Dr. Ambros: It was discovered using forward genetics. We began with a mutant of C. elegans that had an interesting developmental defect; in this case the problem caused really severe morphological defects and therefore was easy to study genetically. We were interested in timing defects that suggested that this gene was some sort of regulator of developmental timing. So that was the basis for the interest in the gene, and the mutant was the basis for the cloning which at that time involved standard physical mapping of the mutation. We narrowed it down to a small piece of DNA. However, the sequence within that piece of DNA was not predicted to encode any protein. We had to do some pretty careful work with the sequence to verify the fact that there were no proteins that could come from it. Once that was established, and once we confirmed the RNA gene product to be so small, the complimentarity between the RNA and its presumed target, Lin-14, was apparent.

Ambion: How long did it take you to convince yourself that the functional gene product was a 21-nucleotide RNA?

Dr. Ambros: What really helped was finding a second mutation that was just a point mutation. The first mutation turned out to be a deletion that removed the whole sequence from which the Lin-4 small RNA was transcribed. The point mutation was located right in that small 22-nucleotide sequence. That helped to reinforce the idea that it was a function gene product. Not only was it complementary to the Lin-14 mRNA, and we knew that Lin-4 was a repressor of Lin-14, moreover now there was a point mutation in that little sequence.

Ambion: Was there skepticism from the research community regarding a small, functional RNA?

Dr. Ambros: No, I think the paper (on lin-4) was pretty solid, there wasn’t any problem with the data, and it was therefore well accepted that this story was true. But I think the concern that all of us had was that this just might be a special case, peculiar to C. elegans or peculiar to nematodes. You know how it is, evolution can cause a lot of interesting situations of bizarre and amazing sorts of mechanisms. We figured this could be one of those bizarre, esoteric kinds of regulatory mechanisms that evolved perhaps only in nematodes. We hoped there were other cases like it, but they were slow in emerging. It was not until Gary Ruvkun’s lab found let-7 did it become apparent that this was a generalizable kind of scheme.

Ambion: You have worked as long as anyone to understand the mechanism by which miRNAs regulate gene expression. Can you share your thoughts on how miRNAs function?

Dr. Ambros: Well, I think that we have some basis for thinking of them as repressors of gene expression. That has come from both cases that have been studied genetically. In each of those cases the micro RNA seems to be a repressor of the expression of mRNA and protein. I think that what people tend to do is generalize and imagine that this will be the general case for miRNAs, that they will function as repressors. But really that remains to be shown. So far we have four out of four, so it seems fair enough to use that as a working hypothesis, and that they work by base pairing with their targets — mismatch base pairing in animals and plants, however there are many miRNAs that match precisely with their target and elicit degradation of that target through RNAi. So those are the two generalizations that we are operating under: in the cases where the micro RNA is a mismatch with its target, it will cause repression of translation, whereas when it is a perfect match, it will cause RNAi based degradation of the mRNA. You know, though, the precise mechanism for the repression of translation is still pretty mysterious at this point, and we need a lot more work to get a grip on that.

Ambion: Do you think that miRNAs directly regulate the expression of a few, many, or most genes?

Dr. Ambros: That is really a hard question, because if we can carry out computational analysis of mRNA sequences to try to identify the sequences that are complimentary to a given miRNA, generating a list of potential targets for that miRNA, the list can be very long depending on the stringency of the match criteria one applies. And the problem is at this point that we don’t really know what the rules are for productively base pairing with a 3' UTR on the target RNA. So it is hard to say how many of the predicted targets will turn out to be real targets. An optimistic or liberal view would have it that most of those predicted targets are real, and therefore almost any mRNA has one or more miRNAs that are regulating its translational efficiency, and that each miRNA may have dozens or hundreds of targets. The more conservative view would be that perhaps miRNAs really do base pair with lots and lots of different mRNAs, but maybe in only a few cases is the context appropriate for them to be really active. I think those are really, really interesting questions, and right now it is sort of a matter of taste whether you tend to be liberal or whether you tend to be conservative.