ABSTRACT
Aberrant cytoplasmic aggregation of FUS, which is caused by mutations primarily in the C-terminal nuclear localisation signal, is associated with 3% of cases of familial amyotrophic lateral sclerosis (ALS). FUS aggregates are also pathognomonic for 10% of all frontotemporal lobar degeneration (FTLD) cases; however, these cases are not associated with mutations in the gene encoding FUS. This suggests that there are differences in the mechanisms that drive inclusion formation of FUS in ALS and FTLD. Here, we show that the C-terminal tyrosine residue at position 526 of FUS is crucial for normal nuclear import. This tyrosine is subjected to phosphorylation, which reduces interaction with transportin 1 and might consequentially affect the transport of FUS into the nucleus. Furthermore, we show that this phosphorylation can occur through the activity of the Src family of kinases. Our study implicates phosphorylation as an additional mechanism by which nuclear transport of FUS might be regulated and potentially perturbed in ALS and FTLD.
INTRODUCTION
Amyotrophic lateral sclerosis (ALS) is a late-onset motor neuron disease, where the loss of upper and lower motor neurons leads to progressive muscle spasticity, wasting, weakness and death 3–5 years from the onset of symptoms (Boillee et al., 2006; Mitchell and Borasio, 2007). Frontotemporal lobar degeneration (FTLD) is a form of dementia that is caused by progressive neuronal loss in the frontal and temporal cortices, and is characterised by personality and behavioural changes and impairment of language skills (Van Langenhove et al., 2012). Mutations in the gene encoding fused in sarcoma (FUS) have been identified in ∼3% of familial ALS (ALS-FUS) cases (Kwiatkowski et al., 2009; Vance et al., 2009). FUS is predominantly a nuclear DNA- and RNA-binding protein with a non-classic PY-type nuclear localisation signal (NLS) at its extreme C-terminus (Fig. 1A). This NLS mediates an interaction with a nuclear import factor transportin 1 (TNPO1), enabling the transport of FUS through the nuclear membrane (Guttinger et al., 2004; Lee et al., 2006; Dormann et al., 2010). The majority of ALS-associated FUS mutations fall within the NLS, which impairs interaction with TNPO1 and nuclear import (Bosco et al., 2010; Dormann et al., 2010; Gal et al., 2011; Niu et al., 2012; Vance et al., 2013). Consequently ALS cases with FUS mutations have FUS immunoreactive cytoplasmic inclusions in neuronal and glial cells (Kwiatkowski et al., 2009; Vance et al., 2009; Blair et al., 2010; Rademakers et al., 2010; Mackenzie et al., 2011).
Cytoplasmic FUS inclusions are also present in 10–15% of FTLD cases that have been previously classified as atypical FTLD with ubiquitylated inclusions (aFTLD-U), basophilic inclusion body disease (BIBID) and neuronal inclusion filament disease (NIFID) (Munoz et al., 2009; Neumann et al., 2009a,,b; Seelaar et al., 2010; Urwin et al., 2010; Snowden et al., 2011), which are grouped together under the designation of FTLD-FUS (Mackenzie et al., 2010). In contrast to ALS-FUS, FTLD-FUS individuals do not have mutations in the gene encoding FUS (Neumann et al., 2009a,,b; Urwin et al., 2010; Lashley et al., 2011; Snowden et al., 2011). Another difference is that cytoplasmic inclusions in FTLD-FUS cases also include the FUS homologues EWS RNA-binding protein 1 (EWSR1) and TATA box binding protein associated factor (TAF15), and nuclear import factor TNPO1, which are absent in ALS-FUS inclusions, implying that there are key differences in the mechanisms underlying inclusion formation in the two disease groups (Brelstaff et al., 2011; Neumann et al., 2011,, 2012; Troakes et al., 2013). FUS, TAF15 and EWSR1 form the FET-family of proteins, which share a homologous C-terminal NLS, characterised by an overall basic charge, a central hydrophobic or basic motif, followed by a C-terminal R/H/Kx(2–5)PY consensus sequence [where x(2–5) is any sequence of 2–5 residues] (Lee et al., 2006; Zakaryan and Gehring, 2006; Marko et al., 2012). Although ALS-FUS cases appear to be caused by mislocalisation and aggregation of FUS alone, all of the FET proteins containing the consensus NLS mislocalise and accumulate in FTLD-FUS (Neumann et al., 2011).
Post-translational modifications of NLS, including phosphorylation, methylation and ubiquitylation, can play essential roles in controlling nuclear import and in regulating the localisation of cargo protein (Terry et al., 2007). It has been previously reported that FUS undergoes arginine methylation in arginine-glycine-glycine (RGG) domains near to the NLS that impairs TNPO1 binding to RGG domains, inhibiting nuclear import (Rappsilber et al., 2003; Hung et al., 2009; Dormann et al., 2012; Tradewell et al., 2012; Yamaguchi and Kitajo, 2012; Scaramuzzino et al., 2013). Finally, phosphorylation of the C-terminal Y656 residue within the NLS of the EWSR1 protein has been shown to inhibit nuclear import (Leemann-Zakaryan et al., 2011). As FUS and EWSR1 have highly homologous NLS domains, we sought to explore the functional effects of post-translational modifications of the NLS of FUS. Here, we report that phosphorylation of the C-terminal Y526 residue abolishes the interaction with TNPO1 and inhibits the transport of FUS into the nucleus. Our study implicates a new post-translational modification as one of the mechanisms by which nuclear transport of FUS might be regulated and potentially perturbed in ALS and FTLD.
RESULTS
The C-terminal tyrosine in FUS is crucial for nuclear import and interaction with transportin 1
It has been shown that residue Y526 in FUS is one of the most important residues for a strong interaction with TNPO1 (Niu et al., 2012; Zhang and Chook, 2012). In order to reassess the role of the C-terminal Y526 residue in the nuclear import of FUS, we generated proteins that had been tagged at the N-terminus with green fluorescent protein (GFP), these included full-length wild-type FUS (FUS-WT), FUS with the deleted C-terminal Y256 residue (ΔY) and C-terminal tyrosine mutant Y526A. As tyrosine residues can undergo post-translational modifications, and phenylalanine and glutamate residues might serve as mimetics of unphosphorylated and phosphorylated tyrosine, respectively, we also generated Y526F and Y526E constructs (Derkinderen et al., 2005; Uezu et al., 2012). In transiently transfected HeLa cells (Fig. 1B) and primary rat cortical neurons (Fig. 1C), GFP–FUS-WT was predominantly nuclear. Mutants ΔY, Y526A and Y526E were excluded from the nucleus, and they formed cytoplasmic aggregates. Mutant Y526F showed nuclear and cytoplasmic localisation, with cytoplasmic staining in more than 50% of transfected cells (Fig. 1B–D). Same mislocalisation of FUS was also observed in SH-SY5Y cells (data not shown).
Because the nuclear import of FUS has been shown to be mediated by the nuclear import factor TNPO1 (Dormann et al., 2010), we sought to determine whether C-terminal tyrosine mutations disrupt interaction of FUS with TNPO1. In order to remove the effects of other FUS tyrosine residues, we constructed N-terminally His6–GFP tagged C-terminal fragments (amino acids 490–526) of wild-type FUS, which has only one FUS tyrosine residue, and mutants ΔY, Y526A, Y526F and Y526E, which lack the C-terminal tyrosine residue (Fig. 2C). We then transiently transfected HeLa cells with these constructs and co-immunoprecipitation was performed with added recombinant glutathione S-transferase-tagged TNPO1 (GST–TNPO1). Recombinant GST–TNPO1 was used because interaction of His6–GFP–FUS-tagged C-terminal fragments (amino acids 490–526) with endogenous TNPO1 from HeLa cells was not detectable by immunoblotting. Consistent with the immunocytochemistry results (Fig. 1B,C), mutants ΔY, Y526A and Y526E bound less GST–TNPO1 compared to the wild-type fragment. The Y526F mutant did not significantly affect binding of GST–TNPO1, which explains the partial nuclear localisation of GFP-tagged FUS-Y526F (Fig. 1E,F).
Residue Y526 in FUS is subjected to phosphorylation
As the substitution of C-terminal Y526 by phenylalanine or glutamate affects the nuclear localisation of FUS (Fig. 1B–D), we sought to determine whether the nuclear import of FUS could be regulated through phosphorylation of Y526. Initially, we determined the phosphorylation state of tyrosine residues in endogenous FUS immunoprecipitated from HeLa cell and rat brain cortex lysates. Western blots probed using antibodies against phospho-tyrosine residues revealed that FUS contains one or more phosphorylated tyrosine residues (Fig. 2A,B). Full-length FUS contains 36 tyrosine residues, any of which could have contributed to the observed signal. In order to determine the phosphorylation state of Y526, we then compared the phospho-tyrosine immunoreactivity of the His6–GFP–FUS-tagged constructs encoding the WT and Y526F C-terminal fragments (amino acids 490–526) (Fig. 2C). After transient transfection, HeLa cells were lysed, and the fragments were isolated using His-tag pull down, and phosphorylation was detected with antibodies against phosphorylated tyrosine residues. The phosphorylation of the wild-type fragment was significantly higher than that of the Y526F mutant, suggesting that Y526 is subjected to phosphorylation (Fig. 2D,E). The weak signal for phosphorylated tyrosine residues in the Y526F construct might be attributable to the possible phosphorylation of the 11 tyrosine residues in the sequence of GFP.
Phosphorylation of Y526 in FUS impairs the interaction with TNPO1
Based on our co-immunoprecipitation results (Fig. 1E,F), where less GST-TNPO1 bound to the mutant fragment Y526E, we speculated that phosphorylation of Y526 in the C-terminus of FUS might impair its interaction with TNPO1. To test this hypothesis, we performed an in vitro pulldown assay using recombinant GST–TNPO1 and two biotinylated synthetic peptides of the C-terminal fragment of FUS – unphosphorylated peptide FUS504-526 and Y526-phosphorylated peptide pFUS504-526 (Fig. 3A). Peptides were immobilised on streptavidin beads and incubated with varying amounts of GST–TNPO1. Consistent with our hypothesis, the phosphorylation of C-terminal residue Y526 completely abolished the binding of TNPO1, whereas the unphosphorylated peptide FUS504-526 efficiently pulled down GST–TNPO1 (Fig. 3B,C). To confirm our findings, we also performed reverse in vitro pull down, where GST–TNPO1 was immobilised on glutathione beads and incubated with varying amounts of synthetic peptide. GST–TNPO1 only pulled down unphosphorylated peptide (Fig. 3D,E), confirming our hypothesis that phosphorylation of Y526 impairs the interaction with TNPO1.
PY-type nuclear localisation signals are overall positively charged, and they possess a hydrophobic or basic motif followed by a C-terminal R/H/Kx(2–5)PY consensus sequence (Lee et al., 2006). The crystal structure of the NLS of FUS (FUS-NLS) bound to TNPO1 has been solved. Residues P525 and Y526 of FUS dock into the hydrophobic pocket in TNPO1 through hydrophobic contacts. In addition, a hydrogen bond is formed between the hydroxyl group of Y526 in FUS and the carboxylic group of residue D384 in TNPO1 (Niu et al., 2012; Zhang and Chook, 2012). Based on this data, we modelled the interaction between FUS-NLS phosphorylated at Y526 and TNPO1 using the software programs Coot (Emsley et al., 2010) and Pymol (DeLano Scientific LLC, San Carlos, CA; http://www.pymol.org) (Fig. 3F,G). Phosphorylation is predicted to introduce a negative charge, leading to electrostatic repulsion between the phosphate group of phosphorylated Y526 in FUS and carboxyl group of D384 in TNPO1. Phosphorylation also enlarges the side chain of Y526, causing steric hindrance between the NLS of FUS and the binding site on TNPO1. Our model is consistent with the results of our pulldown experiments, where phosphorylation of Y526 in FUS abolished binding to TNPO1.
Y526 in FUS is phosphorylated by Src-family kinases
To elucidate which protein tyrosine kinases might be involved in the phosphorylation of Y526, we used the online program NetPhosK 1.0, which predicted the Src family of kinases to be the most likely candidates with a threshold value 0.47 (Blom et al., 2004). To validate this prediction, we transiently transfected HeLa cells with the His6–GFP–FUS WT and Y526F C-terminal fragments (amino acids 490–526) and then treated them with 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2), an inhibitor of the Src-family kinases, followed by treatment with pervanadate, a cell-permeable irreversible inhibitor of protein tyrosine phosphatases, which is used to enhance phosphorylation (Huyer et al., 1997). After cell lysis, the His6–GFP–FUS WT and Y526F fragments were isolated with His-tag pull down. Western blot analysis with antibodies against phosphorylated tyrosine residues revealed that pervanadate treatment increased phosphorylation of both the WT and Y526F fragments, whereas the treatment of cells with PP2 abolished pervanadate-induced phosphorylation of both constructs (Fig. 4A). In order to determine the signal of Y526 phosphorylation at the C-terminus, we subtracted the signal of the His6–GFP-tagged FUS Y526F C-terminal fragment from that of the His6–GFP-tagged FUS WT C-terminal fragment, which revealed a clear reduction in the level of phosphorylation of Y526 after PP2 treatment (Fig. 4B). These results suggest that the Src family, or closely related protein kinases, are able to phosphorylate residue Y526 in FUS.
To determine more specifically which protein tyrosine kinases are responsible for the phosphorylation of Y526, co-transfection experiments were performed in HeLa cells using Fyn and cellular (c-)Src tyrosine kinase expression vectors along with the His6–GFP–FUS WT and Y526F C-terminal fragments. Activity of the transfected kinases was confirmed by western blotting of whole cell lysates with antibodies against phosphorylated tyrosine residues (Fig. 4C). Both the Fyn and c-Src tyrosine kinases increased the phosphorylation of the WT and Y526F constructs. The phospho-tyrosine immunoreactivity of the wild-type fragment was significantly higher after co-transfection with c-Src and Fyn compared to that of the mutant Y526F fragment (Fig. 4D,E). This implicates Fyn and c-Src as candidate protein tyrosine kinases that might phosphorylate Y526 in FUS and inhibit nuclear import.
DISCUSSION
The majority of ALS-linked mutations of FUS are single amino acid substitutions in the C-terminal NLS (Kwiatkowski et al., 2009; Vance et al., 2009). The NLS of FUS has been shown to bind to transportin (Guttinger et al., 2004; Lee et al., 2006), and knockdown of both transportin homologues, TNPO1 and transportin 2 (TNPO2), results in cytoplasmic localisation of FUS (Dormann et al., 2010). The P525L mutation causes one of the most severe forms of juvenile-onset ALS (Baumer et al., 2010; Huang et al., 2010; Mackenzie et al., 2011; Conte et al., 2012). The proline residue introduces a kink into the main chain, enabling hydrophobic interactions with the hydrophobic pocket of TNPO1, which engulfs the proline side chain and phenyl ring of Y526, enabling the formation of a hydrogen bond to occur between Y526 of FUS and D384 of TNPO1. Thus, the tyrosine residue at position 526 might play a central role in mediating the interaction between FUS and TNPO1. To date there are no reported mutations of Y526 in ALS individuals. The severe cytoplasmic redistribution and inclusion formation of our Y526 mutants that we observed in transfected cells suggests that mutations of this residue might be lethal. FUS mislocalisation and inclusion formation has already been observed in several studies of disease-associated FUS-NLS mutants (Bosco et al., 2010; Dormann et al., 2010; Gal et al., 2011; Vance et al., 2013).
Although all ALS-FUS individuals have mutations in FUS, it is curious that no mutations of FUS are reported in FTLD-FUS cases (Neumann et al., 2009a,,b; Urwin et al., 2010; Lashley et al., 2011; Snowden et al., 2011). There seems to be a more general dysfunction of TNPO1-mediated import of the three FET proteins because FUS, EWSR1 and TAF15 colocalise with TNPO1 in the cytoplasmic inclusions. This dysfunction is restricted to interaction between TNPO1 and FET proteins, as other PY-NLS containing RNA-binding proteins (hnRNP A1, SAM68, PABPN1, etc.) are not redistributed into cytoplasmic inclusions (Neumann et al., 2012). Post-translational modification can change the character of amino acid side chains, and in that way mimic amino acid substitutions. In contrast to the majority of other PY-NLS-containing proteins, FET proteins carry an NLS at their extreme C-terminal end. There is a possibility that the C-terminal position of the NLS enables interaction with enzymes that catalyse the post-translational modification of specific amino acids.
It has also been reported that the RGG3 domain of FUS (amino acids 473–503) stabilises the interaction between FUS-NLS and TNPO1 (Dormann et al., 2012). Methylation of RGG repeats interferes with TNPO1 binding, and the inhibition of methylation restores the nuclear localisation of ALS-associated mutants. In ALS-FUS, pathological inclusions contain methylated FUS, whereas in FTLD-FUS cases, FUS inclusions are hypomethylated. Such hypomethylation of all three FET proteins could lead to the enhanced binding of FET proteins to TNPO1 and might reduce their dissociation from TNPO1, and explain the presence of TNPO1 in cytoplasmic inclusions in FTLD-FUS (Dormann et al., 2012).
Our results show that endogenous full-length FUS undergoes phosphorylation on tyrosine residues. The use of the C-terminal fragment of FUS as a probe enabled us to show that phosphorylation of Y526 changes the character of this otherwise neutral tyrosine residue and profoundly affects nuclear import. The predicted electrostatic repulsion and steric hindrance from phosphorylated Y526 impaired the binding of FUS-NLS to TNPO1 in our in vitro pulldown assay. The physiological relevance of Y526 phosphorylation still needs to be elucidated, but it is possible that the phosphorylation of Y526 serves to enhance the dissociation of FUS from TNPO1 once it has been transported into the nucleus. By contrast, phosphorylated Y526 could also prevent a portion of the FUS protein from entering the nucleus in order to fulfil its cytoplasmic functions, where it plays a role in the transport of mRNA to dendritic spines (Belly et al., 2005; Aoki et al., 2012).
Increased or decreased phosphorylation of Y526 could disrupt the fine tuning of nuclear–cytoplasmic shuttling of FUS and potentially lead to cytoplasmic mislocalisation and aggregation. Because phosphorylation of Y526 prevents FUS binding to TNPO1, hyperphosphorylation might lead to its accumulation in the cytoplasm. In the case of FTLD, hyperphosphorylation does not explain the presence of TNPO1 in cytoplasmic inclusions. Hypophosphorylation, however, could prevent the dissociation of the FUS–TNPO1 complex in the nucleus and lead to its re-export into the cytoplasm, explaining the accumulation of TNPO1 in aggregates of FTLD-FUS individuals.
Our results implicate Src-family kinases as kinases that are involved in the phosphorylation of Y526. Treatment of cells with the Src-family tyrosine kinase inhibitor PP2 decreased pervanadate-induced phosphorylation. The Src proteins are a family of non-receptor tyrosine kinases that includes nine members (c-Src, Yes, Fyn, Fgr, Lck, Hck, Blk, Lyn and Frk). Overexpression of two Src-family members (c-Src and Fyn) showed that they can phosphorylate residue Y526 of FUS in cells. Src-family kinases have previously been shown to have an important role in neurodegeneration, as Fyn kinase is responsible for phosphorylated Tau that is present upon post-mortem analysis of brains from Alzheimer's disease individuals (Lee et al., 2004; Scales et al., 2011; Usardi et al., 2011). In the future, the activity and expression levels of Src-family members in ALS and FTLD should be explored.
In conclusion, our results reveal that the phosphorylation state of the tyrosine at position 526 within the NLS strongly influences the interaction of FUS with TNPO1. However, the physiological and pathological roles of this phosphorylation event need to be further elucidated.
MATERIALS AND METHODS
Plasmid DNA
Full-length wild-type FUS and mutants ΔY, Y526A, Y526F and Y526E were cloned into the pEGFP-C1 vector (Clontech) using KpnI and BamHI restriction sites to generate N-terminally tagged fusion proteins. The pcDNA3-His6-GFP vector was generated by inserting the GFP sequence with a Kozak sequence and the N-terminal His6 tag into the HindIII and KpnI sites of the pcDNA3 vector (Invitrogen). C-terminal fragments of FUS (FUS490-526 WT, FUS490-526 ΔY, FUS490-526 Y526A, FUS490-526 Y526F and FUS490-526 Y526E) were cloned using KpnI and BamHI restriction sites into the pcDNA3-His6-GFP vector to generate His6-GFP-FUS490-526 constructs. The sequence integrity of all constructs was verified by sequencing (GATC Biotech). pDEST-GST vector containing TNPO1 has been described previously (Nishimura et al., 2010). pcDNA3 plasmid with c-Src kinase was obtained from Addgene (catalogue number 42202) (Luttrell et al., 1999). Plasmid pcDNA3.1+ with Fyn kinase was a kind gift from Dr Wendy Noble (King's College London, London, UK).
Antibodies
The following commercial antibodies were used: FUS-specific rabbit polyclonal antibody (NB100-565, Novus Biologicals), GFP-specific rabbit polyclonal antibody (ab290, Abcam), phosphotyrosine-specific mouse monoclonal PY20 antibody (sc-508, Santa Cruz Biotechnology), TNPO1-specific mouse monoclonal D45 antibody (ab10303, Abcam), β-tubulin-III-specific rabbit polyclonal antibody (T2200, Sigma-Aldrich) and GAPDH-specific rabbit polyclonal antibody (sc-25778, Santa Cruz Biotechnology).
Cell culture, transfection and treatment with inhibitors
HeLa cells were maintained in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (Gibco), 2 mM GlutaMAX (Gibco) and 100 U/ml penicillin-streptomycin (Gibco). SH-SY5Y cells were maintained in DMEM/F12 (Gibco) supplemented with 10% fetal bovine serum (Gibco), 2 mM GlutaMAX (Gibco) and 100 U/ml penicillin-streptomycin (Gibco). Hela and SH-SY5Y cells were transfected using PolyJet (SignaGen Laboratories) according to the manufacturer's protocol. Primary rat cortical neurons were grown for 5 days before transfection with Lipofectamine 2000 (Invitrogen). Vanadate stock solution was prepared by dissolving sodium orthovanadate (Millipore) in water to concentration of 200 mM. The pH was adjusted to 10.0 and the solution was heated until it turned colourless. After cooling, the pH was readjusted to 10.0 and previous steps were repeated until the pH stabilised at 10.0. The vanadate stock solution was aliquoted and stored at −20°C (Gordon, 1991). Pervanadate was prepared by adding 50 µl of vanadate stock solution and 1.6 µl of 30% (w/w) H2O2 to 948 µl of water, and incubating for 5 min at room temperature. Afterwards, excess H2O2 was removed by adding 200 µg/ml catalase (Sigma-Aldrich), followed by incubating for 5 min at room temperature. Pervanadate was diluted 100× in medium to give a final concentration of 100 µM, and cells were treated for 20 min (Huyer et al., 1997). 4-Amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2, Millipore) was dissolved in DMSO to a concentration of 10 mM. Before use, PP2 stock solution was diluted in medium to give a final concentration of 10 µM, and cells were treated for 1 h.
Immunofluorescence
Cells were grown on glass coverslips and were washed once with PBS before fixation with 4% paraformaldehyde in PBS for 15 min. Following washing with PBS, cells were permeabilised with 0.1% Triton X-100 in PBS for 6 min. After washing with PBS, blocking was performed with donkey serum in PBS. Primary antibodies that had been diluted in blocking medium were incubated over night at 4°C. Following three washes with PBS, cells were incubated in the dark with fluorescence-labelled secondary antibodies (donkey anti-rabbit DyLight 550, Thermo Scientific) for 1 h. After washing with PBS, the cells were stained with DAPI (Sigma-Aldrich), and coverslips were mounted using FluorSave Reagent (Millipore). Images were acquired using a Zeiss LSM 710 inverted confocal laser scanning microscope and ZEN 2010 B SP1 software.
Recombinant TNPO1
For expression of recombinant N-terminally GST-tagged TNPO1 (GST–TNPO1) pDEST-GST-TNPO1 was transformed into Escherichia coli BL21(DE3) (Invitrogen). Bacteria were grown in Luria-Bertani (LB) broth that had been supplemented with ampicillin. GST–TNPO1 protein expression was induced with 1 mM IPTG for 18 h at 20°C. Bacteria were harvested by centrifugation, and then frozen, thawed and resuspended in PBS with 5 mM DTT, 1 mg/ml lysozyme (Sigma-Aldrich), 5 units/ml of DNase (Promega) and 1 unit/ml of RNase A (Qiagen) followed by sonification. GST–TNPO1 was purified with GSTrap FF Columns (GE Healthcare) using a AKTA FPLC System (GE Healthcare) according to the manufacturer's protocol. Eluted protein was dialysed against binding buffer (20 mM sodium phosphate buffer, pH 7.4, 50 mM NaCl, 1 mM EDTA, 1 mM DTT).
Immunoprecipitation
HeLa cells were grown in 6-cm dishes and transfected using PolyJet (SignaGen Laboratories). After 24 h, cells were washed once in PBS and scraped into 300 µl of ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, 400 mM NaCl, 1% Triton X-100, Roche Complete protease inhibitors, Roche PhosSTOP). 100 mg of rat cortex was homogenised with a syringe in 500 µl of ice-cold lysis buffer. After 15 min of incubation on ice, lysates were sonicated and centrifuged at 21,000 g for 5 min for cell lysates and at 21,000 g for 30 min for tissue lysates. Supernatants were transferred into fresh tubes and diluted 1 in 3 with dilution buffer (50 mM Tris-HCl, pH 7.4, Roche Complete protease inhibitors, Roche PhosSTOP). Dynabeads G (Invitrogen) with bound antibodies were added to lysates. After rotation at 4°C for 3 h, beads were washed three times with ice-cold PBS and boiled for 5 min in 1× SDS loading buffer with 100 mM DTT. In the case of immunoprecipitation with recombinant GST–TNPO1, after washing with PBS, 150 pmol of GST–TNPO1 in PBS was added to the beads. After a second rotation at 4°C for 2 h, the beads were washed again three times with ice-cold PBS and boiled for 5 min in 1× SDS loading buffer with 100 mM DTT. Eluted proteins were analysed by immunoblotting.
In vitro pull down using streptavidin beads
Synthetic peptides (FUS504-526 and pFUS504-526) with biotin bound to the N-terminus of the peptide with aminocaproic acid (Ahx) linker were synthesised and purified using high-performance liquid chromatography (ProteoGenix, France) and were dissolved in dissolving buffer (20 mM sodium phosphate buffer, pH 7.4, 50 mM NaCl, 1 mM EDTA, 1 mM DTT). An in vitro pulldown assay was performed according to the protocol described previously (Dormann et al., 2012). Peptides were immobilised on streptavidin beads (Promega, 500 pmol/10 µl). Beads were afterwards blocked in binding buffer (20 mM sodium phosphate buffer, pH 7.4, 150 mM KCl, 0.5 mM EDTA, 5 mM MgCl2, 10% glycerol, 1 mM DTT, 0.5 mg/ml BSA) and incubated with the indicated amounts of recombinant GST–TNPO1 in 100 µl of the same buffer for 2 h at 4°C. Beads were washed three times with washing buffer (20 mM sodium phosphate buffer, pH 7.4, 150 mM KCl, 0.5 mM EDTA, 5 mM MgCl2, 10% glycerol, 1 mM DTT) and boiled for 5 min in 1× SDS loading buffer with 100 mM DTT. Eluted proteins were subjected to SDS-PAGE (10–20%) and visualised by using silver staining. ImageJ software was used to quantify the relative intensities of GST–TNPO1 bands.
In vitro pull down using glutathione beads
Recombinant GST–TNPO1 was immobilised on glutathione beads (Promega, 100 pmol/5 µl). Beads were afterwards blocked in binding buffer and incubated with the indicated amounts of synthetic peptides (FUS504-526 and pFUS504-526) in 100 µl of the same buffer for 2 h at 4°C. Beads were washed three times with washing buffer and boiled for 5 min in 1× SDS loading buffer with 100 mM DTT. Eluted proteins were dot blotted onto nitrocellulose membrane. After blocking with TBS that had been supplemented with 0.5% Tween-20, bound peptides were detected with streptavidin that had been conjugated to horseradish peroxidase (HRP). Chemiluminescent reagent (Millipore) was added, and images were acquired using a GelDoc System (Bio-Rad). ImageJ software was used to quantify the relative intensities of bound peptides.
His-tag pull down
HeLa cells were grown in 6-well plates and transfected using PolyJet (SignaGen Laboratories). After 24–48 h, cells were washed once in PBS and scraped into denaturing binding buffer (100 mM HEPES, pH 7.4, 10 mM imidazole, 8 M urea). After 15 min, cells were sonicated and centrifuged at 16,000 g for 5 min. Supernatant was transferred to fresh tubes, and 10 µl of MagneHis beads (Promega) was added. After rotation at room temperature for 1 h, the beads were washed three times with denaturing washing buffer (100 mM HEPES, pH 7.4, 10 mM imidazole, 8 M urea, 500 mM NaCl). Proteins were eluted from beads by incubating with denaturing elution buffer (100 mM HEPES, pH 7.4, 500 mM imidazole, 8 M urea) for 10 min. 2× SDS loading buffer with 200 mM DTT was added to eluates. Samples were analysed by immunoblotting.
Immunoblotting
Proteins were separated by reducing SDS-PAGE and transferred onto nitrocellulose membrane (GE Healthcare) using wet transfer at 200 mA for 90 min. Membranes were blocked in TBS with 0.5% Tween-20 (TBS-Tween-20) when using an antibody against phosphorylated tyrosine residues (Santa Cruz, sc-508), or in 5% non-fat dry milk in TBS-Tween-20 for any other antibody. Blocking was performed at room temperature for 1 h. Primary antibodies diluted in blocking solution were incubated for 1–4 h at room temperature. Following three washes with TBS-Tween-20, membranes were either incubated with HRP-conjugated secondary antibodies (anti-rabbit–HRP, 1:10,000, Jackson ImmunoResearch or anti-mouse–HRP, 1:5000, Millipore) or in the dark with fluorescent secondary antibodies (anti-rabbit–AlexaFluor555, 1:5000, Life Technologies) in blocking solution. After three further washes with TBS-Tween-20, membranes with HRP-conjugated antibodies were incubated with chemiluminescent reagent (Roche). Images were acquired using a GelDoc System (Bio-Rad). ImageJ software was used to quantify protein bands.
Statistical analyses
Data were analysed using GraphPad Prism 6 (GraphPad Software). Differences between two groups were analysed using Student's t-test (0.05 threshold value). Where differences between multiple groups were analysed, one-way ANOVA followed by Bonferroni post-hoc test was used to assess which groups were significantly different (0.05 threshold value).
Acknowledgements
We thank Dr Wendy Noble (King's College London, UK) for providing plasmid pcDNA3.1+ with Fyn kinase.
Footnotes
Author contributions
S.D., C.E.S., Y.-B.L. and B.R. designed research. S.D., S.P.M. and M.Š. performed the experiments. V.Ž. and G.G. performed the structural analysis. C.E.S., Y.-B.L. and B.R. provided the reagents. S.D., C.E.S. and B.R. wrote the manuscript with extensive input from other authors.
Funding
This work was supported by the Slovenian Research Agency [grant numbers J3-4026, J3-5502, J3-6789 and P4-0127]; Alzheimer's Research UK; the National Institute of Health Research Biomedical Research Centre based at Guy's and St Thomas' National Health Service Foundation Trust and King's College London in partnership with King's College Hospital. M. Mayr is a Senior Fellow of the British Heart Foundation.
References
Competing interests
The authors declare no competing or financial interests.