RNAi Design, Validation and Target Screening

Since Tuschl et al. published the first empirical guidelines on how to design effective siRNA [1], the most significant advancement (based on the understanding of biochemical mechanisms of RNAi such as how RISC is assembled) is the recognition of asymmetric thermostability of the 5’ end of the antisense strand (AS) relative to that of the sense strand (SS) [2, 3]. siRNAs with an A/T-rich AS 5’ end can be more easily integrated into RISC. By biasing against the sense strand for RISC loading, the off-target effects due to the presence of the SS (as one of the sources of off-targets effects) can also be minimized. In recent years datasets of increased number of siRNAs and shRNAs became available and statistical analysis suggested additional rules for RNAi design. These newer rules in general define the siRNA prediction parameters in more detail, for instance, the number of bases of the 5’ ends that should be included when calculating asymmetric thermostability, base preferences at each particular position, and the identity of the 2 nt 3’ overhang [4, 5]. Computer programs and websites are developed based on these features also resulting from NIH funded research through universities and organizations. Among the well-known ones, Design of SIRna (DSIR at biodev.extra.cea.fr/DSIR/DSIR.html) and the shRNA search program at the Broad Institute (broadinstitute.org/genome_bio/trc/publicSearchForHairpinsForm.php) are freely available.

Several companies such as Open Biosystems, System Biosciences, Dharmacon/ThermoFisher, Sigma-Aldrich, Invitrogen/LifeTech, provide premade RNAi reagents against various numbers of human and rodent genes. Although some product lines from these suppliers are labeled as validated RNAi reagents, apparently only one reveals clone sequences and only a few hundred among the claimed 4,500 shRNA clones. It is not possible to find what shRNAs are used against any target gene from most companies even though many of them claim to have a few hundred pre-validated constructs. Some of them may provide additional information upon purchase.

Even with the recent advancement of RNAi design technologies, prediction of effective RNAi is still far from accurate. Depending on the datasets used to score the success rates of the programs at DSIR, Broad or any other software, the general consensus is that about 50% of predicted RNAi target sequences will be effective, resulting in better than 70% gene knockdown. Allele Biotech uses a software that was trained with known RNAi results to predict siRNA target candidates on a given mRNA, and then applies an additional set of rules to pick the most promising candidates. Off-target effects caused by partial-matching between AS strand and untended targets are reduced by searching the chosen site against the NCBI gene base. The basic rules Allele Biotech uses include most currently known ones and are similar to what are listed by The RNAi Consortium (TRC) program at the Broad Institute.

Criteria for RNAi design:
(1) Overall GC content is between 30-55%
(2) The 4 bases at the 5’ of AS is more AT-rich than those of the SS
(3) The first base of AS and SS 5’ is preferably A/T and G/C, respectively
(4) “U” is preferred at the 10th position of the antisense from the 5’ end
(5) “C” is to be avoided as the last base of an overhang
(6) Avoid 4-nt mono-nucleotide regions
(7) Avoid 6-nt GC-rich regions
(8) If possible, do not include those with apparent secondary structures

These selected rules are based on a number of publications (for example, [4-6]), but it is impossible to include all known rules, many of which conflict with each other. In case of conflicting rules we rely more on recent discoveries and our own experience from providing RNAi service during the past 8 years.

Allele Biotech provides RNAi validation and screening services to customers using synthetic siRNA, linear DNA cassettes with engineered Pol III promoter, and shRNA expressing lentiviral vectors in high throughput formats. In a unique design, all RNAi target candidate sequences of a gene transcript are fused consecutively to a bright green fluorescent protein, mWasabi, on a lentiviral vector. Instead of analyzing gene silencing by QPCR, the initial selection of effective RNAi can be performed by measuring fluorescence.

RNAi screening has been conducted to identify correlations between gene functions and cellular phenotypes such as synthetic lethality among DNA damage signaling and repair pathway factors. Successfully performing high throughput screenings requires capabilities of efficient RNAi design, viral packaging, fluorescent proteins, and advanced cell culture and analysis techniques. In addition to these capabilities, Allele’s RNAi services are provided with access to commercial use of Allele’s own patents on Pol III promoter driven shRNA expression, and licensed patents on lentiviral vector, packaging, and fluorescent proteins.

New Product/Service of week Nov 16-22, 09:

RNAi validation/screening service.

1. Tuschl, T., P.D. Zamore, R. Lehmann, D.P. Bartel, and P.A. Sharp, Targeted mRNA degradation by double-stranded RNA in vitro. Genes Dev, 1999. 13(24): p. 3191-7.
2. Khvorova, A., A. Reynolds, and S.D. Jayasena, Functional siRNAs and miRNAs exhibit strand bias. Cell, 2003. 115(2): p. 209-16.
3. Schwarz, D.S., G. Hutvagner, T. Du, Z. Xu, N. Aronin, and P.D. Zamore, Asymmetry in the assembly of the RNAi enzyme complex. Cell, 2003. 115(2): p. 199-208.
4. Vert, J.P., N. Foveau, C. Lajaunie, and Y. Vandenbrouck, An accurate and interpretable model for siRNA efficacy prediction. BMC Bioinformatics, 2006. 7: p. 520.
5. Zhou, H. and X. Zeng, Energy profile and secondary structure impact shRNA efficacy. BMC Genomics, 2009. 10 Suppl 1: p. S9.
6. Ui-Tei, K., Y. Naito, F. Takahashi, T. Haraguchi, H. Ohki-Hamazaki, A. Juni, R. Ueda, and K. Saigo, Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference. Nucleic Acids Res, 2004. 32(3): p. 936-48.

Tags: cell based assays, DNA damage repair, gene knockdown, gene silencing, lentiviral vector, Pol III promoter, RNAi, RNAi design, RNAi screening, RNAi service, RNAi targets, shRNA, siRNA, Synthetic lethal

HPLC Purified siRNA with Known RNAi Effects at $149/12.5nmol

RNA oligo is significantly more difficult to synthesize than DNA oligos, mainly because the efficiency of coupling each new ribonucleotide during RNA synthesis is a few fold lower than deoxyribonucleotide during DNA synthesis. Typically, there is an ~10% chance a DNA oligo of 21 bases will have a mutation (most frequently a deletion mutation); for an RNA oligo of 21 bases, as in an siRNA pair, such chance is much higher. Furthermore, after combining the sense and antisense siRNA strands, some RNA molecules will remain as single-stranded thereby not fitting for the RNAi apparatus.

RNA interference is a dose-sensitive process — specificity of gene silencing is meaningful only relative to the active concentration of siRNA used. When the concentration is too low, even the most effective siRNAs would fail to cause gene expression knockdown; when too high, non-specific effects will be duly observed. Therefore, it is essential that the concentrations of siRNAs are measured correctly. When doing so, one must consider not only what the apparent concentrations are by OD260 reading, but also whether the RNA strands are of full-length and whether only dsRNA molecules are counted. This issue might not affect data interpretation if appropriate controls are included in one set of RNAi experiments, but it could have significant influence on conclusions if data from different experiment sets or labs are compared or combined.

HPLC purification currently provides the best means to remove RNA molecules with deletions or remain single-stranded, however, the price tag added by most reagent providers for such treatment has been prohibiting because manufacturers either need to start synthesis at a much bigger scale to obtain promised amount, or they do not promise the delivery quantity at all. The phosphoramidites (oligo building blocks) for RNA synthesis can be 10 times or more expensive than for DNA. Some companies offer alternative purification methods such as a cartridge type device, but they can only remove salt and small impurities, not RNA oligos of shorter lengths accumulated at each cycle of amide coupling. The AllHPLC siRNAs within Allele’s RNAi product line, pre-validated or custom made, are uniformly HPLC purified with 5 OD or 12.5 nmol of double-stranded, annealed siRNA delivered. Allele passes to customers the cost savings from manufacturing our own RNA amidites and other reagents for oligo synthesis. The pre-validated HPLC purified double-stranded siRNA is offered today at $149/12.5 nmol.

Before purchasing siRNAs, even at a low cost of $29 per pair of HPLC purified control siRNA from Allele, researchers still need to consider how well their cells can be transfected. For hard-to-transfect cells, lentiviral vectors carrying a shRNA expressing cassette is often a better choice. To establish stable cell lines, plasmid vectors should be considered. For low cost target screening, the PCR format linear siRNA expression cassettes have advantages.

Tags: DNA, gene silencing, HPLC, oligo, reagents for chemical synthesis of DNA and RNA oligonucleotides, reagents for oligonucleotide, RNA, RNA interference, RNAi, shRNA, siRNA

What seems to be going on with RNAi related patents in the US

Reciting Table 1 from Ref 1 and Table 3 from Ref 2:

By the end Nov 2008 it appears that Allele’s patent (US 7,294,504 and 7,422,896) are the only currently granted DNA based RNAi patents. The focus of Allele’s technology is siRNA of 21-23, either in separate sense and antisense strands, or shRNA or miRNA format, thus not covered by the Fire patent or the Kreutzer-Limmer patent. Since these RNAi inducers are not synthesized by chemical reactions, or produced with enzymes or cell lysate in vitro, they do not relate to Tuschl I or II patent groups. Allele Biotech can not guarantee that its interpretation is correct or final by any means; commercial user of any of the related technologies should perform own due diligence.

[1] Charlie Schmidt. March 2007 “Negotiating the RNAi patent thicket” Nature Biotechnology 25 (3): 273-275

[2] Dirk Haussecker. May 2008 “The Business of RNAi Therapeutics” Human Gene Therapy 19: 451-462

Have an opinion? Feel free to share it here.

Tags: gene silencing, H1, lentiviral vector, miRNA, PCR cassette, plasmid vector, pol III, retroviral vector, RNAi, shRNA, siRNA, U6

Something you should know about oligos

Oligos are made from 3’ end to 5’ end by nucleotide-wise coupling. Each coupling cycle involves about half a dozen moisture-sensitive chemicals and takes about 15 minutes to complete when 96 oligos are being synthesized at the same time on one machine. Like most chemical reactions, couplings do not reach 100% efficiency; in consequence, about 1% of the oligos would have an unsuccessful coupling at any given position and therefore missing that base. There are “capping” steps designed to prevent oligos having an unsuccessful coupling from continuing to elongate; but in practice, capping can only reduce incomplete oligos in the final pool, not eliminate them.

In PCR reactions, primers with mistakes typically have less chance of pairing with template than those with perfect match. Increasing annealing temperature may prevent primers with deletions from participating in PCR reactions. However, oligos that miss 5’ end restriction site but have no internal deletions will not be selected against by higher annealing temperature since initial annealing is not affected. Purification of oligos by PAGE can effectively remove oligos with deletions in any position, albeit not eliminate them. For cloning purposes, purifying oligos typically makes the post-ligation steps (i.e. inoculation, minipreps, and sequencing) much easier.

At Allele Biotech Oligo Services, we use top quality chemicals from Glen Research and extensive coupling and washing steps in order to synthesize oligos with as few mistakes as chemically possible. Most oligo mistakes occur in individual molecules, which you may encounter by chance. Sequencing a few more colonies for a cloning project is the easiest and fastest way to achieve desired results if a mistake is found in the first round of sequencing. If the customer prefers remaking oligo, we honor our 100% guarantee policy with replacement of all our oligos shorter than 45 bases and the longer ones with purification. We always appreciate our customers’ feedback.

General suggestions for using oligo primers for cloning

* Use more stringent PCR annealing conditions if possible.

* Purify oligos, especially long primers.

* Even though higher cost on oligos, much lower costs and less time on minipreps and sequencing.

* Since mutations happen randomly, sequence a few more colonies could result in identifying desired plasmid

Sometimes it is very difficult to clone PCR products by restriction digestion and ligation. Restriction enzymes do not cut well near the end of linear DNA even with extra bases added 5’ to the restrictions sites. It is almost always helpful to clone the PCR fragment into a PCR cloning first. Cut with designed restriction enzymes and send only those showing insert of correct sizes for sequencing.

Oligo mutations augmented by E. coli during colony selection. The following was a real case using Allele oligos to create genes encoding fusion antibody chains. The strategy was to fuse a humanized antibody heavy chain to a light chain by a pair of oligos that would overlap cDNAs for both chains. The vector was pBluescript II, where insert could disrupt an expressed beta-galatosidase, thus changing the color from blue to white in the presence of IPTG and X-gal. After minipreping and sequencing dozens of colonies, all plasmids had frame-changing mutations in the junction region, which were seemingly introduced by the primers. Worrying about the oligo quality, we checked oligo synthesis records, including reagent log of that run, Trityl color indication record (indicator of coupling efficiency of up until the last base), gel pictures of oligos made in the same batch, feedback record from other customers using oligos from that day. We could not find anything unusual from the records. We decided to remake the oligos. After two weeks of work to regenerate plasmids for sequencing, the same results were obtained—all plasmids had mutations in the same region. There seemed to be nothing else we could do but to remake the oligos with somewhat different designs, e.g. shifting the overlapping regions slightly, or shortening the oligos a little bit, and performed PAGE purification this time. The results were the same once again.Then it occurred to us that maybe the expression of the protein from the pBluescript vector caused toxicity to E. coli and therefore forced the bacterial cells to either select those clones with frame-shifting mutations or create mutations by themselves during growth. Without the option of changing the vector choice, we simply used a different competent cell strain that does not support expression from the promoter on pBluescript. We did it with oligos from different preps, purified and unpurified, in an attempt to obtain as much information as possible about oligo use for our customers and our own future research.The result, all plasmids sequenced were completely correct.

Other Cloning Example Cases:

Case 1:

Aim: To synthesize a 1,650 bp gene from oligos.
Design: Design 36 overlapping oligos of 60 to 80 bases long.
Experiments: Difficult to do PCR in one piece with all oligos. Switched to 3 separate PCR for about 550 bp each.
Results: Sequenced plasmids from colonies with each of the 3 parts cloned into PCR cloning vector: Part I: 2 plasmids sequenced, both with mutations; 1 more sequenced, correct. Part II: 1 plasmid sequenced, wrong; 1 more sequenced, correct. Part III: 1 plasmid sequenced, with mutation; another sequenced, wrong; 3 more sequenced, 2 correct.
Conclusions: Mistakes coming from oligos are random, there is no prediction exactly how many colonies should be sequenced, but normally sequencing 3 in a group is a good practice.

Case 2:

Aim: To synthesize a 300 bp gene from oligos.
Design: Design 16 overlapping oligos of 60 to 80 bases long.
Experiments: PCR in one piece with all oligos.
Results: Sequenced plasmids from colonies cloned into PCR cloning vector: 2 plasmids sequenced, both with deletions; 3 more sequenced, 1 with deletion, 2 with base change; 2 more sequenced, both completely correct.
Conclusions: When luck is not on your side, sometimes a short DNA still requires a good number of colonies to be sequenced.

Case 3:

Aim: To synthesize a hypothetical gene of 1,800bp from oligo overlapping assembly.
Design: 34 oligos of about 55 bases each direction.
Experiments: Difficult to do PCR in one piece with all oligos. Switched to 3 separate PCR for about 550 bp each.
Results: Sequenced 2 colonies, 1 was perfect from one end for about 900 bases, but a number of mutations found when sequenced from the other end. The 2nd plasmid was 100% correct from the first to the last base over the entire 1.8kb region!.
Conclusions: When luck is on your side, you may hit the jack pot blind-folded.

Case 4:

Aim: PCR-clone 3 human cDNAs into a baculovirus expression vector.
Design: PCR with primers that would introduce restriction sites Not I and Xho I.
Experiments: PCR with standard procedure with high fidelity polymerase, PCR products were then run on gel and the desired bands purified by Allele DNA purification kits.
Results: Two of the 3 constructs were all correct in all plasmids sequenced. Plasmids for the other construct all showed PCR products missing half of the Not I site. Sequencing additional plasmids gave the same results. We then gel purified the primer, repeat the process, and all plasmids sequenced were correct.

Have any insights or comments of your own about using oligos? Let’s share them. Thread away.

Tags: dsDNA, fast turnaround, insert, oligo, PAGE, PCR, primer, probe, siRNA, vector