Designing a Better siRNA

The siRNA design algorithm designed by Cenix BioScience, with whom Ambion has partnered to develop a genome wide siRNA library, represents a major improvement over the design rules first described by Tuschl and colleagues. On average, ~50% of the siRNAs featuring the Tuschl design rules (target sites beginning with AA, 3' UU overhangs for both the sense and antisense siRNA strands, ~50% G/C content; 1-3) provide at least 50% reduction in target gene expression (4). In contrast, in an initial screen of 79 human genes, 94% of siRNAs designed by Cenix provided greater than 70% reduction in target mRNA levels (Figure 1).

Figure 1.The Effectiveness of Cenix Designed siRNAs. siRNAs targeting 79 human genes were designed using the algorithm developed by Cenix. The top siRNA candidate for each target was prepared and transfected at a 100 nM final concentration. Forty eight hours post-transfection, target gene expression was quantified by real-time RT-PCR. Relative reduction in mRNA expression was measured against cells transfected with a negative control siRNA. Sample size was normalized by measuring 18S rRNA in the various samples using a real-time PCR.

To develop a rational siRNA design algorithm, Cenix and Ambion tested multiple siRNAs targeting hundreds of different human genes that were expressed at detectable levels in several different cell lines. More than 900 siRNAs have been designed and tested to date. The effective and ineffective siRNAs were used to judge the impact of a number of physical characteristics on siRNA potency and specificity, including Tm, nucleotide content of the 3' overhangs, siRNA length, nucleotide distribution over the length of the siRNA, and presence and location of mismatches. Those characteristics that correlated with siRNA effectiveness and specificity were incorporated into the design algorithm. A key step in the algorithm is a stringent analysis of each siRNA sequence to maximize target specificity. The result is a robust algorithm that uniquely addresses gene-to-gene variability in susceptibility to RNAi. Both Ambion and Cenix are committed to continual refinement of the algorithm as more is learned about the criteria that affect siRNA potency and specificity.

The siRNA design algorithm designed by Cenix is able to identify effective siRNA sequences at a very high rate (Figure 1). The algorithm is also able to identify highly potent siRNA sequences that are effective at low concentrations. Figure 2 shows the potency of six different siRNAs designed using the algorithm. All were as effective at reducing target mRNA levels at 10 nM and 30 nM, as at 100 nM. The use of a highly potent sequence means that less siRNA can be used per experiment, saving reagent costs and permitting the use of multiple siRNAs targeting different genes in a single experiment. More importantly, minimizing the working concentration of siRNA in an experiment has recently been recognized to maximize the specificity of the RNAi effect (5-7).

For your convenience, Ambion can custom synthesize siRNAs pre-designed using the algorithm developed by Cenix. Designs are immediately available for all human, mouse, and rat genes in the RefSeq database maintained by NCBI. Alternatively, Ambion and Cenix have partnered to validate individual siRNAs that have been designed by the algorithm. These Silencer™ Validated siRNAs, available targeting various human genes and in 5 nmol unit sizes, have been verified to reduce target gene expression by >70% using a real-time RT-PCR assay.

Figure 2. Effectiveness of Validated siRNAs at Various Concentrations. The indicated siRNAs were complexed with siPORT™ Lipid and the resulting complexes were added to HeLa cells in twenty-four well plates at the final concentration of siRNA shown. Forty-eight hours after transfection, RNA from the treated cells was recovered using RNAqueous® Mag-96 RNA Isolation Kit and reverse transcribed using the RETROscript® Kit. Target cDNA levels were measured by real-time PCR. Expression of target genes in the transfected cells was compared to cells transfected with an equal concentration of Silencer™ Negative Control #1. Input cDNA in the different samples was normalized using real-time data for 18S rRNA. The bar graphs represent an average of three data points.

back to top