Fig 2: RNase H2 activity is required for productive LINE retrotransposition in HeLa and U2OS cells Schematic of retrotransposition vectors Zfl2-2mneoI (zebrafish LINE-2) and JM101/L1.3 (Human L1.3). The relative position of the EN domain (endonuclease), RT domain (reverse transcriptase) and C domain (cysteine-rich), if present, is indicated. The purple box with a backward NEO label depicts the retrotransposition indicator cassette mneoI. Underneath, representative retrotransposition assays conducted in HeLa parental cells, control (C2-6) and RNASEH2A-KO (KO2-6) clones (used for quantifications in Figure 3C). Cells were transfected with an active human LINE-1 (WT-hL1, element L1.3), an RT-mutant human LINE-1 (RTm-hL1, D702A) or an active zebrafish LINE-2 (WT-zL2, element Zfl2-2) vector.Representative retrotransposition and toxicity assays conducted in parental U2OS cells, a control clone (C1) and two RNASEH2A-KO clones (KO1 and KO2). Cells were transfected with an active human LINE-1 (WT-hL1, element L1.3), an RT-mutant human LINE-1 (RTm-hL1, D702A) or an active zebrafish LINE-2 (WT-zL2, element Zfl2-2) vector. Right, quantification of WT-hL1 (dark grey bars) and WT-zL2 (light grey bars) retrotransposition in U2OS cells, with retrotransposition in parental cells for both elements set at 100% for comparison. Mean ± SD for n = 3 technical replicates (representative of three independent experiments).

Fig 3: Complementation of RNASEH2A-KO cells with wild-type, but not separation of function RNASEH2A rescues LINE-1 retrotransposition A, BRNase H activity assays against single-embedded ribonucleotides using recombinant purified proteins (WT-RNase H2 and SoF-RNase H2). Note that only WT-RNase H2 shows activity in this assay. Plotted, mean ± SEM for three independent experiments.C, DRNase H activity assays against RNA:DNA heteroduplexes using recombinant purified proteins (WT-RNase H2 and SoF-RNase H2). Note that the pattern of products generated by SoF-RNase H2 is different from the wild-type pattern. Plotted, mean ± SEM for three independent experiments.EWestern blot analysis of RNase H2 expression in RNASEH2A-KO HeLa cells complemented with the indicated retroviral vector (EV, empty vector; WT, wild-type RNASEH2A; SoF, RNASEH2A-P40D/Y210A; CD, RNASEH2A-D34A/D169A, see main text for details). Tubulin was used as a loading control.FRNase H activity against single-embedded ribonucleotides in RNASEH2A-KO cells is only rescued by wild-type RNASEH2A (KO1 + WT). DRD:DNA heteroduplex (18 bp; ribonucleotide-containing strand 3'-labelled) was incubated with increasing amounts of whole-cell lysate from the indicated cell line and separated by denaturing PAGE. The graph shows mean values ± SEM for three independent experiments.GRNase H activity against RNA:DNA heteroduplexes in RNASEH2A-KO cells is rescued by wild-type (KO1 + WT), not by CD RNASEH2A (KO1 + CD) or the empty vector (KO1 + EV). Note reduced activity and the difference in cleavage pattern produced by SoF RNASEH2A (KO1 + SoF). RNA:DNA heteroduplex (18 bp; RNA strand 3'-labelled) was incubated with increasing amounts of whole-cell lysate from the indicated cell line and separated by denaturing PAGE. Plotted, mean ± SEM for three independent experiments.HOnly wild-type RNASEH2A rescues the LINE-1 retrotransposition defect in RNASEH2A-KO cells. Left, representative retrotransposition and toxicity assays conducted in the four complemented lines. Cells were transfected with vectors containing an active human LINE-1 (WT-hL1, L1.3), an RT-mutant LINE-1 (RTm-hL1, D702A) or a toxicity control plasmid (CTRL, pcDNA 6.1). Right, quantification of L1-WT retrotransposition. For comparison, the retrotransposition level in KO1 cells complemented with the empty vector (EV) was set at 100%. Mean ± SD for n = 3 technical replicates (representative of six independent experiments). Source data are available online for this figure.

Fig 4: Cellular RNase H activity against RNA:DNA hybrids is important for L1 retrotransposition AIncreased L1 retrotransposition upon RNase H2 SoF overexpression. Left, representative retrotransposition and toxicity assays. HeLa cells were transfected with vectors containing an active LINE-1 (WT-hL1, L1.3), alongside an expression vector for ß-arrestin as a negative control (-ve), the three RNase H2 subunits (RNase H2, at a 1:1:1 ratio; with RNASEH2A-WT or SoF) or a plasmid expressing APOBEC3A, known to restrict LINE-1 retrotransposition, as a positive control (+ve). Right panel, quantification of this retrotransposition assay, with the level in cells co-transfected with ß-arrestin set at 100% for comparison. Values were normalised for transfection efficiency and toxicity. Mean ± SD for n = 2 technical replicates (representative of three independent experiments). Unpaired two-sided t-test; *P < 0.05.BRNase H activity against RNA:DNA heteroduplexes in RNASEH2A-KO cells (clones KO1 and KO2) is partially rescued by overexpression of human nuclear RNase H1 (KO1 + H1 and KO2 + H1 vs. KO1 + EV and KO2 + EV). RNA:DNA heteroduplex (18 bp; RNA strand 3'-labelled) was incubated with whole-cell lysate and speed of cleavage determined using a FRET-based assay. Mean values ± SEM for n = 6 independent experiments.CRNase H activity against single-embedded ribonucleotides in RNASEH2A-KO cells (clones KO1 and KO2) is not rescued by human RNase H1 (KO1 + H1 and KO2 + H1 vs. KO1 + EV and KO2 + EV). Speed of cleavage of an 18-bp substrate was determined using a FRET-based assay. Mean values ± SEM for n = 3 independent experiments.D, ERepresentative retrotransposition and toxicity assays (D) conducted in the RNASEH2-KO clones (KO1 and KO2) complemented with RNASEH2A-WT (+RNASEH2A) or with human RNase H1 (+RNASEH1). Cells were transfected with vectors containing an active LINE-1 (WT-hL1, L1.3), an RT-mutant LINE-1 (RTm-hL1, D702A), or a toxicity control plasmid (CTRL, pcDNA6.1). Quantification of L1-WT retrotransposition (E). For comparison, the retrotransposition level in KO1 or KO2 cells was set at 100%. Mean ± SD for n = 3 technical replicates (representative of three independent experiments). Unpaired two-sided t-test; **P < 0.01; ***P < 0.001FProposed model for TPRT. Degradation of LINE-1 RNA in the RNA:cDNA hybrid by cellular RNase H2 allows completion of LINE-1 insertion.

Fig 5: Complementation of RNASEH2A-KO HeLa cells ASchematic of substrates used in RNase H activity assays. These assays either use an 18-bp RNA:DNA hybrid (left), or a short dsDNA containing a single-embedded ribonucleotide (DRD:DNA). RNASEH2A-WT can cleave both with high efficiency (++++), whereas RNASEH2A-CD (with D34A and D169A mutations) cannot cleave either (-). The separation of function mutant (SoF, with P40D and Y210A mutations) retains some activity against RNA:DNA heteroduplexes (++), but has virtually no activity against single-embedded ribonucleotides (-).BRNase H2 SoF has reduced activity against RNA:DNA heteroduplexes and does not fully process the hybrid, even at high concentration and/or long incubation times. RNase H activity was measured using the 18-bp RNA:DNA substrate, separating products by denaturing PAGE after cleavage with RNase H2. WT, SoF and CD RNase H2 were used at 0.25 nM for 4 h (left) or 2.0 nM for 2 h (right). Note the different pattern of products generated for SoF and WT.CAs expected, the high levels of genome-embedded ribonucleotides in RNASEH2A-KO cells are rescued only by complementation with wild-type RNASEH2A (+WT), not by SoF RNASEH2A (+SoF), CD RNASEH2A-A (+CD) or the empty vector (+EV). Genomic DNA was isolated from parental cells, a control clone (C1) and the four complemented cell lines (+EV, +WT, +SoF and +CD), RNase H2 treated and separated by alkaline gel electrophoresis. Smaller fragments indicate larger numbers of embedded ribonucleotides.D–GRepresentative gels (used for quantifications in Figure 5F and G) with results from RNase H activity against single-embedded ribonucleotides (D and E) and activity against RNA:DNA heteroduplexes (F and G) assays conducted with lysates from the indicated cell lines. Because RNase H1 is expressed in all of these cells, activity measured against RNA:DNA heteroduplex substrate in RNASEH2A-KO cell lysates is not completely absent. In addition, other nucleases present in the cell lysate act (non-specifically) on the substrate, causing further background activity on both substrates.HWild-type RNASEH2A rescues the LINE-1 retrotransposition defect in RNASEH2A-KO2 cells. Representative retrotransposition and toxicity assays conducted in RNASEH2A-KO2 cells and RNASEH2A-KO2 complemented with wild-type RNASEH2A (+WT). Cells were transfected with an active human LINE-1 (WT-hL1, element L1.3), an RT-mutant LINE-1 (RTm-hL1, D702A), or a toxicity control plasmid (CTRL, pcDNA6.1). Right panel, quantification of L1-WT retrotransposition. For comparison, the retrotransposition level in KO2 cells was set at 100%. Mean ± SD for n = 3 technical replicates (representative of three independent experiments).IRNase H2 SoF has reduced RNA:DNA heteroduplex substrate affinity. Initial substrate conversion rates (Vi) by 0.1 nM recombinant RNase H2 were measured at different 18-mer RNA:DNA substrate concentrations. Mean ± SEM for n = 3 independent experiments. Km and kcat ± SEM were calculated in GraphPad Prism 5.04, using non-linear regression. Change for SoF compared to WT indicated between brackets.