M. Kenrick, S. Hancock, S. Stubbs, and N. Thomas
GE Healthcare, The Maynard Centre, Cardiff, UK
Recent advances in siRNA methodologies and the development of high-throughput image analysis platforms such as the IN Cell Analyzer have revolutionized the functional analysis of genes and proteins. Here, we describe the application of two stable cell lines expressing green fluorescent protein (GFP)◊cell cycle sensors to screen a library of siRNAs directed against key cell cycle control genes. Imaging of GFP intensity and distribution within these two cell lines allows cell cycle position to be assigned by automated image analysis procedures and permits their use in the screening of drugs that block the cell cycle.
Recent developments in RNA interference (RNAi) and small interfering RNA (siRNA) techniques for specifically modulating gene expression in a diverse range of cells and organisms (1, 2, 3) have revolutionized the functional analysis of genes and proteins. Advances in synthetic and virally encoded siRNA methodologies (4, 5) have now reached a stage where large scale RNAi screens can be applied to mammalian cells (6, 7, 8). In addition to the provision of large numbers of validated siRNAs, efficient mammalian siRNA functional screens will require information-rich model systems that allow abstraction of multi-parameter data at a level of throughput compatible with large-scale projects. Fortunately, advances in the capabilities of siRNA have been matched by the development of sophisticated fluorescence imagers and software capable of imaging and analyzing cellular events in live cells at high-throughput (9, 10). Such instrumentation enables study of complex systems by combining data from fluorescent cellular sensors with morphological parameters to provide a detailed description of the phenotypic effects of siRNAs in cellular scr eens.
In this study we have used two stable cell lines expressing green fluorescent protein cell cycle sensors (Fig 1) to screen a library of siRNAs directed against key cell cycle control genes.
One cell line (11, 12) reports on G2 phase to M phase cell cycle transition via dynamic expression and degradation of an EGFP sensor that shadows endogenous cyclin B1 levels. The G2/M cell cycle phase marker (CCPM) is switched on in late S phase, switched off at the end of mitosis, and in the intervening period translocates from the cytoplasm to the nucleus.
The second cell line reports on G1 phase to S phase transition via translocation of an EGFP fusion protein incorporating the phosphorylation-dependent sub-cellular location domain (PSLD) of DNA helicase B (13). In this cell line a nuclear localization sequence (NLS) within the PSLD retains the fusion protein within the nuclei of G1 cells. Phosphorylation of serine967 within the PSLD by CDK2/cyclin E during S phase unmasks a previously inactive nuclear export sequence that predominates over the NLS leading to export of the fusion protein into the cytoplasm in S phase cells.
Imaging of GFP intensity and distribution within these two cell lines allows cell cycle position to be assigned by automated image analysis procedures, and permits their use in the screening of drugs that block the cell cycle (Fig 2).
Cell cycle gene knockdowns were carried out using a Dharmacon siARRAY™ of 112 siRNA pools each comprising four siRNAs directed against a single cell cycle related gene. Additional scrambled sequence and Cy™5 labeled siRNAs were used as controls and to determine transfection efficiency. siRNAs were transfected into G2/M CCPM and G1/S CCPM expressing U2OS cells in 96-well plates using Lipofectamine™ 2000 (Invitrogen). Cells were transfected with siRNAs at 1, 5, 50, and 200 nM for 4 h followed by a media change and further incubation for 24–48 h post transfection. Cells were fixed in 4% formalin and nuclei stained with DRAQ5™ (BioStaus).
Cellular imaging was performed on an IN Cell Analyzer 1000 using a 20× objective and 475BP20/535BP50 (GFP) and 620BP60/700BP75 (DRAQ5) excitation/emission filters. Image stacks were converted to IN Cell Analyzer 3000 format and the resulting images analyzed for cell number, cell cycle distribution, and morphology.
Cellular transfection efficiency was determined using a Cy5 labeled siRNA (Fig 3). Image analysis of Cy5 distribution and intensity indicated 90% transfection efficiency. Transfection efficiency was confirmed using an siRNA pool targeting cyclin B1 sequences in the region of cyclin B1 used to construct the G2/M CCPM fusion protein (Fig 4). Treatment of cells with this siRNA pool ablated GFP fluorescence in over 90% of cells. Cells transfected with siRNAs directed against sequences present only in endogenous cyclin B1 retained EGFP expression and their fluorescence increased indicating accumulation of cells in G2 and M phase (see Fig 6, CCNB1). Measurement of cyclin B1 mRNA and EGFP fusion protein mRNA by RT-PCR and micro-array analysis (data not shown) indicated a 70% reduction in mRNA levels for both species, confirming the efficacy of the siRNA transfection.
As would be expected following knockdown of a range of cell cycle control genes, analysis of cell proliferation following siRNA treatment revealed a number of wells in which cell numbers were significantly larger or smaller than in control wells (Fig 5) indicating arrest, slowing, or acceleration of the cell cycle. For example, the effects of treating cells with siRNAs against the DNA replication licensing factor s MCM2-MCM7 (14) (Fig 5 row 7, columns 1–6) resulted in cell cycle arrest very rapidly following siRNA transfection.
Similarly, analysis across all siRNA pools (Fig 6) showed a diverse range of cell cycle distributions following gene knockdown. For example, treatment of cells with cyclin A2 siRNA (CCNA2) resulted in a significant accumulation of cells in prophase and mitosis to a similar degree to that observed for cyclin B1 (CCNB1), corresponding with the requirement for cyclin A for G1/S and G2/M transitions. In confirmation of the specificity of siRNA knockdown, cell cycle perturbation was not observed for the germ line functional homologue cyclin A1 (CCNA1), which is not expressed in differentiated U2OS cells. These changes correlated with additional changes in cellular morphology described below.
Changes in cell cycle distribution around the G1-S boundary following knockdown of cyclin E were not resolvable in G2/M CCPM cells because the cyclin B1 fusion protein is not expressed during this part of the cell cycle. Analysis of cyclin E siRNA treated cells showed 76% of cells in G1 or S phases compared with 74% in control siRNA treated cells. Analysis of G1/S CCPM cells treated under the same conditions (Fig 7) allowed the effects of cyclin E knockdown on G1 to S phase transition to be quantitated. Control G1/S CCPM cells showed the same proportion (76%) of cells in G1 or S phase as control G2/M CCPM cells (74%), which was resolvable in G1/S CCPM cells to 9% G1 cells and 67% S phase cells. On knockdown of cyclin E the proportion of cells in G1 or S phase remained constant at 76%, as observed with G2/M CCPM cells. However the balance between the two phases shifted significantly to 27% G1 cells and 49% S phase cells, reflecting the critical role of cyclin E in G1 to S transition.
Knockdown of Polo-like kinase (PLK) with siRNA has been previously shown to inhibit cell proliferation, arrest cells in mitosis, and induce apoptosis (15). Cell cycle analysis (Fig 6, PLK) of G2/M CCPM cells treated with siRNA directed against PLK showed a dramatic increase in mitotic cells 48 h after transfection with 50 nM siRNA (Fig 8B). Extreme sensitivity to PLK knockdown was confirmed by analysis of G1/S CCPM (Fig 8A) and G2/M CCPM (data not shown) 24 h following transfection with 5 nM siRNA, which showed an increase in G2 and M phase cells and a corresponding decrease in G1 cells.
One key advantage of high content cellular analysis is the ability to analyze high resolution images for multiple parameters. In this case additional morphological analysis of images derived from cells treated with Cyclin A2 siRNA revealed a significant increase in nuclear area (395.3 ± 173.7 µm2) compared with control cells (219.1 ± 95.7 µm2) and Cyclin A1 siRNA treated cells (229.8 ± 98.5 µm2).
Re-analysis of all image data from the siRNA library screen using DNA granularity revealed that PLK siRNA gave the most significant induction of apoptosis across the target genes in this study (Fig 8C).
Perturbation of sensitive and dynamic phenotypic cellular assays via siRNA provides a powerful tool for functional analysis of the cell cycle. High-throughput sub-cellular imaging and automated multi-parameter image analysis provides an information rich environment to screen and study effects of gene knockdown with siRNA.
1. Hannon, G.J. RNA interference. Nature 418 (6894), 244–251 (2002).
2. Paddison, P.J. and Hannon G.J. siRNAs and shRNAs: skeleton keys to the human genome. Curr. Opin. Mol. Ther. 5 (3), 217–224 (2003).
3. M edema, R.H. Optimizing RNA interference for application in mammalian cells. Biochem. J. 380 (3), 593–603 (2004).
4. Kumar, R. Conklin DS, Mittal V. High-throughput selection of effective RNAi probes for gene silencing. Genome Res. 13 (10), 2333–2340 (2003).
5. Luo, B. et al. Small interfering RNA production by enzymatic engineering of DNA (SPEED). Proc. Natl. Acad. Sci. USA 101 (15), 5494–5499 (2004).
6. Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428 (6981), 431–437 (2004).
7. Paddison, P.J. et al. A resource for large-scale RNA-interference-based screens in mammals. Nature 428 (6981), 427–431 (2004).
8. Mousses, S. et al. RNAi microarray analysis in cultured mammalian cells. Genome Res. 13 (10), 2341–2347 (2003).
9. Ramm, P. and Thomas, N. Image-based screening of signal transduction assays. Sci. STKE. 2003 (177), PE14.
10. Price, J.H. et al. Advances in molecular labeling, high-throughput imaging and machine intelligence portend powerful functional cellular biochemistry tools. J. Cell. Biochem. Suppl. 39, 194–210 (2002).
11. Thomas, N. Lighting the circle of life: fluorescent sensors for covert surveillance of the cell cycle. Cell Cycle 2 (6), 545–549 (2003).
12. Thomas, N. et al. Characterization and gene expression profiling of a stable cell line expressing a cell cycle GFP sensor. Cell Cycle 4 (1), (2005).
13. Gu, J. et al. Cell cycle-dependent regulation of a human DN A helicase that localizes in DNA damage foci. Mol. Biol. Cell. 15 (7), 3320–3332 (2004).
14. Bailis, J.M. and Forsburg, S.L. MCM proteins: DNA damage, mutagenesis, and repair. Curr. Opin. Genet. Dev. 14 (1), 17–21 (2004).
15. Liu, X. and Erikson, R.L. Polo-like kinase (Plk)1 depletion induces apoptosis in cancer cells. Proc. Natl. Acad. Sci. USA 100 (10), 5789–5794 (2003).
back to top