The histone methyltransferase SETD2 negatively regulates cell size

Abstract

Cell size varies between cell types but is tightly regulated by cell intrinsic and extrinsic mechanisms. Cell size control is important for cell function, and changes in cell size are frequently observed in cancer. Here, we uncover a role for SETD2 in regulating cell size. SETD2 is a lysine methyltransferase and a tumor suppressor protein involved in transcription, RNA processing and DNA repair. At the molecular level, SETD2 is best known for associating with RNA polymerase II through its Set2-Rbp1 interacting (SRI) domain and methylating histone H3 on lysine 36 (H3K36) during transcription. Using multiple independent perturbation strategies, we identify SETD2 as a negative regulator of global protein synthesis rates and cell size. We provide evidence that overexpression of the H3K36 demethylase KDM4A or the oncohistone H3.3K36M also increase cell size. In addition, ectopic overexpression of a decoy SRI domain increased cell size, suggesting that the relevant substrate is engaged by SETD2 via its SRI domain. These data add a central role of SETD2 in regulating cellular physiology and warrant further studies on separating the different functions of SETD2 in cancer development.

Keywords: Cell size; Histone methyltransferase; SETD2; Translation.

Fig. 1.

SETD2 depletion increases cell size and total protein content of RPE1 cells. (A) Western blot of RPE1 cells with doxycycline (dox)-inducible knockdown of SETD2 using miRNAs (#1, #2) or a non-targeting (NT) miRNA. Cells were treated with doxycycline for 72 h. The amount of cell lysates added to the gel were normalized for total protein content measured by a Lowry assay. The bar plot below the left panel represents genomic DNA levels quantified by qPCR in protein normalized lysates (mean±s.d.; n=3). (B) Quantification of western blot signals in A. Error bars in represent mean±s.d. of three biological replicates. (C) Same lysates as in A, but the amount of cell lysates added to the gel were now normalized for genomic DNA content. (D) 2D cell size (mean±s.d.; n=3) as measured by imaging flow cytometry of RPE1 cells with inducible miRNA-mediated SETD2 depletion. Cells were treated with dox for 72 h for inducible miRNA-based SETD2 knockdown (red). (E) Western blot of RPE1 cells with dox-inducible expression of RfxCas13d and stable expression of guide RNAs (gRNAs; #1, #2) targeting SETD2 or a NT gRNA. (F) Cell volume after RfxCas13d-mediated SETD2 knockdown (+dox condition). (G) Total protein content after RfxCas13d-mediated SETD2 knockdown as measured by flow cytometry of fixed and permeabilized RPE1 cells stained with Zombie NIR amine reactive fluorescent dye. MFI, mean fluorescent intensity. Cells were treated with dox for 96 h in E, F and G. Data in F,G is from four independent biological replicates. (H) Whole-cell lysate western blot of mouse embryonic stem cells (mESCs) with Setd2 endogenously tagged with dTAG. (I) Cell volume of Setd2-dTAG mESCs following SETD2 depletion. Data represents percentage of peak cell volume increase relative to untreated control (day 0); n=3. See Fig. S2 for cell volume changes in femtoliter for individual biological replicates. *P<0.05; **P<0.01; NS, not significant (one-way ANOVA with a Tukey's multiple comparisons test). Blots shown in C are representative of three repeats; those in E and H are representative of two repeats. A.U., arbitrary units.

Fig. 2.

Inducible depletion of SETD2 increases protein synthesis rates but not ploidy, RNA synthesis rates or cell cycle distribution. (A) Chromosome counts in pro-metaphase spreads after RfxCas13d-mediated SETD2 knockdown or non-targeting (NT) control knockdown. Cells were treated with dox for 96 h. Red line indicates the median. (B) Total RNA synthesis rates as measured by 4SU labeling for 20 min. Equal amounts of RNA were loaded for each sample and 4SU labeled RNA was detected by dot blotting (see Materials and Methods). Longer labeling time increases 4SU signal, whereas treating cells with Actinomycin D (ActD), which is a DNA intercalator that blocks RNA polymerases, reduces the signal. ActD was added at the same time as 4SU and might not act immediately. Cells were treated with dox for 96 h before RNA labeling. (C) [³⁵S]Methionine incorporation assay for RPE1 cells 96 h following dox-induced RfxCas13d-based SETD2 knockdown. The [³⁵S]methionine autoradiography signal was normalized for the total protein content of each condition. CHX indicates a control experiment in which cells were treated with cycloheximide for 1 h, which inhibits protein synthesis. (D,E) Cell cycle distribution as measured by flow cytometry of DAPI-stained SETD2-depleted RPE1 cells (D) and mESCs (E). Cells were treated with dox for 96 h in D. Error bars in B–E are mean±s.d. of three biological replicates. *P<0.05; **P<0.01; n.s., not significant (two-tailed unpaired t-test).

Fig. 3.

Expression of cell cycle related genes following SETD2 depletion. (A) RT-qPCR for mRNA expression analysis of genes involved in cell cycle regulation in RPE1 cells, 96 h following doxycycline-induced RfxCas13d-based SETD2 knockdown or non-targeting (NT) control knockdown. (B) The two gRNAs used for targeting SETD2 mRNA are each flanked by a RT-qPCR primer pair used for SETD2 expression analysis in A. Error bars represent mean±s.d. of three biological replicates.

Fig. 4.

Mechanistic insights into cell size control by SETD2. (A) Western blot of RPE1 cells constitutively overexpressing the yeast demethylase Rph1 or human H3K36me3/H3K9me3 demethylase KDM4A. The amount of cell lysates added to the gel were normalized for total protein (left panel) or genomic DNA content (right panel). The bar plot below the left panel represents genomic DNA levels quantified by qPCR in protein-normalized lysates. (B) Western blot of RPE1 cells with doxycycline inducible overexpression of hemagglutinin (HA) epitope-tagged ‘onco’ H3.3 histones. The bar plot represents genomic DNA levels quantified by qPCR in protein normalized lysates. Dots represent the individual values of two biological replicates and blots shown are representative of two repeats (for A,B). (C) Western blot of RPE1 cells with stable overexpression of H3.3–HA and H3.3K36A–HA, with quantification of relative H3K36me3 signal (total signal on H3.3–HA and endogenous H3). Dots represent the individual values of two biological replicates. The arrows 1 and 2 in B and C highlight H3K36me3 on ectopic H3.3–HA and on endogenous H3, respectively. (D) Cell volume of H3.3K36A-overexpressing cells (n=3). P-value was calculated using one-way ANOVA with a Tukey's multiple comparisons test.

Fig. 5.

Ectopic overexpression of the SETD2 SRI domain inhibits H3K36me3 and increases cell size. (A) Cartoon to illustrate how overexpression of a ‘decoy’ SRI domain might specifically interrupt SETD2 activity toward H3K36 (as well as other SRI-dependent SETD2 substrates). (B–D) Live-cell microscopy (B), whole-cell lysate western blotting (C), and cell volume measurements (D) of RPE1 cells expressing RFP-HA-NLS-hSRI or RFP-HA-NLS-hSRI-R2510H. Cells were transduced with empty vector (EV) or hSRI constructs, selected with puromycin for three days and analyzed six days post-transduction. Scale bars: 55 µm. Images shown in B,C are representative of two repeats. Dots in D are independent biological replicates (n=3). **P<0.01; n.s., non-significant (one-way ANOVA with a Tukey's multiple comparisons test). (E) Western blot analysis of ectopically overexpressed SETD2 SRI domains immunoprecipitated from HEK293T cells. HEK293T cells were transiently transfected with HA-NLS-RFP (control), HA-NLS-SRI or HA-NLS-SRI-R2510H encoding plasmids. Cells were lysed 48 h after transfection, and RFP or SRI domains were immunoprecipitated with anti-HA antibody. Blot is representative of two repeats.