Human SKI component SKIV2L regulates telomeric DNA-RNA hybrids and prevents telomere fragility

Abstract

Super killer (SKI) complex is a well-known cytoplasmic 3'-5' mRNA decay complex that functions with the exosome to degrade excessive and aberrant mRNAs, is implicated with the extraction of mRNA at stalled ribosomes, tackling aberrant translation. Here, we show that SKIV2L and TTC37 of the hSKI complex are present within the nucleus, localize on chromatin and at some telomeres during the G2 cell cycle phase. In cells, SKIV2L prevents telomere replication stress, independently of its helicase domain, and increases the stability of telomere DNA-RNA hybrids in G2. We further demonstrate that purified hSKI complex binds telomeric DNA and RNA substrates in vitro and SKIV2L association with telomeres is dependent on DNA-RNA hybrids. Taken together, our results provide a nuclear function for SKIV2L of the hSKI complex in overcoming replication stress at telomeres mediated by its recruitment to DNA-RNA hybrid structures in G2 and thus maintaining telomere stability.

Keywords: Cell biology; Molecular biology.

Graphical abstract

Figure 1

SKIV2L of the hSKI complex is present at telomeres in G2 (A) Subcellular fractionation assay in asynchronous HeLa1.3 cells. (B) Proteomics of isolated chromatin segments analysis showing the binding of hSKI to telomeres throughout the cell cycle. Tables are listing the number of unique peptide numbers isolated, including the fold change values of unique peptides normalized to the asynchronous values (top panel) and the relative LFQ intensity values identified by PICh. AS: asynchronous. Scr: scramble. (C) WB of hSKI (SKIV2L, TTC37, and WDR61) and SKIV2L2 in shCtr (Control) or shSKIV2L HeLa1.3 cells. (D) Immunofluorescence of pre-extracted cells showing co-localization of SKIV2L and TTC37 with TRF2 in asynchronous (AS) and G2-synchronized cells. % of total number of SKIV2L or TTC37 foci colocalizing with TRF2 are normalized to shCtr AS samples (means ± SEM, n = 2 independent experiments, scale bar 15 μm). t test ∗p < 0.05, ∗∗∗p < 0.001. (E) Proximity ligation assay of SKIV2L-TRF2, showing increasing number of foci in G2 synchronized cells (median, Q1 and Q3, 707 (AS) and 638 (G2) cells scored per condition, 3–4 independent experiments, scale bar 10 μm). Mann-Whitney U test ∗∗∗∗p < 0.0001. (F) ChIP-dot blot of SKIV2L and TRF1 in AS and G2-synchronized shCtr and shSKIV2L cells (means ± SEM, n = 3). t test ∗p < 0.05, ∗∗∗∗p < 0.0001. See also Figure S1.

Figure 2

SKIV2L and TTC37 prevents telomere fragility (A and B) Telomere FISH analysis in SKIV2L- and TTC37-depleted HeLa1.3 and HT1080-ST cells using siRNAs: % of telomere fragility (yellow) and loss (purple) per metaphase in siCtr, siSKIV2L and siTTC37 (means ± SD, n > 25 metaphases, 2 independent experiments, scale bar, 10 μm). t test ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. (C) Telomere FISH analysis in shCtr (Control) or shSKIV2L HEK293 cells expressing GFP or SKIV2L (means ± SD, n > 25 metaphases, 2 independent experiments, scale bar, 10 μm). Details are as in (A). (D) Telomere FISH analysis in TTC37-depleted HEK293 cells using siRNAs (means ± SD, n = 25 metaphases). Details are as in (A). (E) Telomere FISH analysis in HeLa1.3 cells using shRNAs: shCtr and shSKIV2L treated with DMSO (−, control) or APH (+) (n > 45 metaphases, 2 independent experiments). Details are as in (A). (F) IF of pre-extracted cells showing co-localization of pS1981 ATM autophosphorylation with TRF2 (telomeres) in shCtr and SKIV2L-depleted (shSKIV2L) HeLa1.3 cells. Quantification of the percentage of cells with co-localization foci and mean number of foci per nucleus is depicted (means ± SEM, n = 200 cells, 2 independent experiments, scale bar 15 μm). t test ∗∗∗p < 0.001. (G) IF-FISH of pre-extracted cells showing co-localization of 53BP1 with telomeres in asynchronous (AS) or G2-synchronized control (shCtr) and SKIV2L-depleted (shSKIV2L) HeLa1.3 cells. Quantification of the percentage of cells with co-localization foci is depicted (means ± SEM, n > 400 cells, 3 independent experiments, scale bar 15 μm). t test ∗p < 0.05. See also Figure S2.

Figure 3

SKIV2L helicase activity is dispensable for the suppression of telomere fragility (A) Schematic diagram to illustrate the domain organization and features of human SKIV2L. The structurally resolved (via cryo-EM) regions of the SKIV2L subunit extracted from the RNA-bound human SKI complex in the closed state (PDB: 7QDY) is represented by a green colored bar. The SKIV2L subunit harbors the following domains: Ski2 N-terminal (Ski2 N) domain, DEAD/DEAH box helicase (DEAH) domain, ATP binding (ATP) domain, Helicase C-terminal (Helicase C) domain, rRNA-processing arch (rRNA proc-arch) domain and DOB1/SK12/helY-like DEAD box helicases C-terminal (DSHCT) domain. The SKIV2L-V341 residue is highlighted in yellow. (B) Overview of the structure of the RNA-bound human SKI complex in the closed state, featuring: the helicase SKIV2L, tetratricopeptide repeat protein 37 (TTC37) and WD repeat-containing protein 61 (WDR61). (C) Telomeric FISH analysis: % of telomere fragility in shCtr and shSKIV2L HEK293 cells expressing GFP or SKIV2L (WT), SKIV2L-K338R, SKIV2L-D423A (means ± SD, n > 20 metaphases, 2 independent experiments). t test ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, significance is not shown for comparisons with p > 0.06. (D) Comparison of the structure of the wild-type (WT) SKIV2L DEAD domain and a mutant structure, generated using Missense3D, featuring a valine to glycine 341 mutant. In the WT structure, V341 forms hydrophobic contacts with: A318, L322, V328 and I451. The hydrophobic surface shown is colored according to the Eisenberg hydrophobicity scale. (E) Telomeric FISH analysis: % of telomere fragility in shCtr and shSKIV2L HEK293 cells expressing GFP or SKIV2L-V341G (means ± SD, n > 20 metaphases, 2 independent experiments). t test ∗∗∗∗p < 0.0001. See also Figure S3.

Figure 4

Purified recombinant hSKI binds preferentially ssDNA and telomeric DNA-RNA hybrids containing 3′ overhangs in vitro (A) hSKI constructs. MBP, Maltose binding protein; His, 6x histidine; Flag, 3x flag. (B) Coomassie blue SDS-PAGE gel showing purified hSKI (lane 10) and different purification fractions. PP, PreScission protease. (C) Electrophoretic mobility shift assays showing binding of hSKI to different RNA and DNA substrates. Blue lines denote RNA, black lines denote DNA and asterisk indicates radioactive ³²P label. hSKI protein amounts are as indicated (nM) (D) Quantification of (C) showing % of nucleic acid binding calculated as protein bound substrate signal relative to free substrate signal (mean ± SEM, n = 2–4). (E) Electrophoretic mobility shift assays showing binding of hSKI complex to different 3′ and 5′ overhang RNA/DNA substrates. Blue lines denote RNA, black lines denote DNA and asterisk indicates radioactive ³²P label. (F) Quantification of (E) showing % of nucleic acid binding calculated as protein bound substrate signal relative to free substrate signal (mean ± SEM, n = 2). See also Figure S4.

Figure 5

SKIV2L regulates telomeric DNA-RNA hybrids in cellulo to prevent telomere fragility (A) Proximity ligation assay (PLA) showing co-localization of SKIV2L and TRF2 in asynchronous (AS) and in G2-synchronized HeLa1.3 cells with and without RNase H1 (RNH1) overexpression (median, Q1 and Q3, at least 600 cells scored per condition, 4 independent experiments, scale bar 10 μm). Mann-Whitney U test ∗∗∗∗p < 0.0001. (B) S9.6 IF in HeLa1.3 cells treated with RNAse III, with and without RNase H1 (RNH1) overexpression (median ± interquartile range, 360 cells scored per condition, 4 independent experiments, scale bar, 10 μm). Mann-Whitney U test ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (C) PLA showing the co-localization of DNA-RNA hybrids (S9.6) and TRF2 in HeLa1.3 cells pre-extracted and treated with RNAse III (median, Q1 and Q3, at least 400 cells scored per condition, 2 independent experiments, scale bar 10 μm). Mann-Whitney U test ∗∗∗∗p < 0.0001. (D) DRIP showing the levels of DNA-RNA hybrids at telomeres in AS and G2-synchronized HeLa1.3 cells, RNH, RNase H treatment (means ± SEM, n = 4). (E) DRIP-qPCR assay of G2-synchronized HEK293 cells overexpressing GFP or SKIV2L at 10q, 13q, 20q, and 22q subtelomeric regions. RNH, RNase H treatment (means ± SEM, n = 5). Percent input values were normalized to the GFP overexpressing condition. (F) Model proposing the function of hSKI at telomeres. Telomeric DNA-RNA hybrid accumulation in late S/G2 phase drives hSKI recruitment to telomeres to regulate physiological DNA-RNA hybrid levels, prevent telomere replication stress and ensure telomere stability. See also Figure S5.