The cryo-EM structure of the SNX-BAR Mvp1 tetramer

Abstract

Sorting nexins (SNX) are a family of PX domain-containing proteins with pivotal roles in trafficking and signaling. SNX-BARs, which also have a curvature-generating Bin/Amphiphysin/Rvs (BAR) domain, have membrane-remodeling functions, particularly at the endosome. The minimal PX-BAR module is a dimer mediated by BAR-BAR interactions. Many SNX-BAR proteins, however, additionally have low-complexity N-terminal regions of unknown function. Here, we present the cryo-EM structure of the full-length SNX-BAR Mvp1, which is an autoinhibited tetramer. The tetramer is a dimer of dimers, wherein the membrane-interacting BAR surfaces are sequestered and the PX lipid-binding sites are occluded. The N-terminal low-complexity region of Mvp1 is essential for tetramerization. Mvp1 lacking its N-terminus is dimeric and exhibits enhanced membrane association. Membrane binding and remodeling by Mvp1 therefore requires unmasking of the PX and BAR domain lipid-interacting surfaces. This work reveals a tetrameric configuration of a SNX-BAR protein that provides critical insight into SNX-BAR function and regulation.

Fig. 1. Mvp1 tetramerization is dependent on its N-terminus.

a Mvp1 deforms PS liposomes containing 5% PI3P into irregular tubes. Representative micrographs of negative-stained liposomes incubated either with buffer (control) or with Mvp1. Scale bar—200 nm. b Mvp1 is a tetramer. The absolute molecular weight of Mvp1 was determined using SEC-MALS. The molecular weight, shown in magenta or teal across the elution peaks, are plotted on the right-hand axis. Theoretical molecular weights of Mvp1 dimers and tetramers, calculated according to the Mvp1 sequence, are shown as dotted gray lines. The differential refractive index of the elutions are plotted on the left-hand axis. Mvp1 was eluted in buffer containing either 150 mM (magenta) or 250 mM (teal) NaCl. c As in b, Mvp1 Δ2-100 was eluted in buffer containing 250 mM NaCl. d Comparison of liposome binding by Mvp1 and Mvp1 Δ2-100. Protein (1.2 μM) was incubated without or with DOPS (PS) or DOPS + 5% PI3P (PS/PI3P) liposomes for 30 min at 21 °C prior to sedimentation. Shown is a representative result. P pellet, S supernatant. Positions of molecular weight markers are shown to the left of the gels. e Quantification of the results presented in d. Individual data points and mean ± s.d. are shown. n = 3 for Mvp1 and 4 for Mvp1 Δ2-100. A 3 × 2 factorial ANOVA was conducted to determine the effects of N-terminus (Mvp1 or Mvp1 Δ2-100) or the absence or presence of liposomes on the % sedimentation of protein. There was a significant interaction term (P < 0.0001) and each main effect was also significant (P < 0.0001). Selected pairs of values significant (Tukey HSD) at the 1% level are shown (**). For d and e, source data are provided as a Source Data file. f As in a but with Mvp1 Δ2-100. The inset shows an example of a rare subpopulation of highly twisted tubes observed only when PS/PI3P liposomes were incubated with Mvp1 Δ2-100. Scale bars—200 nm.

Fig. 2. The cryo-EM structure of the full-length Mvp1 tetramer.

a Single-particle cryo-EM reconstruction of the Mvp1 tetramer. The D2-symmetrized, sharpened (B = −252 Ų) map is shown at a contour level of 4σ. b Map, as in a, overlaid with a real-space refinement model of the Mvp1 tetramer. The upper Mvp1 SNX–BAR dimer is colored lighter and darker blue while the lower dimer is colored lighter and darker pink, to facilitate differentiation of the monomers. c Ribbon diagrams of the Mvp1 tetramer, colored as in b, viewed along each of the three twofold axes in the structure. d Surface models of the Mvp1 tetramer colored by electrostatic potential, colored from red (negative) to blue (positive). The scale is −25 to 25 kT/e. Upper row: the Mvp1 tetramer. Lower row: an isolated Mvp1 dimer. Note the sequestration of the positive charged concave surfaces into the interior of the tetramer. e Contact sites between the Mvp1 dimers responsible for tetramerization. Upper: map segment corresponding to the Mvp1 dimer with the contact sites between dimers colored in tomato. The map segment, extracted from the map presented in a and contoured at 4σ, is oriented with the concave surface of the BAR dimer facing the reader. Lower: ribbon model of the Mvp1 dimer with the contact sites colored in tomato, oriented as for the map segment. f Details of the three tetramerization regions. In these cases, the trans dimer is also shown, in pink, as in c.

Fig. 3. Mvp1 Mut1 is dimeric and is defective in lipid binding and CPY sorting in vivo.

The absolute molecular weights of Mvp1 Mut1 (a) or Mvp1 K198A (purple) or Mvp1 R199A, I200A (pink) (b) were determined using SEC-MALS as in Fig. 1. Proteins were eluted in buffer containing 250 mM NaCl. c Subcellular localization of EGFP-tagged Mvp1 and its mutants in Δmvp1 cells. Vacuoles were stained with the lipophilic dye FM 4–64. Images were obtained by confocal microscopy. Scale bar—5 μm. d CPY secretion assays. W303A or Δmvp1 cells expressing vector or the indicated construct were plated onto SD-Trp to maintain plasmid selection (left) or YPD (right) plates and incubated at 30 °C for 24 h. Plates were then overlaid with nitrocellulose for an additional 16 h. CPY secretion was detected by immunoblotting using an anti-CPY antibody. The leftmost spot in each case is 2 μl of a OD₆₀₀ = 0.5 culture: spots to the right of this are sequential fivefold dilutions. For d source data are provided as a Source Data file.

Fig. 4. Mvp1 Mut1 is defective in membrane binding and remodeling in vitro.

a Comparison of liposome binding by Mvp1 and Mvp1 Mut1. Protein (1.2 μM) was incubated without or with DOPS (PS) or DOPS + 5% PI3P (PS/PI3P) liposomes for 30 min at 21 °C prior to sedimentation. A representative result is shown. P pellet, S supernatant. Positions of molecular weight markers are shown to the left of the gels. b Quantification of the results presented in a. Individual data points and mean ± s.d. are shown. n = 9 for Mvp1 and 3 for Mvp1 Mut1. A 3 × 2 factorial ANOVA was conducted to determine the effects of the mutation (Mvp1 or Mvp1 Mut1) or the absence or presence of liposomes on the % sedimentation of protein. There was a significant interaction term (P = 0.0001) and both effects of liposome and Mvp1 or Mvp1 Mut1 were also significant (P < 0.0001 and P = 0.0002). A selected pair of values significant (Tukey HSD) at the 1% level is shown (**). Please note that the same data set for Mvp1 was used to compare with Mvp1 Mut1 here as was used to compare the different Mvp1 purification protocols presented in Supplementary Fig. 2b. For a and b, source data are provided as a Source Data file. c Mvp1 Mut1 is unable to tubulate PS liposomes containing 5% PI3P. Representative micrographs of negative-stained liposomes incubated either with Mvp1 or with Mvp1 Mut1 are shown. Scale bar—200 nm. d Mvp1 Mut1 is dimeric. Representative cryo-EM 3d class average of Mvp1 Mut1, with an approximate resolution of 8 Å. e Unassigned density. The D2-symmetrized, sharpened map, at contour level 4σ, is colored according to distance from the atomic model of the Mvp1 tetramer: color is assigned if the density is within 3 Å of the atomic model. The unassigned density is therefore shown in gray. f Model of membrane remodeling by Mvp1.