RNA interference (RNAi) is a method that is based on an endogenous cellular response to the presence of long or short double-stranded RNA and is useful to knock down (i.e. reduce but not eliminate) gene function.
The specific RNAi reagent and the method for its delivery to cells is likely to be different for different types of cells, model organisms and assays. But in general, a gene-specific segment of double-stranded RNA is introduced into cells, leading over time to degradation of the endogenous target mRNA. Short segments (~21 bps) are typical for mammalian systems, as longer segments can induce a non-specific interferon response. Longer segments (~500 bps) are typical for model systems that lack an interferon response, such as Drosophila. Similarly, the appropriate delivery systems differ for different systems. Common systems for delivery of the RNAi reagents include viral transduction (shRNAs), transfection or electroporation (shRNAs, siRNAs or dsRNAs) and bathing (dsRNAs).
RNAi can be compared with loss-of-function genetic approaches. When the normal function of a gene is required for a given function, RNAi knockdown may lead to a phenotype detectable in an assay that tests that function, either directly or indirectly. Example assays include activation of a transcriptional reporter; responses to an external stimulus (e.g. stress, drug treatment, pathogen); and sub-cellular localization or morphology of a specific protein or organelle.
Cell-based RNAi screens are typically performed in one of two formats, pooled or arrayed. Both formats are supported by the collaborative DF/HCC Collaborative RNAi Core.
(1) With a pooled approach, all of the gene-specific reagents are pooled together (or synthesized en masse) and added, at random, to cells. This could be compared to transformation of a cDNA library into bacterial cells. You will know that each cell gets about one gene-specific RNAi reagent. But you will not know which cell got which one.
Pooled Selection. In some cases, the researcher next performs a selection; that is, applies a treatment (e.g. a drug, condition or pathogen) that kills most cells, leaving only cells that have an appropriate RNAi knockdown to survive. That is, when the gene-specific RNAi phenotype is viability under conditions in which most cells would die. Subsequent to the selection, a molecular method is used to figure out which RNAi reagent (and thus, which putative gene target) is responsible for “escape” of the lethal treatment.
Pooled Comparison. In other cases, sub-sets of cells (or different cell types) are treated differently before and/or after being subjected to the RNAi library (creating reference and experimental sets). Subsequently, a molecular method such as micro-array analysis is used to compare the two sample sets. The array detects the abundance of the RNAi reagents (such as via detection of the shRNA sequence itself or a molecular barcode in the construct). The array is used to determine which RNAi reagents (thus, which putative gene targets) are under- and/or over-represented in the experimental set as compared with the reference set.
Advantages: a pooled approach is typically more practical to do in a standard lab space and in a relatively small total volume, thus making it more feasible for some groups and less expensive. Disadvantages: deconvolution of the putative “hits” (positive results) requires specialized or custom microarrays and their analysis.
(2) With an arrayed approach, each unique RNAi reagent (or unique set of reagents targeting a single gene, such as for a small pool of independent siRNAs targeting one gene) occupies a unique well in a microtiter plate, such as a 384-well plate. Experiments are done in 384- or 96-well format. Thus, after the assay, you can easily determine which cells got which specific RNAi reagent by looking up the identity of the reagent in a given well using a database or spreadsheet.
An arrayed approach is useful for a wide variety of cell-based assays, including plate-reader assays (i.e. total light output, such as from GFP, FITC or luciferase) and low- or high-content image-based assays. Most cells will have the expected normal response, behavior, shape etc. But some specific reagents will have an abnormal response, behavior, shape etc. (mutant phenotype) that is detected in the assay. For some approaches, such as an imaging approach, multiple read-outs might be detected (i.e. via antibodies or dyes, using multiple fluorescent channels).
Advantages: Arrayed screening arguably opens the door to the widest possible range of cell-based assays that can be performed. You will be able to quickly associate hits with the putative gene target of the unique RNAi reagent that was present in a specific well in which the mutant phenotype was observed. Disadvantages: Typically requires specialized high-throughput equipment for plate processing and assay detection and reagents, media, etc. can be costly.
We recommend also reading the screening guidelines at the ICCB Longwood Screening Facility web site.
Most researchers will probably spend the majority of their time and effort working at or with one of the participating groups. What we emphasize below are those assay pipelines that combine resources from different participating groups and centers. These are just a few examples of how you can design a screen assay pipeline that will take best advantage of the resources available through the DF/HCC Collaborative RNAi Core.
III. RNAi Screening: What Steps are Involved in Preparing to Screen?
At a practical level, you should also ensure that the assay:
After assay development the researcher makes a transition to assay optimization. At this stage, you should be working in the same format in which you will do the screen. And ideally, using the same instruments for assay read-out that you will use in the full screen.
You should ensure that the assay:
Once you have developed and optimized an assay, you are probably ready to screen. Most centers will have a set of criteria that must be fulfilled before a screen is initiated. In addition, most groups will start you off with a test plate and then set of pilot plates, before you start processing a large number of plates.
Please be aware that the hands-on involvement of a researcher from your lab is essential to the process and helps to ensure success. No one will be able to perform the screen better, or more quickly and accurately gauge how thing are going, than those folks who designed and will later analyze the results of the screen.
For the purposes of this text, we will define hit verification as verification of the RNAi results from a primary screen (such as using a different RNAi reagent directed against the same gene), and hit validation as successful testing of the subset of verified hits using one or more independent assay (such as a genetic, biochemical, or proteomics approach). Note that the results of statistical and bioinformatics analyses of your primary screen dataset (see below) will also inform your interpretation of screen results.
In general, the same groups that help facilitate RNAi screening can also provide reagents and help with hit verification. For example, after an arrayed siRNA screen in which each well contained a pool of, say, four siRNAs directed against the same gene target, positive hits are typically repeated with each individual siRNA contained in the primary screening pool. A positive result with just one of the pooled siRNAs suggests the likelihood that the primary result represents an OTE. By contrast, positive results with more than one reagent suggest, but do not guarantee, a gene-specific result. However, as we are still learning about RNAi reagent design, things are not always as cut-and-dried as that suggests, such that validation (see below) is critically important. The participating screening centers can provide reagents for hit verification. You can also utilize reagents from more than one center (e.g. follow up an shRNA screen with siRNAs, or vice-versa).
Regardless of your strategy, planning follow-up often benefits from bioinformatics analysis. You might start by asking some of the following questions:
Commercial software packages such as Microsoft Excel, Spotfire, or Pipeline Pilot; custom databases such as those put together in FileMaker Pro, Microscoft Access, or another database platform; and non-commercial resources such as NIH DAVID or the Boutros group's cellHTS may also be of use for various analyses. In addition, most of the centers have databases and on-line software tools that can help with statistical analysis, viewing analyzed data (such as as a 'heat map'), and provide additional analysis and data management solutions.
For assays that generate image data, image analysis will also be critically important to your screen. Please see the "services" section for information about image analysis support and software available from the participating centers.