VAPB/ALS8 interacts with FFAT-like proteins including the p97 cofactor FAF1 and the ASNA1 ATPase
© Baron et al.; licensee BioMed Central Ltd. 2014
Received: 6 February 2014
Accepted: 12 May 2014
Published: 29 May 2014
Skip to main content
© Baron et al.; licensee BioMed Central Ltd. 2014
Received: 6 February 2014
Accepted: 12 May 2014
Published: 29 May 2014
FAF1 is a ubiquitin-binding adaptor for the p97 ATPase and belongs to the UBA-UBX family of p97 cofactors. p97 converts the energy derived from ATP hydrolysis into conformational changes of the p97 hexamer, which allows the dissociation of its targets from cellular structures or from larger protein complexes to facilitate their ubiquitin-dependent degradation. VAPB and the related protein VAPA form homo- and heterodimers that are anchored in the endoplasmic reticulum membrane and can interact with protein partners carrying a FFAT motif. Mutations in either VAPB or p97 can cause amyotrophic lateral sclerosis, a neurodegenerative disorder that affects upper and lower motor neurons.
We show that FAF1 contains a non-canonical FFAT motif that allows it to interact directly with the MSP domain of VAPB and, thereby, to mediate VAPB interaction with p97. This finding establishes a link between two proteins that can cause amyotrophic lateral sclerosis when mutated, VAPB/ALS8 and p97/ALS14. Subsequently, we identified a similar FFAT-like motif in the ASNA1 subunit of the transmembrane-domain recognition complex (TRC), which in turn mediates ASNA1 interaction with the MSP domain of VAPB.
Proteasome inhibition leads to the accumulation of ubiquitinated species in VAPB immunoprecipitates and this correlates with an increase in FAF1 and p97 binding. We found that VAPB interaction with ubiquitinated proteins is strongly reduced in cells treated with FAF1 siRNA. Our efforts to determine the identity of the ubiquitinated targets common to VAPB and FAF1 led to the identification of RPN2, a subunit of an oligosaccharyl-transferase located at the endoplasmic reticulum, which may be regulated by ubiquitin-mediated degradation.
The FFAT-like motifs we identified in FAF1 and ASNA1 demonstrate that sequences containing a single phenylalanine residue with the consensus (D/E)(D/E)FEDAx(D/E) are also proficient to mediate interaction with VAPB.
Our findings indicate that the repertoire of VAPB interactors is more diverse than previously anticipated and link VAPB to the function of ATPase complexes such as p97/FAF1 and ASNA1/TRC.
FAF1 (also known as UBXN3A or UBXD12) is a p97 cofactor from the UBX-domain family. Humans express 13 UBX-domain proteins, most of which interact with the p97 ATPase via their C-terminal UBX-domain [1, 2]. Exceptionally, UBXD1 does not interact with the p97 N-terminus via its UBX domain, but uses its PUB domain to bind the C-terminus of p97 [3, 4]. Five of these cofactors, including FAF1, have been termed UBA-UBX proteins due to the presence of a ubiquitin-associated (UBA) domain at their N-terminus, which mediates interaction with ubiquitinated proteins. The general function of UBA-UBX proteins is that of ubiquitin-binding adaptors for the p97 ATPase [1, 5–8]. Furthermore, three UBA-UBX proteins – UBXD7 (or UBXN7), UBXD8 (also known as UBXN3B, FAF2 or ETEA) and FAF1 – all have a central UAS domain, which was recently shown to interact with long-chain unsaturated fatty acids thereby mediating UBA-UBX protein oligomerization . However, the various UBA-UBX proteins are not functionally redundant, due, at least in part, to the presence of specific domains that are found in a single UBA-UBX protein. For example, the UIM domain, which is only present in UBXD7, allows this protein to interact specifically with the NEDD8 modification on cullins [10, 11]. Owing to their ATPase activity, p97 complexes function as ‘segregases’, which can dissociate their targets from protein partners [12, 13] or even retrotranslocate proteins from the endoplasmic reticulum (ER) back into the cytosol to allow for their ubiquitination and degradation [14, 15]. This latter function is fundamental for p97’s role in ER-associated degradation (ERAD), a quality control pathway that ensures degradation of ER proteins that are misfolded or misassembled .
Vesicle-associated proteins (VAP) are highly conserved across eukaryotes, from yeast to mammals [17–23]. Humans express two VAP proteins, VAPA and VAPB, whose primary sequence is 63% identical. Both are membrane-anchored at the ER and Golgi via C-terminal transmembrane domains [20, 22] and can exist as homo- or heterodimers . Oligomerization is largely mediated by a cytoplasmic coiled-coil region, with some contribution from the transmembrane domain [24, 25]. VAP proteins contain an MSP domain in their N-terminal half, which has been shown to interact with proteins containing a FFAT motif (two phenylalanines in an acidic tract motif), such as members of the oxysterol-binding protein (OSBP) family or the phosphatidylinositol transfer proteins from the PITPNM family . The consensus sequence for the FFAT motif is EFFDAxE, with an acidic tract immediately upstream of it . The crystal structure of the MSP domain bound to a FFAT peptide suggests that a VAP dimer may bind two FFAT motifs, with the FFATs engaging each other and both MSP domains of the dimer .
A Phe56 to Ser mutation in the MSP domain of VAPB causes amyotrophic lateral sclerosis (ALS) type 8 in humans, transmitted in an autosomally dominant manner . Subsequently, another VAPB mutation in the same region (Thr46 to Ile) was identified in a British ALS patient . ALS is a progressive neurodegenerative disorder, which affects upper and lower motor neurons and is lethal within 5 years of clinical onset . Interestingly, in some cases mutations in p97 can also cause ALS [32, 33].
Our mass spectrometry analysis of Flag-FAF1 immunoprecipitates led to the identification of VAPA and VAPB as potential interaction partners for FAF1. Given the implication of both VAPB and p97 in ALS, we decided to follow up on this interaction, aiming to understand whether FAF1 might link VAPB and p97 in a common functional pathway. We show that VAPB interaction with FAF1 is not mediated by ubiquitin-modification of VAPB, but it is due to the presence of a FFAT-like motif in FAF1. Further work identified a very similar FFAT-like motif in ASNA1, which we demonstrate is another novel binding partner for VAPB. Although VAPB itself is not a ubiquitinated target for FAF1/p97, we found that VAPB does interact with ubiquitinated proteins in a FAF1-dependent manner and, subsequently, identified RPN2 as a common ubiquitinated target for VAPB and FAF1.
VAPA and VAPB were identified by mass spectrometry in Flag-FAF1 immunoprecipitates
Share of spectrum IDsa
Because FAF1 is a ubiquitin-binding adaptor for the p97 ATPase , we initially assumed that VAPB might be targeted for proteasomal degradation and interact with FAF1 in its ubiquitinated form. However, we found that proteasome inhibition with MG132 for 2 or 6 hr had no effect on VAPB levels (Figure 1G, left panel), nor could we detect ubiquitinated forms of VAPB even after a long exposure of Flag-VAPB immunoprecipitates (Additional file 2: Figure S1). Moreover, VAPB interaction with Flag-FAF1 was not affected by MG132 treatment although we could detect an increased binding of ubiquitinated proteins to FAF1 (Figure 1G, right panel). These data suggested that VAPB is not targeted for ubiquitin-mediated proteasomal degradation. Hence, VAPB interaction with FAF1 and p97 appears to serve a function other than facilitating VAPB degradation.
Because the MSP domain of VAPB is known to interact with various proteins carrying a FFAT motif , we searched the FAF1 sequence for the presence of such a motif. There was no typical FFAT motif in FAF1. Yet in the region between the two UBLs and the UAS domain, we identified a short sequence that highly resembled FFAT motifs, the main difference being that in FAF1 the second phenylalanine residue is replaced by an additional acidic residue (Figure 2C). We therefore tested whether deletion of the FFAT-like region in FAF1 has any effect on VAPB binding. Indeed, the removal of only nine amino acids from residue 293 to 301 resulted in a complete loss of VAPB binding, while p97 and ubiquitinated protein binding remained unaffected (Figure 2B, lane 10). To exclude the possibility that this short truncation might affect FAF1 folding, we went on to create a triple mutant (DFE to AAA) and three individual point mutants in that region – D295A, F296A and E297A. Both the triple mutant and the F296A mutation abolished VAPB binding, while the other two point mutants retained the wild-type ability to interact with VAPB (Figure 2D). We therefore conclude that FAF1 interacts with VAPB via a highly conserved FFAT-like motif (Figure 2C), with the single phenylalanine residue present in this region being particularly important for the interaction.
The P56S mutation in VAPB that was identified in ALS patients causes the protein to aggregate and to become insoluble when expressed in mammalian cells in culture . Due to the difficulty of extracting VAPB P56S from human cells, it was not possible for us to study how this mutation might affect FAF1 binding in human cells. However, we found that both wild-type and VAPB P56S can be readily expressed in bacteria. Recombinant VAPB P56S retained the wild-type ability to interact with Flag-FAF1 in vitro (Figure 3B). In contrast, a VAPB double mutant known to be defective in FFAT binding, K87D M89D , could not be co-immunoprecipitated with FAF1 under the same conditions. These results indicate that the P56S mutation as such does not perturb VAPB interaction with FAF1.
TRC complex subunits slightly accumulate in endogenous VAPB immunoprecipitates upon proteasome inhibition
SILAC ratio L/Ha
L + MG 2 hr
H + MG 2 hr
L + MG 6 hr
H + MG 6 hr
1.31 ± 0.08 (31)
0.79 ± 0.07 (27)
1.42 ± 0.10 (32)
0.87 ± 0.06 (35)
1.47 ± 0.18 (36)
0.77 ± 0.11 (36)
1.43 ± 0.19 (52)
0.94 ± 0.10 (43)
1.39 ± 0.11 (8)
0.73 ± 0.06 (11)
1.30 ± 0.08 (9)
0.84 ± 0.03 (13)
1.50 ± 0.10 (7)
0.74 ± 0.04 (9)
1.51 ± 0.15 (7)
0.89 ± 0.04 (10)
TRC subunits were identified by mass spectrometry in Flag-FAF1 immunoprecipitates
Share of spectrum IDsa
MG 2 hr
MG 6 hr
Because the TRC complex has been implicated in ubiquitin-dependent degradation of mislocalized ER membrane proteins , we also checked if ASNA1 might contribute to the binding of ubiquitinated proteins to VAPB. As shown in Figure 4F, ASNA1 depletion by siRNA leads to a reduction in Flag-VAPB interaction with ubiquitinated proteins. We conclude that part of the ubiquitinated targets that interact with VAPB could be TRC quality control targets that are recruited via ASNA1.
Next we wanted to verify whether ASNA1 also interacts with FAF1 as suggested by the mass spectrometry analysis (Table 3). Unlike ASNA1 interaction with VAPB, its interaction with Flag-FAF1 was weak and hard to detect in the absence of proteasome inhibition (Figure 5E, lanes 2 to 7). Upon proteasome inhibition, ASNA1 interaction with Flag-FAF1 became readily detectable (Figure 5E, lanes 9 to 14) and it required the UBA domain of FAF1 (Figure 5E, lane 10). Conversely, Flag-ASNA1 interaction with endogenous FAF1 was only detectable upon proteasome inhibition, which also caused a considerable increase in the binding of ubiquitinated proteins to ASNA1 (Figure 5F, compare lanes 2 and 7). Interestingly, ASNA1 interaction with FAF1 did not depend on the ASNA1 FFAT-motif (Figure 5F, lane 10), but was drastically reduced by a Gly46 point mutation in ASNA1 to either Ala or Arg (Figure 5F, lanes 8 and 9) . We found that these two mutants are completely defective in interacting with the BAG6 subunit of the TRC complex and exhibit reduced binding to ubiquitinated proteins (Figure 5F). Taken together, these findings indicate a correlation between the binding of ubiquitinated proteins to ASNA1 and a detectable interaction with FAF1. They suggest that FAF1 binding to ASNA1 is likely indirect and mediated by the ubiquitinated proteins that associate with ASNA1.
Ubiquitinated targets of VAPB and FAF1 identified by mass spectrometry upon enrichment of ubiquitinated peptides
GlyGly peptidesa(times identified)b
LSK 154 EETVLATVQALQTASHLSQQADLR (4×)
K 244 NFESLSEAFSVASAAAVLSHNR (2×)
LHNQK 442 TGQEVVFVAEPDNK (3×)
APTIVGK 189 SSLNPILFR (2×)
RPN2 and DDOST slightly accumulate in endogenous VAPB immunoprecipitates upon proteasome inhibition
SILAC ratio L/Ha
L + MG 2 hr
H + MG 2 hr
L + MG 6 hr
H + MG 6 hr
1.11 ± 0.02 (4)
0.88 ± 0.00 (2)
1.30 ± 0.10 (4)
0.96 ± 0.02 (6)
1.12 ± 0.00 (2)
0.64 ± 0.00 (1)
RPN2 and DDOST were identified at low levels in Flag-FAF1 immunoprecipitates
Share of spectrum IDsa
MG 2 hr
MG 6 hr
FFAT-like proteins identified in VAPA and VAPB immunoprecipitates by mass spectrometry
Share of spectrum IDsb
Our analysis of VAPB immunoprecipitates from mouse brain also revealed VAPB interaction with syntaxin 1A and B (STX1A and B) (Table 7, Additional file 9: Table S8, Additional file 10: Figure S2A). These proteins harbor a sequence that resembles a FFAT motif (Additional file 10: Figure S2B), but mutating the two phenylalanine residues in either STX1A or STX1B to alanine had no effect on their interaction with VAPB (Additional file 10: Figure S2C). Hence, these are not functional FFAT motifs and are not responsible for syntaxin 1A/B interaction with VAPB.
Humans express five ubiquitin-binding UBX proteins, including FAF1 and UBXD7. Previous work showed that UBXD7 is unique among p97 cofactors due to the presence of an UIM motif that allows it to interact directly with neddylated cullins [10, 11]. Here we show that FAF1 contains an atypical FFAT motif that allows it to associate with the membrane-anchored proteins VAPA and VAPB. Although distinct from the consensus proposed for FFAT motifs  due to the absence of a second aromatic residue, the FFAT motif of FAF1 is still able to mediate the interaction with the MSP domain of VAPB.
The UIM motif of UBXD7 and the FFAT motif of FAF1 represent features unique to each of these UBX proteins and mediate their interaction with specific targets, neddylated cullins and VAPA/B respectively. The identification of these short protein-interaction motifs supports the notion that the general function of UBA-UBX proteins as ubiquitin-binding adaptors for p97 is complemented with specific functions mediated by singular motifs.
The identification of a degenerated, yet functional FFAT motif in FAF1, indicated that other proteins carrying FFAT-like motifs could represent bona fide interaction partners for VAPB. On one hand, non-canonical FFAT motifs may be present in known VAPB-binding partners and on the other hand, searching for FFAT motifs that only partially match the consensus sequence could lead to the identification of novel VAPB-binding partners.
We took the first approach and validated the FFAT-like motifs in ASNA1, RAB3GAP1 and WDR44, proteins that we identified in VAPB immunoprecipitates by mass spectrometry. Mikitova and Levine took the latter approach and computationally identified multiple high-confidence FFAT-like proteins, based on the degree of variation that is tolerated by FFAT motifs in a yeast reporter assay . RAB3GAP1 scored high as a potential binding partner for VAPB in this study, but the more divergent FFAT-like motifs in FAF1 and ASNA1 were not recognized.
FFAT-like motifs of WDR44 and RAB3GAP1 have been picked out by prior computational searches [27, 28, 41] and we confirm that they are functional and truly mediate VAPB binding. Our experiments with syntaxin 1A and B, whose FFAT-like sequences are not required for the interaction with VAPB, further highlight the importance of experimentally testing putative FFAT motifs.
Several observations led us to believe that VAPB itself is not a proteasome substrate: (1) VAPB does not accumulate upon proteasome inhibition with MG132 (Figure 4B); (2) anti-Flag Western blots of Flag-VAPB immunoprecipitates did not reveal any slower migrating, ubiquitinated forms of VAPB, even after proteasome inhibition (Additional file 2: Figure S1); (3) VAPB interaction with FAF1 is not mediated by the ubiquitin-binding UBA domain of FAF1 (Figure 2B), and (4) no ubiquitinated peptide was identified for VAPB in our mass spectrometry analysis of di-glycine peptides from Flag-VAPB/Flag-FAF1 immunoprecipitates.
Kanekura et al. previously observed ubiquitinated species in over-expressed VAPB immunoprecipitates and assumed them to be ubiquitinated VAPB . Instead, we present evidence for VAPB interaction with ubiquitinated proteins: (1) VAPB interacts with ubiquitinated proteins as shown by the anti-ubiquitin Western blotting of Flag-VAPB and endogenous VAPB immunoprecipitates; (2) VAPB interacts directly with a ubiquitin-binding protein, FAF1, and (3) VAPB interaction with ubiquitin is reduced when FAF1 is depleted by treatment with siRNA (Figure 4).One possibility is that the direct binding of FAF1 to VAPB mediates VAPB interaction with ubiquitinated species that are recognized by the UBA domain of FAF1. Alternatively, the proteins that are modified with ubiquitin interact directly with VAPB and FAF1, which is required to protect the ubiquitin chains from deubiquitinating enzymes or from degradation via alternative ubiquitin receptors. This latter possibility is strongly supported by the observation that VAPB variants that cannot bind FFAT proteins interact with ubiquitinated species carrying shorter ubiquitin chains than observed for wild-type VAPB (Figure 4E). Furthermore, RPN2, a novel ubiquitinated target we identified for VAPB and FAF1, interacts more robustly with VAPB than with FAF1 (Figure 6A), suggesting that the interaction is unlikely to be mediated by FAF1.
ASNA1, on the other hand, mediates VAPB interaction with TRC, which is required for membrane insertion of some tail-anchored proteins  and also for the elimination of mislocalized membrane proteins . Interestingly, although VAPB is a tail-anchored protein, it does not interact with the TRC via its C-terminal transmembrane region. We show that ASNA1 interacts only with the MSP domain of VAPB and not with the C-terminal region, which includes the transmembrane domain (Figure 4E). CAML and WRB were identified to be the receptors that recruit the TRC to the ER membrane for insertion of tail-anchored proteins [42, 43]. We show that ASNA1 transiently interacts with ubiquitinated proteins that are targeted for proteasomal degradation and appear to recruit FAF1 (Figure 5F). Furthermore, a fraction of the ubiquitinated proteins that interact with VAPB are recruited via ASNA1 (Figure 4F). It is possible that VAPB may represent an alternative receptor for TRCs whose ubiquitinated targets require p97/FAF1 for degradation (Figure 8C).
Starting from the observation that the p97 cofactor FAF1 interacts with the ER membrane proteins VAPA and VAPB, we identified an atypical FFAT sequence in FAF1 that allows its direct interaction with the MSP domain of VAPA/B. A similar FFAT-like sequence was subsequently recognized in the ASNA1 subunit of the TRC, further expanding the repertoire of proteins that interact with VAPB in this manner. The identification of two subunits of an ER-resident glycosyltransferase – RPN2 and DDOST – as common ubiquitinated targets for VAPB and FAF1, suggests that VAPB might recruit p97/FAF1 to facilitate degradation of a subset of targets located at the ER.
Thus, VAPB dimers emerge as a platform for docking various FFAT and FFAT-like proteins at the ER membrane. These include proteins involved in lipid metabolism as previously noted, but also, as shown in this study, the ubiquitin receptor FAF1, the ASNA1 component of the TRC that mediates/monitors protein insertion in the ER membrane and regulatory enzymes such as RAB3GAP1/2. Perturbations in VAPB interaction with any of these proteins when VAPB aggregates due to the ALS-causing P56S mutation could be responsible for the defects that lead to ALS8 disease.
Human FAF1 [GenBank:NM_007051.2] was amplified from EST IMAGE 5928559. For mammalian expression, wild-type and mutant FAF1 variants (lacking the native methionine) were subcloned as Sal1/Not1 inserts into pCMV5-Flag. For bacterial expression, wild-type FAF1 was subcloned into a modified pGEX6P-1 vector containing a TEV protease site and a Flag-tag downstream of the GST. Human VAPB [GenBank:NM_004738.3] was amplified from EST IMAGE 3543354. For mammalian expression, wild-type, mutant and truncated variants were subcloned as BamH1/Not1 inserts into either pCMV5-Flag or a modified version of pcDNA5-FRT/TO containing an N-terminal Flag-tag. For bacterial expression, ORFs were subcloned as BamH1/Not1 inserts into a modified pGEX6P-1 vector containing a TEV protease site, before removing extraneous linker residues between the glutamine of the TEV and the native methionine of VAPB by site-directed mutagenesis. Human VAPA [GenBank:NM_003574] was amplified from EST IMAGE 4662097 and subcloned as a BamH1/Not1 insert into pCMV5-Flag for mammalian expression. Human RAB3GAP1 [GenBank:NM_001172435.1] was amplified from EST IMAGE 5276867. For mammalian expression, a silent mutation (t909c) was first introduced to disrupt an internal BamH1 site before subcloning as BamH1/Not1 inserts into pcDNA5-FRT/TO-Flag. Human WDR44 [GenBank:NM_019045.4] was amplified by RT-PCR from embryonic umbilical RNA (Agilent Technologies, Wokingham, UK). Human ASNA1 [GenBank:NM_004317.2] was amplified by RT-PCR from total uterus RNA (Agilent Technologies). Wild-type and mutant variants were subcloned as BamH1/Not1 inserts into pCMV5-Flag for mammalian expression.
PCR reactions were carried out using KOD Hot Start DNA Polymerase (Merck Millipore, Darmstadt, Germany). Reverse transcriptase reactions were carried out using Superscript III (Life Technologies). All full-length products were cloned into pSc-B or pSc-A (Agilent Technologies, Carlsbad, USA) and fully sequenced prior to further subcloning or manipulation. All mutations and deletions were made following the Quickchange method (Agilent Technologies), but using KOD Hot Start DNA Polymerase. DNA sequencing was performed by the Sequencing Service at the College of Life Sciences, University of Dundee . For plasmid requests, please contact Medical Research Council Protein Phosphorylation and Ubiquitylation Unit (MRC-PPU) Reagents .
U2OS and HeLa cells were respectively cultured in McCoy’s and MEM media (Gibco, Pailsey, UK) supplemented with 10% fetal bovine serum (FBS) (Thermo Scientific, Fremont, USA), 2 mM L-glutamine (Invitrogen, Camarillo, USA), 100 units/ml penicillin, 100 μg/ml streptomycin (Invitrogen). SH-SY5Y cells were cultured in DMEM/F12 (Gibco) supplemented with 10% FBS, 1% NEAA (Gibco), 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin. Then 10 μM MG132 (Enzo, Farmingdale, USA) was added to the media for the indicated time before harvesting the cells. Tunicamycin (Sigma, Saint Louis, USA) was used at 5 μg/ml for 5 hr.
Flag-VAPB and Flag-FAF1 U2OS cells were generated from Flp-In T-REx U2OS cells, following Invitrogen’s instructions and were maintained in McCoy’s media supplemented with 10% FBS, 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin, 100 μg/ml hygromycin, 15 μg/ml blasticidin (Invitrogen). The expression of Flag-VAPB and Flag-FAF1 was respectively induced by adding 100 ng/ml and 200 ng/ml tetracycline (Bioline, London, UK) for 24 hr.
For SILAC mass spectrometry, low passage U2OS cells were labelled using DMEM high-glucose media without arginine, lysine or methionine (Biosera, Boussens, France) reconstituted in filter-sterilized water supplemented with 3.7 g/l NaHCO3 and 10% dialyzed FBS (Biosera). The media was supplemented with 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin, 30 μg/ml methionine and respectively 146 μg/ml and 84 μg/ml of either lysine K0 and arginine R0 (Sigma) for light (K0R0) labelling or lysine K8 and arginine R10 (Cambridge Isotope Lab, Tewksbury, USA) for heavy (K8R10) labelling. After two weeks of culture in SILAC DMEM, the labelling efficiency was checked by mass spectrometry of random peptides.
Plasmid transfections were performed using TransIT-LT1 (Mirus, Madison, USA), following the manufacturer’s instructions. Briefly, approximately 3 × 106 cells were cultured on a 15-cm plate. Then 24 hr later, the media was refreshed without antibiotics and the cells were transfected with 10 μg plasmid for 24 hr. siRNA transfections were performed for 48 hr, using oligos from Thermo Scientific at a final concentration of 5 nM and Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions. The following siRNA oligos were used: for FAF1 #0 (5′-CCACCUUCAUCAUCUAGUC-3′)  and siGENOME #1 (D-009106-01), #3 (D-009106-03), #4 (D-009106-04); for p97 (D-008727-06); for ASNA1 #3 (D-009666-03) and #18 (D-009666-18), and the Luciferase Duplex (P-002099) as non-targeting control.
For immunoprecipitation, the cells were lysed in buffer A (50 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES)/KOH, pH 7.2; 5 mM Mg(OAc)2; 70 mM KOAc; 0.2% Triton X-100; 10% glycerol; 0.2 mM ethylenediaminetetraacetic acid (EDTA); complete protease inhibitor cocktail (Roche, Mannheim, Germany). The lysates were incubated with anti-Flag beads (Anti-Flag M2 Affinity Gel, A2220, Sigma) or anti-VAPB (R2986, Division of Signal Transduction Therapy Unit (DSTT) at the University of Dundee) antibodies cross-linked to Protein A-Sepharose Fast Flow beads (PAS beads, GE Healthcare Life Sciences, Little Chalfont, UK). Uncoupled PAS beads were used as a control. Endogenous FAF1 immunoprecipitation was performed using anti-FAF1 antibodies (S370D, DSTT) and sheep IgG, without prior cross-linking to PAS beads. Immunoprecipitation was performed at 4°C for 2 hr with rotation. The beads were washed with buffer A and the proteins were eluted by incubating with 3× Laemmli buffer for 10 min in Micro Bio-Spin columns (BioRad, Hemel Hempstead, UK). Mouse brain tissue was homogenized in buffer A using a Polytron (Kinematica, Eschbach, Germany). The lysates were centrifuged for 10 min at 13,000 rpm, filtered using a 0.22 μm sieve and centrifuged again for 10 min at 13,000 rpm. The supernatants were pre-cleared by incubation with uncoupled PAS beads, three times for 20 min each, under rotation. All steps were done at 4°C. For mass spectroscopy, the immunoprecipitation was performed as described with the following changes: after washing with buffer A, the beads were additionally washed with 100 mM Tris-HCl pH 8.5 and the bound proteins were eluted by incubating with 8 M urea in 100 mM Tris-HCl pH 8.5 for 15 min at 37°C. The eluted proteins were reduced by incubation with 3 mM TCEP for 20 min and then with 11 mM iodoacetamide for 15 min, both at room temperature. Samples were pre-digested with 0.1 μg Lys-C (Roche) for 4 hr at 37°C, then diluted and the buffer adjusted so the final digestion buffer contained 2 M urea, 1 mM CaCl2 and 0.5 μg Trypsin (Roche). Trypsin digestion was performed at 37°C for 16 hr. The digested peptides were acidified to pH < 3 using 1% TFA, purified using C18 Silica microspin columns (The Nest Group, Southborough, USA), eluted in 0.1% TFA, 50% ACN and stored dry at -80°C until mass spectroscopy.
For SILAC immunoprecipitation, light- or heavy-labelled cells were treated with either MG132 or dimethyl sulfoxide as a control. The cells were independently lysed in buffer A and equal amounts of heavy and light lysate were mixed before incubation with the appropriate beads. The immunoprecipitation was performed as described above. For lysine-ϵ-GlyGly peptide purification, light-labelled cells were transfected with Flag-VAPB and heavy-labelled cells with Flag-FAF1, both treated with MG132 for 2 hr. The cells were independently lysed in buffer A, Flag-immunoprecipitated and trypsin digested as described above. The digested peptides were pooled together and the lysine-ϵ-GlyGly peptides were immunoaffinity purified using the PTMScan ubiquitin remnant motif (K-ϵ-GG) kit (Cell Signaling, Leiden, The Netherlands; 5562) as described by Udeshi et al..
The cells were plated on cover glass (thickness No 1.5) for 24 hr, followed by Flag-ASNA1 plasmid transfection or Flag-FAF1 tetracycline induction for 24 hr, and treated with MG132 for 2 hr where indicated. Cells were then fixed in 3.7% paraformaldehyde/PBS at room temperature for 15 min, permeabilized in 0.2% Triton X-100/PBS for 5 min, and blocked in 3% FBS in PBS/0.05% Tween for 45 min. Antibodies diluted in 3% FBS in PBS/0.05% Tween were sequentially added to the cells and incubated for 1 hr at room temperature followed by three washes with PBS/0.05% Tween. Cells were then incubated for 5 min with 4',6-diamidino-2-phenylindole (DAPI). Finally, the samples were washed three times with PBS/0.05% Tween, three times with PBS, and twice with water before mounting on microscope slides with Mowiol 4-88 (Polysciences, Eppelheim, Germany). Images were obtained with a DeltaVision Spectris microscope (Applied Precision, Issaquah, USA), using a CoolSNAP HQ camera (Roper Scientific, Martinsried, Germany) and a 100× 1.4 NA objective (Olympus, Southend-on-Sea, UK). The SoftWorx software (Applied Precision) was used for image acquisition and deconvolution.
GST-fusion proteins were expressed in BL21 cells at 15°C for 16 hr upon induction with 0.1 mM isopropyl β-D-1-thiogalactopyranoside. GST-TEV-VAPB proteins were captured on GSH-Sepharose, washed and then either eluted with 10 mM reduced glutathione or recovered by incubating the resin with TEV-His6 protease, which was subsequently removed by affinity chromatography. To remove the GST tag, GST-TEV-Flag-FAF1 was cleaved using AcTEV protease (Invitrogen) overnight at 4°C. Then 2 μg Flag-FAF1 was mixed with 0.5 μg of the indicated form of VAPB in a final volume of 400 μl buffer B (50 mM HEPES/KOH, pH 7.5; 60 mM KOAc; 5 mM MgCl2; 5% glycerol; 0.1% Triton X-100). BSA was added to control samples to reach the same final protein concentration. The protein mixtures were pre-incubated for 30 min at 4°C, then anti-Flag beads (Sigma) were added and incubated for another 30 min at 4°C, under rotation. The beads were washed and the bound proteins were eluted by incubation with 3× Laemmli buffer for 10 min at 70°C.
The dried peptides were resuspended in 20 μl 0.1% (v/v) TFA and separated on a Dionex UltiMate 3000 LC system (Thermo Scientific) using a 25 cm column packed with 3 μm Magic C18 material (Michrom Bioresources, Auburn, USA). Mass spectra were acquired on an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific) operating in data-dependent mode. After conversion to mzXML, the raw data were searched using Comet against version 3.87 of the IPI human or mouse protein database using static carboxamidomethylation of cysteine residues, variable oxidation of methionine residues and accounting for up to two missed tryptic cleavages. For SILAC samples, variable modification of lysine and arginine residues was also specified. The Trans-Proteomic Pipeline was used to assign peptide and protein probabilities and to filter results at a 1% false discovery rate. Data from the anti di-glycine remnant antibody analysis were searched using X!!TANDEM (v. 2010.12.01.1) with the k-score plugin. Two sets of static modifications were used: (1) carboxamidomethylation of cysteine and (2) carboxamidomethylation of cysteine, SILAC heavy Lys and Arg. Additionally, variable oxidation of methionine and GlyGly modification of lysine residues were accounted for.
The following antibodies were used: mouse anti-Flag M2 (Sigma, A8592), mouse anti-ubiquitin FK2 (Enzo, PW8810), rabbit anti-ubiquitin (Dako, Ely, UK; Z0458), mouse anti-p97 (Fitzgerald, North Acton, USA; 10R-P104A), rabbit anti-VAPA (Epitomics, Burlingame, USA; S1706), mouse anti-FAF1 (Abnova, Taipei City, Taiwan; H00011124-A01), rabbit anti-FAF1 (courtesy of Millipore, Billerica, USA), rabbit anti-GAPDH (Cell Signaling, 2118), rabbit anti-RAB3GAP1 (Proteintech, Manchester, UK; 21663-1-AP), rabbit anti-RAB3GAP2 (Abgent, Maidenhead, UK; AP9635B), rabbit anti-WDR44 (Bethyl, Montgomery, USA; A301-441A), mouse anti-ASNA1 (Abnova, H00000439-M03), rabbit anti-Syntaxin 1A (GeneTex, Irvine, USA; GTX113559), rabbit anti-BAG6 (Cell Signaling, 8523S), mouse anti-RPN2 (Abnova, H00006185-B01), rabbit anti-CD147 (Abcam, Cambridge, UK; ab108317), Alexa Fluor® 488 goat anti-rabbit (Invitrogen, A11008) and Alexa Fluor® 594 chicken anti-mouse (Invitrogen, A21201). Anti-VAPA (Epitomics, S1706) antibodies recognize both VAPA (upper band) and VAPB (lower band) in human cell extracts. The following antibodies were raised and affinity purified using the appropriate antigen by the DSTT: sheep anti-FAF1 (S370D; antigen human FAF1, 1-650), rabbit anti-VAPB (R2986 and R2987; antigen human VAPB, 1-210).
amyotrophic lateral sclerosis
bovine serum albumin
Dulbecco's modified Eagle's medium
Division of Signal Transduction Therapy Unit
fetal bovine serum
two phenylalanines in an acidic tract motif
VAPB double mutant K87D M89D
modified Eagle's medium
Medical Research Council Protein Phosphorylation and Ubiquitylation Unit
polymerase chain reaction
reverse-transcription-polymerase chain reaction
stable isotope labeling by amino acids in cell culture
small interfering RNA
transmembrane-domain recognition complex
We thank the MRC-PPU DSTT antibody production team (coordinated by Hilary McLauchlan and James Hastie), the MRC-PPU tissue culture team (coordinated by Kirsten McLeod) and the DNA Sequencing Service (coordinated by Nicholas Helps) for their excellent technical support. This work was supported by a grant from the Scottish Government. The CLS microscopy facility is supported by the Wellcome Trust Strategic Award 097945/Z/11/Z.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.