Stable RNA interference rules for silencing

Abstract

RNA interference has become an indispensable tool for loss-of-function studies across eukaryotes. By enabling stable and reversible gene silencing, shRNAs provide a means to study long-term phenotypes, perform pool-based forward genetic screens and examine the consequences of temporary target inhibition in vivo. However, efficient implementation in vertebrate systems has been hindered by technical difficulties affecting potency and specificity. Focusing on these issues, we analyse current strategies to obtain maximal knockdown with minimal off-target effects.

Figure 1

Discovery and development of RNA interference (RNAi). Ever since the discovery of RNAi as an endogenous mechanism that fine-tunes gene expression, efforts have been made to exploit it experimentally to silence genes of choice for both research and therapeutic purposes. Pigmentation defects observed in petunias engineered to overexpress key enzymes in the flavonoid biosynthesis pathway were among the first experimental hints of a post-transcriptional gene silencing mechanism mediated by mRNA degradation^,. Several years later, lin-4, a gene known to time Caenorhabditis elegans larval development, was found to code for a pair of small RNAs that regulate LIN-14 protein levels^,. Although regulatory roles for RNAs had been proposed earlier^,, lin-4 and its prototype counterpart Xist, a long non-coding RNA involved in mammalian X chromosome inactivation, were traditionally regarded as exceptions to the prevailing model of gene expression control by protein transcription factors. lin-4 is now recognized as the founding member of a class of small non-coding RNAs called microRNAs (miRNAs) that, in turn, were found to regulate the expression of at least one third of all human genes^–. Decisive for the boom of experimental RNAi was the discovery of a sequence-specific gene silencing mechanism mediated by double-stranded RNA (dsRNA) in the nematode C. elegans. Genetic and biochemical studies subsequently revealed insight into the molecular underpinnings of the RNAi pathway and uncovered analogous mechanisms in many eukaryotic organisms, enabling their exploitation for loss-of-function genetics in vertebrate systems. Shown here are selected theoretical, experimental and technical achievements.

Figure 2

Endogenous RNAi pathways and their use as tools for gene silencing. The RNAi machinery can be engaged at different points of the pathway by various forms of endogenous or exogenous double-stranded RNA (dsRNA) molecules to trigger the suppression of homologous genes. Like most endogenous miRNAs, synthetic miRNA-based shRNAs are expressed from RNA polymerase II (Pol II) promoters as primary miRNA transcripts (pri-miRNAs) that are recognized as natural substrates of the RNAi pathway. In contrast, stem-loop shRNAs are expressed as shorter precursor miRNAs (pre-miRNAs) mostly from promoters dependent on the activity of RNA polymerase III (Pol III) and thus skip the initial Drosha-mediated step of miRNA biogenesis. Different types of chemically synthesized short interfering RNAs (siRNAs) are transduced into cells as small RNA duplexes and enter the pathway at or after the Dicer stage (see Table 1 for details) to be directly incorporated into the RNA-induced silencing complex (RISC). Largely based on the degree of homology between the RNAi trigger and its target transcript, target knockdown occurs either through Ago-mediated mRNA degradation or translational repression and deadenylation.

Figure 3

Best practices for efficient and specific stable RNAi. For loss-of-function studies, RNAi has become an indispensable tool that enables functional gene annotation through stable and reversible gene silencing. RNAi can mirror gene loss during disease progression or mimic pharmacological target inhibition even where no such drug currently exists. Despite this potential, the design of potent and specific shRNAs is not trivial and careful experimental design is required to obtain robust results. For single-gene studies, the use of multiple shRNAs targeting the same gene can help experimentally reject phenotypes associated with off-target silencing. Likewise, multiple RNAi transgenic animals expressing shRNAs targeting the same gene can confirm on-target specificity of phenotypes. With pooled RNAi screens, special attention must be given to the complexity of shRNA libraries to ensure sufficient representation of each shRNA throughout the experimental procedure. Furthermore, for pooled screens and transgenic animals, shRNAs must be potent enough to induce a phenotype when expressed from a single genomic integration (that is, ‘single-copy’). Ultimately, the successful design of an experimental RNAi system includes selection of the right target transcript (region), choice of an appropriate expression system, informed selection of candidate shRNAs, and functional validation of knockdown efficiency. Using both biological and computational tricks to minimize system noise, RNAi results can be enhanced even following the experimental assay (for example, by analysing only those cells that productively express an shRNA). However, as most safeguard mechanisms are system-dependent, they require their implementation to be planned already during the experimental setup.

Figure 4

Ramifications of off-target effects. When interpreting RNAi results, off-target effects remain a key concern. However, these can be minimized with an appropriate shRNA design and implementation of certain experimental precautions. Most off-target effects fall into one of three large categories: (1) sequence-based homology with non-target transcripts; (2) aberrant miRNA biogenesis of endogenous or exogenous RNAi triggers; and (3) general perturbations affecting cell homeostasis. The diagram depicts the experimental precautions (red circles) that can minimize the various ramifications of each class of off-target effects.