Tips For Designing Your Own siRNAs

You need to start with the DNA sequence of your gene of interest from the relevant species (most siRNA design tools include help for mouse, rat, and human sequences). The initial work on RNA interference from the labs of Tom Tuschl and others suggests that silencing is most effective when using double-stranded siRNA that is comprised of a 19-nt region that is complementary to the target gene, plus a 2-nt overhang at each 3’ end. But how to choose the sequence of this siRNA?

Another good criterion for suitable sequences is placement within the gene of interest. Generally, try to select sequences from different regions of the mRNA, as some regions may be bound by regulatory proteins or folded in such a way as to hinder access. However, this may not be as important as once thought, according to the Tuschl lab, which finds that the secondary structure of the target mRNA has little effect on RNA interference.

Another criterion in choosing your siRNAs is avoiding unwanted homology to other sequences, in order to reduce non-specific mRNA interference. It’s a good idea to use BLAST (http://www.ncbi.nlm.nih.gov/blast/) to check your potential target sequences in their species database, to eliminate from consideration any sequences having considerable homology to other, unrelated coding sequences (some say that considerable homology comprises more than 16 contiguous base pairs, but others are more cautious). Also be aware of another unlikely but possible landmine—single-nucleotide polymorphisms (also known as SNPs). These are random point mutations in the genetic code that can show up in your target mRNA sequence, and render your siRNAs ineffective. The bioinformatics group at the Whitehead Institute for Biomedical Research has an online siRNA design tool that, among other nice features, allows you to exclude sites containing single-nucleotide polymorphisms from your siRNAs (see their web address at the end of the article). This online tool is free after registration.

With your target sequences in hand, the next step is to design some control sequences. One common type of control is a negative control consisting of your siRNA with a scrambled nucleotide sequence (with the nucleotide composition retained). The scrambled control should differ in at least four or five nucleotides from your original siRNA. After scrambling the sequence, make sure that you do a BLAST search using your scrambled control sequences to check for homology to other genes. Another type of control is to use multiple siRNA sequences that target the same mRNA (as mentioned above). Your RNA interference results will be stronger if you can show that more than one siRNA independently reduce the target mRNA levels.

Cenix BioScience, for instance, claims to have developed a design algorithm that resulted in 94% of their siRNAs yielding greater than 70% reduction in their target mRNA expression. The algorithm is also able to identify siRNAs that are quite potent, and thus can be used at lower concentrations. Cenix BioScience has joined forces with Ambion to develop mouse and human genome-wide siRNA libraries. Also, if you use the design algorithm offered by Cenix BioScience to design your siRNAs, you can order them from Ambion. Or, if you don’t want or need to design your own siRNAs, Cenix and Ambion offer Silencer Validated siRNAs for rat, mouse, and human genes. These are individual siRNA duplexes designed using the Cenix algorithm and then experimentally verified to reduce levels of their target mRNA. In fact, each one is guaranteed to reduce target expression by at least 70% within 48 hours after transfection. Ambion also offers companion products: the Silencer Positive and Negative Control siRNAs.

Researchers at Dharmacon developed another kind of siRNA design algorithm. It is offered as a free online tool from Dharmacon’s website (see below), and is also available as an Excel template written by an individual, Maurice Ho (http://boz094.ust.hk/RNAi/siRNA). The algorithm is known as rational siRNA design and is based on a point system that the researchers developed after examining how efficiently 180 different siRNAs silenced the target mRNA of two genes. They correlated the silencing efficacy with sequence characteristics of the individual siRNAs. These characteristics are embedded in the design algorithm, with the result that it ranks targeted sequences using a point system to assign scores to each one. The higher a sequence’s score, the higher its chances of success in RNA interference. Characteristics that earn a sequence points are: having a GC content between 30-52%; at least three A/Us at nucleotide positions 15-19; a lack of internal repeats; an A at nucleotide position 3; an A at nucleotide position 19; and a U at nucleotide position 10. Characteristics that lose points for a sequence are: having neither G nor C at nucleotide position 19, and having no G at position 13. In general, siRNAs scoring higher than six are worthy candidates for RNA interference experiments.

Dharmacon’s SMARTselection rational design algorithm underlies their SMARTpool Reagent, which together, they claim, offer a guaranteed 75% minimum target knock-down efficiency, with effectiveness reaching 95% in most cases (this guarantee does not apply to the free online design tool, but they maintain that siRNAs designed with the free tool will still be more effective than those designed using conventional algorithms). The SMARTpool Reagent is a mixture of four individual siRNA duplexes, each one designed by the SMARTselection algorithm and targeted to a different region of the same target. The individual siRNAs are also checked for sequence homology to non-targeted genes to avoid unwanted interference.

Invitrogen offers an online design tool called the BLOCK-iT RNAi Designer that allows you to select the design algorithm that you want to use—either their proprietary algorithm or one or more patterns based on Tom Tuschl’s work. Ordering siRNAs that result from the algorithm is optional; Invitrogen guarantees that if you order the two most highly-ranked siRNAs, one of them will give more than 70% knock-down of your target mRNA. If both fail, they will design and ship you a third for free.

In