At A Glance
Name: Craig Hunter
Position: Professor, Harvard University
Background: Assistant professor, Harvard University — 1997-2003; Postdoc, University of California, San Francisco — 1992-1996; PhD, developmental genetics, University of Colorado at Boulder — 1990; BS, chemistry, University of Oregon, Eugene — 1984
As a recently tenured professor at Harvard, Craig Hunter is continuing the RNAi work in C. elegans that he began before the technology had even been named. This week, Hunter took some time to speak with RNAi News about his efforts.
How did you get started with RNA interference?
I may have done one of the first purposeful RNAi experiments in 1995. The phenomenon had been described in C. elegans, but not completely yet — we didn’t know that the actual trigger was double-stranded RNA — we just had the primary observation that either in vitro synthesized sense or antisense RNA seemed to interfere with gene function specifically. It was only a few years later that [Craig] Mello and [Andy] Fire showed definitively that the actual trigger of both RNA preparations was double-stranded RNA.
At that same time, in the 1998 Nature paper, they also showed you didn’t have to be very good at your injections in order to get the gene silencing to work; that is, you didn’t have to hit the target tissue. Anywhere you could inject it, the double-stranded RNA was sufficient. This suggested that, somehow, most likely the double-stranded RNA itself was being transported between cells and tissues.
When you did your first purposeful experiment, how did that come about?
I’d been working on a specific gene and it was expressed both in the mother and in the early embryo. It’s also an essential gene, so you could never get a mother that was homozygous mutant, and therefore you could never get an embryo that didn’t have any gene product because it always inherited some from its heterozygous mother.
We jumped through the gymnastic hoops and made mothers that were homozygous deficient in the germ line only, and we could get about one of those embryos a week. Or, we could do this RNA injection — at that time either sense or antisense — and eliminate both the maternal and the embryonic function for a whole brood of animals. All of a sudden, we had dozens or hundreds of embryos to work with in an experiment instead of one a week. That’s why we tried [RNAi].
The irony is that the very first prep of RNA I made for this was contaminated with lots and lots of double-stranded RNA, and I decided to use it anyway just to find out if it was going to work. For all the other experiments I did for [a] paper I went back and got very clean single-stranded RNA so that it wouldn’t be a dirty experiment. It didn’t quite work as well, but I didn’t make the connection.
Where did that lead in terms of what you’re doing?
We used RNAi both before and after it was named quite extensively as a tool to investigate early embryonic development in C. elegans. Then, Craig Mello described at a meeting the systemic effect — the fact that you could inject a double-stranded RNA anywhere in the animal and that was sufficient. So we started talking about it and thinking about how that could possibly work, and we tried to design genetic screens that could help identify components of that process.
Then, about a year later, it was described that not only could the double-stranded RNA spread between cells, but you could simply soak the animal in double-stranded RNA or feed it bacteria expressing double-stranded RNA, and somehow that double-stranded RNA was able to trigger silencing in the animal. It was going into the animal from the environment, as well as between cells.
So, there were double the number of interesting questions and [the systemic effect] provided the means to do a real selective screen because you could just feed thousands and thousands of animals a particular double-stranded RNA — in fact that’s what Craig Mello did in his first screen for RNA-defective mutants. We set up a similar screen to look for RNA-defective mutants, but we arranged the screen in such a way that if they were simply defective for the mechanics of RNAi — chewing up the RNA, dicing the long double-stranded RNA into siRNAs — all of those things we’d recognize and discard immediately. We were only interested in mutations that disrupted the cell-to-cell transport or the import into the animal.
We screened through a little over a half a million animals, and found 200 mutations [that] defined five genes. We called the genes SID, for systemic interference defective. The first gene that we cloned [called SID-1]… encoded a large, predicted multi-transmembrane protein [that] was required in the cells [that] were taking up double-stranded RNA from the environment or from their neighbors — it wasn’t clear which.
It was required for uptake and it was expressed in all the cells that are sensitive to systemic RNAi, and its expression was not detected in the cells that are resistant.
In [another] paper, we showed that it is a transmembrane protein — we determined the topology of about half of the protein, identifying which particular putative transmembrane domains spanned the membrane of their orientation — and showed that when this protein is expressed in a heterologous system it was sufficient to allow the uptake of double-stranded RNA into cells.
You’re also doing a lot of work with … SID-2, is that right?
Right. We’re just getting ready to publish the SID-2 work … [and] I can give you a brief synopsis.
SID-2 turned out to be completely different than SID-1. It disrupts the second process — that is, the transport of double-stranded RNA from the environment into the animal. The clearest demonstration of that is, injection of double-stranded RNA anywhere into a SID-2 mutant completely bypasses the defect, so they’re fully sensitive to systemic RNAi if the double-stranded RNA is injected into any cell in the animal. But, they’re completely resistant to soaking and feeding means of delivering double-stranded RNA.
Where do you go from here?
There are multiple avenues. One major question is: Why does the animal do this? We see no overt phenotypes in the animal other than this resistance to systemic RNAi — there are no developmental or behavior phenotypes that we’ve been able to observe.
We hypothesize that there may be a role in accrued immune response against viruses, but no viruses for nematodes are known. [Systemic RNAi resistance] may have other roles in responding to environmental cues. It may be that double-stranded RNA is not the natural substrate for either of these proteins — it’s just a lucky break for a scientist that it works. So, SID-1 and SID-2 may not really be involved in transporting double-stranded RNA — it may be something else.
These kinds of questions we’re trying to address by examining the specificity and activity of these two proteins in vitro. The other obvious direction to take it is as a tool for molecular biology. It looks like expression SID-1 in nearly any cell system is sufficient to allow uptake of double-stranded RNA from the culture media of those cells. That’s potentially a powerful tool for doing systematic, high-throughput RNAi screens in tissue culture.
A really interesting observation is that there are SID-1 homologs in the sequence databases, but they’re restricted to vertebrates. So, we don’t find them in arthropods, with one exception [that] may be [due to] contamination — the source of DNA for sequencing that arthropod was an adult animal, so you don’t know that you didn’t have a nematode infection.
So, we’re investigating the function of the mammalian homologs, and we’re investigating the consequences of expressing the C. elegans protein in mammalian cells and in mice.
This potential research tool is being pursued currently?
That ties together with the [SID-1 and SID-2] patent application (see RNAi News, 9/8/2003)?
Is this something you are pursuing just within Harvard? Do you have collaborators?
We definitely have collaborators. We’re a C. elegans lab, and to work with mammalian tissue culture is one level of complexity that we’re tackling, but to work with mice is yet another level we’re getting ready to tackle. But we’re not quite ready yet. We want to characterize these genes and proteins in tissue culture first.
Is the interest in collaborations primarily academic at this point? Has there been any interest from companies?
I’m not directly involved in that. People have been contacting our technology transfer office — they’re handling those requests and negotiations.
Are there other areas where you’re applying RNA interference that we didn’t touch on?
We’re focusing on the systemic RNAi. There are plenty of people investigating RNAi mechanisms and applications. The systemic bit is a little niche that not a lot of people are working on.
The big problem with RNAi as a therapeutic is delivery, and investigating how C. elegans has solved that problem naturally may be informative for how we can solve [it] in the therapeutic setting. The function of the mammalian homologs is one avenue towards that.
It may be possible to modify [the homologs’] activity, but we need to discover what their activity is — whether they have the same activity [as SID-1] and then whether that can be modulated, for example, pharmacologically.
It’s possible you can take a pill and turn the pores on, then suck up RNA from the bloodstream — this is just fantasizing at this point, but the potential is out there.
If you manage to somehow crack the delivery problem you’ll have a lot of interest from the industry.
What we need right now is a lot of money and a lot of hands, so I’m hoping that somebody in the industry will get interested enough to get involved at an early stage — what may be considered an embryonic stage.
Have you ever considered … [licensing the technology from Harvard and starting your] own company?
It has definitely crossed my mind. It’s a matter of time. I have no experience in running a company. Colleagues of mine have experience in starting companies but not necessarily running them. I’m hoping that some far-sighted company will step up and help in the development.
If not, [launching a start-up] is always an option.