CD95/Fas ligand mRNA is toxic to cells

Figures

• 

  Figure 1.. The CD95L mRNA is toxic to cells. (A) Left: Schematic of the different CD95L mutants used. Right: Percent cell confluence over time of HeyA8 parental cells in the absence (left panel) or in the presence of 20 μM zVAD-fmk (center panel), or CD95 k.o. cells (right panel) after expression of CD95L constructs. Data are representative of one to three independent experiments. Values were calculated from samples done in triplicate or quadruplicate shown as mean ±SE. (B) Left: Western blot analysis of HeyA8 cells overexpressing different CD95L mutant RNAs. Cells expressing CD95LMUT or CD95L were pretreated with 20 μM zVAD-fmk. Note the small amount of truncated CD95L in cells infected with CD95L MUTNP does not have CD95 binding activity. Very similar data were obtained when the constructs were expressed in either CD95 k.o. HeyA8 cells (clone #11) or NB-7 cells, which lack expression of caspase-8, both without treatment with zVAD (data not shown). Right: RT-qPCR analysis for CD95L of the same samples. Data are representative of two independent experiments. Each bar represents mean ±S.D. of three replicates. (C, D) Quantification of cell death (C) and ROS production (D) in CD95 k.o. HeyA8…

• 

  Figure 1—figure supplement 1.. Generation of complete CD95 k.o. MCF-7 cells. (A) Schematic of the genomic locations and sequences of the gRNAs used to excise the entire CD95 gene in MCF-7 cells. PAM sites are underlined. (B) PCR with flanking (top panels) and internal (bottom panels) primers used to confirm the absence of the CD95 gene in MCF-7 clones. Parental (Par.) cells and three clones infected with Cas9 only (Cas9) and two homozygous complete CD95 k.o. clones (F2 and FA4) are shown. (C) RT-qPCR analysis of the indicated clones using primers spanning either exon 1/2 or exon 2/3 of the CD95 gene. (D) Surface staining for CD95 of one wt and one k.o. clone. (E) Western blot analysis of all clones.

• 

  Figure 1—figure supplement 2.. Mutation of the alternative start codon in CD95LMUTNP construct. (A) Left: Schematic of the different CD95L mutants used. For the M86I mutant protein the point mutation on the DNA level is in parentheses. (B) Western blot analysis of HeyA8 CD95 k.o. cells overexpressing different CD95L mutants. Note, for unknown reasons in this experiment expression of the CD95LMUTNP protein was more efficient than in the experiment shown in Figure 1B. (C) Percent cell confluence over time of HeyA8 CD95 k.o. (cl. #11) cells after expression of the different CD95L constructs. Values were calculated from samples plated in triplicate shown as mean ±SE.

• 

  Figure 1—figure supplement 3.. Toxicity of CD95L mRNA is independent of CD95L protein expression. (A) Schematic showing the positions of the codon optimized silent mutations of the CD95LSIL mutant compared to wild type CD95L (mutations highlighted in blue). (B) Percent cell confluence over time of HeyA8 CD95 k.o. cells (cl. #11) over-expressing empty pLenti vector (vec), wild-type CD95L (from two separately cloned viruses), or the CD95LSIL. (C) RT-qPCR analysis and Western blot (inset) of wild type CD95L and CD95LSIL mutant mRNAs in the over-expressing cells shown in B. (D) Probability density plot comparing the toxicity of all possible 6mer seeds located in either the WT or SIL CD95L mRNA. p-value was calculated using a two-sample two-sided K-S test.

• 

  Figure 2.. Toxicity induced by CD95L overexpression is reminiscent of DISE. (A) Phase-contrast images of HeyA8 and HeyA8 CD95 k.o. cells (cl. #11) after infection with pLKO-shScr/shL3 or pLenti (vec)/pLenti-CD95L, respectively, at the indicated time point. (B) Gene set enrichment analysis for the 1846 survival genes (top panel) and the 416 nonsurvival genes (bottom panel) identified in the Sabatini study (Putzbach et al., 2017; Wang et al., 2015) of mRNAs downregulated in CD95L expressing HeyA8 CD95 k.o. cells compared to HeyA8 CD95 k.o. cells infected with pLenti virus. p-values indicate the significance of enrichment. (C) Common genes downregulated in all RNA-Seq experiments from (HeyA8) cells treated with either one of four si/shRNAs (Putzbach et al., 2017) derived from either CD95 or CD95L (see Figure 2D) and cells overexpressing CD95L ORF as described in (B). (D) Metascape analysis of 5 RNA Seq data sets analyzed. The boxed GO term clusters were highly enriched in all five data sets.

• 

  Figure 3.. Small RNAs generated in cells expressing CD95L mRNA are loaded into the RISC. (A) Percent cell confluence over time of HCT116 parental (left) or Drosha k.o. (right) cells after infection with CD95MUTNP. Data are representative of three independent experiments. Each data point represents the mean ±SE of three replicates. Inset: Phase contrast images of Drosha k.o. cells 9 days after infection with either empty vector or CD95LMUTNP. (B) Percent cell confluence of HeyA8 CD95 k.o. cells transfected with either non-targeting siRNA (siCtr) or a pool of 4 siRNAs targeting AGO2 following subsequent infection with either empty pLenti (vec) or pLenti CD95L. Inset: Western blot showing knock-down of human AGO2. (C) Top: autoradiograph on RNAs pulled down with the Ago binding peptide. Bottom: Western blot analysis of pulled down Ago proteins. v, pLenti; L, pLenti-CD95L expressing cells. (D) Pie charts showing the relative ratio of sRNAs pulled down with the Ago proteins in wt and Drosha k.o. cells. Depicted are all the amounts of all sRNAs that contributed at least 0.01% to the total RNA content. Only in the Drosha k.o. cells was a significant amount of CD95L derived Ago bound reads found.…

• 

  Figure 3—figure supplement 1.. All CD95L mRNA mutants are toxic through RNAi. (A) Percent cell confluence of HeyA8 CD95 k.o. cells (cl. #11) transfected with either non-targeting siRNAs (siCtr) or a pool of 4 siRNAs targeting AGO2 following subsequent infection with either empty pLenti (vec), pLenti CD95LMUTNP (left), or pLenti CD95LSIL (right). (B) Percent cell confluence of parental HeyA8 cells transfected with either a pool of 4 non-targeting siRNA (siCtr) or a pool of 4 siRNAs targeting AGO2 following subsequent infection with either empty pLenti (vec), pLenti CD95L WT (left), or pLenti CD95LSIL (right). (C) Percent cell confluence of parental HeyA8 cells transfected with either a pool of 4 non-targeting siRNA (siCtr) or a pool of 4 siRNAs targeting AGO2 following subsequent infection with either empty pLenti (vec), or pLenti CD95LMUTNP. In A cells were plated after puromycin selection. In (B) and (C) cells were plated directly after viral infection and puromycin was added 50 hr after infection (arrow).

• 

  Figure 4.. The entire CD95L mRNA gives rise to sRNAs that bind to the RISC. (A) Length distribution of CD95L derived reads in various analyses. (B, C) Read alignment with CD95LMUTNP ORF of analyses of sRNAs pulled down with Ago proteins from HCT116 wt (B, top) and Drosha k.o. (B, bottom) cells and of total sRNAs from HCT116 Drosha k.o. cells (C) after infection with CD95LMUTNP. (D, E) Read alignment with wt CD95L ORF of analyses of sRNAs pulled down with Ago proteins (D) or total sRNAs (E) from HeyA8 CD95 k.o. cells after infection with wt CD95L. (F) Percent cell confluence over time of HCT116 parental (top) or Dicer k.o. (clone #43) (bottom) cells after infection with CD95MUTNP. (Dicer k.o. clone #45, gave a similar result, data not shown). Data are representative of two independent experiments. Each data point represents the mean ±SE of three replicates. (G) RT-qPCR analysis of clusters 8 and 21 in HCT116 parental, Dicer k.o. (clone #43), and Drosha k.o. cells after infection with CD95MUTNP. Each bar represents mean ± S.D. of three replicates. v, vector, L, CD95L expressing cells.

• 

  Figure 4—figure supplement 1.. Predicted secondary structure of CD95L ORF and toxicity of CD95L-derived sRNAs after conversion to siRNAs. (A) The CD95LMUTNP RNA was subjected to a RNA secondary structure analysis (http://rna.tbi.univie.ac.at) using default settings. The locations of 22 reads representative of the 22 read clusters are shown. Regions with potential duplex formation are boxed. The oligonucleotides that were found to be toxic when expressed as siRNAs are circled. (B) Toxicity of the eight siRNAs designed using the CD95L-derived small RNA fragments bound to Ago as the antisense strand sequences 96 hr post-transfection in the indicated cell lines. Each data point represents the mean ±SE of three replicates.

• 

  Figure 4—figure supplement 2.. CD95L fragments are less toxic than full length CD95L mRNA. (A) Schematic of the different CD95L fragments used. (B) Left: Western blot analysis of HeyA8 CD95 k.o. cells overexpressing different CD95L mutant RNAs. Right: RT-qPCR analysis for CD95L of the same samples using primers detecting either the 5' or the 3' half of the mRNA. Data are representative of two independent experiments. Each bar represents mean ±S.D. of three replicates. (C) Percent cell confluence over time of HeyA8 CD95 k.o. cells after expressing the CD95L mutant or fragments. Data are representative of two independent experiments. Values were calculated from samples done in triplicate shown as mean ±SE. (D) Left: Western blot analysis of HCT116 Drosha k.o. cells overexpressing different CD95L mutant RNAs. Right: RT-qPCR analysis for CD95L of the same samples using primers detecting either the 5' or the 3' half of the mRNA. Data are representative of two independent experiments. Each bar represents mean ±S.D. of three replicates. (E) Percent cell confluence over time of HCT116 Drosha k.o. cells after expressing the CD95L mutant or fragments. Data are representative of two independent experiments. Values were…

• 

  Figure 5.. Endogenous mRNAs are processed and loaded into the RISC. (A) Length distribution of reads derived from five of the top ten most abundant genes loaded into the RISC of CD95L expressing HCT116 Drosha k.o. cells. The numbers in parentheses indicate the ranking in the top ten most abundant genes with Ago bound reads. (B) Alignment of the reads from the five genes shown in A with horizontal blue lines representing the mapped positions of the reads. Each blue line represents an individual read, with its length in the plot proportional to the read length. Small RNAs pulled down with Ago proteins (top) or total sRNAs (bottom) from HCT116 and Drosha k.o. cells after infection with wt CD95L. *This stack contains 14899 reads of which 3000 were randomly chosen and plotted. (C) All 4262 genes in HCT116 Drosha k.o. cells expressing CD95L ranked according to highest expression with more than 10 reads expressed as reads per kb per million (RPKM). The abundance of the six genes shown in A and B is labeled. (D) Genes ranked according to highest abundance in the RISC of Drosha k.o. cells. Reads derived from the five genes in A are labeled as well as the location of the reads derived from CD95L. (E)…

• 

  Figure 5—figure supplement 1.. Endogenous mRNAs are processed and loaded into the RISC - additional genes. (A) Length distribution of reads derived from 5 of the top ten most abundant genes loaded into RISC of CD95L expressing HCT116 Drosha k.o. cells. The numbers in brackets indicate the ranking in the top ten most abundant genes with Ago bound reads. (B) Alignment of the reads from the five genes shown in A with their respective mRNAs derived from small RNAs pulled down with Ago proteins (top) or total small RNAs (bottom) from HCT116 and Drosha k.o. cells after infection with WT CD95L.