Methylation of ESCRT-III components regulates the timing of cytokinetic abscission

Abstract

Abscission is the final stage of cytokinesis, which cleaves the intercellular bridge (ICB) connecting two daughter cells. Abscission requires tight control of the recruitment and polymerization of the Endosomal Protein Complex Required for Transport-III (ESCRT-III) components. We explore the role of post-translational modifications in regulating ESCRT dynamics. We discover that SMYD2 methylates the lysine 6 residue of human CHMP2B, a key ESCRT-III component, at the ICB, impacting the dynamic relocation of CHMP2B to sites of abscission. SMYD2 loss-of-function (genetically or pharmacologically) causes CHMP2B hypomethylation, delayed CHMP2B polymerization and delayed abscission. This is phenocopied by CHMP2B lysine 6 mutants that cannot be methylated. Conversely, SMYD2 gain-of-function causes CHMP2B hypermethylation and accelerated abscission, specifically in cells undergoing cytokinetic challenges, thereby bypassing the abscission checkpoint. Additional experiments highlight the importance of CHMP2B methylation beyond cytokinesis, namely during ESCRT-III-mediated HIV-1 budding. We propose that lysine methylation signaling fine-tunes the ESCRT-III machinery to regulate the timing of cytokinetic abscission and other ESCRT-III dependent functions.

Fig. 1. SMYD2 localizes at the midbody and methylates the CHMP2B protein.

A Representative cell images of live HeLa cells during cytokinesis, which were transiently transfected with GFP_SMYD2 and treated with siR tubulin (magenta) to stain for tubulin cytoskeleton. GFP_SMYD2 localizes to the midbody (circle). Scale bar = 5 um (N = 3). B The Venn diagram showing common proteins found between the 489 proteins identified in remnant midbodies from HeLa cells and the 198 reproducible SMYD2 dependent Kme1 sites detected through SILAC analysis conducted on KYSE-150 cells. C HeLa cells were co-transfected with the constructs as indicated on the figure. Protein extracts were immunoprecipitated using a GFP antibody and the interaction between GFP_CHMP2B and Flag_SMYD2 was assessed by western blot (N = 3). D Recombinant SMYD2 (WT or catalytic dead Y240A mutant) were incubated with recombinant CHMP2B. Samples were then subjected to SDS-PAGE, and the total proteins were visualized using Coomassie staining. Gel was later dried and methylated proteins were detected by autoradiography (16 h exposure). The experiment displayed here is representative of 3 different replicates. E Tandem mass spectrometry (MS/MS) analysis showing mono-methylation of K6 on endogenous CHMP2B after immunoprecipitation from HeLa cells. Sequence of the chymotryptic endogenous mono-methylated CHMP2B peptide is displayed in inset. Spectrum of the corresponding synthetic CHMP2B peptide is shown inverted. The single- and double-charged fragment ions labeled in the spectrum represent cleavage of the peptide bond, recording sequence information from the N and C termini (b- and y-type ions, respectively). F Boxplot representation of log2 transformed intensities of peptide CHMP2B K6me1 after immunoprecipitation, gel purification and chymotrypsin digestion of GFP_CHMP2B co expressed with SMYD2 or SMYD2 Y240A in HeLa cells (N = 5 for GFP_CHMP2B co-expressed with Flag_SMYD2, and N = 4 for GFP_CHMP2B co-expressed with Flag_SMYD2 Y240A). The boxplot indicates the 25% (bottom), 50% (center) and 75% quartiles (top). Whiskers represent the minimum (bottom) and the maximum (top). Two-sided unpaired t-test. G Recombinant SMYD2 (WT or catalytic dead Y240A mutant) were incubated in the presence of recombinant CHMP2B (WT or K6A mutant). Samples were then subjected to SDS-PAGE, and the total proteins were visualized using Coomassie staining. Gel was later dried and methylated proteins were detected by autoradiography (16 h exposure). The experiment displayed here is representative of 3 different replicates.

Fig. 2. SMYD2 methylates CHMP2B K6 at the intercellular bridge of dividing cells.

A Immunodetection of CHMP2B K6 methylation on GFP_CHMP2B using CHMP2B K6me1 antibody in HeLa cells transfected with Flag_SMYD2 or SMYD2 Y240A and GFP_CHMP2B or CHMP2B K6A mutant as indicated on the figure (N = 3). B Protein extracts from HeLa cells treated with DMSO or 10 uM of BAY-598 (SMYD2 inhibitor) were immunoprecipitated using CHMP2B antibody and the CHMP2B K6 methylation was assessed by western blot using CHMP2B K6me1 antibody (N = 3). C Left panels: Representative cell images of CHMP2B K6me1 (green) and alpha-tubulin (magenta) immunostaining in HeLa cells treated with Ctrl or SMYD2 siRNAs. The white circle shows the midbody. Scale bar = 5 um. Right panel: quantification of the CHMP2B K6me1 signal at the midbody of HeLa cells treated with Ctrl or SMYD2 siRNAs as indicated (N = 3, n = 125 for siCtrl cells and n = 140 siSMYD2 cells midbodies quantified by condition, mean ± SD, Two-sided Mann–Whitney test). D Western blots of protein extracts from HeLa treated with control (siLuciferase referred to siCtrl) or SMYD2 siRNAs, revealed with SMYD2 and tubulin antibodies (N = 3). Loading controls: tubulin and Ponceau red staining. E Left panels: Representative cell images of CHMP2B K6me1 (green) and alpha-tubulin (magenta) immunostaining in HeLa cells treated with DMSO or 10 uM of BAY-598. The white circle shows the midbody. Scale bar = 5 um. Right panel: quantification of the CHMP2B K6me1 signal at the midbody of HeLa cells treated with Ctrl or SMYD2 siRNAs as indicated (N = 3, n = 121 for DMSO and n = 125 for BAY-598 midbodies quantified by condition, mean ± SD, Two-sided Mann–Whitney test).

Fig. 3. CHMP2B K6 is involved in the timely progression of CHMP2B to the abscission site.

A Western blots of protein extracts from HeLa cell lines stably expressing GFP or siRNA resistant (siR) CHMP2B_GFP or CHMP2B K6A_GFP treated with control or CHMP2B siRNAs revealed with CHMP2B antibody (N = 3, CHMP2B exo. = CHMP2B_GFP protein, CHMP2B endo. = endogenous CHMP2B). Loading control: Ponceau red staining. B Selected frames from phase-contrast time-lapse videomicroscopy of HeLa cells stably expressing GFP treated with control or CHMP2B siRNAs entering in cytokinesis and completing their abscission. Time is expressed in minutes, and time 00 is set at the time point prior the intercellular bridge formation. Last time point corresponds to the ICB cut indicated by the empty arrow, considered here as the abscission (N = 3). Scale bar = 10 um. C Distribution of the abscission time measured by phase-contrast time-lapse microscopy in the indicated cells (N = 3, n > 109 cells, Two-sided Kolmogorov–Smirnov test, ns non-significant (p > 0.05)) and the mean abscission duration (minutes, N = 3, mean ± SD, t-test, ns non-significant (p > 0.05)) are shown. D Quantification of the abscission time of HeLa cell lines stably expressing CHMP2B_GFP or CHMP2B K6A_GFP (from furrow onset to the microtubule cut), each dot represents one dividing cells. (N = 3, n > 50, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test). E Left panel: live cell imaging of HeLa cell lines stably expressing CHMP2B_GFP (top) or CHMP2B K6A_GFP (bottom) treated with siR-tubulin (magenta). Time is expressed in minutes, and time 00 is set at the time point prior the intercellular bridge formation. The last time point corresponds to the microtubule cut, considered here as the abscission. Scale bar = 5 um. Right panel: duration of each abscission stage measured from live cell imaging performed in CHMP2B_GFP or CHMP2B K6A_GFP expressing cells (N = 3, n > 50, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test, ns non-significant (p > 0.05)).

Fig. 4. SMYD2 regulates the localization of CHMP2B and the timing of abscission.

A Western blots of protein extracts from HeLa cell lines stably expressing CHMP2B_GFP and Cherry_SMYD2 treated for 24 h with DMSO or SMYD2 inhibitor (BAY-598 10 µM) were blotted for CHMP2B K6me1 or CHMP2B antibodies (N = 3). Loading control: Ponceau red. B Abscission time (from furrow onset to the microtubule cut) of HeLa cell line stably expressing CHMP2B_GFP and treated as indicated. (N = 3, n > 65 cells, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test). C Western blots of protein extracts from HeLa cell line stably expressing CHMP2B_GFP treated with control or SMYD2 siRNAs were blotted as indicated (N = 3). Loading control: Tubulin and Ponceau red. D Abscission time was determined as in (B). (N = 3, n > 75, median (red lines) and quartiles (dotty lines), Mann–Whitney test). E Left panels: live cell imaging of CHMP2B_GFP HeLa cell line treated as indicated. Time 00 is set prior the intercellular bridge formation. Last time point corresponds to the microtubule cut. Scale bar = 5 um. Right panel: Duration of each abscission stage (N = 3, n > 65 cells, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test, ns non-significant (p > 0.05)). F Left panels: live cell imaging of CHMP2B_GFP HeLa cell line treated as indicated. Time 00 is set prior the intercellular bridge formation. Last time point corresponds to the microtubule cut. Scale bar = 5 um. Right panel: Duration of each abscission stage (N = 3, n > 75 cells, median, Two-sided Mann–Whitney test, ns non-significant (p > 0.05)). G Western blots of protein extracts from CHMP2B K6A_GFP HeLa cell line treated with control or SMYD2 siRNAs were blotted as indicated (N = 3). Loading controls: Tubulin and Ponceau red staining. H Abscission time of CHMP2B K6A_GFP HeLa cell line treated with control or SMYD2 siRNAs was determined as in (B). (N = 3, n > 41 cells, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test, ns non-significant (p > 0.05). I Left panels: live cell imaging of CHMP2B K6A_GFP HeLa cell line treated as indicated. Time 00 is set prior the intercellular bridge formation. Last time point corresponds to the microtubule cut. Scale bar = 5 um. Right panel: Duration of each abscission stage (N = 3, n > 41 cells, median (red lines) and quartiles (dotty lines), Two-sided Mann–Whitney test, ns non-significant (p > 0.05)). J Recombinant CHMP3, CHMP2B-∆C and CHMP2B-∆C K6me1 proteins were purified and visualized using Coomassie (left) and Ponceau red staining (right). CHMP2B K6 methylation was assessed by western blot using CHMP2B K6me1 antibody (N = 2). K Left panel: Sorting coefficient plot showing preferential assembly of fluorescently labeled CHMP3 on positively curved membranes when in the presence of methylated CHMP2B. (N = 2, median (red lines) and quartiles (black lines), Two-sided t-test, (p < 0.05)). Right panel: Representative fluorescence micrograph used to calculate sorting coefficient values showing recruitment of fluorescently labeled CHMP3 on pulled lipid tubes from GUVs. Magenta is membrane signal, cyan is CHMP3 signal. Scale bar = 5 um.

Fig. 5. SMYD2 accelerates abscission timing when cytokinesis is challenged.

A Western blots of protein extracts from HeLa cell lines stably expressing GFP or SMYD2_GFP treated with control or Nup153 siRNAs (N = 3). Loading control: Ponceau red. B Left panels: representative cell images. Yellow arrows show cytokinetic cells connected by an ICB. Scale bar = 10 um. Right panel: quantification of cytokinetic cells (N = 3, n > 300 cells, mean ± SD, Two-sided multiple unpaired t-test, ns non-significant (p > 0.05)). C Left panels: frames of the indicated HeLa cell lines. Time 00 is set prior the ICB formation. Last time point corresponds to the ICB cut, indicated by the arrow. Scale bar = 10 um. Right panel: distribution of the abscission time measured by phase-contrast time-lapse microscopy (N = 3, n > 126 cells, Two-sided Kolmogorov–Smirnov test, ns non-significant (p > 0.05)) and mean abscission duration (N = 3, mean ± SD, t-test, ns non-significant (p > 0.05)) are shown. D Quantification of ICBs with either no CHMP2B, CHMP2B in rings at the midbody, and CHMP2B both at the midbody and at the abscission site (N = 5, n > 40 ICBs counted/N, mean ± SD, Two-sided multiple unpaired t-test, ns non-significant (p > 0.05)). E Quantification of ICBs with either no VPS4A, VPS4A at midbody rings, and VPS4A both at the midbody and at the abscission site (N = 3, n > 50 ICBs counted/N, mean ± SD, Two-sided multiple unpaired t-test, ns non-significant (p > 0.05)). F Left panels: selected phase contrast time-lapse and fluorescent videomicroscopy frames of HeLa cell lines with Lap2_ β_RFP that stains for chromatin bridges (indicated by 3 arrows). Last time point corresponds to the ICB cut. Scale bar = 10 um. Right panel: distribution of the abscission time measured by phase-contrast time-lapse microscopy plotted as individual points. (n = 74, n = 18, for GFP cells with or without chromatin bridge and n = 53; n = 62 for GFP-SMYD2 cells with or without chromatin bridge, N = 3 experiments, mean ± SD, Two-sided Mann–Whitney test, ns non-significant (p > 0.05)). G Left panels: selected frames of HeLa cell lines treated with Rab35 siRNA. Time 00 is set prior the ICB formation. Last time point corresponds to the ICB cut. Scale bar = 10 um. Right panel: distribution of the abscission time measured by phase-contrast time-lapse microscopy (N = 3, n > 166 cells, Two-sided Kolmogorov–Smirnov test, ns non-significant (p > 0.05)) and mean abscission duration (N = 3, mean ± SD, t-test, ns non-significant (p > 0.05)).

Fig. 6. CHMP2B K6 and its methylation contribute to HIV particles release.

HeLa cells expressing GFP, CHMP2B_WT-GFP, CHMP2B K6R-GFP and CHMP2B K6A-GFP were transfected with pNL4-3 HIV-1 pro-viral DNA. Forty-eight hours later, cells were lysed and virus contained in supernatants were pelleted by ultracentrifugation. Viral production was evaluated by western blot and by ELISA quantification of HIV-1 CAp24. A Western blot analysis of HIV-1 Gag and CAp24 products, GFP and GAPDH in transfected cells and pelleted viruses. B HIV-1 release index was calculated as the ratio between released CAp24 and cell-associated CAp24. Results were normalized to control cells (set as 100%). C The virus titer was scored by infection of HeLa P4R6 indicator cells, followed by β-galactosidase activity quantification in the cells. (Statistical analysis using two-tailed unpaired Student’s t test, ns non-significant (p > 0.05), N = 4, mean ± SD. All western blots are representative of at least three independent experiments). D HeLa cells transfected with indicated siRNA were transfected with pNL4-3 HIV-1 pro-viral DNA. Forty-eight hours later, cells were lysed, and virus contained in supernatants were pelleted by ultracentrifugation. Viral production was evaluated by western blot and by ELISA quantification of HIV-1 CAp24. Western blot analysis of HIV-1 Gag and CAp24 products, SMYD2 and GAPDH in transfected cells and purified viruses. E HIV-1 release index was calculated as the ratio between released CAp24 and cell-associated CAp24. Results were normalized to control cells (set as 100%). F The virus titer was scored by infection of HeLa P4R6 indicator cells, followed by β-galactosidase activity (Statistical analysis using two-tailed unpaired Student’s t test, N = 4, mean ± SD. All western blots are representative of at least three independent experiments).