RNAi: the gentle art of shredding genes
This essay was written by Jocelyn Downey and was first published in the 2003 Mill Hill Essays.
At the beginning of this millennium a newly discovered genetic phenomenon was catapulted into the limelight and hailed by some as the most significant research in years. This research identified a novel mechanism by which cells can switch off their genes. The phenomenon was initially observed by a few different groups working in very diverse areas of biology. Plant biologists gave it the descriptive, if not entirely concise, title, post-transcriptional gene silencing, while groups working with fungi called it, more succinctly, quelling. However, when the phenomenon was found to exist in species as diverse as flies, worms, and mammals it became more widely known as RNA interference, or RNAi for short.
In order to understand this in more detail, and to see why this has generated so much interest both in medical research and other branches of biology, we must first look at the role of RNA inside cells. To this end, it’s useful to summarise a few fundamental aspects of how cells work. Every cell within any animal or plant contains a nucleus that stores all the information required for the cell to function, and this information consists of strings of the molecule, DNA, an acronym for deoxyribonucleic acid.
DNA as we often find it in cells is really two strands of DNA twined together into a long helix. Each strand consists of only four different types of molecule joined to the deoxyribose support, called bases, and usually referred to individually by their abbreviations, A, C, G, and T, for adenine, cytosine, guanine and thymine, respectively. These are linked together in an ordered manner to produce a single strand, and the order in which these occur on a DNA strand is commonly referred to as the sequence. The strands are held together by weak bonds between bases on opposite strands, and these bonds occur in a very specific manner: G will always pair-up with C on the opposite strand, and should never bond with A, or T, or to even itself. T, on the other hand, will always pair-up with A, and not with G, C, or another T. All of this information can be found on any two-pound coin celebrating James Watson and Francis Crick’s model of DNA, on tee-shirts, and in various science fiction movies. The structure of DNA is very revealing and its discovery provided important insights into how the molecule actually functions inside cells. The main function that we are going to be concerned with here is the role of DNA in the manufacture of proteins.
Proteins are involved in just about everything we do – they allow us to digest our food then turn it into energy, they hold our cells together, and give us our sight. The information to make every protein in our bodies is contained on the DNA strands. DNA serves as a long code with different regions coding for different proteins. A region of DNA encoding a protein is called a gene. Since every cell in the body shares the same sequence of DNA, there must be some way for each cell to control which bit of the DNA gets deciphered into protein at any given time. If this weren’t the case, all those proteins designed for a specific part of your body would be found in every other part of your body. For example it would be very inconvenient if your eyes were produced on your toes!
So, to prevent us turning into something resembling The Blob, cells do not make protein directly from DNA. Instead, whenever any particular protein is required, a cell will make a copy of the section of DNA required for that protein. This copy then moves out of the nucleus of the cell to be read and deciphered. The copy differs from the original in that it is made not of DNA, but of a related molecule, ribonucleic acid, or RNA. This molecule, rather than being double-stranded like DNA, generally has only one strand. It also differs from DNA in one other way: instead of containing thymine (T) it contains another base called uracil (U).
The copy is made through a process involving the RNA bases pairing-up with their complementary base on one of the DNA strands, so that the copy is produced in the opposite orientation to the original. In other words, a DNA sequence reading AATTCCAGGG will give a sequence of RNA that reads UUAAGGUCCC. If the opposite DNA strand (reading CCCTGGAATT, since opposing strands read in opposite directions) were to be copied into RNA this sequence would then read GGGACCUUAA, and this would not make the same protein, if it even encodes a protein at all. The RNA sequence that is read into protein is called the sense strand and the RNA sequence from the opposing strand is the antisense strand. Generally it is only the sense strand that is produced by cells, but if both strands are present together it is possible for them to bind together and form a double-stranded RNA molecule (usually abbreviated to dsRNA).
ince genes must be copied into an RNA molecule before the sequence is deciphered into protein, protein production can be stopped by interfering with either the DNA sequence, or the RNA sequence for that gene. Methods of stopping the production of a specific protein are highly sought after, because they have applications in human illness. It may be possible to fight some infections by targetting the genes of a particular virus or bacterium, in order to kill it or stop it causing disease. It may also be possible to treat a number of genetic diseases that are caused by the overproduction of a given protein, as is the case in a number of autoimmune diseases. Of course not all biologists’ work is focused on fighting disease, but even those biologists focusing on basic science would love to have the ability to disrupt a gene’s function. Disrupting a gene is an important tool to uncover how that gene functions. In fact, tools that prevent the production of a specified protein within cultured cells are especially desirable because this has proven notoriously difficult to do.
Not surprisingly, then, when it was discovered that introducing antisense RNA sequences into cells could inhibit the production of the protein encoded by the sense RNA strand, antisense RNA technology became all the rage. It was generally agreed that this worked because the antisense RNA strand could bind to the sense RNA strand, and in doing so, prevent it from being deciphered by the cell. Antisense RNA seemed to be relatively straightforward and antisense technology seemed to be an ideal method for blocking protein synthesis from any targetted gene. Its proposed mechanism was relatively simple and obeyed the law of common sense. It was almost inconceivable that it could ever fail to work. However, real life being what it is, people began to notice that antisense RNA did not live up to expectations. A number of genes failed to show any sensitivity to antisense RNA and performing experiments with it became a very hit and miss thing. No-one seemed to be able to give a satisfactory explanation as to why it would work for one gene but not for another.
Just to throw some more mud over the picture, some researchers even found that sense RNA, used as a negative control in antisense RNA experiments, was as effective as antisense RNA at preventing the expression of a gene – which made no sense whatsoever. It seemed completely counter-intuitive that adding more protein-encoding RNA into the cell should actually reduce the amount of protein produced from that gene.
Then, in 1998, while performing an antisense RNA experiment, a group of scientists working in the USA noticed that their RNA preparations were often contaminated with a small amount of RNA from the opposite strand. Since these opposing strands would be able to pair-up to each other, the scientists began to wonder whether the small amount of double-stranded RNA within the mixture might be exerting some effect. So they fed double-stranded RNA to Caenorhabditis elegans, a small worm sometimes used in experiments to study development, and noticed that it not only inactivated its target gene in the adult worm, but it also inactivated the target gene in its progeny. Indeed, only a very small amount of dsRNA was required to achieve this phenomenon, and it effectively prevented production of the target protein in several subsequent generations of the worm.
This was very odd. Antisense RNA was thought to work by each copy of antisense RNA binding to a copy of sense RNA, forming a complex that cannot be read by the cell. This should mean that each single antisense molecule blocks only one copy of the gene. Because cells make lots of copies of a gene when they make a protein, antisense RNA effects tend to be short lived. However, a small amount of dsRNA can effectively knock out a gene through several generations. This implies that the dsRNA does not work by simply binding to its target gene and inhibiting its decipherment into protein. It suggests that the inhibitory signal was somehow being amplified. The US group’s experiments also indicated that dsRNA did not just block the production of protein from the target gene but actually caused a decrease in the amount of the target RNA within the cell. When groups began to report that other organisms also responded in a similar manner to dsRNA, there was a lot of excitement in the scientific community.
Some researchers, however, advised caution before raising expectations too high. It had been known for some time that mammalian cells possess a self-defence mechanism against viral infections, called the interferon response. If cells detect the presence of virus inside themselves, they can react by shutting down protein manufacture, effectively committing suicide and stopping the infection dead (with the minor side-effect that the cell is also dead). This directly affects RNAi experiments because one way that cells detect the presence of invading viruses is by scanning for dsRNA. Double-stranded RNA is thought to be uncommon in mammalian cells, but can occur during viral infections. The danger of attempting to stop the production of a single protein in a mammalian cell using dsRNA is that its introduction into the cell may activate the interferon response and injure the cell in the process by globally affecting protein production. It came as no surprise, then, when it initially proved difficult to induce RNAi in mammalian systems.
However, as the examination of RNAi in different species continued, it emerged that dsRNA introduced into cells was quickly reduced into smaller fragments, around 21 nucleotides in length. And, since it seemed that dsRNA was indeed processed into smaller pieces in order to function, then that strongly implied that the small fragments of dsRNA could themselves be used to induce RNAi in cells. Sure enough, when these 21 nucleotide fragments were introduced into cells they were effective. Crucially these so-called small interfering RNAs (siRNAs) also worked in mammalian cells. It seemed as if the small size of siRNAs allowed them to bypass the interferon response. This, however, may itself be an oversimplification, since there is now evidence to indicate that even short fragments of dsRNA are capable of inducing the interferon response. Nevertheless, RNAi does seem to occur in mammalian cells without resulting in the death of the cells. So the questions remain: how exactly does RNAi work in the first place, and what is it there for?
When dsRNA is introduced into cells, it is broken down into short fragments. It was found early on in RNAi studies that not only is the dsRNA broken down, but also the RNA produced from the targetted gene. Furthermore, it seemed that only the target gene was affected and not other genes. This raised the possibility that the short dsRNA fragments might act as templates for the destruction of the target RNA molecule. Since then, it has emerged that dsRNA is indeed processed by a cell protein called Dicer, which cleaves the double-stranded molecule into the smaller fragments. Once cleaved it is less clear what happens to the short dsRNA molecules, though in general it is thought that they associate with a group of proteins, together known as RISC (for RNA-induced silencing complex). The dsRNA then separates into single strands within RISC. These single-stranded siRNA molecules are used as templates to direct the RISC to its target RNA and to destroy that RNA.
However, it has recently been found that dsRNAs do more than just initiate the degradation of RNA molecules. Double-stranded RNA, it seems, may also silence genes by affecting the DNA itself. Or it may produce changes in the proteins that normally coat DNA within the nucleus, preventing access of the specialised proteins which normally read the DNA sequence. Since changes to this DNA-protein complex, called chromatin, once initiated, can persist long enough to be inherited, this may provide part of the explanation as to why dsRNA can switch off a gene in successive generations.
The ability of cells to degrade RNA in this way may be another weapon in their armoury against disease, enabling them to react to the dsRNA produced by viruses and selectively degrade it. But, it is very likely that it has other functions. Small RNA molecules that do not code for protein have been found to be naturally produced within cells. They form siRNAs to degrade other RNA molecules. If cells can naturally produce their own dsRNA, that implies RNAi may also be involved in the day-to-day regulation of protein production inside cells. These RNA species, which have been termed micro-RNA molecules, are around 70 nucleotides long and have sequences that can fold back on themselves to form hair-pin like structures. They can be processed by Dicer to produce even shorter dsRNAs, that can associate with RISC.
As is clear from the above, many of the details of RNAi function remain sketchy, and there is still a lot to learn. In particular the roles that RNAi plays in cells and its mechanism of action are not fully understood. In addition, the fact that not all siRNA molecules are capable of knocking out their target gene, and the reasons for this are unclear at present.
With all the questions and possibilities that RNAi brings, a lot of excitement surrounds the subject at the moment, both in terms of understanding how it works, and in developing new technologies that can be applied to the growing industry of biotechnology and therapeutics. Indeed, RNAi has already been used to produce a strain of decaffeinated coffee bean, and many academics and companies are exploring new applications for RNAi. While it may not be necessary to fully elucidate the mechanism of RNAi before it is applied, there are still issues to be resolved before we see human therapeutics arising from RNA technology. We know that siRNA can activate not only RNAi responses in mammalian cells but also the interferon response, which may result in injury to the cells. More studies must therefore be undertaken to ensure that the siRNA is safe to use medically, and that the interferon response does not interfere with the intended result.
Much development work on delivery systems also remains to be done, since delivering RNAi to the correct site in a patient is not a straightforward matter. Experiments have shown that siRNA molecules do not seem to be particularly effective if they are directly injected into the bloodstream, or when applied to the skin. Scientists are therefore busy exploring different methods that might be used to carry the siRNA to its target, either by creating carrier molecules to piggy-back the siRNA, or by using something like a specially adapted virus to deliver it into selected cells. A number of challenges seem to lie ahead, and bitter memories of antisense technology still hang in the air. But the power to control genes is a rich prize, and a cautious optimism prevails over the field of RNAi research. There is the hope that the future might bring a string of technologies to treat a myriad of diseases but only hard work and time can tell if this hope will be realised.