5.0 build 4769 (AB Sciex) at "thorough" settings. MS/MS spectra were searched in the UniProtKB using Homo sapiens as reference proteome (revision 04/2018, 93661 entries) at a false discovery rate (FDR) of 1%. Spectral library and SWATH peak extraction were obtained in PeakView Software version 2.1 build 11041 (AB Sciex). Following retention time correction using the iRT standard, peak areas were extracted using information from the MS/MS library at 1% FDR. The peak areas were summed up to peptide and finally corresponding protein area values that were used for downstream statistical and functional analysis.

Differential enrichment analysis of RBPome
For a detailed analysis of the isolated RNA-binding proteomic candidates, two different approaches were used, here labeled approach A and B.
Approach A: to find out differentially enriched proteins in disease groups, pairwise t-tests (two-sided) were performed for all disease group combinations using Perseus software (version 1.5.0.31). Missing or Zero values are observed in the data for several proteins in MS-based proteomics data. These values appear when the mass spectrometer cannot detect peptides having abundances below the censoring cutoff of the mass spectrometer. Such values are informative because they are below the lowest abundance observed for a peptide. In such situations, when quantitative values are missing in one group but are present in other groups, this more likely represents differences in abundance between groups, which might indicate interesting features specific to that group.
Therefore, zero values were imputed by half of the minimum value of total spectrum count values, to have statistical analysis of the proteomic candidates (Approach A). Fold change (FC) for all comparisons was set at ±1.5 and a p-value < 0.05 for significance. Proteins which were identified as significantly enriched were used to make heatmaps using Perseus software (version 1.5.0.31). Volcano plots were also calculated using Perseus software, where FC was log2-transformed, so that the data are centered on zero, while the p-values were transformed into −log10.
Approach B: given the exploratory nature of this discovery-based proteomics work-flow, the criteria were relaxed and proteins with a single quantitative value were included in the analysis to compare all the disease groups. To find similar and unique proteins, venn diagram was prepared using functional enrichment and analysis tool (FunRich).

Gene Ontology analysis and functional network mapping
To gain functional insights from proteomics data, three different enrichment strategies were used. Functional enrichment analysis was initially performed using Perseus software for significantly-enriched Gene Ontology (GO) processes including the Biological Process and Molecular Function by Fisher's exact-test. Then, overrepresentation enrichment analysis was performed using the web-based Gene SeT Analysis Toolkit (WebGestalt), in the domains of Biological and Molecular Functions to have a GO Slim summary of enriched terms. Finally, an Ingenuity Pathway Analysis (IPA, Qiagen, USA) was performed to find out canonical pathways associated with up-and down-regulated proteins after tau-expression as described below.

Ingenuity Pathway Analysis (IPA)
The proteomic dataset containing the fold change and p-values of significantly regulated proteins was uploaded to IPA for core analysis (Qiagen, USA). The protein candidates from the submitted dataset generated top molecular networks based on Molecular and Biological Functions including canonical pathways, potential upstream regulators, and disease-based networks. The settings for analysis were based on direct and indirect relationships between differentially expressed proteins (DEPs), and were supported by experimentally reported data from human, mouse and rat studies [27]. Potential upstream regulators were designated as inhibited or activated according to the fold change and p-values (-log10-p-values) [65] of the DEPs.

Fig. 7f uncropped blots:
For TIA-1, two banding patterns have been reported, in 12 % gels, a double band pattern and in gradient gels (4-12 %), a single band pattern. We have run TIA-1 in both gradient (Gel 1, representative blot in the Manuscript) and in 12% gel (Gel 2) for confirmation. Abcam images for the antibody are also displayed on the right side. Gel 3: Tau-5 antibody, Tau-5 detects total tau both phosphorylated and unphosphorylated form of tau protein. Different isoforms of tau are reported between 75-50 kDa and tau cleavage products below 50 kDa. Star * represents high molecular weight phosphorylated oligomeric species of tau as Tau-5 also detects phosphorylated tau.   ble band pattern was observed. For SFPQ, Gel_1: As the expression of SFPQ was very low in particularly MM1 and VV2, and bands were visible only in controls (refer to arrows in Gel-1). Another gel is displayed (Gel-2), where concentration of secondary antibody was increased to enhance detection of SFPQ, which has resulted in high background bands. A third gel (Gel-3) is displayed as well for reference, where upper part of the membrane was probed with SFPQ to increase specificity, showing a good expression of SFPQ.