microRNAs as an experimental tool: RNAi

In Section 12-6, we describe how microRNAs regulate gene expression by knocking down the expression of specific genes.  This knockdown can occur either by targeting an mRNA for degradation or by blocking its translation or both.  Somewhat independently of the identification of microRNAs as a natural cellular process, a laboratory technique called RNA interference (RNAi) was developed that relies on the same biological mechanisms as microRNAs.  RNAi and microRNAs are two faces of the same phenomenon, one being a laboratory technique while the other occurring in nature.   The connection between microRNAs and RNAi is both mechanistic and personal.  Not only do microRNAs and RNAi work by some of the same mechanisms, but the experiments were often done by the same laboratories. The significance of this work was recognized by the awarding of the 2006 Nobel Prize to Andrew Fire and Craig Mello, who were among those who developed the connection between microRNAs and RNAi. 

Figure A

An overview of RNAi.  RNA interference (RNAi) is done by introducing double-stranded RNA corresponding to part of the gene.  In this drawing, part of the coding region of the myosin heavy chain gene is microinjected into worms.  The double-stranded RNA (dsRNA) blocks the expression of the normal gene, and thus mimics the mutant phenotype of the gene.  In this example, the dsRNA blocks expression of the myosin gene, resulting in a paralyzed worm.  The effect is usually not heritable, so in the next generation, the worm has normal movement.

Geneticists are always interested in knocking out a gene or knocking down its expression in order to understand its function.  As sequences of individual genes and complete genomes have become available, this knockdown can be specifically targeted to individual genes using RNAi.  In a typical RNAi experiment, double-stranded RNA (dsRNA) corresponding to a portion of the transcript from a gene is introduced into the cell or organism.  The dsRNA specifically blocks or reduces expression of that gene, thereby producing a phenotype, as illustrated in Figure A.

For example, when dsRNA that corresponds in sequence to part of the mRNA from the myosin heavy chain gene is introduced into C. elegans, it causes the worms to be paralyzed. This phenotype resembles that seen for actual mutations in the myosin heavy chain gene, although is usually not quite as severe. But (with some exceptions) the effects of RNAi are not heritable because it is the mRNA that has been disrupted rather than the DNA. The DNA sequence remains intact to be passed to the next generation. Strictly speaking, RNAi treated individuals should not be called mutants (as that implies heritable change), though they are commonly referred to that way. A non-heritable mutant phenotype induced by environmental agents like RNAi is more accurately, but less commonly, referred to as a phenocopy.

The antecedents of RNAi lie in other experiments. RNAi is rooted in an older approach known as antisense technology.  Antisense techniques involve producing a single-stranded RNA, or another modified nucleic acid, that is complementary to the mRNA or sense strand. The single-stranded RNA is predicted to form a double-stranded hybrid with the mRNA and thereby block its translation or splicing. Indeed antisense techniques using modified nucleic acids known as morpholinos are still used in this way to block expression of mRNAs in a variety of organisms such as Xenopus and zebrafish. The mode of action for these antisense experiments and for RNAi experiments is shown in Figure B

Figure B

Modes of actions of dsRNA, antisense RNA, and microRNAs.  A single-stranded RNA is generated which makes a double-stranded RNA hybrid with the target mRNA.  This dsRNA may target the mRNA for degradation, block its translation, or both.

Many antisense experiments have been done over the years, and it is now clear that some of the results that were first thought to be mediated by single-stranded antisense nucleotides were in fact RNAi experiments arising from small amounts of double-stranded RNA in the preparations of antisense RNA. Indeed, careful follow up experiments in C. elegans showed dsRNA gave stronger effects than the single-stranded antisense RNA. Interestingly, double-stranded RNA mediated reductions in gene expression had also been previously reported in plants where the process is referred to as post-transcriptional gene silencing (PTGS). Despite the different terminology, the general mechanism for gene silencing by dsRNA appears to be the same in plants and animals.  Thus, although the experimental history of RNAi is ecent, the evolutionary history of the response to foreign dsRNA pre-dates the divergence of plants and animals.  It has been widely postulated that the RNAi response is an ancient type of immune response, protecting eukaryotic organisms from RNA viruses and the effects of transposable elements, both of which involve dsRNA molecules. 

As discussed in Section 12-6, it is now understood that small non-coding RNAs are important naturally occurring regulators of gene expression for nearly all eukaryotes, and certainly all multicellular organisms. The normal cellular response to these microRNAs involves many of the same biochemical steps exploited in RNAi experiments.  Since the cellular machinery used in microRNA regulation is evolutionarily conserved and found in nearly all eukaryotes, RNAi can also be done in nearly all eukaryotes. In a few organisms, such as C. elegans, microRNAs and also the RNAi effect can diffuse between cells so that the gene knockdown occurs throughout the organism.  In most organisms, the effect of RNAi occurs only in the cells into which it was introduced, so the experiments are usually done with tissue-culture cells. 

The mechanism of RNAi relies on normal cellular functions. How then does the dsRNA produce its silencing effect?  Just like microRNAs, the dsRNA molecule for RNAi knocks down the expression of the mRNA and thus the function of the gene by two different mechanisms, blocked translation or targeted mRNA degradation.  In either case, the introduced dsRNA has one strand identical in sequence to the mRNA (the sense strand) whereas the other is the complementary sequence (the antisense strand).  The introduced dsRNA is cleaved into fragments by the enzyme Dicer.  As its name implies, Dicer cuts the dsRNA into smaller fragments of 19-24 nucleotides in length.

The RNA fragments produced by Dicer have a characteristic structure—a central dsRNA region of about 19 nucleotides with a two or three base overhang at each end, as depicted in Figure C.  These fragments are called short interfering RNAs (siRNAs), and their appearance in the process is common to all organisms.  The sense strand of the siRNA is then degraded, leaving the antisense strand, which can form the double-stranded RNA hybrid with the target mRNA.  This dsRNA hybrid may block translation of the mRNA, as antisense molecules, morpholinos, and microRNAs also do. 

Figure C

Short interfering RNAs.  When the dsRNA is introduced, in Figure A, one strand is degraded leaving only a single strand.  This single strand is known as the short interfering or siRNA.  The siRNA, about 22 nucleotides in length with the overhang structure shown, forms the double-stranded hybrid with the target mRNA.

Alternatively, the antisense strand can become incorporated into a protein complex referred to as RISC (for RNA induced silencing complex); the presence of the siRNA in this complex targets it to the corresponding complementary mRNA.  RISC degrades the corresponding mRNA without degrading the antisense strand of the siRNA so the same complex can target and degrade many copies of an mRNA; this targeted degradation is also a mechanism used by microRNAs.  Because the same RISC is used repeatedly, the pool of mRNA from a gene can be degraded and no protein is produced.  Both translational blocks and degradation likely occur for nearly all RNAi experiments. 

Introducing the dsRNA into the cells.  All RNAi experiments require that dsRNA is introduced into cells or the organism, and the methods to introduce dsRNA depend on the organism.  In many cases, the dsRNA is directly injected into the cells or the organism, which is the method indicated in Figure A.  In others, such as mammalian cells, the sequence for the dsRNA is cloned into a viral vector that is introduced into the cells where it is expressed; the vector can also be based on a microRNA gene rather than a virus.  In C. elegans, which eat E. coli, the sequence for dsRNA is cloned into an E. coli plasmid, expressed in the bacteria, and then fed to the worm. For many organisms, libraries with dsRNA for most genes in the genome cloned into the appropriate vector have been constructed and are available commercially. These have been widely used for genome-wide mutant screens, as described in Box 15-2.

RNAi is proving to be a powerful method to knock down gene expression in many organisms, including those for which few traditional genetics techniques are available.  The process requires none of the standard tools of traditional genetic screens described in Chapter 14, such as mutations or even a genetic map, so it is easy to see why RNAi approaches are now common in the genetics literature.  RNAi is also thought to hold promise as a potential therapeutic agent for knocking down gene expression, such as blocking the genes involved in macular degeneration by direct injection of the corresponding siRNA into the eye.  Even if RNAi produces no new therapeutic agents, it has provided an easy and powerful means to silence gene expression in diverse experimental systems, opening them up to the power of genetic analysis.