It’s been postulated that the mechanism of RNAi developed in cells as a means to fight off viral invaders and transposable elements. Essentially, this makes it a kind of cellular immune system, according to the authors of a paper published in this week’s Proceedings of the National Academy of Sciences. But this hypothesis, while reasonable, raises the question of how the RNAi system is able to deal with the equivalent of an autoimmune disorder, a system error where silencing is directed to a cell’s own genes.
Such a situation does not appear to be a problem for humans, which do not have RNA-directed RNA polymerase (RdRp) and appear to lack any alternate amplification pathway. However, it could be disastrous for other kinds of organisms that have both RdRp and systemic silencing; in such cases, a potentially fatal gene-silencing accident would not necessarily be isolated to one or two individual cells, but could spread indefinitely.
To sort this issue out, the University of Washington’s Carl Bergstrom and colleagues developed mathematical models to “explore how the RNA-silencing system tackles the central challenge faced by any immune mechanism: the need to rapidly generate specific responses to foreign pathogens while guarding responses against [the] self,” according to the PNAS article.
The first model the researchers developed, called A Basic Model of RNA Silencing, is composed of four steps: the RNAi system uses the presence of dsRNA as a warning signal; after Dicer cleaves the enzyme into fragments roughly 22 nucleotides in length, the fragments become associated with RISC units and are unwound into a single strand form; the RISC units target and bind with mRNA that has the same sequence as the single-strand of short RNA; either the mRNA is degraded in a sequence-specific manner, or the dsRNA that is complementary in sequence to the mRNA and therefore similar to the dsRNA that initiated the process is synthesized by RdRp.
According to the researchers, this model is able to successfully demonstrate the detection and degradation of the initial dose of dsRNA; the rapid generation of sequence-specific siRNAs by the rapid rise of RISC in the early stages of the reaction; the amplification of response and the production of secondary dsRNAs and siRNAs; and the silencing of the target mRNA.
However, they noted that the model also makes two predictions that don’t gibe with certain aspects of empirical RNAi studies.
Specifically, the model predicts that any starting dose of dsRNA will lead to the same level of silencing as long as the reaction is able to get going in the first place. This lack of dose dependence not only contradicts experimental studies, the researchers wrote, but it also implies that large quantities of dsRNA, typically a reliable indicator of the presence of foreign genes, would not be a factor in preventing self-directed gene silencing.
The model also predicts that gene silencing is self-perpetuating even in the absence of additional dsRNA. If this were the case, they wrote, “once a self-directed reaction is initiated, the cell will be unable to correct the mistake and attenuate the response. Taken together with the prediction of no dose dependence, this is a major problem: Even a small mistake will generate a full-scale and permanent silencing reaction.”
“This suggested to us that there’s a real autoimmune problem here if RNA silencing works as the standard picture shows,” Bergstrom told RNAi News. “Presumably, there has been quite ample time to evolve mechanisms to prevent these kinds of self-directed reactions. What are they?
“The answer was right there already in the literature.”
A paper that Titia Sijen of the Hubrecht Laboratory published in Cell two years ago [Cell 2001 Nov. 16;107:465-76] explained how an analysis of siRNA produced during RNAi in C. elegans “revealed a substantial fraction that cannot derive directly from input dsRNA.” Instead, Sijen wrote, these so-called secondary siRNAs appeared to derive from RdRp on the mRNAs targeted by RNAi. “The distribution of the secondary siRNAs exhibited a polarity (5’ to 3’ on the antisense strand), indicating a cyclic amplification process, in which RdRp is primed by existing siRNAs.”
“We saw this [article], and at some point it just clicked,” Bergstrom said. The researchers then extended their basic RNA silencing model to incorporate polar, or unidirectional, amplification.
In the resulting paradigm, the silencing reaction takes off, significantly amplifying the target-specific siRNAs. Eventually, the composition of the siRNA population changes such that the upstream siRNAs increasingly make up the total siRNA population, and because these upstream siRNAs are not able to prime further RNA polymerization, the silencing reaction eventually ends on its own.
“The unidirectional nature of RNA amplification limits the extent to which a single copy of dsRNA can be amplified,” Bergstrom wrote in the PNAS paper. “Hence, unidirectional amplification restricts the size and duration of most accidental responses directed against cellular mRNA but not those responses directed to viruses or transposons that continually generate dsRNA.”
Bergstrom compared unidirectional amplification to “the perfect microphone. You’d want this microphone that you could whisper into and then your voice would go booming throughout the room you were in,” he said. “But the microphone wouldn’t pick up any of that booming; it would just continue to hear your low-level whispering.”
Bergstrom and his colleagues caution in their paper that unidirectional amplification is probably not the only mechanism controlling self-directed gene silencing. “As with most potentially damaging processes, we expect multiple and redundant safeguards,” they wrote.
Among the other possibilities they suggest: chromosome-level silencing that would cause mRNA levels to fall below the threshold required to keep the silencing reaction going, cellular cooperativeness that would require multiple intercellular signals to start a secondary silencing reaction, and the localized enzymatic modification of dsRNA produced within the nucleus.
“All that our paper does is put a hypothesis out there,” he said. “[Unidirectional amplification] may or may not be the dominant way that self-directed reactions are avoided. We may or may not have properly interpreted the experimental evidence to date.
“But what it does is it makes a suggestion and makes a set of rigorous quantitative predictions about what will happen in other experiments,” he said. “We can follow up and see whether this is a key component of [the RNAi] process that in some ways needs to be built into our basic conceptual model of how RNA silencing is working.”
—DM