Professor, molecular/cellular biology, Harvard University
• Assistant professor, Harvard University — 1997-2003
• Postdoc, University of California, San Francisco — 1992-1996
• PhD, developmental genetics, University of Colorado, Boulder — 1990
Researchers from Harvard University this month reported genetic evidence that RNAi-triggering double-stranded RNA, as well as at least one intermediate RNA produced during the RNAi process, can serve as mobile silencing RNAs in Caenorhabditis elegans.
“This dsRNA intermediate requires the long dsRNA-binding protein RDE-4, the endonuclease DCR-1, which cleaves long dsRNA into double-stranded short-interfering RNA, and the putative nucleotidyltransferase MUT-2,” they wrote in Nature Structural & Molecular Biology.
“However, single-stranded siRNA and downstream secondary siRNA produced upon amplification by the RNA-dependent RNA polymerase RRF-1 do not generate mobile silencing RNA. Restricting intertissue transport to long dsRNA and directly processed siRNA intermediates rather than amplified siRNA may serve to modulate the extent of systemic silencing in proportion to available dsRNA.”
This week, Gene Silencing News spoke with Craig Hunter, the paper's senior author, about the findings.
Let's start with some background to set the stage for the paper's findings.
In Andy Fire and Craig Mello's initial discovery of RNAi in C. elegans in 1998, they showed that the double-stranded RNA that was injected into the animal did not have to be injected into the target tissue, but injection into the body cavity, the head, [or] the tail could still cause silencing throughout the animal and among its progeny. [This indicated] that RNA or an RNA-derived signal was global — it could move between cells and tissues.
Our [genetic] screen for mutants that disrupted this process identified SID-1, a large transmembrane protein, [as required for such mobility] and we have since shown that SID-1 is a double-stranded RNA channel protein. Its activity in heterologous systems for transport of RNA across membranes is specific for double-stranded RNA; DNA, RNA/DNA hybrids, and single-stranded RNA transport is not detected.
And, its activity appears to be gated. When we look at the electrical activity of these cells, we can show double-stranded RNA-specific changes in membrane conductance, suggesting that the SID-1 channel is a double-stranded RNA-gated double-stranded RNA transporter.
Despite all this evidence, we've never been able to demonstrate that the mobile signal in the animal itself is an RNA. The work that was just published … is genetic evidence that strongly implicates double-stranded RNA as the mobile signal, specifically long double-stranded RNA, which is what is typically introduced into C. elegans ...[and acts as a] substrate for Dicer cleavage and the generation of primary siRNAs. [The results from this work also indicate that] a product after Dicer cleavage is also mobile, but apparently RNA-processing products that are generated subsequently are not mobile.
Specifically, we are unable to detect mobility of secondary siRNAs … [and] these are likely the RNAs that carry out the bulk of the silencing in C. elegans. Since they are not mobile, they are trapped in the cells in which they are generated, presumably by an RNA-dependent RNA polymerase activity.
Thus, the less abundant primary siRNAs, along with the introduced long double-stranded RNA, are mobile, while the abundant secondary siRNAs function only in the cell where they are produced.
Is this lack of mobility a mechanism to ensure that specific silencing occurs only where it needs to?
We can imagine that it's an insurance protocol. A small amount of double-stranded RNA in one cell is not sufficient to spread throughout the organism. For example, accidental silencing of an essential gene in one cell triggered by a small amount of dsRNA may damage that cell, but since the amplified products can't move, the accidental silencing can't spread to other cells, which could kill the organism. The limiting factor for mobile silencing or systemic silencing is the amount of double-stranded RNA.
In hindsight, that the limiting factor for mobile silencing or systemic silencing is the amount of double-stranded RNA makes sense; we know there is a threshold for effective RNAi silencing. If it were all left up to the amplification system, we might expect a little bit of RNA to be just as effective as a larger amount, at least on an individual animal basis.
The other important finding is that transport is regulated. Only long double-stranded RNA and a form of Dicer-processed siRNAs are mobile. These findings also indicate that there are distinctions between primary and secondary siRNAs, [but what these are is unclear]. We don't know precisely what the mobility signal is. Is it just length? Is it double-stranded character? Is it modifications to the RNA? These results indicate there is something that makes these RNAs mobile.
Can you give a snapshot of the experiments you conducted to get these findings?
We created animals that were mosaic — composed of wild-type cells and mutant cells — for a variety of genes essential for RNAi processing and silencing. For example, animals that were genetically mosaic for Dicer had some cells that were Dicer plus and some that were Dicer minus. We could introduce double-stranded RNA into, specifically, the Dicer minus cells and see silencing in the Dicer plus cells. So we knew that the long double-stranded RNA had to move to the Dicer plus cells in order for silencing to occur.
When we introduced the long double-stranded RNA into a Dicer plus cell and asked if we could get silencing in a Dicer minus cell, we saw silencing in the Dicer minus cell. Therefore, a product of Dicer plus processing moved from that cell to the Dicer minus cell. That's our evidence that Dicer-processed products are mobile.
A similar analysis with other genes in the RNAi pathway allowed us to deduce that other products are not mobile in the same way. For example, mutations that disrupt RNA-dependent RNA polymerase activity prevent an effective RNAi silencing procedure. Introducing double-stranded RNA into an RNA-dependent RNA polymerase plus cell did not allow silencing to occur in neighboring RNA-dependent RNA polymerase minus cells. Therefore, an RNA-dependent RNA polymerase product cannot move from a plus to a minus cell and cause silening.
What are the next steps going forward?
In combination with our in vitro work, we're really interested in identifying the structural features that promote or restrict transport of RNAs.
SID-1 is an important part of this mobility system; it is highly conserved. But the transport of RNA into vertebrate cells or mammalian cells has been problematic, or inefficient, at best. Our hope is that genetically dissecting the mobile silencing system in C. elegans will give us clues to the modifications of RNAs that may enable transport in mammalian systems.
The pessimistic observation from our earlier work is that if the specificity for double-stranded RNA that we observed for SID-1 holds true in mammalian systems, any modification to the RNA may interfere with the ability of the mammalian SID-1 homologs to transport it. Even endogenous modifications such as methylation may act to modify mobility, much less the dramatic modifications the therapeutic industry is making to siRNAs.
[Yet] the clear prediction from this work is that there are modifications to the RNA that controls its mobility. If we can identify those modifications, we may be able to apply them to mammalian systems to enable more efficient transmembrane transport of RNAi therapeutics.