Dysferlin Binds SNAREs (Soluble N-Ethylmaleimide-sensitive Factor (NSF) Attachment Protein Receptors) and Stimulates Membrane Fusion in a Calcium-sensitive Manner*

Resealing of tears in the sarcolemma of myofibers is a necessary step in the repair of muscle tissue. Recent work suggests a critical role for dysferlin in the membrane repair process and that mutations in dysferlin are responsible for limb girdle muscular dystrophy 2B and Miyoshi myopathy. Beyond membrane repair, dysferlin has been linked to SNARE-mediated exocytotic events including cytokine release and acid sphingomyelinase secretion. However, it is unclear whether dysferlin regulates SNARE-mediated membrane fusion. In this study we demonstrate a direct interaction between dysferlin and the SNARE proteins syntaxin 4 and SNAP-23. In addition, analysis of FRET and in vitro reconstituted lipid mixing assays indicate that dysferlin accelerates syntaxin 4/SNAP-23 heterodimer formation and SNARE-mediated lipid mixing in a calcium-sensitive manner. These results support a function for dysferlin as a calcium-sensing SNARE effector for membrane fusion events.

More recent studies have determined that dysferlin contributes to cytokine secretion, lysosome exocytosis, acid sphingomyelinase secretion, and phagocytosis (11)(12)(13)(14). These reports suggest a wider role for dysferlin in calcium-sensitive membrane trafficking events at the cell membrane. However, the exact functions and mechanisms by which dysferlin operates remain unclear.
One proposed function for dysferlin is as a calcium-sensitive scaffold for the recruitment of other proteins involved in membrane trafficking. Support for this comes from studies that have reported that the C2 domains of dysferlin bind annexins and caveolin (15)(16)(17). Dysferlin may also act as a regulator of calcium influx through interaction with T-tubule dihydropyridine receptors (18 -20). However, several studies have reported that deficiencies in dysferlin result in attenuated exocytosis and the accumulation of unfused vesicles at membrane lesions, suggesting a function in membrane fusion events that occur during repair of cell membrane wounds (9,12). In agreement with this, dysferlin has been reported to bind calcium and the cell membrane lipids phosphatidylserine and phosphatidylinositol bisphosphate and to co-localize with the SNARE protein syntaxin 4 (21)(22)(23)(24). Knockdown of dysferlin also results in a reduction in lysosome exocytosis and delayed release of acid sphingomyelinase, in agreement with dysferlin's proposed role in membrane fusion (11,13).
Direct evidence establishing the function of dysferlin in membrane trafficking is lacking, and despite the requirement of SNAREs for membrane repair and exocytosis, no study has directly tested whether dysferlin functions as a SNARE effector for membrane fusion (25). This has been due in part to the complexity of membrane trafficking processes, which can lead to ambiguity in the interpretation of results. In this study we use cell-based techniques and well defined reconstituted systems to characterize the relationship between dysferlin and SNARE proteins and establish a function for dysferlin in membrane fusion events.

Results
Dysferlin Binds Syntaxin 4 and SNAP-23-Dysferlin has been reported to co-localize with syntaxin 4, a ubiquitously expressed SNARE protein that contributes to lysosomal exocytosis (21). To determine whether dysferlin interacts with syntaxin 4 as well as the cognate SNARE SNAP-23, dysferlin-immunoprecipitated samples from serum-starved differentiated C2C12 cells were probed by Western blot. As shown in Fig. 1A, syntaxin 4 and SNAP-23 co-immunoprecipitated with dysferlin in a calcium-independent manner. By contrast, actin did not co-immunoprecipitate. Subsequent analysis of reciprocal coimmunoprecipitation samples generated with an anti-syntaxin 4 antibody also supported an association between syntaxin 4 and dysferlin in cell lysate (Fig. 1B). To ascertain whether dysferlin and the SNAREs syntaxin 4 and SNAP-23 reside in close proximity within cells, we utilized a proximity ligation assay on serum-starved differentiated C2C12 cells (26). Fixed cells incubated with anti-dysferlin and either anti-syntaxin 4 or anti-SNAP-23 resulted in 9.32 Ϯ 3.86 or 1.55 Ϯ 0.56 fluorescent puncta per cell nuclei, respectively (Fig. 1, D and E). By contrast, fixed cells incubated with anti-dysferlin and anti-histone h3 or anti-TFAM resulted in 0.39 Ϯ 0.24 and 0.39 Ϯ 0.21 puncta per nuclei. Based upon analysis of Western blot and proximity liga-tion assay data, we conclude that dysferlin associates with syntaxin 4 and SNAP-23.
We next tested whether the co-immunoprecipitation and proximity ligation assay results were due to a direct interaction between dysferlin and SNAREs by conducting fluorescence anisotropy measurements on a fluorescein-5-maleimide (FITC; ThermoFisher)-labeled syntaxin 4 single cysteine construct (syntaxin 4 ⌬TM amino acids 1-274) in solution. When a construct composed of the entire cytoplasmic region of dysferlin lacking only the transmembrane domain was added to samples containing labeled syntaxin 4, a concentration-dependent increase in anisotropy was observed (Fig. 2B). This observed increase in anisotropy was independent of the presence of calcium with an apparent K d of 0.094 Ϯ 0.020 M.
Syntaxin is composed of a single-span C-terminal transmembrane helix, a single SNARE domain (H3 domain), and an N-terminal ␣-helical regulatory domain (Habc domain) (27)(28)(29). Previous reports have established that the C2 domains of the dysferlin homologue otoferlin bind specifically to the H3 domain of syntaxin 1 (30). To determine if the C2 domains of dysferlin interact with the H3 SNARE domain of syntaxin 4, anisotropy measurements of a FITC-labeled truncated form of syntaxin 4 lacking the Habc and transmembrane regions (amino acids 163-274 with residue 275 changed to a cysteine) were collected. The addition of the cytoplasmic region of dysferlin increased the anisotropy value of the H3 domain of syntaxin 4 in a calcium-insensitive dose-dependent manner, with an apparent K d of 0.022 Ϯ 0.001 M. Similarly, dysferlin con-FIGURE 1. Dysferlin interacts with syntaxin 4 and SNAP-23. A, Western blot of dysferlin immunoprecipitation samples from differentiated C2C12 cells probed for syntaxin 4, SNAP-23, or actin. Syntaxin 4 and SNAP-23, but not actin, co-immunoprecipitate. Total refers to the total soluble cell lysate; EDTA and Ca 2ϩ refer to immunoprecipitation samples in the presence of EDTA or calcium. B, Western blot of syntaxin 4 immunoprecipitated samples from differentiated C2C12 cells probed for dysferlin. C, quantitation of Western blot results. n ϭ 3. D, dysferlin co-localizes with syntaxin 4 and SNAP-23 but not histone h3 as determined by in situ proximity ligation. E, quantitation of proximity ligation results. Error bars represent S.D. n ϭ 12. *, p Ͻ 0.05. ns, not significant. structs composed of the first three N-terminal (C2ABC) or four C-terminal (C2DEFG) C2 domains of dysferlin also increased anisotropy values when added to labeled syntaxin H3 domain, with apparent K d values of 0.38 Ϯ 0.04 M for C2ABC and 0.32 Ϯ 0.02 M for C2DEFG (Fig. 2C). To identify the specific C2 domains of dysferlin that mediate binding, titrations were conducted between the syntaxin 4 H3 domain and each individual C2 domain of dysferlin. As with the larger multidomain constructs, the addition of the individual C2 domains increased the anisotropy values of the FITC-labeled H3 domain of syntaxin 4 in a calcium-independent manner. In all cases the increase in SNARE anisotropy upon the addition of dysferlin constructs show cooperativity. Such cooperative interactions have been noted for other protein-protein interactions; however, the basis for the observed cooperative interactions between C2 domains and SNAREs is not well understood. Apparent K d values for the individual domains ranged from 2.5 to 10.4 M, indicating that the isolated domains bind with lower affinity than the multidomain constructs. The apparent K d value for a two-domain construct (C2AB) was also determined ( Table 1). Anisotropy values did not change significantly with the addition of either recombinant glutathione S-transferase (GST) or maltose-binding protein (MBP) to FITC-labeled syntaxin 4 samples, which served as negative controls. In addition, no changes in anisotropy were observed when titrations were conducted in the presence of 8 M urea, which disrupts proteinprotein interactions. An increase in anisotropy was also observed when the dysferlin C2ABC or C2DEFG constructs were titrated against a FITC-labeled SNAP-23 single cysteine mutant (SNAP-23 C79A,C80A,C83A,C85A), with apparent K d values of 0.69 Ϯ 0.05 and 0.39 Ϯ 0.01 M, respectively (Fig. 2D). A summary of apparent K d values is listed in Table 1. We con-clude that the C2 domains of dysferlin bind both syntaxin 4 and SNAP-23.
Dysferlin Stimulates the Assembly of SNARE Heterodimers-Assembly of syntaxin 4 with the cognate SNARE SNAP-23 is thought to be a necessary prerequisite for the formation of membrane fusion-competent SNARE complexes (27)(28)(29)(31)(32)(33). This assembly proceeds through interaction between the SNARE motifs of syntaxin 4 and SNAP-23 and is thought to be inhibited by the Habc domain of syntaxin 4 (34 -36). Having established that dysferlin binds SNAP-23 and the SNARE H3 domain of syntaxin 4, we sought to determine if dysferlin facilitated assembly of the syntaxin 4/SNAP-23 heterodimer. To monitor SNARE heterodimer assembly, we modified a method used previously to detect SNARE interaction by measuring FRET values between FITC-labeled full-length syntaxin 4 C278A proteoliposomes and a soluble form of SNAP-23 C79A,C80A,C83A,C85A labeled with an Alexa-fluor 546 C5 maleimide, as depicted in   (35). Relative to samples lacking dysferlin, the addition of either the C2ABC or C2DEFG constructs enhanced the FRET value as determined by a decrease in donor fluorescence and concomitant increase in acceptor fluorescence (Fig. 3, B and C). The observed dysferlin-dependent increase in FRET was significantly enhanced by the presence of 500 M calcium (Fig. 3C). Dysferlin Stimulates SNARE-mediated Lipid Mixing-Having established that dysferlin facilitates SNARE heterodimer formation, we sought to determine what effect dysferlin has on SNARE-mediated membrane fusion using an in vitro reconstituted lipid mixing assay (31-33, 37, 38). In this assay, VAMP 2 liposomes containing a FRET pair were mixed with wild type full-length syntaxin 4 containing liposomes that do not harbor fluorophores. When a soluble form of SNAP-23 is added to the liposome mixture, the three SNAREs assemble into a ternary complex that drives fusion between the liposomes, resulting in lipid mixing, dilution of the FRET pair, and dequenching of the NBD 2 fluorophore (Fig. 4A). Previous studies using this assay have determined that syntaxin 4/SNAP-23 heterodimer assembly is both a prerequisite and rate-limiting step in SNARE-mediated lipid mixing between the two populations of proteoliposomes regardless of whether the proteoliposomes are aggregated (32,33). To determine if dysferlin can accelerate SNARE assembly and lipid mixing, we monitored the dequenching of the NBD signal in proteoliposome samples in the presence or absence of the full-length cytoplasmic region of dysferlin. In the absence of dysferlin, a low basal level of fusion was observed (Fig. 4B). In the absence of calcium, the addition of dysferlin did not enhance lipid mixing. However, in the presence of calcium, dysferlin enhanced both the rate and final extent of fusion (Fig. 4B). Consistent with the requirement of SNARE assembly, the observed stimulatory effect of dysferlin was SNARE-dependent, as membrane fusion was not observed in the absence of either SNAP-23, syntaxin 4, or VAMP 2 regardless of the presence of dysferlin or calcium (Fig. 4C). Dysferlin also stimulated lipid mixing in assays conducted with dithionite-treated VAMP 2 liposomes, suggesting that fusion extends beyond the hemifused state and that the inner leaflets of the SNARE liposomes mix (33). We next tested the shortened forms of dysferlin for the ability to stimulate membrane fusion and found that truncated constructs composed of either the first three (C2ABC) or last four (C2DEFG) C2 domains stimulated SNARE-mediated membrane fusion in a calcium-enhanced manner (Fig. 4D). In summary, these results suggest that dysferlin promotes lipid mixing between SNARE proteoliposomes in a calciumsensitive manner.
To ensure that the lipid mixing events involved ternary SNARE complex formation, aliquots from the lipid mixing assays were analyzed by Western blot (Fig. 4E). SNARE proteoliposome samples containing calcium and the C2ABC or C2DEFG constructs that were allowed to react for 1 h were found to contain high molecular weight SDS-resistant syntaxin 4 bands, consistent with the product of SNARE complex formation (Fig. 4E, left panel) (39,40). Western blot analysis of SNARE proteoliposomes lacking dysferlin constructs showed little-to-no high molecular weight complexes, consistent with the low rate and extent of NBD dequenching observed in the lipid mixing assays (Fig. 4E, left panel). When dysferlin-containing aliquots were boiled for 10 min, the high molecular syntaxin 4 weight bands were no longer apparent, consistent with previous studies which have demonstrated that SNARE complexes are SDS-resistant and require boiling to dissociate (Fig. 4E, right panel).

Discussion
Dysferlin has been implicated in lysosome exocytosis, cell membrane repair, and vesicle-vesicle fusion (11)(12)(13)41). In this study we used in vitro approaches to test the dysferlin effects on membrane fusion. Analysis of co-immunoprecipitation samples indicated that dysferlin associates with syntaxin 4 and SNAP-23, two ubiquitously expressed SNAREs previously implicated in lysosome exocytosis. Using a FRET-based assay, dysferlin was found to facilitate assembly of syntaxin 4/SNAP-23 heterodimers. Using a reconstituted membrane fusion assay, we found that dysferlin promoted SNARE-mediated lipid mixing in a calcium-sensitive manner. These results suggest dysferlin could act as a calcium sensor for membrane repair, cytokine secretion, and lysosome exocytosis (11)(12)(13)(14)42). The results of our studies may also explain why dysferlin deficiencies result in attenuated rates of exocytosis and an increase in unfused vesicles (9). Our findings that regions in both the N and C termini of the protein engage SNAREs and stimulate fusion may explain cell studies that have implicated both the C2A domain and the C-terminal C2 domains in dysferlin function (11,(43)(44)(45).
Previous studies have determined that the dysferlin homologue otoferlin also stimulates SNARE-mediated membrane fusion and binds plasma membrane lipids (22,33,46,47). Similarly, defects in Fer1, the Caenorhabditis elegans ferlin homologue, results in defective calcium-dependent membranous organelle fusion with the sperm plasma membrane during spermatogenesis (48). That multiple ferlin proteins in both vertebrate and invertebrate species have been determined to promote calcium-dependent membrane fusion suggests a common function among members of this family. However, members of the ferlin family appear to have distinct binding partners and calcium sensitivities, and ferlins are not believed to be completely functionally redundant (49). We speculate ferlin proteins share a conserved function as calcium-sensing membrane fusion proteins that have diversified to regulate different exocytotic events.

Molecular Biology and Generation of Recombinant Protein
Constructs-cDNA encoding human full-length dysferlin (accession number AF075575) was used as a template for clon-  JULY 8, 2016 • VOLUME 291 • NUMBER 28 ing and was a gift from Dr. Kate Bushby (Newcastle University). The single and multi-C2 domain (C2ABC, C2DEFG) dysferlin constructs were described previously (22,23). A soluble dysferlin construct encoding the entire cytosolic region (residues 1-1965) in pMCSG9 was generated via ligation-independent cloning with forward primer TACTTCCAATCCAATGCAA-TGCTGAGGGTCTTCATCCTC and reverse primer TTATC-CACTTCCAATGCTATTCTGCTACACAGGGCCAC. Rat SNARE constructs were a gift from Edwin Chapman (University of Wisconsin, Madison, WI) and were expressed in either pGEX2T, pGEX6p3, or pET28a. Rat full-length syntaxin 4 in pGEX2T encoding residues 1-298 was subjected to site-directed mutagenesis to change a cysteine in the transmembrane domain at residue 279 to an alanine via forward primer GAA-AAAGGTCATTGCCATCGCGGTTTCTGTCACTGTTCT-CATC and reverse primer GATGAGAACAGTGACAGAAA-CCGCGATGGCAATGACCTTTTTC and is denoted as syntaxin 4 C279A. To generate a soluble syntaxin 4 with a single cysteine, the syntaxin 4 cytosolic region composed of residues 1-274 was amplified out of the pGEX2T vector with forward primer CGGGGCGGATCCCGCGACAGGACCCATGA-GTTG and reverse primer CCGGCGGAATTCTCAGACCT-TTTTCTTCCTCGCCTTCTTCTG. The amplicon was restriction-cloned into pET28a with BamHI and EcoRI and denoted syntaxin 4 ⌬TM. To generate a syntaxin 4 H3 construct with a single cysteine residue position number 275, the syntaxin 4 H3 domain encoding residues 163-274 was amplified out of the pGEX2T plasmid with forward primer CGGGGCGGATCCA-CCAATGCTGGAATGGTGTCTGACG and reverse primer ccggcgGAATTCTCAGCAGACCTTTTTCTTCCTCGCCTT-CTTCTGATTC. The amplicon was restriction cloned into pET28a with BamHI and EcoRI.

Dysferlin Stimulates SNARE-mediated Membrane Fusion
To generate a SNAP-23 construct with one remaining endogenous cysteine at residue 87, denoted SNAP-23 C79A,C80A, C83A,C85A, rat SNAP-23 was subjected to two sequential rounds of site-directed mutagenesis, first changing cysteine residues 79 and 80 to alanine utilizing forward primer AACA-GAACTCAACAAGGCTGCTGGCCTCTGCGTCTGCC-CTTG and reverse primer CAAGGGCAGACGCAGAGGCC-AGCAGCCTTGTTGAGTTCTGTT followed by a second round changing cysteine residues 83 and 85 to alanine utilizing forward primer CAAGGCTGCTGGCCTCGCCGTCGCCCC-TTGTAATAGGACC and reverse primer GGTCCTATTAC-AAGGGGCGACGGCGAGGCCAGCAGCCTTG. The wild type SNAP-23 and cysteine mutant SNAP-23 C79A,C80A, C83A,C85A were amplified out of pGEX2T via forward primer CGGGGCGGATCCATGGATGATCTATCACCAGAAGA-AATT CAGCTTC and reverse primer ccggcgGAATTCTCA-GCTGTCAATGAGTTTCTTTGCTCTTGTATTGG and restriction-cloned into pGEX6p3 with BamHI and EcoRI (New England BioLabs). All site-directed mutagenesis was carried out utilizing the QuikChange II mutagenesis kit from Agilent. All constructs, generated or received, were sequenced verified by Genscript or The Center for Genome Research and Biocomputing at Oregon State University.
Recombinant Protein Purification-The dysferlin fusion protein constructs were purified using methods described previously (22,23). In brief, the pMCSG9 vector, which encodes a polyhistidine-maltose binding domain tag with a tobacco etch virus protease cleavage site was used to generate dysferlin fusion proteins in Escherichia coli. The dysferlin single domain proteins were expressed in BL21 cells that were cultured in Luria-Bertani broth in the presence of 100 g/ml ampicillin at 37°C and induced for 3 h with 0.3 mM isopropyl ␤-D-1-thiogalactopyranoside after optical density reached A 600 0.6. The dysferlin C2ABC, C2DEFG, and seven C2 domain protein constructs were expressed in Rosetta™ (DE3) pLysS cells (EMD Millipore) that were cultured in terrific broth with 100 g/ml ampicillin and 34 g/ml chloramphenicol. Protein expression was induced at A 600 ϭ 0.9 with 0.3 mM isopropyl ␤-D-1-thiogalactopyranoside for 14 h at 18°C. All cell pellets were isolated by centrifugation at 4000 rpm at 4°C for 20 min and then resuspended in lysis buffer, 50 mM Tris base pH 7.5, 150 mM sodium chloride, 20 mM imidazole in the presence of 1 mM phenylmethanesulfonyl fluoride (PMSF) and 1 M each of the following protease inhibitors; leupeptin, pepstatin, aprotinin (Research Products International). The cells were lysed by sonication. The soluble fraction was isolated by centrifugation in a Beckman J2-21 centrifuge at 20,000 ϫ g at 4°C for 20 min and then incubated with Ni-NTA (ThermoFisher) beads in lysis buffer for 1 h with rocking at 4°C. The beads were washed with 20 column volumes of lysis buffer, and the protein was then eluted with lysis buffer plus 500 mM imidazole. The pure fractions were checked via SDS-PAGE, pooled, and extensively dialyzed into 50 mM Tris base pH 7.5, 100 mM sodium chloride using dialysis tubing 4000 -6000 molecular weight cutoff (Spectrum Labs).
The SNARE protein constructs were expressed and purified as described previously (50). The pGEX2T vector, which encodes an N-terminal glutathione S-transferase protein domain tag with a thrombin protease cleavage site was used to express the full-length wild type syntaxin 4 and syntaxin 4 C279A. pGEX6p3 SNAP-23 constructs were expressed with a N-terminal glutathione S-transferase with a precision protease cleavage site. The GST fusion SNARE constructs were expressed in BL21 cells cultured in Luria-Bertani broth in the presence of 100 g/ml ampicillin at 37°C and induced for 3 h after A 600 0.6 with 1 mM isopropyl ␤-D-1-thiogalactopyranoside. All cell pellets were isolated as noted above. The pellets were resuspended in 50 mM Hepes, pH 7.5, 150 mM sodium chloride. 10% glycerol was also added to the full-length syntaxin 4 suspension buffer. Protease inhibitors were subsequently added to the cell suspension buffer, and the cells were lysed by sonication. Triton X-100 to 1% V/V was added to the fulllength syntaxin 4 lysates, and all lysates were rocked for 30 min at 4°C to promote protein solubilization. The soluble fraction of the lysates were isolated by centrifugation and then incubated with glutathione-Sepharose high performance beads (GE healthcare) for 1 h with rocking at 4°C. The beads were washed with 20 column volumes of lysis buffer (plus 0.1% Triton X-100 for full-length syntaxin 4), and the protein was eluted in lysis buffer with 200 mM reduced glutathione (GoldBio) and dialyzed extensively into 50 mM Hepes, pH 7.5, 100 mM sodium chloride, and 0.1% W/V CHAPS (Research Products International).
The syntaxin 4 ⌬TM and H3 domain constructs were expressed as a pET28a vector construct in Luria-Bertani broth in the same fashion as the GST fusion SNAREs. Cell pellets were isolated and lysed as described above. The protein was purified similarly to full-length syntaxin 4, with a lysis buffer of 50 mM Tris, pH 7.5, 20 mM imidazole, 150 mM sodium chloride, and a protein solubilization step after the addition of Triton X-100 to 0.1% (v/v) concentration. The soluble fraction was isolated as above by centrifugation and then incubated with pre-equilibrated Ni-NTA beads for 1 h with rocking in a cold room. The beads were washed with 20 column volumes of lysis buffer plus 0.1% Triton X-100, and the protein was eluted in lysis buffer with 500 mM imidazole. The protein fractions were checked for purity via SDS-PAGE and dialyzed extensively into 50 mM Hepes, pH 7.5, 100 mM sodium chloride, and 0.1% W/V CHAPS (Research Products International).
The VAMP 2 pET28a construct was expressed, cell pellets were collected, and the soluble fraction was isolated as described for syntaxin 4 and SNAP-23. The lysis buffer consisted of 50 mM Hepes, pH 7.5, 150 mM sodium chloride, 20 mM imidazole, 10% glycerol, 2 mM ␤-mercaptoethanol. The soluble lysates were incubated for 1 h with Ni-NTA beads. The beads were then washed with 10 column volumes of lysis buffer followed by 10 column volumes of wash buffer 1 (50 mM Tris base, pH 7.5, 300 mM sodium chloride, 20 mM imidazole, 2 mM ␤-mercaptoethanol) and 10 column volumes of wash buffer 2 (50 mM Tris base pH 7.5, 300 mM sodium chloride, 40 mM imidazole). The protein was eluted in 50 mM Tris base pH 7.5, 150 mM sodium chloride, 500 mM imidazole, and 1.5% CHAPS. All SNAREs were checked by SDS-PAGE and pooled as above, and the proteins were extensively dialyzed into 50 mM Tris base pH 7.5, 100 mM sodium chloride or 50 mM Hepes, pH 7.5, 100 mM sodium chloride (plus 0.1% CHAPS for syntaxin 4 ⌬TM, syntaxin H3 domain, and full-length VAMP 2) if the protein was to be utilized in a fluorescence-labeled assay.
Enzymatic removal of fusion protein tag glutathione S-transferase was achieved by the addition of 10 units of thrombin (Sigma) per milligram of protein and incubation at 4°C overnight (syntaxin 4) or the addition of 10 units of precision protease per milligram of protein and incubation at 4°C (SNAP-23) overnight with rocking. Thrombin was removed with p-aminobenzamidine-agarose (Sigma) beads. Cleaved protein was then added to glutathione-Sepharose beads. The beads were rocked for 1 h in at 4°C and the flow-through was collected.
Proteins were concentrated if necessary with a VivaSpin Turbo 10,000 molecular weight cutoff spin column and subsequently used for studies. The protein concentrations were measured using a nanodrop ND-1000 and calculated using extinction coefficient based on protein sequence.
Fluorescence Anisotropy-Syntaxin 4 ⌬TM, SNAP-23 C79A,C80A,C83A,C85A, and syntaxin 4 H3 proteins were labeled with FITC. FITC was dissolved into sterile DMSO (Sigma) at a concentration of 10 mg/ml, and 50 l was slowly added with gentle agitation to protein with a concentration of ϳ4 mg/ml in 50 mM Hepes, pH 7.5, 100 mM sodium chloride, pH 7.5 (plus 0.1% CHAPS and 1 mM tris(2-carboxyethyl)phosphine for the syntaxin 4 ⌬TM and H3 domain). The dye was allowed to react with SNAP-23 and syntaxin 4 at 23°C for 1 h isolated from light, whereas the syntaxin 4 ⌬TM and syntaxin 4 H3 domain required 14 or 3 h of incubation at room temperature, respectively, with dye for efficient labeling. The labeled protein was then extensively dialyzed in 4000 -6000 molecular weight cutoff tubing into buffer composed of 50 mM Tris base, pH 7.5, 100 mM sodium chloride (TBS) or Hepes, pH 7.5, 100 mM sodium chloride, (with 0.1% CHAPS, 1 mM tris(2-carboxyethyl)phosphine for syntaxin 4 ⌬TM and syntaxin 4 H3 domain) with two dialysis buffer exchanges to remove excess dye. The degree of labeling was determined by the manufacturer's protocol, and all absorbances were measured using a nanodrop ND-1000. The degree of labeling was calculated as (A max ϫ M r )/([protein] ϫ 68,000). The degree of labeling was determined to be ϳ0.39 labels per syntaxin 4 ⌬TM, 0.9 labels per full-length syntaxin 4 C279A, 0.6 for SNAP-23 C78A,C80A,C83A,C85A, and 0.25 labels per syntaxin 4 H3 domain. Dysferlin constructs were added to the labeled SNARE at room temperature in phosphate-buffered saline (20 mM phosphate, pH 7.5, 100 mM sodium chloride) with a total reaction volume of 60 l. Measurements were collected after a 1-min equilibration time at room temperature with a PTI fluorimeter in a 200-l quartz cuvette with excitation of ϭ 494 nm, emission of ϭ 518 nm. Anisotropy was calculated using the equation r ϭ (IVV Ϫ IVH)/(IVV ϩ 2IVH) where IVV and IVH correspond to the parallel and perpendicular fluorescence emission intensities. The fraction of protein bound, P, was calculated as p ϭ (r Ϫ r f )/(r b Ϫ r f ), where r is the observed anisotropy, r f is the anisotropy in the absence of dysferlin, and r b is the anisotropy of 100% bound SNARE. (r b was set to the maximum anisotropy value at the plateau phase of the titration, where the further increase in dysferlin concentration resulted in little to no increase in the anisotropy value under the concentrations tested). The reported values represent the mean Ϯ propagated error for three or more samples.
SNARE Heterodimer Assembly Assay-The FRET assembly experiments were adapted from a previously described method (35). GST cleaved full-length syntaxin 4 C278A and GSTcleaved SNAP-23 C78A,C80A,C83A,C85A were labeled via cysteine-reactive FITC or cysteine-reactive Alexa-fluor 546 (Invitrogen), respectively, for 1 h at 23°C in 50 mM Hepes. pH 7.5, 100 mM sodium chloride as described above. Excess dye was removed by extensive dialysis in TBS. Labeling efficiency was calculated as described above (FITC) and determined to be ϳ0.8 labels per syntaxin 4 and 0.9 labels per SNAP-23. The labeled syntaxin 4 protein was subsequently reconstituted into liposomes composed of 25% POPS, 75% POPC at ϳ100 copies per liposome as described previously (33,37,38). Briefly, lipids (Avanti Polar Lipids) were evaporated overnight and resuspended with labeled full-length syntaxin 4 C278A (1.2 mg) and brought to 1.5-ml total volume with reconstitution buffer. Two volumes of reconstitution buffer (25 mM Hepes, pH 7.8, 100 mM KCl) were added dropwise with light agitation, and the proteoliposome sample was dialyzed overnight in reconstitution buffer with 1 mM DTT. The reconstituted liposomes were subsequently mixed in an eqi-volume of 80% Accudenz solution in a thin walled ultracentrifuge tube and layered with 2.25 ml of 30% Accudenz followed by 0.750 ml of reconstitution buffer. The liposomes were then floated on the Accudenz gradient via centrifugation at 41,000 rpm in a SW-41 (Beckman) rotor with adaptors at 4°C. Proteoliposomes were isolated from the 0/30% Accudenz interface.
FRET samples were 100 l in total volume and composed of 5 l of syntaxin 4 proteoliposomes, 5 M SNAP-23, and 1 M dysferlin. The assay was run in the absence or presence of calcium (500 M) with an excitation wavelength of 494 nm and a FRET emission spectrum collected over 500 -600-nm wavelength. FRET was calculated as the ratio of donor to acceptor emission intensities. The criterion for FRET was a concomitant reduction in donor emission intensity and increase in acceptor emission intensity.
Immunoprecipitation-Cultured C2C12 cells obtained from ATCC were grown to 70% confluence in a T75 flask in high glucose plus pyruvate DMEM (Invitrogen), 10% fetal bovine serum (Invitrogen), and 20 units/ml penicillin-streptomycin (Invitrogen) at 37°C with 5% CO 2 . The cells were then changed into a serum-deficient media to promote differentiation per ATCC protocol, (high glucose plus pyruvate DMEM, 2% horse serum (Invitrogen), 20 units/ml penicillin-streptomycin) and cultured for 7 days with serum changes every 2 days. The medium was removed, and the cells were washed twice with sterile PBS (50 mM phosphate, pH 7.5, 100 mM sodium chloride), and 1 ml of lysis buffer (sterile 50 mM phosphate, pH 7.5, 100 mM sodium chloride, and 0.2% Triton X-100 and Roche mini complete protease inhibitors) was added to the flask. Cells were incubated for 30 min with rocking at 4°C. The total lysate was removed from the flask and split into 2 samples; to 1 sample 3 mM calcium chloride (Sigma) was added, and to the other 3 mM EDTA (Sigma) was added. The samples were then centrifuged at 14,000 ϫ g for 5 min at 4°C to remove cell debris. The soluble lysate fraction was subsequently isolated, and 50 l of the Hamlet mouse anti-dysferlin (Abcam) antibody was added to the sample. The samples were rocked at room temperature for 1 h followed by the addition of 50 l of fresh PBS-washed protein A/G beads. After 1 h of incubation at room temperature with gentle agitation, a magnetic rack was used to pellet the beads, and the supernatant was removed. The beads were then washed with three sample volumes of PBS. To dissociate bound proteins from the beads, one sample volume of 2ϫ Laemmli sample buffer was added, the sample was then boiled for 2 min, and the sample was loaded onto a 12% SDS-PAGE gel. The proteins were then transferred using a wet tank transfer blot box (Invitrogen) in cold Tobin buffer (2.5 mM Tris, pH 8.3, 19.2 mM glycine) with 20% methanol onto a PVDF membrane (Immobilon-P membrane, PVDF, 0.45 m, Invitrogen) for 90 min at 200 mA (51). The membrane was blocked with 2% nonfat milk in PBS plus 0.05% Tween 20 for 1 h at room temperature, and rabbit primary antibodies against SNAP 23 (Sigma) or syntaxin 4 at 1:1000 dilution (Sigma) were added. After incubating overnight at 4°C with rocking, membranes were washed in triplicate for 10 min in PBS then incubated with goat antirabbit HRP secondary 1:5000 dilution (Sigma) in PBS at room temperature for 1 h. The syntaxin 4 co-immunoprecipitation protocol was performed similarly to that described above, with anti-syntaxin 4 antibody (Abcam). Samples were run on a 8% gel, blotted, and probed with Hamlet mouse anti-dysferlin (1:500, Abcam) and anti-mouse HRP secondary (1:5000, Sigma). Blots were developed with SuperSignal West Pico Chemiluminescent Substrate (ThermoFisher) per the manufacturer's protocol and imaged with a Kodak Digital Science Image Station 440CF. Quantification of co-immunoprecipitation was conducted by comparison of Western blot band intensities using Image J, with the total lysate sample set to 1 and the co-immunoprecipitation band as a fraction of the total lysate. Statistical significance corresponding to p Ͻ 0.1 were indicated by an asterisk (*). Negative control immunoprecipitations using 50 l of protein G beads coupled to IgG from mouse serum were conducted similarly to the dysferlin immunoprecipitation described above. Briefly, after the addition of cell lysate to the IgG-beads, the sample was rocked at room temperature for 1 h with gentle agitation, and a magnetic rack was used to pellet the beads, with subsequent removal of supernatant. The IgG-coupled beads were then washed with three sample volumes of PBS. To dissociate any bound proteins from the beads, 2ϫ Laemmli buffer was added to the sample, and the sample was then boiled for 2 min and loaded onto a 10% SDS-PAGE gel.
Duolink-C2C12 cells grown to 70% confluence in full serum medium as described above and treated with 2 ml of 0.25% trypsin (Life Technologies) for 30 min followed by the addition of 2 ml of complete medium to neutralize trypsin, and 200 l of cell suspension was subsequently plated onto square ethanolsterilized coverslips (Fisher) in a sterile 6-well plate (Sigma). Cells were allowed to adhere for 1 h with incubation followed by the addition 2 ml of full serum medium. After the cells reached 70% confluence, the medium was changed to serum-deficient medium, and cells were differentiated for 7 days. Cells were washed twice in PBS and then fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. Fixed cells were washed twice with PBS, permeabilized with 0.2% Triton X-100 in PBS for 10 min, and then blocked with 3 drops Duolink blocking buffer for 1 h at 37°C. Each sample was then incubated with two primary antibodies in parallel, overnight, in a total volume of 250 l of Duolink diluent at 4°C (Hamlet mouse anti-dysferlin 1:250 (Abcam), rabbit anti-syntaxin 4 1:500 (Sigma), rabbit anti-SNAP-23 1:500 (Sigma), rabbit anti-histone H3 1:500 (Sigma), rabbit anti-TFAM 1:500 (Bethyl Laboratories)). Samples were subsequently washed twice with PBS and incubated with 8 l of Duolink anti-mouse minus and 8 l of anti-rabbit plus proximity ligation probes for 1 h at 37°C. Cells were removed from 37°C and washed twice with PBS, and two unique oligonucleotides complimentary to the individual proximity ligation assay probes provided by Sigma were added in the presence of a ligation enzyme per the manufacturer's protocol. The samples were incubated for 30 min at 37°C. The samples were subsequently washed with twice with PBS and subjected to rolling circular amplification for 100 min at 37°C in the presence of labeled complimentary oligonucleotide probes per the Duolink instructions. Samples were washed twice in PBS and incubated with Hoechst stain at a 1:5000 dilution for 30 min. After incubation the samples were washed and mounted on glass slides (Fisher) with Fluoromount (Life Technologies). Images were captured using a Zeiss Axiovert S100TV fluorescent microscope and Metamorph software version 6.3r7. At least six images per sample were analyzed, and each sample was repeated in duplicate. The number of puncta-per-nuclei was recorded. A paired t test was performed on the data, and p values are reported. Each antibody used in our study was tested for specificity using standard immunofluorescence. The highest dilution of the antibody that gave specific, detectable immunofluorescence was used for Duolink measurements so as to minimize nonspecific antibody interactions.
Reconstituted Fusion Assays-Reconstitution of proteoliposomes was conducted as previously reported and as noted above (33,37,38). VAMP 2 proteoliposomes were composed of 72% POPC, 25% POPS, 1.5% rhodamine-phosphoethanolamine (PE) (acceptor), and 1.5% NBD-PE (donor). Syntaxin 4 proteoliposomes were composed of 75 mol % POPC and 25 mol % POPS. Proteoliposomes contained ϳ100 copies of VAMP 2 or syntaxin 4 proteins per proteoliposome. All fusion assays were conducted using a PTI fluorimeter with a 200-l quartz cuvette. Samples contained a total volume of 75 l with 30 l of purified syntaxin 4 proteoliposomes, 15 l of purified VAMP 2 proteoliposomes, 0 or 15 M SNAP-23, and varying concentrations of dysferlin protein in TBS. Samples were mixed in a sequential order, with syntaxin 4 proteoliposomes first incubated overnight with SNAP-23 and dysferlin at 4°C followed by the addition of VAMP proteoliposomes just before measurements. For samples containing calcium, the free calcium concentration was 500 M. Measurements were collected over 60 min, and the dequenching of NBD was monitored. At the end of each experiment, 0.5% w/v n-dodecyl-maltoside was added to determine the maximal dequenched NBD signal. Raw fluorescence was normalized to obtain the percentage of maximum fluorescence as described previously (31,50).
Author Contributions-C. P. J. conceived and coordinated the study. C. P. J. and S. J. C. wrote the paper. S. J. C. and C. P. J. performed and analyzed the experiments. N. A. and N. M. provided technical assistance. The authors reviewed the results and approved the final version of the manuscript.