siRNA用户手册

The siRNA user guide (revised May 6, 2004)

Selection of siRNA duplexes from the target mRNA sequence

Using Drosophila melanogaster lysates (Tuschl et al. 1999), we have systematically analyzed the silencing efficiency of siRNA duplexes as a function of the length of the siRNAs, the length of the overhang and the sequence in the overhang (Elbashir et al. 2001c). The most efficient silencing was obtained with siRNA duplexes composed of 21-nt sense and 21-nt antisense strands, paired in a manner to have a 2-nt 3' overhang. The sequence of the 2-nt 3' overhang makes a small contribution to the specificity of target recognition restricted to the unpaired nucleotide adjacent to the first base pair. 2'-Deoxynucleotides in the 3' overhangs are as efficient as ribonucleotides, but are often cheaper to synthesize and probably more nuclease resistant. We used to ｓｅｌｅｃｔ siRNA sequences with TT in the overhang.

The targeted region is selected from a given cDNA sequence beginning 50 to 100 nt downstream of the start codon. Initially, 5' or 3' UTRs and regions nearby the start codon were avoided assuming that UTR-binding proteins and/or translation initiation complexes may interfere with binding of the siRNP or RISC endonuclease complex. More recently, however, we have targeted 3'-UTRs and have not experienced any problems in knocking down the targeted genes. In order to design a siRNA duplex, we search for the 23-nt sequence motif AA(N19)TT (N, any nucleotide) and ｓｅｌｅｃｔ hits with approx. 50% G/C-content (30% to 70% has also worked in our hands). If no suitable sequences are found, the search is extended using the motif NA(N21). The sequence of the sense siRNA corresponds to (N19)TT or N21 (position 3 to 23 of the 23-nt motif), respectively. In the latter case, we convert the 3' end of the sense siRNA to TT. The rationale for this sequence conversion is to generate a symmetric duplex with respect to the sequence composition of the sense and antisense 3' overhangs. The antisense siRNA is synthesized as the complement to position 1 to 21 of the 23-nt motif. Because position 1 of the 23-nt motif is not recognized sequence-specifically by the antisense siRNA, the 3'-most nucleotide residue of the antisense siRNA, can be chosen deliberately. However, the penultimate nucleotide of the antisense siRNA (complementary to position 2 of the 23-nt motif) should always be complementary to the targeted sequence. For simplifying chemical synthesis, we always use TT. Should an siRNA be expressed from pol III expression vectors, it is preferable that the first transcribed nucleotide is a purine. Upgraded selection rules suggest to bias the stability of the siRNA duplex in a manner that the 5' portion of the antisense siRNA is paired less stably to the sense siRNA than its 3' portion (Khvorova et al. 2003; Schwarz et al. 2003).

We always design siRNAs with symmetric 3' TT overhangs, believing that symmetric 3' overhangs help to ensure that the siRNPs are formed with approximately equal ratios of sense and antisense target RNA-cleaving siRNPs (Elbashir et al. 2001b; Elbashir et al. 2001c). Please note that the modification of the overhang of the sense sequence of the siRNA duplex is not expected to affect targeted mRNA recognition, as the antisense siRNA strand guides target recognition. In summary, no matter what you do to your overhangs, siRNAs should still function to a reasonable extent. However, using TT in the 3' overhang will always help your RNA synthesis company to let you know when you accidentally order a siRNA sequences 3' to 5' rather than in the recommended format of 5' to 3'. You may think this is funny, but it has happened quite a lot.

Compared to antisense or ribozyme technology, the secondary structure of the target mRNA does not appear to have a strong effect on silencing. We say that, because we have already knocked-down more than 20 genes using a single, essentially randomly chosen siRNA duplex (Harborth et al. 2001). Only 3 siRNA duplexes have been ineffective so far. In one or two other cases, we have found siRNAs to be inactive because the targeting site contained a single-nucleotide polymorphism. We were also able to knock-down two genes simultaneously (e.g. lamin A/C and NuMA) by using equal concentrations of siRNA duplexes.

We recommend to blast-search (NCBI database) the selected siRNA sequence against EST libraries to ensure that only one gene is targeted. In addition, we also recommend to knock-down your gene with two independent siRNA duplexes to control for specificity of the silencing effect. If selected siRNA duplexes do not function for silencing, please check for sequencing errors of the gene, polymorphisms, and whether your cell line is really from the expected species. Our initial studies on the specificity of target recognition by siRNA duplexes indicate that a single point mutation located in the paired region of an siRNA duplex is sufficient to abolish target mRNA degradation (Elbashir et al. 2001c). Furthermore, it is unknown if targeting of a gene by two different siRNA duplexes is more effective than using a single siRNA duplex. We think that the amount of siRNA-associating proteins is limiting for silencing rather than the target accessibility.

A siRNA search engine has recently been developed by Bingbing Yuan and Fran Lewitter in the Bioinformatics group of the Whitehead Institute for Biomedical Research. The program has a web interface and can be accessed after registration. Its use is free of charge for academic users. The program output is ranked by the degree of specificity of the predicted siRNAs. The user is able to define his own sequence search patterns and can also exclude single-nucleotide polymorphic sites from siRNA predictions. Here is a short description on how to access siRNA at Whitehead . If you experience difficulties with this first link, try the following: (1) Go to , (2) click onto the Biocomputing link on the left side, (3) go to the Bioinformatics pull-down link in the lower left side, and ｓｅｌｅｃｔ Access siRNA, (4) register at the site and then use it for prediction of siRNAs.