Article Figures & Data
Figures
- Figure 1. Reconstituted Xenopus Lnp forms stacked membrane discs.
(A) Domain organization of Lnp. CC1, CC2, coiled-coil domains 1 and 2; TM1 and TM2, transmembrane segments 1 and 2; P, phosphorylation domain; Zn2+, Zn2+-finger domain. Domains are colored according to the degree of sequence conservation. (B) Purified His10-tagged Xenopus Lnp was reconstituted with phospholipids at a protein-to-lipid molar ratio of 1:5 and visualized by negative-stain EM. The lower panel shows a magnified view. Scale bar, 20 nm. (C) As in (B), but with different protein-to-lipid ratios. Scale bar, 20 nm.
- Figure S1. Purity of Lnp and its derivatives.
(A) SEC of His10-tagged Xenopus Lnp. Shown is the absorbance at 280 nm. (B) As in (A), but for human Lnp. The second peak originates from the detergent Triton X-100 used for the purification of the protein. (C) Purified wild-type and mutant Lnp proteins were analyzed by SDS-PAGE and Coomassie blue staining. Lines in between lanes indicate slices that were merged from different gels. The proteins shown all contain the two TM segments. For purity of the cytosolic fragments, see Wang et al (16).
- Figure S2. Lnp protein requires phospholipids to form stacked discs.
(A) Proteoliposomes containing full-length Xenopus Lnp were subjected to Nycodenz gradient centrifugation. Fractions (F1–F6) were collected from the top and analyzed by SDS-PAGE and Coomassie blue staining. (B) Non-reconstituted, purified full-length Xenopus Lnp was visualized by negative-stain EM at different concentrations. Scale bar, 20 nm. (C) Purified full-length Xenopus Lnp was reconstituted with phospholipids at a 1:5 protein-to-lipid ratio. Detergent was added after reconstitution and the sample was imaged by negative-stain EM. Scale bar, 20 nm.
- Figure S3. Untagged Lnp proteins form stacked discs.
Full-length Xenopus Lnp, lacking the His tag, was reconstituted with phospholipids. The sample was analyzed by negative-stain EM. Scale bar, 20 nm.
- Figure S4. Lnp antibodies disrupt the formation of stacked bicelles.
(A) Purified full-length Xenopus Lnp was reconstituted with phospholipids at a 1:5 protein-to-lipid ratio and analyzed by negative-stain EM. Scale bar, 20 nm. (B) As in (A), but affinity-purified Xenopus Lnp antibodies were added after reconstitution at a molar ratio of 1:1 with respect to Lnp protein. (C) As in (B), but with affinity-purified Xenopus Lnp antibodies preincubated with a 2.5-fold excess of cytLnp protein before addition to the reconstituted sample. (D) As in (B), but with affinity-purified Xenopus ATL antibodies. (E) As in (B), but with affinity-purified Xenopus Rtn4 antibodies.
- Figure 2. Reconstituted human Lnp forms stacked discs.
(A) Human Lnp was expressed in E. coli as a His10-tagged protein, purified, and reconstituted with phospholipids at different protein-to-lipid ratios. The samples were visualized by negative-stain EM. Scale bar, 20 nm. (B) As in (A), but with untagged human Lnp expressed in, and purified from, S. cerevisiae. (C) As in (A) with His-tagged protein reconstituted with lipid at a ratio of 1:20. Shown is a magnified view of a branch in which one disc is connecting two stacks (arrow). (D) As in (C), but with a branch point containing a ring-like structure (arrow). (E) As in (A) with a protein-to-lipid ratio of 1:200, showing two examples in which neighboring discs are laterally connected (arrows).
- Figure 3. Cryo-ET of stacks.
(A) Human Lnp was reconstituted with phospholipids at a 1:5 molar ratio. The sample was analyzed by cryo-ET at −3 μm defocus. Shown is a slice through the tomogram. Scale bar, 50 nm. (B) As in (A), but at a higher zoom level. Scale bar, 20 nm. (C) 3D rendering of the stack shown in (B), generated by subtomogram averaging of 115 discs. The discs are shown in light yellow and the central low-density region in dark yellow. (D) Plot of the density profile of three individual discs (in orange, blue, and green). The distance between the high-density regions of a disc is about 5 nm. (E) As in (A), but with human Lnp reconstituted at a 1:20 protein-to-lipid ratio and with images taken at −5 μm defocus. The plot of the density profile shows that the discs are separated by 12 nm and are connected by low-density material. (F) Model of Lnp-induced bicelles, with the lipid bilayer in yellow and the detergent monolayer at the edge in orange. Lnp molecules sit in each membrane disc facing opposite directions. They interact through the P and Zn2+-finger domains (Figs 4, 5, 6, and 7).
- Figure 4. Interference of Lnp domains with the formation of stacked bicelles.
(A) Purified full-length Xenopus Lnp was mixed with phospholipids and a five-fold molar excess of purified cytoplasmic domain of Lnp (cytLnp) or buffer. After removal of detergent with Bio-Beads, the sample was visualized by negative-stain EM. Scale bar, 20 nm. (B) As in (A), except that cytLnp was added after reconstitution of full-length Lnp. (C) As in (A), except that cytLnp was replaced by a C-terminal fragment containing the Zn2+-finger. (D) As in (A), except that cytLnp was replaced by an N-terminal fragment containing the CC2 and P domains. (E) As in (A), except that cytLnp was replaced by an N-terminal fragment containing only the P domain.
- Figure 5. The coiled-coil domains of Lnp are required for bicelle stacking.
(A) Purified full-length human Lnp or an Lnp mutant carrying an insertion after TM2 was reconstituted with phospholipids at a 1:5 ratio. The samples were visualized by negative-stain EM. Scale bar, 20 nm. (B) As in (A), but with either wild-type Xenopus Lnp or a mutant that contains several point mutations in CC1 and CC2 (M21I, I25L, L28I, F31I, N35L, L38I, T102I, L112I, K116L, V123L, and K126I). The control on the left shows the same image as in Fig 1B. (C) As in (B), but with a mutant that contains different mutations in CC1 and CC2 (L18I, M21I, L28I, F31I, N35I, L38I, T102I, L109I, L112I, K116I, V123I, and K126I).
- Figure 6. The Zn2+-finger domain is not essential for the formation of stacked bicelles.
(A) Purified Lnp lacking the Zn2+-finger domain (ΔI275–F302) was reconstituted with phospholipids at a 1:5 protein-to-lipid ratio and the sample was analyzed by negative-stain EM. Scale bar, 20 nm. (B) As in (A), but with Lnp lacking both the Zn2+-finger and C-terminal regions (ΔM235–E428). (C) Purified SBP-tagged full-length human Lnp was incubated with either His10-tagged full-length human Lnp or Lnp lacking the Zn2+-finger domain at a molar ratio of 1:3 or 1:5. The samples were incubated with streptavidin resin, washed, and eluted with biotin. They were analyzed by SDS-PAGE followed by Coomassie blue staining. An aliquot of the input material was analyzed in the lower panels.
- Figure 7. Lnp-mediated stacked bicelle formation requires the P domain.
(A) Purified full-length human Lnp or Lnp lacking the P domain (ΔL134–G234) was reconstituted with phospholipids at a 1:5 protein-to-lipid ratio and the samples were analyzed by negative-stain EM. Scale bar, 20 nm. (B) As in (A), but with full-length Xenopus Lnp or Lnp carrying phosphomimetic mutations in mitotic phosphorylation sites (16) (S177E, S179E, S188E, S192E, T198E, S206E, S215E, T219E, S222E, and S231E) in the P domain (XlLnp-E-His10). (C) Quantification of the average number of discs per stacked structure for samples shown in (A) and (B). The data plotted reflect the mean ± standard deviation from three independent experiments. About 100 stacked structures per sample were quantified in each independent experiment.
