Mutations in BIN1 associated with centronuclear myopathy disrupt membrane remodeling by affecting protein density and oligomerization

Abstract

The regulation of membrane shapes is central to many cellular phenomena. Bin/Amphiphysin/Rvs (BAR) domain-containing proteins are key players for membrane remodeling during endocytosis, cell migration, and endosomal sorting. BIN1, which contains an N-BAR domain, is assumed to be essential for biogenesis of plasma membrane invaginations (T-tubules) in muscle tissues. Three mutations, K35N, D151N and R154Q, have been discovered so far in the BAR domain of BIN1 in patients with centronuclear myopathy (CNM), where impaired organization of T-tubules has been reported. However, molecular mechanisms behind this malfunction have remained elusive. None of the BIN1 disease mutants displayed a significantly compromised curvature sensing ability. However, two mutants showed impaired membrane tubulation both in vivo and in vitro, and displayed characteristically different behaviors. R154Q generated smaller membrane curvature compared to WT N-BAR. Quantification of protein density on membranes revealed a lower membrane-bound density for R154Q compared to WT and the other mutants, which appeared to be the primary reason for the observation of impaired deformation capacity. The D151N mutant was unable to tubulate liposomes under certain experimental conditions. At medium protein concentrations we found 'budding' structures on liposomes that we hypothesized to be intermediates during the tubulation process except for the D151N mutant. Chemical crosslinking assays suggested that the D151N mutation impaired protein oligomerization upon membrane binding. Although we found an insignificant difference between WT and K35N N-BAR in in vitro assays, depolymerizing actin in live cells allowed tubulation of plasma membranes through the K35N mutant. Our results provide insights into the membrane-involved pathophysiological mechanisms leading to human disease.

Figure 1. Location of CNM-associated mutations in BIN1 protein.

A) Schematic location of K35N, D151N, R154Q, and two nonsense mutation in SH3 domain found in CNM patients in human BIN1 protein. B) Cartoon structure of human BIN1 N-BAR domain (PDB ID code: 2FIC). Residues D151 and R154 are shown as sticks at the distal arm of a BAR domain dimer.

Figure 2. Expressing CNM-related mutants in C2C12 cells leads to loss of membrane tubulation.

A) BIN N-BAR* WT and mutants are fused to mKate at C-terminus and transiently transfected in C2C12 myoblasts. Cells are imaged by confocal fluorescence microscopy. Scale bar: 20 µm. B) Quantification of three types of membrane deformations by BIN1 N-BAR* variants. Clusters are defined as structures with lengths not longer than five times of their width. Over 50 cells were analyzed in each separated experiments. Error bars: standard error of the mean in black and standard deviation in light grey. Student t-test for statistical significance: n.s: p>0.05, *: p<0.05, **: p<0.01, ***: p<0.001.

Figure 3. TIRF micrographs show homogeneous fluorescence of CNM-related mutants on plasma membrane.

BIN N-BAR* WT and mutants are fused to EGFP at N-terminus and transiently transfected in C2C12 myoblasts. Cells are imaged by TIRF fluorescence microscopy. Scale bar: 20 µm. All variants of BIN1-N-BAR* proteins show binding ability on plasma membrane. However, only WT BIN1-N-BAR* generates fluorescent puncta on the membrane.

Figure 4. On membrane with high negative charge, only R154Q shows mild reduction in tubulation.

A) Electron micrographs of liposomes (100% DOPS) tubulated by BIN1 N-BAR and mutants. Membrane tubules are observed for all BIN1 N-BAR variants. Scale bar: 200 nm. B) Quantifications of tubule diameter and length by BIN1 variants. R154Q mutation leads to increased tubule diameters and decreased tubule length comparing to WT BIN1 protein. C) Averaged occurrence of membrane tubules and vesiculated liposomes in each micrograph. K35N and D151N mutant are similar to WT N-BAR in membrane tubulation. All the disease mutants cause decrease in vesiculation to varied extent. Over 30 images were analyzed for quantifications. Error bars: standard error of the mean in black and standard deviation in light grey. Student t-test for statistical significance: n.s: p>0.05, *: p<0.05, **: p<0.01, ***: p<0.001.

Figure 5. Lowering membrane negative charge further reduces tubulation ability for the D151N and R154Q mutants.

A) Electron micrographs of liposomes (60% DOPC/20% DOPS/10% DOPE/10% PI(4,5)P₂) incubated with BIN1 N-BAR and the mutants. WT and K35N still strongly deform vesicles. Short tubules are observed in R154Q samples. Almost complete loss of curvature generation is observed in D151N. Scale bar: 200 nm. Electron micrograph in B) is the zoom-in image boxed in A). Uniform small vesicles with 28±7 nm diameters are found after incubation with N-BAR proteins. Scale bar: 50 nm. C) Quantifications of tubule diameter and length by BIN1 variants. R154Q mutation shows compromised curvature generation. D) Quantification of the averaged occurrence of membrane tubules and vesiculated liposomes in each micrograph. K35N shows similar number of tubules per frame and mild decrease in vesiculation. Both D151N and R154Q lead to significant decrease in tubulation and vesiculation. Over 30 images were analyzed for quantifications. Error bars: standard error of the mean in black and standard deviation in light grey. Student t-test for statistical significance: n.s: p>0.05, *: p<0.05, **: p<0.01, ***: p<0.001.

Figure 6. R154Q mutation lowers membrane association density.

Protein adsorption isotherm on GUVs composed of A) 100% DOPS and B) 60% DOPC, 20% DOPS, 10% DOPE, and 10% PI(4,5)P₂. Protein and vesicles are incubated in 20 mM Hepes, 50 mM NaCl, 1 mM DTT, pH 7.4 buffer at room temperature for 30 minutes before imaging on confocal fluorescence microscopy. Green fluorescence intensities on vesicle equators are measured and normalized to the total membrane area. Absolute molecule density is calculated according to the calibration curve. Data are fitted by the Langmuir isotherm model to obtain K_d and B_max. Membrane bound density of R154Q dramatically decreases on biomimetic membrane. The error bars are the standard error of the mean of 10 measured GUVs. C) Quantitative analysis of the curvature sorting abilities of BIN1 N-BAR domain and its mutants. A membrane tether is pulled from a micropipette-aspirated GUV (60% DOPC, 20% DOPS, 10% DOPE, 10% PI(4,5)P₂, labeled by Texas-Red lipid dye) by a polystyrene bead and imaged by confocal fluorescence microscopy. At 100 nM, WT N-BAR fluorescence on GUV and tether is brighter than R154Q N-BAR as shown in the upper panel. WT (100 nM), K35N (100 nM), D151N (100 nM) and R154Q (600 nM) are pre-incubated with GUVs in 20 mM Hepes, 50 mM NaCl, 1 mM DTT, pH 7.4 buffer at room temperature for 30 minutes before transfer to aspiration chamber. By increasing membrane tension controlled by aspiration pressure, protein partitioning on tubular membrane tether and decrease in lipid fluorescence on tether is observed. Curvature coupling parameter / from images recorded by Kalman-averaged confocal xz line-scan images is plotted with square root of membrane tension, revealing a linear relationship. Data from six vesicles was binned (vertical error bars represent standard error of the mean of / and horizontal error bars show standard error of the mean of square root of tension). Inset graph demonstrates similar fluorescence intensity on GUVs analyzed for four BIN1 variants. Error bars are standard errors of the mean. D) Equilibrium force of tethers (pulled from GUVs with composition (60% DOPC, 20% DOPS, 10% DOPE, 10% PI(4,5)P₂) measured by optical trap at membrane tension of 0.21±0.003 mN/m. Error bars: standard error of the mean in black and standard deviation in light grey. Student t-test for statistical significance: n.s: p>0.05, *: p<0.05, **: p<0.01, ***: p<0.001.

Figure 7. Spontaneous tubulation by BIN1 N-BAR includes ‘membrane budding’ step which is compromised in the D151N mutant.

A) Electron micrographs of liposome tubulation by BIN1 N-BAR WT at protein concentrations ranging from 200 nM to 5 µM. Scale bar: 200 nm. At low protein concentration, no morphology deformation is observed. At 500 nM, membrane showed waviness at liposome edges. Further increasing protein concentration results in the growth of longer membrane tubules. B) Titrating D151N N-BAR in tubulation assay (200 nM–15 µM) does not change liposome morphologies.

Figure 8. Chemical crosslinking reveals that D151N mutation impairs protein oligomerization upon membrane binding.

A) WT, B) K35N, C) R154Q and D) D151N are incubated in absence or presence of 100% DOPS LUVs (0.1 mg/mL, final concentration) at room temperature for 30 mins. Indicated amount of BS3 (Bis[sulfosuccinimidyl] suberate) is added in each sample followed by incubation at 37°C for 2 min. Samples are analyzed by SDS-PAGE gel and stained by Coomassie staining. In absence of liposome, N-BAR domain proteins are crosslinked into dimer with MW around 66 kDa. In presence of liposome, WT and K35N show oligomers bands at 0.5 mM BS3 concentration. Further increasing crosslinker concentration yields species which cannot enter the resolving gel. R154Q in C) shows weaker crosslinked bands while D151N in D) shows major dimer bands in presence of membrane. The higher molecular weight band is absent even at 5 mM BS3 concentration in the D151N sample. Buffer: 20 mM Hepes, 150 mM NaCl, pH 7.4.

Figure 9. D151N/H155R double mutant rescues tubulation ability in C2C12 cells.

BIN N-BAR* WT and its mutants are fused to GFP and transiently transfected in C2C12 myoblasts. Cells are imaged by confocal fluorescence microscopy. Scale bar: 20 µm. Expression of BIN1 N-BAR* WT causes great membrane tubulation in C2C12 cells. H155 is identified as a key residue in regulating N-BAR* domain tubulation ability. Mutating H155 to alanine abolishes tubulation in C2C12 cells while double mutant-D151N/H155R deforms membrane into tubular structures.

Figure 10. K35N does not change membrane insertion and depolymerizing actin in cells allows membrane deformation.

A) WT and K35N BAR domain insertion was studied by monolayer composed of DOPC/DOPS/PI (4,5)P₂/DOPE (60∶20∶10∶10). π_c was determined by extrapolating the Δπ versus π₀ plot to the abscissa. π_c is 26.2 and 25.6 dyne/cm for WT and K35N respectively. Error bars represent standard error of the mean from three independent experiments. An F-test of the fittings indicates two linear regressions are not significantly different. B) TIRF images of C2C12 myoblasts transfected BIN1 N-BAR* WT or K35N maintained at 37°C. At t = 0 s, cell culture medium containing 1 µM Latrunculin A are added into the culture dish and time-lapse images are taken. New membrane tubules are generated after actin is depolymerized as the arrows indicate. No increased tubulation was observed in cells transfected with D151N and R154Q N-BAR* (data not shown).

Figure 11. Proposed schematic illustration of BIN1 N-BAR domain tubulating membrane.

1) N-BAR domains bind on membrane; 2) Proteins oligomerize locally to create membrane buds; 3) N-BAR domains sense local curvature and diffuse onto the tubule to further elongate membrane tubule and stabilize the highly curved membrane; 4) Finally, the tubules are fully decorated by N-BAR domains. Electron micrographs are N-BAR domain WT (200 nM–5 µM) incubated with DOPC/DOPS/PI (4,5)P₂/DOPE (60∶20∶10∶10) at room temperature. Scale bar: 200 nM.