Mutations of the human valosin-containing protein gene cause autosomal-dominant inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia. We identified strumpellin as a novel valosin-containing protein binding partner. Strumpellin mutations have been shown to cause hereditary spastic paraplegia. We demonstrate that strumpellin is a ubiquitously expressed protein present in cytosolic and endoplasmic reticulum cell fractions. Overexpression or ablation of wild-type strumpellin caused significantly reduced wound closure velocities in wound healing assays, whereas overexpression of the disease-causing strumpellin N471D mutant showed no functional effect. Strumpellin knockdown experiments in human neuroblastoma cells resulted in a dramatic reduction of axonal outgrowth. Knockdown studies in zebrafish revealed severe cardiac contractile dysfunction, tail curvature and impaired motility. The latter phenotype is due to a loss of central and peripheral motoneuron formation. These data imply a strumpellin loss-of-function pathogenesis in hereditary spastic paraplegia. In the human central nervous system strumpellin shows a presynaptic localization. We further identified strumpellin in pathological protein aggregates in inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia, various myofibrillar myopathies and in cortical neurons of a Huntington’s disease mouse model. Beyond hereditary spastic paraplegia, our findings imply that mutant forms of strumpellin and valosin-containing protein may have a concerted pathogenic role in various protein aggregate diseases.
The late-onset multisystem disorder inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia (IBMPFD; OMIM 605382) is caused by mutations of the valosin-containing protein (VCP; also known as p97, TER ATPase and Cdc48p) gene on chromosome 9p13-p12 (Watts et al., 2004; Schröder et al., 2005). VCP is a ubiquitously expressed member of the AAA-ATPase family, which has a tripartite structure comprising an N-terminal domain (CDC48) involved in ubiquitin binding, and two central D1 and D2 domains which bind and hydrolyse ATP (DeLaBarre and Brunger, 2003). VCP assembles into functional hexamers with the D-domains forming a central cylindrical shape that is surrounded by the N-terminal domains (Rouiller et al., 2002). To date, 17 unique VCP missense mutations have been reported. Apart from the R191Q, A232E and A439S mutations, which reside in the N-D1-linker region, D1 domain and D1–D2-linker region, respectively, all other pathogenic IBMPFD mutations described thus far are located in exons coding for the CDC48 domain of the VCP protein with codon 155 being a mutation hot spot (Kimonis et al., 2008; Djamshidian et al., 2009; Stojkovic et al., 2009).
VCP has been associated with a wide variety of essential cellular processes comprising nuclear envelope reconstruction, the cell cycle, postmitotic Golgi reassembly, suppression of apoptosis, DNA damage response, the ubiquitin proteasome protein degradation system and aggresome formation (Kondo et al., 1997; Rabouille et al., 1998; Meyer et al., 2000; Hetzer et al., 2001; Rabinovich et al., 2002; Ju et al., 2008). VCP along with its co-factors Udf1 and Npl4 as well as Derlin-1, E3 and E4 ubiquitin ligases, and de-ubiquitinating enzymes has been implicated in a central role in endoplasmic reticulum-associated protein degradation, a process that removes improperly folded proteins from the endoplasmic reticulum to the cytosol for degradation by the 26S proteasome (Dai et al., 1998; Ye et al., 2001, 2004; Rabinovich et al., 2002; Lilley and Ploegh, 2005; Halawani and Latterich, 2006; Alexandru et al., 2008).
Within the large group of human protein aggregation diseases, IBMPFD has a unique role, because pathological protein aggregation occurs in both neuronal and striated muscle cells. While nuclear VCP and ubiquitin positive inclusions are found in striated muscle cells and neurons, the sarcoplasmic IBMPFD skeletal muscle pathology is characterized by the presence of VCP, ubiquitin, desmin and αB-crystallin-positive inclusions, rimmed (autophagic) vacuole formation and signs of myofibrillar degeneration. The latter findings link the VCP-related muscle pathology to the characteristic morphological findings in myofibrillar myopathies (Hübbers et al., 2007).
However, VCP positive inclusions are not specific for IBMPFD and have been documented in a wide variety of neurodegenerative disorders comprising Parkinson’s disease, Lewy body disease, Huntington’s disease, amyotrophic lateral sclerosis and spinocerebellar ataxia type III (Machado-Joseph disease) (Hirabayashi et al., 2001; Mizuno et al., 2003; Nan et al., 2005), as well as in various protein aggregation myopathies, including myofibrillar myopathies and inclusion body myositis (Greenberg et al., 2007; Schröder and Schoser, 2009). Furthermore, VCP in association with the cytoplasmic histone deacetylase 6, an aggresome shuttling protein, provides a link between the inhibition of the proteasome degradation system and a consecutive upregulation of the autophagy lysosomal pathway. Here, overexpression of histone deacetylase 6 has been shown to decrease protein aggregate toxicity, to enhance the autophagic degradation of proteins and to rescue model organisms from neurodegeneration (Kawaguchi et al., 2003; Pandey et al., 2007; Ju et al., 2008). In addition, studies using Drosophila demonstrated that VCP, which directly interacts with ataxin-3 (the protein mutated in spinocerebellar ataxia type III), selectively modulates aggregation and neurotoxicity induced by pathogenic ataxin-3 (Böddrich et al., 2006).
Here we report on the identification of a direct VCP–strumpellin interaction. Strumpellin (KIAA0196) in its mutant form causes a pure motor form of hereditary spastic paraplegia (Valdmanis et al., 2007) and also has been implicated in the pathogenesis of prostate cancer (Porkka et al., 2004; van Duin et al., 2005; Jalava et al., 2009). Furthermore, strumpellin has been identified as a subunit of the Wiskott Aldrich Syndrome protein and scar homologue complex, which activates Arp2/3-mediated actin polymerization and has a function in the fission of endosomes (Derivery et al., 2009). We provide a characterization of the strumpellin gene and protein and demonstrate its presence in protein aggregate diseases affecting the striated muscle and the CNS.
Materials and methods
Generation of mouse monoclonal VCP and rabbit polyclonal strumpellin antibodies
Generation of the VCP monoclonal antibody was performed as described earlier (Kohler and Milstein, 1975, 2005; Niebuhr et al., 1998) with modifications. Four 6- to 10-week-old Balb/C female mice were injected in their hind footpads with a mixture of a human VCP peptide (‘Results’ section) and Freund’s complete adjuvant (100 µl total volume containing 50 µg VCP peptide for four mice). For the following five injections the antigen was mixed with Freund’s incomplete adjuvant. One day after the final injection, mice were killed by cervical dislocation and popliteal lymph nodes were isolated for lymphocyte extraction. For hybridoma production, lymphocytes were fused with mouse myeloma cell lines Ag8 and Pai, using PEG4000. Out of 960 hybridoma clones, 248 clones were screened by western blot strip method employing recombinant glutathione S-transferase (GST)-VCP as probe. The consistent hybridoma clones secreting anti-VCP were subjected to single cell sub-cloning and the best sub-clone (K76-318-1) was chosen for further studies based on the specific detection of human VCP and the Dictyostelium discoideum orthologue cdcD (‘Results’ section). The monoclonal VCP antibody was purified from hybridoma supernatant using protein A-coated sepharose beads, dialysed against phosphate buffered saline and stored in aliquots at either −20 or 4°C containing 0.02% sodium azide.
For generation of rabbit polyclonal antibodies directed against human strumpellin a highly conserved epitope was chosen (‘Results’ section). The peptide was synthesized, coupled to the keyhole limpet haemocyanin carrier protein, and used for rabbit immunization (Speedy 28-day programme; Eurogentec S.A., Belgium). After three additional boosts, anti-sera were collected and the polyclonal strumpellin antibodies were affinity purified against the antigen. Antibody EP085378/SY1593 was chosen for further studies based on the specific detection of human strumpellin and its D. discoideum orthologue.
Co-immunoprecipitation experiments and mass spectrometry
Immunoprecipitations were performed as previously described (Dalal et al., 2004). In brief, cells were solubilized in lysis buffer containing 30 mM 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid (HEPES) pH 7.4, 100 mM NaCl, 5 mM MgCl2, 2 mM ATP, 1 mM dithiothreitol, 2 mM benzamidine, 1 mM phenyl-methanesulphonyl fluoride, 10 µg/µl aprotinin, 10 µg/µl leupeptin, 10 µg/µl pepstatin and 0.5% Triton X-100. The cell suspension was homogenized by 15 strokes of a tightly fitting dounce homogenizer, pre-cleared with Protein A-sepharose beads for 1 h and spun for 20 min at 15 000 g to remove cell debris. VCP monoclonal antibody (K76-318-1) (600 µg) or a 100 µl strumpellin polyclonal antibody solution was incubated with 100 µl Protein A-sepharose beads for 1 h and then blocked with 5% bovine serum albumin overnight. The pre-cleared cell lysate was incubated with the anti-VCP bound beads for 1 h, washed with lysis buffer (three times, 1 ml per washing) and then finally washed with buffer (five times) containing 30 mM HEPES pH 7.4, 100 mM NaCl, 5 mM MgCl2, 2 mM ATP, 1 mM dithiothreitol and 0.5% Triton X-100. All the steps were carried out either on ice or at 4°C. Proteins were eluted with 2 × sodium dodecyl sulphate (SDS)-sample buffer and separated by 10 or 12% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE) gels. Gels were either used for immunoblotting or stained with Coomassie blue to cut protein bands for mass spectrometry.
In-gel digestion of proteins for mass spectrometry was performed on a robotic platform using parameters described earlier (Staubach et al., 2009). Liquid chromatography–mass spectrometry data were acquired on an HCT ETD II ion trap mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with a nano electrospray ionization source (Bruker Daltonics, Bremen, Germany). Samples were introduced by an easy nano liquid chromatography system (Proxeon, Odense, Denmark) using a vented column setup comprising a 0.1 mm × 20 mm trapping column and a 0.075 mm × 100 mm analytical column, both self-packed with ReproSil-Pur C18-AQ, 5 µm (Dr Maisch, Ammerbuch, Germany). Samples of 5–18 µl were aspirated into the sample loop and a total of 25 µl was loaded onto the trap column using a flow rate of 6 µl/min. Loading pump buffer was 0.1% fatty acid. Peptides were eluted with a gradient of 0–35% acetonitrile in 0.1% fatty acid over 20 min and a column flow rate of 300 nl/min. Subsequently the acetonitrile content was raised to 100% over 2 min and the column was regenerated in 100% acetonitrile for an additional 8 min.
Data-dependent acquisition of mass spectrometry and tandem mass spectrometry spectra were controlled by the Compass 3.0 software. MS1 scans were acquired in standard enhanced mode. Five single scans in the mass range from m/z 400 to 1400 were combined for one survey scan. Up to three doubly and triply charged ions rising above a given threshold were selected for tandem mass spectrometry experiments. Ultrascan mode was used for the acquisition of MS2 scans in the mass range from m/z 100 to 1600 and three single scans were combined. The ion charge control value was set to 250 000 for all scan types. Peak lists in mascot generic format were generated from the raw data by using the Data Analysis software module (Bruker Daltoniks, Bremen, Germany).
Proteins were identified by searching the Swiss Prot 50.5 or 56.4 (human) or NCBInr release 20080210 (Dictyostelium) using a local installation of MASCOT 2.2 (Matrix Science Ltd, London, UK). Searches were submitted via Proteinscape 2.0 (Bruker Daltoniks, Bremen, Germany) with the parameter settings enzyme = ‘trypsin’, species = ‘human’ or ‘all entries’, fixed modifications = ‘carbamidomethyl (Cys)’, optional modifications = ‘methionine oxidation’ and missed cleavages = 1. The mass tolerance was set to 0.3 Da for peptide and fragment spectra. Peptide charges were 2+ and 3+. The tandem mass spectrometry tolerance was 0.3 Da.
Initial protein identifications in the course of this project were performed by peptide mass fingerprint analyses. Automatic acquisition of matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry spectra was performed on a Bruker Reflex IV mass spectrometer running under Flexcontrol 1.3. Methods for automatic acquisition were defined in the Autoexecute module integrated in FlexControl. Briefly, all spectra were acquired in the positive ion mode using an external calibration. The spectra were acquired in 20 acquisition cycles. In each cycle, five shots were stored in the temporal acquisition buffer for evaluation by the software. Spectra with a minimum signal to noise ratio of 10 and a minimum resolution of 4500 for the most intense peak in the mass range from m/z 1200 to 3000 were added to the sum. If the spectra did not meet these criteria, the software tried to optimize the laser intensity and if this failed five times, the acquisition continued to another position on the same spot. The software moved on to the next position when 100 single shots had been summed or when >80 consecutive acquisition cycles failed. The raw spectra were processed by Flexanalysis 2.4 and the generated peak lists were transferred to Biotools 3.0, which triggered batch searches in NCBInr release 20080210 using MASCOT 1.9. Searches were restricted to Homo sapiens (fixed modifications = ‘carbamidomethylated Cys’, variable modifications = ‘oxidized Met’) and used trypsin specificity with one missed cleavage allowed. The maximum mass error was 100 ppm for externally calibrated spectra.
Western blotting and antibodies
Protein samples were separated by 10, 12 or 15% SDS–PAGE under reducing conditions and transferred onto nitrocellulose membranes by the semi-dry (Laemmli, 1970; Towbin et al., 1979) or wet-blotting method (Villanueva, 2008). The membranes were blocked with Tris-buffered saline-Tween-20 (TBS-T) buffer (10 mM Tris–HCl pH 8.0, 150 mM NaCl, 0.2% Tween-20) containing 5% milk powder. Human multiple tissue western blot membranes were purchased from BioCat/BioChain (Human Adult Normal Tissue, #W1234404) and ProSci (Human Normal Tissue customized selection of human brain regions, #1541N, as well as Human Normal Tissue II, #1522, and III, #1523). Two-dimensional gel electrophoresis in conjunction with immunoblotting was performed as previously described (Clemen et al., 2005). Incubation with primary antibody was followed by incubation with goat anti-mouse or anti-rabbit IgG coupled to horseradish peroxidase (Sigma); visualization was done by enhanced chemiluminescence and images were recorded and analysed using the imaging system Fluorchem SP (Alpha Innotech).
VCP was detected with monoclonal antibody K76-318-1 (dilution 1:3000) and strumpellin was recognized by polyclonal antibody EP085378/SY1593 (dilution 1:200). Calnexin was recognized by anti-Calnexin polyclonal rabbit (dilution 1:1000; C4731, Sigma), Golgi 58K protein by anti-58K-9 mouse monoclonal (dilution 1:500; ab27043, Abcam), mitochondrial complex IV by COX1 antibody (dilution 1:3000; A-6403, Molecular Probes), clathrin heavy chain by anti-clathrin heavy chain monoclonal mouse (dilution 1:1000; 610499, BD Transduction Laboratories), synaptophysin by anti-synaptophysin (dilution 1:2000; RM-9111, Thermo Scientific), poly-ubiquitin by multi ubiquitin monoclonal antibody (FK2) (dilution 1:1000; SPA-205, Assay Designs) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was detected by monoclonal GAPDH-71.1 (dilution 1:10 000; Sigma). Furthermore, a monoclonal antibody against mRFPmars also recognizing mCherry (A Müller-Taubenberger and AA Noegel, unpublished data) was employed.
Expression constructs and in vitro mutagenesis
Plasmids pGEX-6P-1 and pEGFP-N1 carrying the cDNA encoding human wild-type and R155C mutant VCP were generated and verified as described in our previous study (Hübbers et al., 2007). The cDNA coding for the human VCP peptide used for immunization (see above) was amplified by polymerase chain reaction (PCR) (primer pair: VCPmAbFor including BamHI restriction site: 5′-GTGGATCCATTGCTCGAGCTGTAGC-3′, VCPmAbRev including Stop codon and EcoRI restriction site: 5′-CGAATTCTCAAGCGATGGCATCTAGC-3′), cloned into pGEX-4T-1 and verified by DNA sequencing. Plasmid DSRed2-ER containing the endoplasmic reticulum targeting sequence of calreticulin fused to the 5′ and the endoplasmic reticulum retention sequence KDEL fused to the 3′ end of DSRed2 was used as luminal endoplasmic reticulum-marker (Clontech, #632409).
Human strumpellin cDNA (accession no. NM_014846.3) in pReceiver-B05 (EX-T0444-B05) for bacterial expression of GST-strumpellin as well as in pReceiver-M55 (EX-T0444-M55) and pReceiver-M56 (EX-T0444-M56) for expression of mCherry–strumpellin and strumpellin–mCherry, respectively, in mammalian cells was purchased from GeneCopoeia. The primers, forward 5′-GGAGAAAAATGAGGACCTTCAAGCTTGG-3′ and reverse 5′-CCAAGCTTGAAGGTCCTCATTTTTCTCC-3′ were used to introduce the c.1411A>G (p.Asn471Asp, N471D) mutation and a silent EcoO109I restriction site for control digestion into the human strumpellin cDNA by PCR-based site-directed mutagenesis (Quickchange kit, Stratagene). The full-length sequence of human strumpellin in the constructs and the presence of the point mutation were confirmed by sequencing.
Transfection of a small hairpin RNA cloned into pLKO.1-puro (BioCat/SBI/Open Biosystems) was used to transiently reduce the expression of strumpellin in mammalian cells (most efficient small hairpin RNA oligo TRCN0000129686: CGAAGATTCAAGATTGGCAAA). Knockdown efficiency was determined from the reduction of the fluorescence intensity of a co-transfected mCherry–strumpellin expression construct in a separate experiment.
VCP expression and purification
The vector pGEX-4T-1 and pGEX-6P-1 carrying the cDNA encoding the VCP immunization peptide and VCP full-length protein (wild-type and R155 C mutant), respectively, were transformed into an Escherichia coli arctic express RIL strain (Stratagene). Three litre fresh lysogeny broth medium containing ampicillin (100 μg/ml) and gentamycin (20 μg/ml) was inoculated with 1% of overnight grown culture. The cells were grown at 30°C to an A600 of 0.6 and were induced by 0.3 mM isopropyl-1-thio-β-d-galactopyronoside (IPTG) followed by overnight incubation at 10°C. The cells were harvested, resuspended in lysis buffer containing 50 mM Tris–HCl pH 8.0, 150 mM NaCl, 2 mM ethylene glycol tetraacetic acid, 1 mM phenylmethylsulphonyl fluoride, 2 mM benzamidine, 2 mM ATP, 5 mM MgCl2, 5 mM dithiothreitol, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin, 5% glycerol and 100 µg/ml lysozyme, and sonicated. Cell debris was removed at 20 000 g for 15 min at 4°C. The prepared cell homogenates were loaded onto glutathione–sepharose columns, which were equilibrated with wash buffer. The columns were washed with wash buffer containing 50 mM Tris–HCl pH 7.4, 150 mM NaCl and 2 mM ethylene glycol tetra-acetic acid to an A280 of 0.0, and further washed with 1× phosphate buffered saline in case of the VCP immunization peptide and with PreScission protease buffer containing 50 mM Tris–HCl pH 7.0, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid and 1 mM dithiothreitol for the VCP full-length protein purification.
The GST-VCP-monoclonal antibody peptide bound beads were incubated with thrombin [Sigma; for 1 ml of glutathione–sepharose bed volume 50 µl of resuspended thrombin (1 NIH unit/μl) and 950 μl of 1× phosphate buffered saline were pre-mixed] at room temperature overnight. The thrombin cleaved VCP immunization peptide subsequently was eluted with 20 ml of 1× phosphate buffered saline and concentrated to 500 μl using 5 kDa centripreps (Vivaspin). The VCP peptide obtained was further purified by reversed phase high-performance liquid chromatography, lyophilized, and kept at −20°C prior to its use for the immunization procedure.
The GST-VCP full-length protein (wild-type and R155 C mutant) bound beads were incubated with PreScission protease (GE Healthcare) for 4 h at 4°C and the cleaved protein was eluted with the same buffer as above followed by overnight dialysis against a buffer containing 30 mM HEPES pH 7.4, 150 mM NaCl, 2 mM ATP, 5 mM MgCl2 and 1 mM dithiothreitol. The VCP wild-type and R155 C mutant proteins obtained were stored at −20°C. Samples were analysed by SDS–PAGE and mass spectrometry, and the formation of VCP hexamers was verified by size exclusion chromatography.
Strumpellin protein expression and purification
pReceiver-B05 carrying the cDNA encoding wild-type and N471D mutant strumpellin were transformed into E. coli BL21-Gold (DE3) pLysS (Stratagene) and BL21 (DE3) (Stratagene), respectively. Three litre fresh lysogeny broth medium containing ampicillin (100 μg/ml) and chloramphenicol (35 μg/ml) for GST–strumpellin wild-type, and ampicillin for GST–strumpellin N471D mutant was inoculated with 1% overnight grown culture. The cells were grown at 37°C to an A600 of 0.6 and were induced by 0.6 mM isopropyl β-d-1-thiogalactopyranoside at room temperature for 6 h. Cells were harvested and resuspended in lysis buffer containing 30 mM HEPES pH 7.4, 150 mM NaCl, 2 mM ethylene glycol tetraacetic acid, 2 mM ethylenediaminetetraacetic acid, 1 mM phenylmethylsulphonyl fluoride, 2 mM benzamidine, 5 mM dithiothreitol, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin, 5% glycerol, 0.5% Triton X-100 and 100 µg/ml lysozyme, sonicated and spun at 20 000 g for 15 min at 4°C. The cell lysates obtained were incubated with glutathione–sepharose beads (batch assay) and washed with buffer containing 30 mM HEPES pH 7.4, 300 mM NaCl, 2 mM ethylene glycol tetraacetic acid, 2 mM ethylenediaminetetraacetic acid, 1 mM dithiothreitol and 0.5% Triton X-100. The GST–strumpellin proteins were eluted with buffer containing 30 mM HEPES, pH 8.0, 100 mM NaCl, 1 mM dithiothreitol, 5% glycerol and 20 mM glutathione followed by overnight dialysis against 30 mM HEPES pH 7.4, 150 mM NaCl and 1 mM dithiothreitol. Proteins were stored in aliquots at −20°C. Samples were analysed by SDS–PAGE and mass spectrometry.
Strumpellin–VCP pull-down assays
The purified GST–strumpellin wild-type and N471D proteins, as well as purified GST as a control, were immobilized on 100 µl of glutathione–sepharose beads and washed with buffer A containing 30 mM HEPES pH 7.4, 100 mM NaCl, 2 mM ATP, 1 mM dithiothreitol and 5 mM MgCl2. Either GST-cleaved VCP wild-type or R155C mutant was added to the beads and incubated for 2 h at 4°C on a rotating wheel. Beads were then washed with buffer A containing 0.5% Triton X-100, and bound VCP protein was eluted with 2× SDS-sample buffer followed by SDS–PAGE and analysis of protein–protein interactions by immunoblotting.
Ubiquitin dot-blot overlay assay
One microgram of either K48-linked ubiquitin chain mixture (Ub1-7; Biomol, UC-240) or K63-linked ubiquitin chain mixture (Ub1-7; Biomol, UC-340), bovine serum albumin, purified recombinant VCP and strumpellin were spotted onto nitrocellulose membranes (Whatman GmbH). The membranes were dried at room temperature for 15 min and washed with Tris-buffered saline–Tween-20 buffer (10 mM Tris–Cl pH 8.0, 150 mM NaCl, 0.2% Tween-20) for 5 min. The dot blots were blocked in Tris-buffered saline–Tween-20 buffer containing 5% milk powder for 1 h at room temperature and washed. Blots were incubated with either VCP or strumpellin in 2 ml of protein-binding buffer containing 30 mM HEPES pH 7.4, 100 mM NaCl, 2 mM ATP and 5 mM MgCl2 as previously described (Fenner et al., 2009). Binding was detected with either VCP or strumpellin antibody followed by secondary antibody conjugated with peroxidase.
Strumpellin reverse transcription polymerase chain reaction and polymerase chain reaction analyses
Total RNA from human brain regions (total brain, #636530, cerebellum, #636535, spinal cord, #636554, cerebral cortex, #636561, medulla oblongata, #636562, frontal lobe, #636563, temporal lobe, #636564, parietal lobe, #636571, pons, #636572, substantia nigra, #636560, occipital pole, #636570, paracentral gyrus, #636574; all from Clontech) and human skeletal muscle (#636534, Clontech) were used to generate cDNA employing p(dN6)-oligomers and Moloney murine leukaemia virus reverse transcriptase (Promega) at 37°C for 1 h. The human multiple tissue cDNA Panels 1 (#636742) and 2 (#636743) (Clontech) were used for expression profiling.
Strumpellin semi-quantitative PCRs (25 cycles; annealing temperature 58°C) were performed using three different primer pairs (hStrumPCR1for: 5′-ATGTTGGACTTTCTAGCCGAG-3′, hStrumPCR1rev: 5′-CTGAGTCTGCTGTATGAAGC-3′, product size 1150 bp; hStrumPCR2for: 5′-CCTGTGACCCAAACAACAAAC-3′, hStrumPCR2rev: 5′-CTTCGTTCTTAGAAAGTTATTAC-3′, product size 1229 bp; hStrumPCR3for: 5′-CATGTACCAGTCCACTCATATTC-3′, hStrumPCR3rev: 5′-CAGCACTGTTCTGAACTCATC-3′, product size 1081 bp). For control a GAPDH primer pair was used (hGAPDHfor: 5′-GCCGTCTAGAAAAACCTGCCAAATATGATG-3′, hGAPDHrev: 5′-GTGAGGGTCTCTCTCTTCCTCTTGTGCTCT-3′, annealing temperature 71.5°C, product size 324 bp).
Cell culture, transfection and in vitro assays
HEK293 human embryonic kidney cells (ATCC CRL-1573), 293TN cells (BioCat/SBI LV900A-1), Pop10 cells (Quasdorff et al., 2008) and HEK293 cells stably expressing VCP wild-type or R155C mutant were grown in standard culture medium as described earlier (Hübbers et al., 2007). The neuroblastoma cell line SH-SY5Y (ECACC 94030304) was grown in a 1:1 mixture of Ham’s F12 and Eagle’s minimum essential medium supplemented with 15% foetal bovine serum, 2 mM l-glutamine, 1% non-essential amino acids, 100 U/ml penicillin G and 100 µg/ml streptomycin at 5% CO2 in a 37°C incubator. Wild-type AX2 and cdcD(VCP)-green fluorescent protein expressing (The Dictyostelium strain expressing GFP-tagged VCP is part of another study and will be published separately.) D. discoideum amoeba were cultivated as previously described (Na et al., 2007).
For wound healing assays, 70 µl of 293TN cell suspensions were seeded into each half-chamber in a 24-well culture-insert plate (Ibidi; #80241). Cells were grown to 60–70% confluence and transiently transfected with mCherry–Strumpellin wild-type, N471D mutant, or strumpellin-small hairpin RNA vectors using the lipofectamine 2000 method (Invitrogen). To create a wound, the culture-insert was removed after 24 h of transfection and wound closure was monitored at different time periods, i.e. at 5 and 26 h after insert removal, using a Leica DMIRE2 microscope in phase contrast mode and Leica FW4000 software.
For analysis of axonal outgrowth, SH-SY5Y cells were seeded into µ-dishes (Ibidi, #81156, 35 mm) and transfected with the strumpellin small hairpin RNA expression plasmid. After 24 h of transfection, cells were induced to differentiate into neuron-like cells by addition of 10 µM retinoic acid to the growth medium for another 24 h. Images of axonal outgrowth were recorded by using a Leica DMIRE2 microscope in phase contrast mode. Statistical significance of wound closure and axonal outgrowth were analysed using the mean value, standard deviation and Student’s t-test. The efficiency of strumpellin knockdown and axonal outgrowth defects were confirmed by semi-quantitative reverse transcription polymerase chain reaction of strumpellin and western blot analysis of synaptophysin expression, respectively.
Direct and indirect immunofluorescence imaging of cultured cells
Cells were grown on glass coverslips in 6-well culture plates and transiently transfected with the plasmid DNA indicated. The following reagents were prepared in 1× phosphate buffered saline: 4% paraformaldehyde, 0.2% saponin, 0.02% saponin and 0.15% glycine; all steps were carried out at room temperature. After 24 h of transfection, cells were washed with 1× phosphate buffered saline and subsequently fixed with 4% paraformaldehyde for 20 min. Cells were then permeablized with 0.2% saponin for 5 min and incubated with 0.15% glycine for 10 min followed by three washes with 0.02% saponin for 5 min each. For nuclear staining, 4′,6-diamidino-2-phenylindole was diluted to 1 : 100 in 0.02% saponin and incubated for 30 min.
For indirect immunostaining, cells were further blocked with 1% foetal bovine serum in 0.02% saponin for 30 min and washed again three times with 0.02% saponin. For endoplasmic reticulum staining, cells were incubated with polyclonal rabbit anti-calnexin (dilution 1 : 200; Sigma) for 1 h and washed three times with 0.02% saponin. Secondary Alexa488-labelled donkey anti-rabbit antibody (dilution 1 : 500; Invitrogen) was incubated with cells for 1 h to detect the primary antibody in conjunction with nuclear staining. All cell images were captured on a Leica TCS SP5/AOBS/tandem scanning system equipped with the Leica LAS-AF software (version 2.2.1 build 4842).
Indirect immunofluorescence imaging of human skeletal muscle specimen
Muscle specimens were obtained by standard open biopsy during diagnostic work-up for neuromuscular complaints. Control muscles were obtained through diagnostic biopsies of patients complaining of muscle symptoms who were ultimately deemed free of muscle disease. The use of these human muscle specimens for protein expression analyses was approved by the local ethics committees of the Universities of Bonn and Munich.
Muscles were frozen in melting isopentane and stored in liquid nitrogen. Cryosections were cut at 6 µm thickness and transferred onto silaned glass slides. The sections were then fixed in a 1:1 methanol acetone mixture for 5 min at −20°C and air dried either to be stained directly or stored at −20°C for subsequent staining. Isolation of human myofibres was performed by manual separation of fresh muscle biopsies in ‘relaxation solution’ (Vielhaber et al., 2000). The fibres were then dried on glass slides, fixed and stained the same way as the cryostat sections.
For immunohistochemical analysis of the muscle specimens, the polyclonal strumpellin antibody EP085378/SY1593 (dilution 1:40), the VCP antibody K76-318-1 (dilution 1:150) and a monoclonal mouse antibody against desmin (dilution 1:300; D33, DAKO) were used. Secondary antibodies were Alexa594-conjugated goat anti-mouse IgG (dilution 1:800), as well as isotype specific Alexa594-conjugated goat anti-mouse IgG1 (dilution 1:800), Alexa488-conjugated goat anti-mouse IgG2b (dilution 1:400) and Alexa488-conjugated goat anti-rabbit IgG (dilution 1:400; all Molecular Probes/Invitrogen). The secondary antibodies were diluted in phosphate buffered saline and applied for 1 h at room temperature. Nuclei were stained by 2 min incubation in a solution of bisbenzimide (0.5 g/ml; Sigma-Aldrich) in phosphate buffered saline at room temperature prior to the final wash. The sections were finally mounted in a Mowiol 4-88 (Calbiochem/Merck Chemicals) and glycerol mix in Tris buffer pH 8.5 with 0.1% 1,4-diazabicyclo(2,2,2)octane (Sigma-Aldrich). A Nikon H800 microscope with a SPOT FLEX 64 Mp Shifting Pixel CCD-camera (model #15.2, Diagnostic Instruments Inc., Visitron Systems GmbH) and SPOT software (version 4.6, Visitron Systems GmbH) was used for analysis and documentation.
Indirect immunofluorescence imaging of human brain specimen
Post-mortem human hippocampus (n = 3) or spinal cord (n = 3) was obtained from six patients without neurological disorders. The use of these human brain specimens for protein expression analyses was approved by the local ethics committee of the University of Erlangen-Nuremberg.
All tissues were fixed in 10% formalin for 2 weeks and processed into liquid paraffin. Paraffin-embedded tissue blocks were cut at 4 μm on a rotation microtome (Microm). Double immunofluorescence staining with antibodies directed against strumpellin (dilution 1:75) and synaptophysin (mouse-anti-synaptophysin; dilution 1:50; Dako) was performed with Cy2 and Cy3 labelled secondary antibodies (Dianova). Nuclei were counterstained with Hoechst 33342 (Sigma-Aldrich). All images were recorded with appropriate fluorescence filter combinations.
Immunohistochemical detection of strumpellin in mouse and rat brain
Animals were deeply anaesthetized by an intraperitoneal injection of ketamine–hydrochloride (10 mg/kg) and xylazinhydrochloride (100 mg/kg) (both Pharmingen), transcardially perfused through the ascending aorta with 100 ml ice-cold 0.9% NaCl followed by 300 ml ice-cold 4% paraformaldehyde dissolved in 0.1 M phosphate buffered saline. Brains were postfixed in 4% paraformaldehyde, cryoprotected in 30% (v/v) sucrose solution and shock-frozen in −60°C petrolether. Serial one-to-four 30 µm coronal cryosections were prepared from the rostro-to-caudal dimension of the mouse/rat brain according to the anatomical atlases of the mouse/rat brain (Paxinos and Watson, 2007; Franklin and Paxinos, 2008). Sections were collected in 0.1 M phosphate buffered saline or stored in cryoprotectant solution at −20°C.
Substrate-based, single-labelling immunohistochemistry of VCP and strumpellin were performed on mouse and rat brain cryosections according to the free-floating method. All washing steps were done in 0.1 M phosphate-buffered saline, 3 × 10 min. Endogenous peroxidase was blocked by incubation in 1% H2O2, 20% v/v methanol, 0.1 M phosphate-buffered saline (5 min, room temperature). Then sections were pre-incubated in 5% bovine serum albumin (0.3% Triton, 0.1 M phosphate-buffered saline, 45 min, room temperature) and incubated either with rabbit polyclonal anti-VCP antibody (dilution 1:500–1:2000; NB100-1558, Novus Biologicals), rabbit polyclonal anti-strumpellin (dilution 1:50) or rabbit monoclonal anti-huntingtin antibody (1:500; clone EP867Y, Epitomics) diluted in 2% bovine serum albumin, 0.3% Triton, 0.1 M phosphate-buffered saline (48 h, 4°C). The huntingtin antibody detects a sequence of human huntingtin corresponding to the apopain cleavage site and reacts with human, mouse and rat huntingtin. Sections were rinsed in 0.1 M phosphate-buffered saline and incubated in biotinylated donkey anti-rabbit antibody (dilution 1:500; Santa Cruz Biotechnology; diluted in 0.1 M phosphate-buffered saline, 2% bovine serum albumin for 1 h at room temperature). Sections were rinsed, incubated in the avidin–biotin–peroxidase reaction according to the manufacturer’s recommendation (Vector Laboratories) and reacted with 3,3′-diaminobenzidine as a chromogen. Finally, sections were mounted onto object slides, air-dried and cover-slipped with DPX Mountant (Fluka Chemie GmbH).
For single immunofluorescence labelling, adjacent sections at the level of the striatum were removed from the cryoprotectant solution, washed and pre-incubated in 5% bovine serum albumin, 0.3% Triton and 0.1 M phosphate buffered saline for 45 min at room temperature. Sections were then incubated with primary antibodies, all diluted in 0.3% Triton and 0.1 M phosphate buffered saline (48 h, 4°C). Sections were rinsed in 0.1 M phosphate buffered saline and incubated with the appropriate Cy2- and Cy3-conjugated, affinity-purified secondary antibodies (Jackson ImmunoResearch) (both diluted 1 : 400 in 0.1 M phosphate-buffered saline, 1 h, room temperature). After washing, sections were counterstained with 4′,6-diamidino-2-phenylindole or DRAQ5 (Biostatus) (both diluted 1 : 1000), mounted and cover-slipped with an aqueous fluorescence mounting medium (Sigma).
For double-labelling immunofluorescence histochemistry for strumpellin with VCP or huntingtin, adjacent series of sections were used. Sections were pre-treated by antigen retrieval, peroxidase inhibition and biotin was blocked using the TNB blocking buffer (NEL700, Perkin Elmer Life Sciences; 30 min) and no peroxidase inhibition was done. Subsequently, anti-strumpellin (dilution 1:50) was diluted in 2% bovine serum albumin, 0.4% Triton X-100 and incubated for 72 h at 4°C. After rinsing, the sections were transferred to secondary biotinylated donkey anti rabbit (dilution 1:500; SC-2098, Santa Cruz Biotechnologies) diluted in 2% bovine serum albumin, phosphate-buffered saline (1 h, room temperature). Subsequently, sections were incubated with streptavidin-conjugated horseradish peroxidase (dilution 1:100) followed by biotinyl tyramide (dilution 1:100; 5 min, dark), according to the kit’s instructions (NEL700). Then, sections were washed and immunoreactivity was visualized using Cy2-conjugated streptavidin (Dianova), diluted 1:800 in phosphate-buffered saline (1 h, room temperature). For co-localization, sections were washed thoroughly and incubated with the polyclonal/monoclonal rabbit anti-VCP or anti-huntingtin (diluted 1 : 500 in phosphate-buffered saline and 2% bovine serum albumin overnight at 4°C in the dark). Sections were rinsed, incubated in Cy3-conjugated anti-rabbit (Jackson ImmunoResearch) and diluted 1:400 in 1 M phosphate-buffered saline. Finally, sections were coverslipped with a fluorescent mounting medium (Sigma-Aldrich). Controls included: (i) omission of the primary antibody; (ii) replacing the primary antiserum with normal non-immune rabbit serum; and (iii) omitting the biotinylated secondary antibody, and did not show any immunoreactivity.
For microscopic analysis, labelled sections were examined with a Zeiss Axio Imager Z1 equipped with an ApoTome (Zeiss). Images were taken with a 20× (0.8, Apochromat) or a 100× (1.3 oil, Plan-Neofluar) objective as stacks of multiple optical sections, and projections were calculated with AxioVision 4.8 software (Zeiss).
Differential centrifugation and sucrose density gradient assays
Separation of total lysates from cultured cells and murine skeletal muscle tissue into 100 000 g supernatants and pellets was carried out as previously described (Spoerl et al., 2002). Isopycnic separation on discontinuous sucrose gradients was performed as previously described (Clemen et al., 1999). The sedimentation coefficient in Svedberg units was calculated by dividing the minimal sedimentation velocity by the centrifugal force. The molecular mass of the complex was estimated based on the obtained Svedberg constant.
Zebrafish injection procedures, fractional shortening measurements, histology, in situ hybridizations and immunostainings
Care and breeding of zebrafish Danio rerio was as previously described (Westerfield, 1995). Morpholino modified oligonucleotides were directed against the splice donor site of intron 3 of strumpellin (MO-strumpellin, 5′-GATTTGAAATGAACACACACCTTA-3′). A standard control oligonucleotide (MO-control) (GENETOOLS, LLC) was injected at the same concentration as a negative control. To inhibit pigmentation, 0.003% 1-phenyl-2-thiourea was added to the embryo medium. Pictures and movies were recorded at 24, 48 and 72 h after fertilization. To analyse the contractile force of the strumpellin morphant and control-Morpholino injected embryos, we performed fractional shortening measurements with the help of the zebraFS software application (www.benegfx.de). For histology, embryos were fixed in 4% paraformaldehyde and embedded in JB-4 (Polysciences Inc.). Five micrometre sections were cut, dried and stained with haematoxylin/eosin. Whole mount antisense RNA in situ hybridization was carried out as described (Rottbauer et al., 2005) using a digoxigenin-labelled antisense probe for zebrafish MyoD. For immunostainings, zebrafish embryos were fixed in Dent’s fixative and monoclonal antibodies MF 20 and S 46 were used (Bader et al., 1982).
Twelve-week-old male R6/2 mice, transgenic for exon 1 of the truncated 5′-end of the human Huntington’s disease gene carrying ∼120 ± 5 CAG repeat expansions (Mangiarini et al., 1996) and 15-month-old YAC-128 mice, containing a YAC full-length human huntingtin gene with a 128 CAG repeat expansion in exon 1 (Slow et al., 2003) including littermate wild-type controls were kindly provided by Dr Huu-Phuc Nguyen (University of Tübingen, Germany). Twelve-month-old homozygous male transgenic Huntington’s disease rats with a huntingtin cDNA fragment with 51 CAG repeats under control of the native rat huntingtin promoter have been previously described (von Hörsten et al., 2003), expressing a truncated gene product corresponding to 22% of the full-length huntingtin (cDNA position 324–2321, exon 1–16). Animals were genotyped as previously described (Nguyen et al., 2006). All animals were kept within a 12–12 h light–dark cycle, under controlled room temperature (22 ± 0.5°C) and free access to standard laboratory food and water. All animal procedures were approved by the District Government of Mittelfranken, Bavaria, Germany, and performed in compliance with international animal welfare standards.
Identification of putative VCP binding partners in human and D. discoideum cells
In order to identify novel and disease-related VCP binding partners, we performed co-immunoprecipitation experiments employing a newly generated mouse monoclonal VCP antibody (K76-318-1). For the immunization, we chose a highly conserved peptide (amino acid 254-IARAVANETGAFFFLINGPEIMSKLAGESESNLRKAFEEAEKNAPAIIFIDELDAIA-amino acid 310) residing in the D1 domain of the human VCP protein, which is 94% identical to the D. discoideum orthologue cdcD. In silico structural analysis showed that residues 278–297 are surface exposed in the hexameric quaternary structure of VCP. Thus generated VCP antibody specificity was validated by immunoblotting, indirect immunofluorescence, immunohistochemistry and immunoprecipitation assays. Our VCP antibody specifically detects recombinant GST-tagged VCP, green fluorescent protein-tagged overexpressed VCP from HEK293 cells and endogenous VCP from D. discoideum, various mammalian cell lines and tissues.
Coomassie stained SDS–PAGE gels of samples from VCP co-immunoprecipitation experiments using soluble cell lysates from HEK293 cells resulted in multiple distinct bands, which were excised and identified by mass spectrometry (Fig. 1A, upper panel). As a positive control, samples were used for VCP and clathrin heavy chain [an established VCP-binding partner (Pleasure et al., 1993)] immunoblotting (Fig. 1A, lower panel). Corresponding experiments were performed on soluble cell lysates from stably transfected HEK293 cells overexpressing VCP-green fluorescent protein, and wild-type and VCP-green fluorescent protein expressing D. discoideum cells. The identified putative VCP binding partners from these four types of experiments are summarized in Supplementary Tables 1–4 and illustrated in a Venn diagram (Fig. 1B). The presence of strumpellin and clathrin heavy chain was identified by mass spectrometry in all four co-immunoprecipitation experiments. The identification of strumpellin as a putative novel VCP binding protein is of particular interest, because of its recently described association with a variant of pure hereditary spastic paraplegia (Valdmanis et al., 2007).
Strumpellin is a direct VCP interaction protein
To further analyse the VCP–strumpellin interaction, we generated a rabbit polyclonal strumpellin antibody using a highly conserved peptide (amino acid 458-SGVKPLTRVEKNENL-amino acid 472) which is 80% identical to the D. discoideum orthologue (Eichinger et al., 2005). The specificity of this antibody was validated by immunoblotting (recombinant GST-tagged strumpellin and mCherry-tagged strumpellin expressed in 293TN cells as probes), indirect immunofluorescence and immunoprecipitation. In contrast to the sequence derived molecular mass of 134 kDa, our strumpellin antibody consistently detected a single band corresponding to 110 kDa in SDS–PAGE using lysates of human and murine skeletal muscle tissues. Similarly, corresponding experiments using protein extracts from 293TN cells expressing mCherry–strumpellin and bacterially expressed GST–strumpellin with calculated molecular masses of 162 kDa also indicated lower apparent molecular masses of 145 kDa (Fig. 2A). To further analyse the aberrant migration in SDS–PAGE we subjected Coomassie stained bands of recombinant bacterially expressed GST–strumpellin and its degradation products to mass spectrometry. Only the masses obtained from the analysis of the 145 kDa band were distributed over the entire strumpellin amino acid sequence including peptides at the very N- and C-terminal positions (sequence coverage 53%). Thus, the 110 and 145 kDa signals represent the full-length strumpellin and GST–strumpellin, respectively.
Employing reverse co-immunoprecipitation experiments using our novel strumpellin antibody and lysates of 293TN cells overexpressing mCherry–strumpellin resulted in the identification of VCP (Fig. 2B, upper panel). In addition, blots were probed with the strumpellin antibody and a monoclonal antibody against mRFPmars, which recognizes the mCherry tag. The latter two immunoblots verified a specific co-immunoprecipitation of VCP and strumpellin (Fig. 2B, middle and lower panels).
These findings provide strong evidence for VCP–strumpellin interactions; however, it is not clear whether VCP directly interacts with strumpellin. To address this issue, we performed pull-down experiments employing bacterially expressed recombinant GST–VCP and GST–strumpellin proteins. Using GST–strumpellin as bait, soluble VCP was identified as a directly interacting protein, as illustrated in Fig. 2C. In addition, dot-blot overlay assays (immobilized VCP/overlaid GST–strumpellin and vice versa) were performed, which also confirmed a direct interaction (data not shown).
Strumpellin is an evolutionarily highly conserved protein containing ‘spectrin-like’ repeats
We performed a detailed bioinformatic analysis of the strumpellin gene with special emphasis on its structural and evolutionary conservation. Sequence evidence was obtained for a single strumpellin gene in all phyla studied, except for its apparent absence from bacteria and fungi, partial absence from certain protozoa (e.g. Giardia) and some plants (e.g. Arabidopsis), and a gene duplication in Paramecium (Supplementary Fig. 1). The human gene structure with 29 exons was preserved in all vertebrate genes examined (Fig. 3A). Three known sequence variations cause hereditary spastic paraplegia phenotypes (N471D, L619F, V626F). A splicing anomaly was observed in 17 out of 500 (3%) human expressed sequence tags, which lacked the first coding exon 2. Such a transcript, if expressed in vivo, would potentially encode a truncated protein lacking the initial 148 amino acids preceding the next in-frame Met-149 within the highly conserved amino terminal domain. It would be expected to have a calculated molecular mass of 117 kDa. However, no such band was identified in our protein expression studies.
Blast analysis (http://blast.ncbi.nlm.nih.gov) of the translated open-reading frame confirmed that the entire strumpellin gene product represents a unique conserved protein. Strumpellin amino acid sequences from 100 representative species were aligned for phylogenetic analysis and statistical modelling as a profile hidden Markov model to characterize the rate and pattern of evolution in functionally important regions (Supplementary Fig. 2). The Pfam database (http://pfam.sanger.ac.uk/) entry for strumpellin lists only one known structural fold for this protein that is supposed to contain a ‘spectrin-like’ repeat (Yan et al., 1993). Analysis of the predicted secondary structure of human strumpellin suggests that the protein can be divided into three parts (Fig. 3B). The N-terminal part is constituted mainly of six α-helices and two short β-strand segments, and extends from residue 1 to ∼240. We have found neither structural similarity nor homology to any known domain. The central ‘spectrin-like’ repeat extends from residues 241–791 and is comprised of five ‘spectrin-like’ units, each consisting of three α-helices. The tryptophan typically present at position 17 of the first α-helix is conserved in repeats 2 and 4, and substituted by phenylalanine in repeats 1, 3 and 5. Using PSIPRED (McGuffin et al., 2000), the ‘spectrin-like’ region of strumpellin has been found to possess structural similarity to the proteins Exo70 (protein data bank accession number 2pft; P-value 7e−5; 11% identity) as well as importin beta-2 (protein data bank accession number 2z5k; P-value 9e−5; 11% identity). In the same manner, structural similarity seems to exist between the α-helical C-terminal part of strumpellin (residues 792–1158) and exportin-5 (protein data bank accession number 3a6p; P-value 8e−6; 14% identity) as well as importin β-1 (protein data bank accession number 2bpt; P-value 5e−5; 7% identity).
Strumpellin is a ubiquitously expressed dual compartment protein
Our reverse transcription polymerase chain reaction data demonstrated ubiquitous strumpellin mRNA expression in all human tissues analysed (Fig. 4A). In the analysis of various human central nervous system regions, strumpellin was present in all brain regions and in the spinal cord (Fig. 4B). Analogous experiments at the protein level demonstrated the presence of strumpellin in human brain, skeletal muscle, lung and gastrointestinal tissues with highest expression in brain and lung tissues (Fig. 4C). It is important to note that freezing and thawing of SDS samples induced the degradation of strumpellin and had a marked effect on the immunoblot signal intensity. In analogy to calpain-3 immunoblots, this may lead to difficulties in the interpretation of immunoblot data.
Highly efficient transient strumpellin expression could only be achieved in 293TN cells and to a much lesser extent in Pop10 and Cos7 cells. In contrast, virtually no ectopic expression of strumpellin was obtained in other human cell lines, such as the SK-N-DZ, SK-N-BE(2) and SH-SY5Y neuronal as well as the U373 glioblastoma cells, using various transfection methods comprising electroporation [GenePulserXcell (Bio-Rad) and Nucleofector (Amaxa/Lonza)], as well as calcium phosphate and liposome-based protocols. In 293TN and Pop10 cells strumpellin exhibited a cytoplasmic staining pattern indistinguishable from the one observed for VCP (Fig. 5A). In addition to a diffuse cytosolic distribution, co-staining of mCherry-strumpellin transfected cells with the endoplasmic reticulum-marker calnexin demonstrated that strumpellin is also present at the endoplasmic reticulum (Fig. 5B). In contrast, no co-localization with either mitochondria or the Golgi-apparatus was detected (data not shown).
To address the subcellular localization in more detail we performed cell fractionation experiments. Differential centrifugation assays using total lysates from mCherry-strumpellin transfected 293TN cells and murine skeletal muscle tissue demonstrated that strumpellin is exclusively present in the pellet fraction. The addition of Triton X-100 to the lysate, which solubilizes membranes, resulted in a very limited shift of the strumpellin signal to the supernatant (Fig. 6A). Further treatment of the lysates with a combination of Triton X-100 and saponin, which induces the solubilization of membranes including Triton X-100 insoluble cholesterol-rich lipid rafts (Chamberlain et al., 2001), resulted in a partial solubilization of strumpellin and in a complete solubilization of the lipid raft marker flotillin-1 (Fig. 6B, upper panel). The distribution of strumpellin is mirrored by that observed for VCP. When the lysates were treated with a mixture of Triton X-100, saponin and SDS, which solubilizes membranes and protein complexes, an almost complete shift of strumpellin and VCP into the soluble fraction were observed (Fig. 6B, lower panel). Thus, the obtained data strongly indicate that strumpellin and its binding protein VCP are components of a high molecular mass complex with a sedimentation coefficient larger than 60S, which corresponds to a molecular mass of at least 2.5 MDa.
Discontinuous sucrose density gradient fractionation of nucleus-free organelle-preserved lysates from mCherry-transfected 293TN cells demonstrated the presence of strumpellin in fractions 6–22 (Fig. 6C). Fractions 6–9 contain the cytosolic marker GAPDH, fractions 10–22 the endoplasmic reticulum marker calnexin, and fractions 19 and 20 the mitochondrial marker COX1. The Golgi 58K protein is present in fractions 5–7 and 13–21. Additionally, the strumpellin-rich fractions 9, 15, 16 and 19 were subjected to SDS–PAGE in conjunction with mass spectrometry and resulted in the identification of cytosolic, endoplasmic reticulum and mitochondrial markers, respectively (data not shown). The observed strumpellin distribution mirrors to a large extent that observed for VCP (Fig. 6C). Together, the transfection and biochemical results strongly indicate that both VCP and strumpellin are cytosolic and endoplasmic reticulum-associated proteins.
The N471D strumpellin mutation does not alter strumpellin’s subcellular distribution and its binding to VCP
To evaluate the effect of the previously described hereditary spastic paraplegia-causing N471D strumpellin missense mutation, we performed transfections and protein binding studies. Transfections of 293TN and Pop10 cells with mCherry–N471D-strumpellin and N471D-strumpellin–mCherry, as well as equivalent wild-type expression constructs showed neither a visible effect on the subcellular distribution of the mutant strumpellin nor its co-localization with VCP in immunofluorescence, differential centrifugation and sucrose density gradient fractionation assays. In addition, pull-down assays using recombinant wild-type and N471D-strumpellin as well as wild-type and R155 C-VCP showed no differences in their binding properties. Since VCP is a known poly-ubiquitin binding protein, we tested a putative interaction of strumpellin with ubiquitin chain mixture. Dot-blot overlay assays employing soluble wild-type strumpellin protein in combination with either K48-linked or K63-linked immobilised ubiquitin (each a mixture of monomers and dimers to heptamers) revealed no interaction of ubiquitin with strumpellin (data not shown).
Wound healing assays imply a loss-of-function of the N471D mutant strumpellin
In vitro wound healing assays after overexpression of wild-type and mCherry-N471D-strumpellin, as well as small hairpin RNA mediated strumpellin knockdown in human 293TN cells were carried out to address the functional role of wild-type and N471D mutant strumpellin. Both ablation and overexpression of wild-type strumpellin showed a marked effect on wound closure velocity. While untreated control cells displayed a mean velocity of 17.9 µm/h [standard deviation (SD) ± 5.2], strumpellin knockdown and mCherry–wild-type-strumpellin overexpressing cells showed significantly reduced mean velocities of 11.2 µm/h (SD ± 6.7; n = 18; t-test: P = 0.0025; strumpellin knockdown efficiency 80%) and 12.9 µm/h (SD ± 4.8; n = 18; t-test: P = 0.0062), respectively. In contrast, cell layers expressing mCherry–N471D-strumpellin showed velocities of 15.2 µm/h (SD ± 6.2; n = 18; t-test: P = 0.2147) which are in the range of untreated controls.
Strumpellin knockdown causes a marked reduction of axonal growth in neuroblastoma cells
With respect to a putative function of strumpellin in neurons, we performed small hairpin RNA-mediated strumpellin knockdown experiments in human SH-SY5Y neuroblastoma cells. Cells were initially transfected with the strumpellin-specific small hairpin RNA expression plasmid, followed by induction of differentiation by addition of retinoic acid. Measurements of axonal length of randomly chosen single cells demonstrated a highly significant reduction of the axonal outgrowth in strumpellin knockdown cells (strumpellin knockdown: mean axonal length 83 µm; SD ± 48, n = 100; control: mean axonal length 166 µm; SD ± 66; n = 100; t-test: P = 1.9 × 10−19). This finding was further corroborated in immunoblots, which demonstrated a significant reduction of the synaptophysin expression level in strumpellin knockdown cells (Fig. 7).
Strumpellin knockdown in zebrafish induces cardiac contractile dysfunction and loss of motoneuron formation
To elucidate the role of strumpellin in vivo we inactivated the zebrafish orthologue of strumpellin by the injection of Morpholino-modified antisense oligonucleotides directed against the splice donor site of intron 3 (MO-strumpellin) into one-cell stage zebrafish embryos. Control Morpholino-injected zebrafish embryos (n = 147; 4 ng/embryo) displayed normal morphology at 72 h postfertilization (Fig. 8A), whereas 73% of MO-strumpellin-injected zebrafish embryos (n = 159; 4 ng/embryo) (Fig. 8B) developed severe cardiac contractile dysfunction, tail curvature and impaired motility (Supplementary Movies 1 and 2). To demonstrate the effectiveness of the strumpellin knockdown, we analysed strumpellin mRNA by reverse transcription polymerase chain reaction and sequencing. MO-strumpellin led to an abnormal splice product (integration of 127 bp from intron 3) and therefore to premature termination of translation of zebrafish strumpellin (Fig. 8C).
Strumpellin morphants showed a severe pericardial oedema that serves as a hallmark of cardiac failure and blood congestion at the sinus venosus suggestive for low contractile performance. Determination of the ventricular fractional shortening between 48 and 96 h after fertilization demonstrated a progressive reduction of systolic heart function (MO-control: 57 ± 8.5% versus MO-strumpellin: 26 ± 12.3% at 96 h postfertilization) (Fig. 8D). However, strumpellin deficiency did not interfere with the development and differentiation of cardiac structures (Fig. 8E).
Strumpellin morphant zebrafish embryos also displayed severe skeletal muscle dysfunction. Tactile stimulation only resulted in a short shiver response that does not generate enough force for a directed flight response (Supplementary Movie 3). By contrast, touch-stimulation of a MO-control injected embryo led to an instant and powerful response (Supplementary Movie 4). Since early somitogenesis markers such as MyoD were normally expressed and the myopathological examination did not reveal any obvious difference in form and structure between MO-control and MO-strumpellin injected zebrafish (data not shown), we conclude that the loss of strumpellin implies either a neuronal or neuromuscular junction defect. Immunostainings of zebrafish morphants using an antibody against acetylated tubulin, a marker for growing axons, demonstrated a dramatic reduction of the signal intensity of the caudal neuronal tube (Fig. 8F). In addition, the formation of ventral and caudal motoneurons was completely abolished.
Strumpellin is a novel component of pathological protein aggregates in IBMPFD and myofibrillar myopathies
Skeletal muscle pathology in IBMPFD and myofibrillar myopathies is characterized by the presence of VCP and desmin positive cytoplasmic protein aggregates. In addition nuclear aggregates are present in IBMPFD (Hübbers et al., 2007). In cross sections of normal human skeletal muscle strumpellin showed a sarcoplasmic and sarcolemmal staining pattern (Supplementary Fig. 3). Indirect immunofluorescence of strumpellin in muscle biopsies from two patients with genetically proven VCP mutations [R155C (Hübbers et al., 2007) and R93C (Krause et al., 2007)] revealed that strumpellin is present in VCP positive cytoplasmic protein aggregates and in VCP positive nuclei (Fig. 9). Moreover, strumpellin was present in pathological protein aggregates in all 10 cases of myofibrillar myopathies due to desmin (n = 6; 1 × K240del, 2 × R350P, 1 × D339Y, 2 × E245D/D214_E245del), αB-crystallin [n = 1; 1 × G154 S (Reilich et al., 2010)], myotilin (n = 2; 2 × S55F) and FHL1 [n = 1; 1 × C224W (Schoser et al., 2009)] mutations (Fig. 9). It is noteworthy that our strumpellin antibody also labelled protein aggregates that were not VCP or desmin immunoreactive.
In order to analyse if the increased strumpellin immunostaining was also mirrored on the protein expression level, we performed strumpellin immunoblotting of total protein extracts derived from normal, IBMPFD and desminopathy skeletal muscle tissue. Remarkably, immunoblotting revealed a single band of 110 kDa in all probes with a significant decrease (approximately one-third in desminopathy and half in IBMPFD) of the total amount of strumpellin when normalized to GAPDH (data not shown).
Strumpellin is a pre-synaptic protein in the human central nervous system
We studied the expression patterns of strumpellin in the human central nervous system, i.e. the hippocampal formation and the spinal cord. In both areas, double-immunofluorescence staining revealed co-localization of strumpellin with synaptophysin (Fig. 10A–C). This co-localization was most evident in the external molecular layer of the dentate gyrus, which receives synaptic input from the perforant path. A similar finding was observed for motoneurons in the ventral horn of the human spinal cord, in which dot-like strumpellin immunoreactivity was detected at the cellular membrane and co-localized with the pre-synaptic protein synaptophysin (Fig. 10D–F).
Beyond muscle protein aggregates: strumpellin shows a partial co-localization with huntingtin in cortical neurons of a Huntington’s disease mouse model
Immunofluorescence analysis of the cortex of wild-type mice demonstrated a cytoplasmic and dendritic staining pattern of strumpellin that was present in almost all pyramidal neurons (Fig. 11A). The VCP signal appeared in the cytosol and nuclei (Fig. 11B and C). In the cortex of R6/2 Huntington’s disease transgenic mice, the dendritic staining for strumpellin seemed more pronounced while the cytoplasmic staining was attenuated. However, strumpellin was present in cortical neurons with multiple huntingtin immunoreactive foci (Fig. 11D–F). At higher magnification large pyramidal neurons were identified that displayed an inhomogeneous cytoplasmic strumpellin immunoreactivity (Fig. 11G). Here, huntingtin was found in foci within the nucleus, cytoplasm and neurophil (Fig. 11H). Merged images demonstrated a close association of strumpellin and huntingtin in this cytoplasmic aggregation pathology (Fig. 11I). This strumpellin pattern largely concurred with the one observed in YAC-128 mice and transgenic Huntington’s disease rats (data not shown).
The primary aim of our work was to identify novel and disease related VCP binding partners. Our co-immunoprecipitation experiments in human and D. discoideum cell lysates led to the identification of strumpellin. Pull-down experiments using recombinant proteins proved a direct interaction between VCP and strumpellin. The identification of strumpellin is of particular interest, since mutant strumpellin variants are known to cause a severe and relatively pure motor form of hereditary spastic paraplegia (SPG8 or Strümpell-Lorrain-Disease, autosomal-dominant inheritance, OMIM #603563) (Behan and Maia, 1974; Valdmanis et al., 2007; Salinas et al., 2008;). This progressive neurodegenerative disorder is clinically characterized by central motor system deficits leading to spastic paraparesis of the lower limbs (Fink, 2003).
The human strumpellin gene (KIAA0196, NC_000008.9), located on chromosome 8p24.13, is a single copy gene consisting of 29 exons encoding an 1159 amino acid protein (Q12768-1, NP_055661.3). Our in silico gene and protein analyses revealed that the strumpellin gene structure is preserved in all vertebrate genomes and that the gene product represents a unique conserved protein. The central ‘spectrin-like’ region as well as the C-terminal part of strumpellin possesses structural similarity to Exo70 and importin beta-2, as well as exportin-5 and importin beta-1, respectively. In addition, distinct regions of the strumpellin protein exhibit an unusually high degree of conservation implying an association with functional significance. Strumpellin has recently been identified as a component of the Wiskott Aldrich Syndrome protein and scar homologue complex involved in actin nucleation and endosome fission (Derivery et al., 2009). The conserved domains of the strumpellin protein are likely to participate in molecular interactions with one or several members of the Wiskott Aldrich Syndrome protein and scar homologue complex.
In accordance with previously published northern blot data (Nagase et al., 1996) our reverse transcription polymerase chain reaction experiments demonstrated ubiquitous strumpellin mRNA expression in human tissues. Using a newly generated strumpellin polyclonal antibody, we demonstrated the ubiquitous expression of strumpellin. The antibody detected a single band corresponding to an apparent molecular mass of 110 kDa. The subcellular distribution of strumpellin mirrors to a large extent that reported for VCP (Hirabayashi et al., 2001; Kobayashi et al., 2002; Partridge et al., 2003; Hübbers et al., 2007; Lim et al., 2009). Our transfection and biochemical analyses showed that strumpellin is primarily present in cytosolic and endoplasmic reticulum membrane fractions.
In addition to its association with hereditary spastic paraplegia, previous studies indicated that strumpellin has a pathogenic role in prostate cancer. mRNA expression studies demonstrated an upregulation of strumpellin in PC-3 and other prostate cancer cell lines, as well as in prostate carcinoma tissue specimens. This upregulation of strumpellin and other gene products is due to an increased copy number of the chromosomal region 8q associated with prostate cancer progression and poor prognosis (Porkka et al., 2004; van Duin et al., 2005). Overexpression as well as knockdown of strumpellin in prostate cancer cells affected neither cell proliferation nor invasion (Jalava et al., 2009).
In our in vitro wound healing assays the knockdown, as well as the overexpression of wild-type strumpellin, caused significant reductions in wound closure velocities. So far three heterozygous disease causing missense mutations have been described [protein: Asn471Asp (N471D), p.Leu619Phe (L619F), p.Val626Phe (V626F); cDNA NM_014846.3: c.1411A>G, c.1857G>C, c.1876G>T] (Valdmanis et al., 2007). The functional properties of the N471D strumpellin mutation residing in the proximity of the spectrin repeat domain have not yet been studied. This mutant showed no effect in our wound healing assays when compared to non-transfected cells. Overexpression of L619F and V626F mutant strumpellin in zebrafish similarly showed no pathology (Valdmanis et al., 2007). This study reported an enlarged heart cavity, a curly-tail phenotype and shorter and abnormally branched axons of spinal cord motoneurons and interneurons in strumpellin knockdown zebrafish (Valdmanis et al., 2007). To specifically analyse the role of strumpellin in neuronal cells, we employed axonal outgrowth experiments in a human neuroblastoma cell line. Here, knockdown of strumpellin resulted in a dramatic decrease of axon formation associated with a marked decrease in synaptophysin expression. Our zebrafish knockdown studies mirror the phenotype reported by Valdmanis et al. (2007). According to our more detailed phenotypic analysis, the observed pericardial oedema is not due to abnormal development of cardiac structures, but has to be attributed to a reduction of systolic heart function. Analysis of the severe motor deficits showed normal somitogenesis and muscle structures, whereas the formation of ventral and caudal motoneurons was severely compromised.
The corroborative data strongly indicate that disease associated strumpellin mutations do not inflict a dominant negative effect, but are associated with a loss-of-function, i.e. haploinsufficiency, of the strumpellin protein. Thus the expression level of wild-type strumpellin seems to be of critical importance in the pathogenesis of strumpellin associated hereditary spastic paraplegia. The pathogenesis of the clinically and genetically extremely heterogeneous group of hereditary spastic paraplegia variants is very complex. Identified gene defects have been related to a variety of disturbed cell functions, e.g. membrane trafficking, axonal transport, mitochondrial function, endoplasmic reticulum morphology and myelination (Soderblom and Blackstone, 2006; Salinas et al., 2008). The presence of strumpellin in the Wiskott Aldrich Syndrome protein and scar homologue complex at endosomes (Derivery et al., 2009) in conjunction with its observed presence in endoplasmic reticulum fractions makes it tempting to speculate that the pathophysiology of strumpellin-related hereditary spastic paraplegia is primarily associated with defective actin dynamics in membrane organization and trafficking.
VCP is a known component of pathological protein aggregates in a wide variety of neurodegenerative disorders and myofibrillar myopathies (Hirabayashi et al., 2001; Mizuno et al., 2003; Nan et al., 2005; Greenberg et al., 2007; Hübbers et al., 2007; Schröder and Schoser, 2009). We demonstrate that strumpellin is present in pathological aggregates in IBMPFD and myofibrillar myopathies due to desmin, myotilin, αB-crystallin and FHL1 mutations. In the human central nervous system strumpellin is localized in pre-synaptic, synaptophysin positive structures. In a transgenic Huntington’s disease mouse model we further demonstrated the presence of strumpellin in cortical neurons where strumpellin and huntingtin foci were closely associated.
Beyond its role in hereditary spastic paraplegia, our findings in IBMPFD, myofibrillar myopathies and transgenic huntingtin mice indicate that VCP–strumpellin interactions may play a pivotal role in the molecular pathogenesis of human protein aggregate diseases, thereby interlinking motoneuron diseases, frontotemporal dementias and protein aggregate myopathies.
The German Research Foundation (DFG: SCHR 562/7-1, CL 381/1-1; DFG/FOR1228: SCHR 562/9-1, CL 381/3-1, EI 399/5-1); Fritz-Thyssen-Foundation (grant 10.07.1.165 to R.S. and C.S.C.); German ministry of education and research, (BMBF) (via the German network on muscular dystrophies (MD-NET2) to R.S. and C.S.C.); Spanish Ministry of Science and Innovation (MICINN) (grant BFU2007-67876 to R.O.M. and M.-P.F.).
Supplementary material is available at Brain online.
R.S. and C.S.C. are members of the German network on muscular dystrophies (MD-NET).
glyceraldehyde 3-phosphate dehydrogenase
inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia
sodium dodecyl sulphate polyacrylamide gel electrophoresis
- protein binding
- osteitis deformans
- central nervous system
- frontotemporal dementia
- huntington's disease
- cell motility
- endoplasmic reticulum
- fluorescent antibody technique
- motor neurons
- muscle contraction
- skeletal muscles
- spastic paraplegia, hereditary
- mass spectrometry
- wound healing
- diagnostic imaging
- wound closure
- protein overexpression
- neuroblastoma cell
- inclusion body myopathy
- protein aggregation, pathological
- protein aggregates