Targeted comparative RNA interference analysis reveals differential requirement of genes essential for cell proliferation

Abstract

Differences in the genetic and epigenetic make up of cell lines have been very useful for dissecting the roles of specific genes in the biology of a cell. Targeted comparative RNAi (TARCOR) analysis uses high throughput RNA interference (RNAi) against a targeted gene set and rigorous quantitation of the phenotype to identify genes with a differential requirement for proliferation between cell lines of different genetic backgrounds. To demonstrate the utility of such an analysis, we examined 257 growth-regulated genes in parallel in a breast epithelial cell line, MCF10A, and a prostate cancer cell line, PC3. Depletion of an unexpectedly high number of genes (25%) differentially affected proliferation of the two cell lines. Knockdown of many genes that spare PC3 (p53-) but inhibit MCF10A (p53+) proliferation induces p53 in MCF10A cells. EBNA1BP2, involved in ribosome biogenesis, is an example of such a gene, with its depletion arresting MCF10A at G1/S in a p53-dependent manner. TARCOR is thus useful for identifying cell type-specific genes and pathways involved in proliferation and also for exploring the heterogeneity of cell lines. In particular, our data emphasize the importance of considering the genetic status, when performing siRNA screens in mammalian cells.

Figure 1.

TARCOR analysis in human cells. (A) Experimental scheme of TARCOR analysis of genes essential for cell proliferation. (B) BrdU ELISA in MCF10A cells. Values are mean ± SD (n = 3). Black and white columns represent negative (GL2) and positive (ORC2) controls, respectively. (C) mRNA levels of a standard gene, ORC2, in MCF10A and PC3. Cells were transfected with control (GL2) or ORC2 siRNAs for 48 h, and the relative amount of ORC2 mRNA was measured by real-time PCR. Amount of ORC2 mRNA was normalized to that of β-actin. Values are mean ± SD (n = 3). (D and E) Elimination of biologically nonreproducible genes. Inhibition indices from two screens (Experiments 1 and 2) in MCF10A (D) and PC3 (E) are plotted. Dashed lines show the cutoff for elimination of biologically nonreproducible genes. Eliminated genes are shown by white circles.

Figure 2.

Comparison of genes essential for proliferation between MCF10A and PC3. (A and B) Scatter plot of inhibition indices of selected genes in MCF10A cells (A) and PC3 (B). (C) Scatter plot of inhibition indices of selected genes in MCF10A and PC3. Average of two biological replicates in each cell line were plotted. Genes (n = 172) that showed reproducible inhibition indices in both MCF10A and PC3 are plotted in A–C. (D) A heat map of a hierarchical cluster analysis of the inhibition indices of 172 genes. Yellow and blue represent inhibition and stimulation of BrdU incorporation after RNAi, respectively. Clusters of genes whose knockdown have differential effects between MCF10A and PC3 are indicated by green and red lines.

Figure 3.

mRNA levels after RNAi-mediated gene knockdown. (A–L) MCF10A and PC3 cells were transfected with control (GL2) or indicated siRNAs for 48 h. Relative amount of target mRNAs was quantitated by real-time PCR and shown after normalization to β-actin mRNA levels. *Cell line where RNAi did not inhibit BrdU incorporation.

Figure 4.

p53 protein levels after knockdown of genes that affect MCF10A but not PC3. (A) MCF10A cells were transfected with indicated siRNAs for 48 h and cell lysates analyzed by Western blotting for p53 and β-actin (loading control). (B) Quantitation of Western blot performed in triplicate. p53 signal normalized with β-actin is shown.

Figure 5.

Analysis of cells after RNAi against genes involved in ribosome biogenesis. (A) FACS analysis of DNA content after RNAi against EBNA1BP2 (indicated as EBP2) and DDX21. MCF10A was transfected with indicated siRNAs and examined by FACS after 48 h. (B) Western blot analysis of p53 and p21 after RNAi (48 h) against EBNA1BP2 and DDX21in MCF10A. (C) Cyclin E–associated kinase activity after EBNA1BP2 RNAi. Cyclin E protein levels in cell lysates and incorporation of ³²P in the substrate (pRB-C) in the in vitro kinase assay is shown. (D) p53-dependent p21 induction after EBNA1BP2 RNAi. MCF10A cells were transfected with GL2 or p53-1 siRNAs for 48 h followed by transfection with GL2 or EBNA1BP2–1 siRNAs (48 h). Cell lysates were analyzed by Western blotting for indicated proteins. (E) p53-dependent G1 arrest in EBNA1BP2-depleted cells. MCF10A cells treated as in D were analyzed by FACS, and the percentage of cells in the indicated phases of the cell cycle is shown.