Name: David Corey
Title: Professor, pharmacology, University of Texas Southwestern Medical Center
Background: Associate professor, UTSMC — 1998-2003
Assistant professor, UTSMC — 1992-1998
Postdoc, University of California, San Francisco — 1990-1992
PhD, chemistry, University of California, Berkeley — 1990
BA, chemistry, Harvard University — 1985
In 2004, University of Texas Southwestern Medical Center researcher David Corey spoke with RNAi News about his efforts to modify RNAi molecules in order to improve gene silencing in vivo (see RNAi News, 8/13/2004).
Earlier this month, Corey and his colleagues published a paper in Nature Chemical Biology showing that the same double-stranded RNAs that could be used to knock down gene expression could also be used to activate expression if targeted properly.
This week, RNAi News spoke again with Corey about his latest findings.
When last we spoke, a lot of the discussion focused on your work exploring siRNA modifications and locked nucleic acids. Could you give an update on your research in that area?
The work [detailed in Nature Chemical Biology] is really an outgrowth of that [earlier research]. Part of our interest was in seeing how we could improve recognition of any cellular target with chemically modified nucleic acids. That would include improving double-stranded RNAs, but also included trying to improve single-stranded oligomers and trying to have them recognize more targets.
One of the ones we looked at was peptide nucleic acids, and we were trying to target those to DNA. Toward the end of that work, we became aware of a couple of papers where double-stranded RNA was being used to target DNA. We thought that was something we needed to look into because it seemed like it would provide a totally new perspective on RNA-mediated recognition and new avenues for optimizing it.
That was how we got into [antigene RNA].
Are there ongoing RNA interference-specific projects happening in your lab at this point?
I think this is RNA interference. It’s just a different name. The mechanism is very similar.
I meant conventionally what people think of as RNAi.
Oh, sure. We’re continuing to do some of that as well.
One of the things we’re looking at now is targeting the 3’ UTR. We think that’s an interesting region to look at, especially because people are thinking that’s where microRNAs target. We want to see what happens when you try to target it deliberately. … Surprisingly, there hasn’t been a lot done on [this].
Touching on the antigene RNA stuff, can you give some background on the thought process that led to deciding to look into this?
We had success targeting peptide nucleic acids to the transcription start sites of genes, showing that we could put synthetic molecules into cells, have them recognize chromosomal DNA, [and] have them block gene expression, which was pretty exciting.
Then two papers came out in 2004 [from Kevin Morris et al. in Science and Kazunari Taira et al. in Nature] suggesting that RNA might be able to do the same thing. [Editor’s Note: The paper from Taira . was later retracted due to the lack of supporting data.] We were hesitant at first because it seemed really unlikely that double-stranded RNA, which is negatively charged and we think of as being synonymous with targeting messenger RNA, could actually target chromosomal DNA.
But we tried doing that because we had a good experimental system in which to test it from the PNA work. We made RNAs that were the same sequences as those we’d used for the PNAs targeting transcriptional start sites. And it worked beautifully. In the course of those experiments, we tested over 20 RNAs. Some were highly active and inhibited expression, but there were a few that reproducibly caused very slight increases in expression. That hinted that maybe some kind of activation was taking place, which is a crazy idea, but we couldn’t dismiss it because that’s what our experiments were telling us.
About this time last year, Bethany Janowski, who is a research assistant professor in the lab, had the hypothesis that if we did these experiments in a cell line where expression of our target gene — progesterone receptor — was much lower, we might be able to see the activation better because we wouldn’t be looking at it against a high background of basal expression. She tried doing the experiment in MCF7 [human] breast cancer cells [in which] progesterone receptor expression can be detected but is very low.
The first experiment came back, and it was beautifully activated. The rest of the [Nature Chemical Biology]paper is doing the controls and convincing ourselves that the result was real.
Do you have any sense of what mechanistically is going on?
We can make a good guess at this. There are two models that have been proposed for how the antigene RNAs work. The first is that they bind directly to DNA. We don’t think that is happening because we know from our previous work that we published last year [in Nature Structural & Molecular Biology]that Argonaute proteins are involved. And Argonaute proteins are known to be able to cause RNA to bind to RNA, not RNA to bind to DNA.
So what we think is happening is that there are rare transcripts that are being produced at promoters. These are not the transcripts that make protein; these are ones that may be produced one-one thousandth as much as the mRNA that produces a protein, but they are there. It’s becoming more and more appreciated that the majority of the genome is actually transcribed, except it’s just transcribed very rarely. The purpose of those rare transcripts so far has been unknown.
So the most likely mechanism [of how the antigene RNAs work] is that Argonaute helps these antigene RNAs bind to the rare transcript. The rare transcript is right in proximity to the promoter, so that this Argonaute/RNA/rare transcript complex is then able to associate with proteins at the promoter, alter their characteristics, and tilt the balance: If they’re activated before, it causes inhibition, or … if they’re inactive it can tilt the balance toward activation — basically perturb the sensitive mix of proteins at a promoter.
That would be my best guess right now, and I emphasize that it’s a guess.
I’d imagine one of the next steps is to see if that guess is correct.
Yes, and that’s what we’re doing right now.
Can you give an overview of the next experiments you’re working on trying to figure this out more?
Well, I don’t want to reveal too much, but obviously one wants to detect the rare transcript and one wants to do experiments to see whether its expression is important for the antigene silencing or activation that we observe.
These were 19- to 21-nucleotide long oligos [in the Nature Chemical Biology paper]. Did you try looking at different lengths?
We tried a little bit with different lengths, but it didn’t help and it worked worse, so we didn’t put a lot of effort into pursuing it beyond that. The beauty of this is these are regular RNAs, regular transfection protocols, you do nothing differently except you target them upstream of where the plus-one side is.
We might be wrong about the mechanism, but the phenomenon is there. We’ve tried two genes so far, and in both cases we were able to up-regulate them. So however it’s happening, it’s real. I think that if people have a gene that they want to see activated, there is no reason to stop them from trying this.
Do you suspect that this kind of activity might be occurring endogenously?
Yes. In fact, I’d be willing to bet money on it. The reason for my confidence is that it just works so well. I’m very experienced using silencing strategies, especially with antisense oligonucleotides where they can work well but it’s difficult to get them to [do so]. The beauty of standard RNAi is that you have nature working for you: the RISC helps your molecule find its target. That’s why it’s relatively easy to get it to work.
Here we have the same thing. Either with the repression that we’ve shown in previous papers or the activation we’re showing now, it’s easy. New students can get this to work practically the first time. We’re using the activating RNA right now as the positive control to train new students on, so this is not something you’ve got to push to get to work. That tells me that nature is our friend here; this is a natural mechanism, and for something to be happening that is this divergent from standard RNAi where you’re recognizing messenger RNA, that can’t be happening by accident. There has to be a reason why that’s occurring.