RNA interference (RNAi) is a process that inhibits the flow of genetic information to protein synthesis, which normally proceeds from deoxyribonucleic acid (DNA) to messenger ribonucleic acid (mRNA) to proteins - a process known as gene expression (see figure, top right). mRNA, a single-stranded polymer, is the key molecule whose sequence of nucleotides transports the information encoded in DNA to regions of the cell where that DNA-genetic information is used for synthesizing proteins. Understanding that flow of information is necessary for understanding the RNAi mediated feedback loops that inhibit gene expression.
Specifically, RNAi is a mechanism in eukaryotic cells that is triggered when such cells are exposed to certain double-stranded ribonucleic acid (dsRNA) molecules. The process has been detected in many organisms, including animal, plant and protist cells. The distinguishing characteristic of RNAi is destruction of mRNA molecules that share at least some of the sequence characteristics of dsRNA trigger molecules to which the cells have been exposed. RNAi has important biological roles in gene regulation and in protection of organisms against genetic parasites such as viruses and transposons.
The discovery of RNAi is a major technological breakthrough in biological research, perhaps as important as the development of the polymerase chain reaction (PCR), an in vitro technique that enables even tiny amounts of specific mRNAs to be measured easily. In experiments using RNAi in the fruit fly Drosophila or in the roundworm Caenorhabditis elegans, the effect of the loss of function of every known gene on a molecular pathway, cellular structure, or organism phenotype can now be determined rapidly and easily.[1]
The revolutionary discovery of RNAi resulted from the unexpected outcome of experiments by plant scientists in the USA and the Netherlands.[2] They were attempting to enhance the colors of petunia petals by introducing extra copies of a gene encoding an enzyme for flower pigmentation. While some plants with extra copies of the gene did show more intense colors, others surprisingly lost pigmentation (see figure). Analysis of the white petals showed that mRNA from both the endogenous gene and the newly introduced transgenes was absent (The term transgene refers to genes introduced into a plant by genetic manipulation from another species). Accordingly, the phenomenon was first called 'co-suppression of gene expression' but the molecular mechanism was unknown.
![]() |
A few years later, plant virologists made a similar observation during experiments aimed at increasing plants' resistance to viruses. They knew that plants expressing virus-specific proteins could have enhanced tolerance, or even resistance, to viral infection, but to their surprise they found the same effect using transgenes that only contained short regions of viral RNA sequences, too short to code for any viral protein. They concluded that viral RNA produced by transgenes could suppress virus activity and stop them from spreading throughout the plant. Reversing the experiment, by splicing short pieces of a plant gene into a plant virus, led to the discovery that these modified viruses could initiate the silencing of the specific plant gene. This 'virus-induced gene silencing' ('VIGS'), along with the cosupression phenomena have been collectively called post transcriptional gene silencing (PTGS), but currently the more usual scientific term for such phenomena is RNA interference.[3]
Subsequently, many laboratories began to look for this phenomenon in other organisms and cosuppression was documented in C. elegans and Drosophila.[4] In 1998, Andrew Fire and Craig Mello coined the term RNAi when they made a particularly notable discovery that the injection of double-stranded RNA into C. elegans led to a potent and specific gene silencing effect.[5] This represented the first identification of the causative agent (double stranded RNA) of this hereto inexplicable phenomenon. In October 2006, Fire and Mello won the Nobel Prize in Physiology or Medicine for their discoveries on gene silencing by RNAi.[6]
RNAi is but one of a group of mechanistically related gene silencing phenomena which share a number of distinctive features. In these 'RNA silencing' mechanisms, small single strands of RNA, appropriately called 'guide' RNA, silence gene expression by guiding protein complexes to the final target sites where gene expression is altered. The steps leading to generating these 'guide' molecules, all involve an early initiation event triggered by double-stranded regions in larger RNA molecules, and as discussed here involve the same cellular proteins (including ribonuclease enzymes named 'Dicer' and 'Argonaute').
RNA silencing pathways exist in animals, plants, protists, and fungi, and there are many variations in the final outcome, but they share a common remarkable mechanistic feature of a partnership between a guide RNA molecule and an Argonaute-like protein that ultimately modifies gene activity. The final outcomes vary, for instance in which other cell components are involved and whether DNA or mRNA is the main target for silencing. Despite their confusing variety (made more confusing by the different naming conventions used in different organisms), these pathways are all initiated by a larger double-stranded RNA trigger molecule, all involve the same types of specialized ribonuclease enzymes in generation of the small guide RNA, and all involve participation of an Argonaute-like protein guided by a small single strand of guide RNA to recognize a particular target, suggesting a common origin early in evolution.
To better understand current scientific discussions of RNAi it is helpful to remember that there are two mechanistically distinct classes of RNA-mediated gene silencing events recognized today. The first type is where a block to gene activity prevents formation of mRNA. Gene silencing by methylation of DNA[7] or histones are processes in this class.[3]
In the second class of RNA-mediated gene silencing mechanism, gene silencing results from interference with mRNA's cellular functions. Although in the past such events have been denoted as 'post-transcriptional gene silencing' (PTGS), they are now simply called RNAi. (See RNA silencing, Cellular and molecular mechanisms, and Matzke and Matzke (2004)[3] for details.)
The deliberate use of RNAi by plant scientists to reduce gene expression in plants (now usually called gene 'knockdown') has become common in recent years. Single-stranded antisense RNA that hybridized to a complementary, single-stranded, sense mRNA was deliberately introduced into plant cells to achieve knockdown. While plant biologists first believed that the resulting dsRNA helix could not be translated into a protein, it is now clear that the dsRNA triggered a RNAi response. The experimental use of dsRNA in biological research became more widespread after the discovery of the RNAi machinery, first in petunias and later in C. elegans.
It has been suggested that the basic RNAi machinery was present is the common ancestor of all eukaryotic cells with an ancestral role in defence against genetic parasites.[8] In current day organisms the RNAi pathway does indeed play a role in defending against viruses and other foreign genetic material, both in animals (as part of the immune response) but especially in plants where it may protect against the self-propagation of parasitic or selfish DNA such as transposons. The pathway is conserved across all eukaryotes, although it has been independently recruited to play other functions such as histone modification, the reorganization of genomic regions with complementary sequence to induce heterochromatin formation, and maintenance of centromeric heterochromatin.[9]
The native cellular RNAi machinery is used by both plants and animals to regulate groups of tens or even hundreds of cellular genes via key regulatory genes that produce natural substrates for the Dicer enzyme. These natural Dicer substrates are called primary-micro RNAs. For producing micro RNAs (miRNAs), certain parts of the genome are transcribed into relatively short single-stranded RNA molecules that fold back on themselves in a hairpin shape to create a region with double stranded primary miRNA structure (pri-miRNA). By 2006, thousands of miRNAs had been identified in plants and animals, including more than 470 in humans.
The Dicer enzyme then cuts 20-25 nucleotides from the base of the pri-miRNA hairpin to release the mature miRNA. If base-pairing with the target is perfect or near-perfect this may result in cleavage of messenger RNA (mRNA). This is quite similar to the events triggered in the cell by siRNA, however many miRNA's will base pair with mRNA with an imperfect match. In such cases, the miRNA causes the inhibition of translation and prevents normal function. Consequently, the RNAi machinery is important to regulate endogenous gene activity. This effect was first described for the C. elegans in 1993 by R. C. Lee et al of Harvard University. In plants, this mechanism was first shown in the "JAW microRNA" of Arabidopsis; it is involved in regulating several genes that control the plant's shape. Genes have been found in bacteria that are similar in the sense that they control mRNA abundance or translation by binding an mRNA by base pairing, however they are not generally considered to be miRNA's because the Dicer enzyme is not involved.[10]
RNAi has recently been used to study the function of genes in model organisms. Double-stranded RNA for a gene of interest is introduced into a cell or organism, where, by RNAi, it often drastically reduces production of the protein that the gene codes for. Studying the effects of this can yield insights into the protein's role and function. As RNAi may not totally abolish expression of the gene, this is sometimes referred as a 'knockdown', to distinguish it from 'knockout' in which expression of a gene is entirely eliminated by removing or destroying its DNA sequence.
Most functional genomics applications of RNAi have used the nematode C. elegans and the fruit fly Drosophila melanogaster, both of which are commonly used as 'model organisms' in genetics research.[11] C. elegans is particularly useful for RNAi research because the effects of the gene silencing are generally heritable and because delivery of the dsRNA is exceptionally easy. Via a mechanism whose details are poorly understood, bacteria such as Escherichia coli that carry the desired dsRNA can be fed to the worms and will transfer the RNA to the worm via the intestinal tract. This 'delivery by feeding' yields essentially the same magnitude of gene silencing as do more costly and time-consuming traditional delivery methods, such as soaking the worms in dsRNA solution and injecting dsRNA into the gonads.[12]
The dsRNAs that trigger RNAi might be usable as drugs. The first application to reach clinical trials is in the treatment of macular degeneration. RNAi has also proved able to completely reverse induced liver failure in mouse models (laboratory mice in which liver failure has been induced experimentally). Another speculative use of dsRNA is in the repression of essential genes in eukaryotic human pathogens or viruses that are dissimilar from any human genes; this would be analogous to how existing drugs work.
RNAi interferes with the translation process of gene expression and appears not to interact with the DNA itself. Proponents of therapies based on RNAi suggest that the lack of interaction with DNA might alleviate some patients' concerns about alteration of their DNA (as practiced in gene therapy), and suggest that this treatment would likely be no more feared than taking any prescription drug. For this reason RNAi and therapies based on RNAi have attracted much interest in the pharmaceutical and biotechnology industries. More recently, RNAi researchers have used RNAi to silence the expression of the human immunodeficiency virus (HIV) in mice.
RNAi is widely used to probe gene functions in plants. An example is a recent study from the laboratory of Jorge Dubcovsky where it was necessary to determine the function of a gene GPC-B1 that was thought to be involved in regulating wheat leaf senescence (and to affect cereal protein content).[13] This laboratory used RNAi to knockdown expression of the GPC-B1 gene in cereal wheat and found spectacular changes in grain germination after knockdown. The GPC-B1 knockdown wheat variety showed 30% less grain protein, zinc and iron, without differences in grain size, and verified that a single gene was responsible for all the effects. They suggest that increased expression of the gene can improve grain protein content and nutritional value.
More practically, Australia's CSIRO has developed a new experimental wheat variety with the potential to provide benefits in the areas of bowel health, diabetes and obesity. In this case, RNAi was used in wheat to increase the content of amylose, a form of starch that is more resistant to digestion.[14]
CSIRO and others have argued that cisgenic plants - that is plants created by genetic manipulation using RNAi (which they dub 'GM-lite')- pose less risks than addition of genes from other species (transgenics) as (they argue) new proteins are unlikely to be produced.[15]
![]() |
![]() |
There are two types of triggers for RNAi. These are: (1) fully double stranded RNA molecules (dsRNA), and (2) 'hairpin' forms of single-stranded RNA molecules in which the 'stem' of the RNA hairpin structure has complementary RNA base-pairing between two different parts of the same RNA strand (see figure to the left). These two types of trigger structure can both be substrates for a ribonuclease enzyme that is appropriately named 'Dicer'.
The RNAi process requires active participation of cellular machinery, and the properties of the Dicer enzyme are important for understanding this process.
Dicer has two active sites, both able to cleave its substrate RNAs. These active sites are separated some distance from one another in the Dicer enzyme (see figure to the right) and this explains the size of the RNA fragments they produce when they act on a double stranded RNA (dsRNA) target, a dicing process which is an early step in the cellular pathway that brings about RNAi.[17]
Dicer binds to and cleaves short double-stranded trigger RNA molecules (dsRNA) in two positions to produce double-stranded fragments of 21-23 base pairs with two-base single-stranded overhangs on each end. The short double-stranded fragments produced by Dicer when it acts on dsRNA are called small interfering RNAs (siRNAs), or miRNA when it acts on the cell's own transcribed, hairpin-forming miRNA precursors. One RNA strand - the guide strand - from these siRNAs.[18] is generated by an aggregate of several proteins, including DICER, called the RNA-induced silencing complex (RISC).[19]
In plants, protozoa, fungi, and nematodes, a cell-encoded RNA-dependent RNA polymerase (cRdRp, RDR; see figures) produces fully dsRNA trigger molecules for RNA from single-stranded RNA. The structure of cell-encoded RNA-dependent RNA polymerase from Neurospora crassa (shown to the right) is a relatively compact dimeric molecule and its core is a catalytic apparatus and protein folding that is strikingly similar to the catalytic core of the DNA-dependent RNA polymerases responsible for transcription.[16]
The RISC complex containing Dicer and Argonaute has an important role in processing trigger molecules to generate the short single stranded effectors of interference, termed the 'guide strand', and in degrading the other strand, termed the 'passenger strand'. The human RISC complex shows a nearly 10-fold greater activity using the pre-miRNA Dicer substrate over duplex siRNA. RISC can distinguish the guide strand of the siRNA from the passenger strand, and specifically incorporates only the guide strand.[20]
After integration into the RISC, single guide strands of RNA, either from siRNAs or miRNAs, bound to Argonaute-like proteins, can move to the target sites for gene silencing. One outcome that can occur at the target site is that the guide RNA can base pair to target mRNA and induce the RISC component protein Argonaute to cleave it, thereby preventing it from being used as a translation template. Proteins with a similar sequence to Argonaute (e.g. RDE-1, P-element associated wimpy testes (Piwi)) are present in nearly every eukaryote, from fungi to plants, flies, and mammals, often as gene families.[21]
Organisms vary in their cells' ability to take up foreign dsRNA and use it in the RNAi pathway. The effects of RNAi are both systemic and heritable in plants and in C. elegans, although not in Drosophila or mammals because of the absence of RNA replicase in these organisms. In plants, RNAi is thought to propagate through cells via the transfer of siRNAs through plasmodesmata.[19]
As this article emphasizes, the finding that various small molecules of RNA can inhibit the final expression of genes to generate the proteins they encode introduced a new paradigm in the biology of gene expression regulation. The possibility that the new paradigm will expand to include activation of gene expression by small RNAs now appears on the biological horizon. In November, 2006, Long-Cheng Li and collaborators[22] reported on several synthetic small, double-stranded RNAs, similar to small interfering RNAs, that activated the expression of human genes (specifically, E-cadherin, p21, and VEGF) — 2 to 10 fold increases in induction, with detection of protein levels — through sequence-specific effects on noncoding regulatory regions in the genes’ promoters. The authors could not identify the exact mechanism of gene expression activation, except to note the involvement of some of the same auxiliary proteins involved in RNA interference.
A subsequent study by another research group also found evidence of gene expression activation by small double-stranded RNA molecules.[23] A group of scientists at Yale University have also demonstrated that microRNAs can up-regulate translation, and also found that the microRNAs can switch from from repression to activation in certain circumstances.[24]
If unequivocally established and mechanistically defined, RNA activation (RNAa) might prove as therapeutically applicable as RNA interference. Detailed expositions of the research results of the above-mentioned scientists, current work in the field, and the implications of RNAa require a separate treatment (see RNA activation)