Many researchers now use small interfering RNAs (siRNAs) to reduce the expression of specific mammalian genes. Here we describe five methods for producing siRNAs for use in mammalian RNAi experiments. Each of these methods has its advantages and drawbacks. The best method for generating siRNAs will depend on the goals of the experiment. This article briefly describes the five methods, presents their advantages and disadvantages, and discusses the types of applications for which they are best suited.

Currently, there are five methods for generating siRNAs for gene silencing studies:

  1. Chemical synthesis
  2. In vitro transcription
  3. Digestion of long dsRNA by an RNase III family enzyme (e.g. Dicer, RNase III)
  4. Expression in cells from an siRNA expression plasmid or viral vector
  5. Expression in cells from a PCR-derived siRNA expression cassette

The first three methods involve in vitro preparation of siRNAs that are then introduced directly into mammalian cells by lipofection, electroporation, or other technique. The last two methods rely on the introduction of DNA-based vectors and cassettes that express siRNAs within the cells. Table 1, below, summarizes these methods.

In vitro
digestion of
Expression Cassette
Requirements (2) 21-
mer RNA 
(2) 29-mer DNA oligos Transcription template (200-800 bp region flanked by T7 promoters) (2) 55-60-mer DNA oligos (2) ~50-mer DNA oligos
Turnaround time (total preparation/
synthesis time)
4 days to
2 weeks*
24 hours + DNA oligo 1 day + transcription template preparation time 5+ days + DNA oligo ~ 6 hours + DNA oligo
Hands on time Little to none* Moderate High Moderate Moderate
Testing to find optimal siRNA sequence Required Required Not Required Required Required
Ability to label siRNA (i.e., for analyzing siRNA uptake or localization by fluorescence microscopy) Yes Yes Yes No No
Relative ease of transfection Good Good Good Fair Fair
Selectability (i.e, antibiotic selection) No No No Yes No
Useful for long term studies No No No Yes, with selection No
Ability to scale up synthesis Yes Limited Limited Yes Limited
Monitor transfection efficiency of entire population No No No Yes No
Relative cost per gene (not including labor) High Moderate Low Moderate Moderate
Ambion Solution Ambion's Custom siRNA Synthesis Service Silencer siRNA Construction Kit Silencer siRNA Cocktail Kit (RNase III) pSilencer siRNA Expression Vectors Silencer Express siRNA Expression Cassette Kits
*Depends on purification/deprotection options selected and format (e.g., annealed and ready to transfect versus single strands supplied lyophilized)

siRNA Design

All of the above methods, except digestion of long dsRNA by RNase III, require the design of individual siRNA sequences prior to siRNA preparation. High throughput screening of potential siRNAs and subsequent analysis for knockdown has recently led to the development of predictive algorithms for identifying functional siRNAs.

In Vitro Preparation

Method #1: Chemical Synthesis
Although more expensive than any of the other methods, the production of chemically synthesized siRNAs requires almost no effort by the researcher. Ambion, and several other companies, provide high quality, chemically synthesized siRNAs on a custom basis. One of the major benefits of chemical synthesis is the large yield of high purity siRNA. Drawbacks include the price and turnaround times (typically 4–12 days depending on synthesis and purification options). Because of the relatively high price tag, many researchers find it beneficial to screen siRNA sequences using a less expensive preparation method, such as in vitro transcription, and then have the most effective sequence(s) synthesized chemically. A robust design algorithm that results in a high proportion of effective siRNA sequences is available from Ambion. Improved siRNA design criteria decreases the number of sequences that need to be tested, lowering the costs associated with this option.

Best for:

Studies that require large amounts of a defined ultrapure siRNA sequence

Not suitable for:
Long term studies

Ambion's Solution: Custom, Pre-designed and Validated siRNAs
Custom siRNAs: Ambion provides premium quality, made-to-order siRNAs with your choice of synthesis options. Our standard purification procedure includes deprotection and column purification, and typically yields siRNA that is 90% pure (guaranteed >80% pure). HPLC and PAGE purification is also available, yielding siRNAs that are guaranteed >97% pure. All RNA oligonucleotides are assessed by MALDI-TOF (matrix-assisted laser desorption-ionization – time-of-flight) mass spectrometry and annealed siRNAs are analyzed by nondenaturing gel or capillary electrophoresis to confirm that the strands are annealed properly.

Pre-designed siRNAs: A design algorithm developed by Ambion's partner, Cenix BioScience, predicts potent and specific siRNA sequences with an impressive success rate. Multiple siRNA designs are available for the human, mouse, and rat genes listed in the public RefSeq database maintained by NCBI. We can also use the algorithm to design siRNAs to other organisms or genes not in the RefSeq database.

Validated siRNAs: Silencer™ Validated siRNAs are single siRNA duplexes that have been verified experimentally to reduce the expression of individual human genes. Each siRNA has been shown to reduce target gene expression by at least 70% forty-eight hours post transfection by real time RT-PCR. Most of the siRNAs were found to reduce target gene expression by 90% or more. Every Silencer Validated siRNA strand is purified by HPLC and subjected to rigorous quality control measures. The result is the highest quality siRNA with a sequence verified to reduce gene expression.

Method #2: In Vitro Transcription
siRNAs can be readily prepared by in vitro transcription. siRNAs produced by this method are considerably less expensive than their chemically synthesized counterparts, making them a more cost-effective choice for screening siRNA sequences. In addition, they can be produced more quickly than chemically synthesized siRNAs. Once template deoxynucleotides are obtained, the procedure takes about 24 hours, with little hands on time. Disadvantages of this method include the limited scale up potential (although each reaction produces enough siRNA for hundreds of transfections) and the fact that it requires more hands-on time from the researcher compared to chemically synthesized siRNAs which can simply be purchased. It should be noted that in vitro transcribed siRNAs work as well as chemically synthesized siRNAs and usually at lower concentrations -- 0.5–20 nM vs. 50–100 nM concentration per transfection (Figure 1).

Best for:
Screening siRNA sequences or when the price of chemical siRNA synthesis is an obstacle

Not suitable for:
Long term studies or studies that require large amounts of a single siRNA sequence

Ambion's Solution: Silencer™ siRNA Construction Kit
The Silencer siRNA Construction Kit produces transfection-ready siRNA at a fraction of the cost of chemical synthesis. The kit is based on a patent-pending in vitro transcription method and can generate up to 15 purified siRNAs in less than 24 hours. The cost and time saved using the Silencer siRNA Construction Kit enable screening sequences of more genes and more potential targets within the target gene to find the most potent siRNA.

Figure 1. Use of Chemically Synthesized and in Vitro Transcribed siRNAs to Induce Gene Silencing. siRNAs targeting ß-actin were prepared by chemical synthesis (Ambion) or by in vitro transcription using Ambion's Silencer siRNA Construction Kit. HeLa cells were plated at 30,000 cells per well in a 24 well tissue culture plate containing glass slides. The cells were transfected 24 hours after plating, using 2 µl siPORT™ Lipid (Ambion) according to the manufacturer's protocol, at a final siRNA concentration of 75 nM. Immunofluorescence analysis was performed 96 hr post transfection using mouse anti-human ß-actin primary antibody and a FITC conjugated anti-mouse IgG secondary antibody. Photographs were taken using the appropriate fluorescent filters and quantified using MetaMorph software. Note that both siRNA preparation methods resulted in > 95% reduction in ß-actin protein levels.

Method #3: Digestion of Long dsRNAs to Create an siRNA Cocktail
One of the major drawbacks of all the other methods of siRNA production is the need to design and test several siRNA sequences before an effective one can be identified. Preparation of siRNA cocktails overcomes this limitation. In this method, long dsRNAs are prepared by in vitro transcription using a template that typically encodes a 200–1000 nt region of the target mRNA. The dsRNA is then digested in vitro with RNase III (or Dicer) to produce a population, or cocktail, of siRNAs. Since the cocktail contains many different siRNAs, efficient gene knockdown is virtually guaranteed. A representative of one of these experiments is depicted in Figure 2.
The major benefit of this approach is the ability to bypass the testing steps involved in selecting an effective siRNA sequence, saving researchers both time and money (note that RNase III reactions are typically less expensive than those performed with Dicer). One downside of this approach, however, is the theoretical potential for nonspecific silencing effects, particularly for closely related genes. Most research to date indicates that this does not pose a problem (1-4).

Best for:
Fast and inexpensive analysis of loss of function phenotypes

Not suited for:
Long term studies or studies that require a single, defined siRNA sequence

Ambion's Solution: Silencer™ siRNA Cocktail Kit (RNase III)
With the Silencer siRNA Cocktail Kit (RNase III), a population of siRNAs to a specific target can be prepared in a quick, simple procedure. The population of siRNAs produced by the Silencer siRNA Cocktail Kit (RNase III) elicits gene silencing effects comparable to well designed individual siRNAs, without the need to design and screen individual siRNAs. Ambion scientists have tested many genes and they have yet to see an increase in cytotoxicity. They have also not observed any nonspecific effects associated with the use of siRNA populations as compared to individual, chemically synthesized siRNAs targeting the same gene. To date, Ambion scientists have used this kit successfully with NIH-3T3, HeLa, S3, 293 and BJ cell lines and have knocked down the expression of numerous genes, including c-fos, GAPDH, La, ß-actin, and Ku-70.

Figure 2. Gene Silencing with the Silencer siRNA Cocktail Kit. A population of siRNAs targeting 200 nt of the Ku-70 mRNA was prepared with the Silencer siRNA Cocktail Kit (RNase III) and transfected into HeLa cells at a final concentration of 100 nM. Cells were analyzed 48 hours later by immunofluorescence. Ku-70 levels were reduced 86% in cells transfected with the siRNA cocktail, compared to non-transfected controls.

In Vivo Expression

All of the methods described so far rely on the in vitro preparation of siRNAs. The use of siRNA expression vectors and PCR-based expression cassettes, however, relies on in vivo transcription of siRNAs from DNA templates introduced into cells. One advantage of these two approaches is that there is no need to work directly with RNA.

Method #4: siRNA Expression Vectors

Most siRNA expression vectors rely on an RNA polymerase III (pol III) promoter to drive the expression of a small hairpin siRNA in mammalian cells (1–4). RNA pol III was chosen to drive siRNA expression because it naturally expresses relatively large amounts of small RNAs in mammalian cells, it terminates transcription upon incorporating a string of 1-4 uridines, and its transcripts lack poly(A) tails.

To use siRNA expression vectors, two oligodeoxynucleotides encoding the desired short hairpin RNA sequence are ordered, annealed, and cloned into the vector downstream of the promoter. Because cloning is involved, the procedure takes several days, and sequencing the region containing the insert is required. However, this limitation is balanced by the ability to produce large quantities of vector once the vector is shown to work well in gene silencing experiments.

Without a question, the main advantage of siRNA expression vectors is that they are amenable to long term studies. Vectors with antibiotic resistance markers can be used to reduce the expression of targeted genes for several weeks or longer. Transient selection of cells transfected with selectable marker containing plasmids also permits the enrichment of cells that have taken up the plasmid. This can help compensate for low transfection efficiencies in difficult to transfect cells. Recently, several groups including Ambion have begun preparing adenoviral, retroviral, and other viral vectors for siRNA expression (see pSilencer adeno 1.0-CMV). These vectors offer the added advantage that cells can be transduced for gene silencing studies, thus reducing problems associated with inefficient plasmid transfection.

Best for:
Long term and other studies in which antibiotic selection of siRNA containing cells is desired. Retroviruses allow for efficient integration of the siRNA expression cassette.

Not suitable for:
Screening siRNA sequences (note: screening siRNA sequences is possible, but is time and labor intensive with vectors).

Ambion's Solution: pSilencer™ siRNA Expression Vectors

pSilencer siRNA Plasmid Expression Vectors feature U6 and H1 pol III promoters, an ampicillin resistance gene; an E. coli origin of replication; and one of three antibiotic selectable markers (neomycin, puromycin, or hygromycin resistance genes). An siRNA template is cloned into the vector resulting in siRNA expression once the plasmid is delivered to cells.

pSilencer-adeno siRNA Expression Vectors are also available, making expression vector delivery much easier in otherwise hard to transfect cells.

Figure 3. Long Term Silencing of GFP with pSilencer 2.1-U6 hygro. HeLa cells expressing cycle 3 GFP were transfected with pSilencer 2.1-U6 hygro containing an insert encoding an siRNA targeting cycle 3 GFP or pSilencer 2.1-U6 hygro without an siRNA-encoding insert. Following transfection, the cells were selected with hygromycin. Three weeks following selection, the cells were analyzed for GFP expression by fluorescence microscopy. Green: GFP. Blue: DAPI stained nuclei. GFP levels were remarkably reduced (94%) in cells transfected with the GFP siRNA-encoding pSilencer 2.1-U6 hygro siRNA Expression Vector as compared to those transfected with an "empty" siRNA expression vector.

Method #5: siRNA Expression Cassettes
siRNA expression cassettes (SECs) are PCR-derived siRNA expression templates that can be introduced into cells directly -- without first being cloned into a vector. Initially described by Castanotto and colleagues (5), SECs include an RNA pol III promoter, a sequence encoding an siRNA hairpin, and an RNA pol III termination site. In contrast to siRNA expression vectors, which require cloning and sequencing prior to use and thus take 1–2 weeks to prepare, SECs can be generated by PCR in less than a day. SECs thus provide an excellent screening tool to find the most effective siRNA sequence, or to identify the most effective combination of promoter and siRNA in a given cell line. In fact, SECs provide the perfect complement to siRNA expression vectors. By incorporating restriction sites at their ends, SECs found to effectively elicit gene silencing can be readily cloned into a plasmid or viral vector to create an siRNA expression vector. The siRNA expression vector can then be used for stable expression and long term studies. One disadvantage of SECs is that they are not as easily transfected into cells as siRNAs. As new transfection agents and protocols are developed, however, SEC transfection efficiency should increase. Because they are generated by PCR, SECs are not amenable to scale up without being cloned into plasmids. Figure 3 shows data from an experiment in which SECs each containing one of three different promoters were tested for their ability to drive siRNA expression and induce knockdown of c-fos expression.

Best for:
Screening siRNA sequences and testing promoters before preparing vectors

Not suitable for:
Long term studies until cloned into vectors containing selectable markers

Figure 4. Variable Reduction in Target Gene Expression Using SECs with Different Promoters. siRNA Expression Cassettes featuring the mouse U6 (Mo-U6), human U6 (Hu-U6), and human H1 (Hu-H1) promoters and encoding a c-fos-specific hairpin siRNA were transfected into HeLa cells. 72 hours post-transfection, the cells were assessed using nuclear staining with DAPI and immunofluorescence using a c-FOS antibody. Non-transfected cells (NT) as well as cells transfected with an SEC expressing a negative control siRNA (scramble) demonstrate wild-type levels of c-FOS. The relative level of reduction in gene expression was quantified and is provided in the bar graph.

siRNA and siRNA Expression Vector Delivery

Once siRNAs or siRNA expression vectors are obtained, they must be delivered to cells. Poor delivery efficiency is one of the main reasons RNAi experiments fail (or give false negative results) and deserves much attention.

Summing It Up

The use of RNAi to induce gene silencing is an exciting new method for creating loss of function phenotypes. The table above summarizes the five different methods for producing siRNAs. More information about each of these methods can be found throughout this issue.