MYH9 mutations differentially stabilize non-muscle myosin II filaments and induce distinct cellular aggregation phenotypes

Figures

• 

  Fig. 1. NM2-A N93K is a bona fide molecular motor with ATPase and in vitro actin motility activity. A Left, model of an assembly-competent (6S) NM2-A, highlighting the functional domains of the heavy chain as indicated. This conformation is induced by phosphorylation of the RLC in Ser19 when NM2-A is in the compact, filament assembly-incompetent (10S) conformation (top right). Extended NM2-A monomers quickly form bipolar filaments (bottom right). Dephosphorylation of filamentous NM2-A triggers filament disassembly and induces a rapid return to the 10S conformation. B Actin-activated steady-state ATPase activity of NM2-A WT (grey) and NM2-A N93K (orange). The measured k100 of NM2-A N93K and NM2-A WT are 0.07 ± 0.008 s−1 and 0.09 ± 0.098 s−1, respectively. The kcat and Kapp lie outside the assessable F-actin concentration range and thus could not be determined. n = 3. C Inhibition of the actin-activated steady-state ATPase (100 µM F-actin) of NM2-A WT (grey) and NM2-A N93K (orange) by blebbistatin. Blebbistatin inhibits the enzymatic activity of NM2-A N93K and NM2-A WT with an IC50 of 1.33 ± 0.25 µM and 2.7 ± 0.85 µM, respectively. n = 3. D The in vitro gliding speed of actin filaments…

• 

  Fig. 2. A molecular chimera of MHCII-B and -A devoid of ATPase activity does not restore cellular contractility but accumulates at the rear of polarized cells. (A) Cartoon depicting the wild type and chimeric proteins used in Fig. 2B-G. Blue represents NM2-A and violet is NM2-B. Mutations assessed in these experiments (N93K and R709C) are also indicated. (B-C) Representative COS7 cells were transfected with EGFP-fusion proteins as indicated in left, allowed to adhere to fibronectin for > 4 h and stained for F-actin (B) or paxillin (C). In (C), yellow outline is the estimated position of the leading edge based on the rim of nascent adhesions that define the lamellipodium, as reported elsewhere [29]. Note the proximity to the leading edge in wild type (WT), N93K and the BH-AT chimera, and the posterior accumulation of the BHR709C-AT chimera. Scale bar = 20 µm. (D-F) Quantification of the percentage of adhesive area (D), number of adhesions per µm.2 (E), and area per individual adhesion (F). Data was calculated as indicated in Material and Methods. n > 55 cells from three independent experiments. Significance was calculated using non-parametric Mann Whitney test. * p < 0.05; ** p < 0.01;…

• 

  Fig. 3. Mutations causing MYH9-RD in the head or the assembly domain of MHCII-A display a similar subcellular distribution but different filament kinetics. A COS7 cells were transfected with Halo-MHCII-A wild type (WT) and EGFP-MHCII-A carrying the indicated mutation (indicated as GFP). After 24 h, cells were allowed to spread for 90 min on fibronectin (2 µg/mL)-coated dishes, fixed and stained for endogenous MHCII-B. In the magenta/green composite images, Halo-MHCII-A WT is shown in magenta, and each NM2-A variant is green. In the green/red composite images, the NM2-A variant (or WT) is green and endogenous MHCII-B is red. Representative cells are shown. Scale bar = 10 µm. (B-D) NM2-A N93K and NM2-A E1841K, but not NM2-A R705H, display increased stability within actomyosin filaments. B FRAP curves of EGFP-MHCII-A decorated actomyosin filaments. COS-7 cells, transfected with either EGFP-MHCII-A WT or EGFP-MHCII-A bearing the indicated mutations were allowed to adhere on fibronectin-coated (8 μg/mL) coverslips for 2 h and then imaged. FRAP was performed always in comparable regions across mutants in well-delineated fibers. Data is the average of > 50 filaments from three independent…

• 

  Fig. 4. UNC45a displays an increased association with NM2-A N93K in COS7 cells. A COS-7 cells were co-transfected with pEGFP-N1 or UNC45-A-(E)GFP plasmids and FLAG-MHCIIA WT or N93K as indicated. After 48 h, cells were lysed and incubated with GFP-trap beads and separated using SDS-PAGE. Top panels represent input; bottom panels represent the immunoprecipitated fraction. B Densitometric quantification of experiments as in (A), corrected by the amount of UNC45a-(E)GFP detected in the immunoprecipitated fraction. Data represents the mean ± SD of nine independent experiments. * p < 0.05. C COS7 cells were co-transfected with UNC45a-(E)GFP and the indicated FLAG-tagged MHCII-A construct. After 48 h, cells were lysed and incubated with anti-FLAG-conjugated magnetic beads and separated using SDS-PAGE. Top panels represent input; bottom panels represent the immunoprecipitated fraction. D Densitometric quantification of experiments as in (D), corrected by the amount of FLAG-MHCII-A detected in the immunoprecipitated fraction. Data represents the mean ± SD of ten independent experiments. ** p < 0.01. E IBAQ-based quantification of the interaction of endogenous UNC45a with WT and N93K MHCII-A.…

• 

  Fig. 5. In silico modeling of the autoinhibited state of NM2-A N93K mutant. Homology model of NM2-A in the autoinhibited state suggests altered interactions between N93K in the blocked head and R1417 in segment 2 of the folded tail. Notably, the respective mutation in the free head does not affect interactions with the tail or other portions of the protein

• 

  Fig. 6. MYH9-RD-causing NM2-A variants form aggregates with distinct molecular composition in COS7 cells. A Representative immunofluorescence images of COS7 cells transfected with Halo-tagged-MHCII-A variants, adhered to fibronectin for 4 h and stained with anti-MHCII-A. In all panels, JF635 denotes the Halo-ligand used for visualization. Blue arrows/arrowheads point to aggregates visualized with Halo-ligand; magenta arrows/arrowheads point to aggregates visualized with anti-MHCII-A. Scale bar = 20 µm. Top cartoon represents the localization of Halo tag and the antibody binding site with respect to NM2-A, in the N-terminus and C-terminus, respectively. B Quantification of cytoplasmic aggregates in COS7 cells. Data is the mean ± SD of four independent experiments (n > 50 in each case). C Representative immunofluorescence images of COS7 cells co-transfected with Halo-tagged-MHCII-A variants and UNC45a-(E)GFP. Cells were stained with AlexaFluor568-conjugated phalloidin to visualize F-actin. Overlay represents the localization of the NM2-A variant and UNC45a-(E)GFP. Yellow arrows and arrowheads indicate NM2-A N93K and R705H aggregates, respectively. Scale bar = 20 µm. D Representative…

• 

  Fig. 7. MYH9-RD-causing NM2-A variants induce different types of aggregates in MEG-01 cells. A FCS gating strategy for flow cytometry sorting. MEG-01 cells were co-transfected with mNeonGreen-MHCII-A WT and Halo-MHCII-A WT, N93K, R705H or E1841K. After 48 h, cells were incubated with JF635, sorted by flow cytometry and allowed to adhere to fibronectin (25 µg/mL)-coated coverslips, fixed and imaged using STED. X axis represents mNeonGreen levels; Y axis represents Halotag-JF635 levels. Low expressors (red box) and high expressors (green box) were collected separately and used as indicated. B Representative STED images of MEG-01 cells collected from the low expressors box. Scale bar = 5 µm. Red boxes represent the magnified region shown in the zoom panels. C Representative STED images of MEG-01 cells collected from the high expressors box. Scale bar = 5 µm. In the composite images, magenta represents the Halo-tagged mutant and green is WT. Yellow boxes represent the magnified region shown in the zoom panels. D Quantification of the percentage of cells displaying aggregates in experiments as in (B, C). Each dot represents a count of at least 50 cells in 5–10 fields across three independent…

• 

  Fig. 8. UNC45a co-localizes with amorphous aggregates containing NM2-A N93K but is mostly excluded from worm-like aggregates containing NM2-A E1841K. A Representative images of MEG-01 cells transfected as indicated in Fig. 7, collected from the green box in Fig. 7A (high expressors) and stained for endogenous UNC45a. Scale bar = 5 µm. In composite images, red represents the Halo-mutant, green is WT and blue is endogenous UNC45a. Colored boxes on right represent the zoomed regions indicated in the composite images. Note the color pattern, which is similar in wild type (first row) and R705H (third row), with white (co-localization) dominating N93K (second row), and the segregation of red (E1841K) and blue (UNC45a) in the fourth row. B Pearson correlation was calculated using the Coloc2 algorithm in Fiji in images as in (A). Data represents 4–5 non-overlapping regions from 4–5 cells (high level of expression, as described in Fig. 7A) per condition. * p < 0.05; **** p < 0.0001; n.s., not significant

• 

  Fig. 9. Analysis of particle intensity, density and degree of aggregation by STED-ICS in MEG-01 cells expressing MYH9-RD-causing NM2-A variants and mNeonGreen-MHCII-A wild type reveals diverse degrees of sub-resolution aggregation. Images similar to those shown in Fig. 7B (low expressors) were analyzed by ICS as described in Material and Methods. Data represents the intensity (A-D), degree of aggregation (E–H) and particle density (I-L) of Halo-MHCII-A WT (A, E, I), N93K (B, F, J), R705H (C, G, K) and E1841 (D, H, L) against mNeonGreen-MHCII-A WT in the same region. (M) represents the comparison of the degree of aggregation of the Halo-tagged mutants from one condition to another (columns on left in panels E–H), whereas (N) represents mNeonGreen-MHCII-A WT in cells expressing the indicated Halo-tagged mutant (columns on right in panels E–H). Analysis was performed on 70–120 120 × 120 pixels ROIs of 20–30 MEG-01 STED images per condition from three independent experiments. Statistical significance was determined with the nonparametric Mann–Whitney U test. ** p < 0.01; *** p < 0.001, **** p < 0.0001, n.s., not significant. In (M), * means significance vs. WT; # represent significance vs.…