Earlier this year, ScienceWatch.com
published an analysis of high-impact research in
Molecular Biology & Genetics over the past five
years. The analysis ranked David Bartel #2 among
authors publishing high-impact papers in this field,
based on 19 papers cited a total of 4,542 times.
In Essential Science Indicators fromThomson
Reuters, Dr. Bartel's record includes 76
original articles and review papers published between
January 1, 1998 and February 29, 2008, with a total of
10,386 cites. These papers can be found in the fields
of Molecular Biology & Genetics, Biology &
Biochemistry, and Plant & Animal Science.
Dr. Bartel is a Howard Hughes Medical Institute investigator whose lab
is based at the Whitehead Institute for Biomedical Research in Cambridge,
Massachusetts. He is also a Professor in the Department of Biology at
In the interview below, he talks with
correspondent Gary Taubes about his highly cited
What motivated your research originally on
these small, non-coding RNAs, and particularly the 2000 Cell
paper (Zamore PD, et al., RNAi: double-stranded RNA directs the
ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals,"
101: 25-33, 31 March 200), which is your most-highly cited paper
that’s not a review?
Tom Tuschl and
[see also] were postdocs in my lab working in
[see also] on the biochemistry of RNAi. In 1999, they
developed a cell-free system for studying RNAi, and they started using
that system to see exactly what was happening. They put double-stranded
RNA into the system and saw that it was getting processed down to the
21- to 23-mers.
Later, working independently in his own lab, Tom was able to chemically
synthesize those 21- to 23-mers and confirm that they could direct the
cleavage of the messenger RNA. He then showed that he could deliver them to
mammalian cells to specifically silence mammalian genes. That was a very
important technical advance, which has dramatically altered the way that
biologists working in mammalian systems do their experiments. His results
also implied that mammalian cells had the biochemical machinery to use
these small RNAs to direct the repression of messenger RNAs. Reasoning that
this machinery was not there just to help gene-knockdown experiments, my
lab started exploring whether endogenous small RNAs might be regulating
Are you surprised that paper has been so
influential, or is this what you expected?
"I think that
virtually every gene, at some point
in the growth and development of
the animal, is going to be found to
be influenced by microRNAs."
Well, it seemed very interesting to us at the time, so I can’t say
I’m surprised. And there was a lot of interest in RNAi at the time,
as there still is.
What prompted you to write the 2004 review in
Cell ("MicroRNAs: genomics, biogenesis, mechanism, and
function," 116: 281-97, 23 January 2004), which has garnered over
1,300 citations in just four years?
That was for a special issue commemorating some of the classic
Cell papers. These included the 1993 papers from the labs of
Victor Ambros and Gary Ruvkun. So Cell was looking for a review of
the microRNA research to go with reprints of those papers, and I thought
that those papers deserved commemoration.
There was also a lot of new information about microRNAs and ideas about
microRNAs that I thought would interest people. Clearly a lot of people
were going to start looking at microRNAs, and a lot of people were asking
me questions about them.
Did you see the review as a guide to researchers
who were now getting into microRNA research?
I was mainly trying to answer the types of questions my colleagues were
asking about microRNAs. A lot of people were curious about them. And of
course people in my lab had been thinking a lot about microRNAs, so I tried
to lay out what the understanding was at that time.
What was the understanding circa 2004?
We were just getting hints of the widespread targeting by mammalian
microRNAs. At that point we knew what was happening with plant microRNAs,
in terms of each plant microRNA apparently targeting just a few genes. But
it looked like the recognition by the mammalian microRNAs was going to be
much less precise—they could be regulating many more genes.
How has that story played out in the last four
Since then, there have been all kinds of data showing that in animals each
microRNA—at least, each highly conserved microRNA—is targeting
hundreds of messenger RNAs. In 2005, for instance, we showed that for each
microRNA or microRNA family—there are about 80 or 90 of these
families in mammals—there are hundreds of messenger RNAs that are
preserving their pairing to the microRNA over the course of evolution.
They’re preserving at a frequency that is much higher than
you’d expect by chance. That was the first clear indication that it
was true that each microRNA was regulating many messenger RNAs.
How big are these microRNA families?
Some families in mammals just have one member. For instance, miR-223 is the
only member of its family, whereas the let-7 microRNA family has at least
12 members—12 different hairpins making microRNAs that appear to have
largely overlapping targets.
Can you go into a little more detail about how
microRNA regulation differs in plants and animals?
ScienceWatch.com (only) general search results
with the term
"MicroRNAs" as the keyword.
In flowering plants, you have about 20 families of very highly conserved
microRNAs; some are conserved throughout all land plants, and they each
regulate a few genes. Usually those genes are highly related to each other
and encode transcription factors important for plant development. The
microRNAs direct the cleavage of the messages.
In animals, you have more families of conserved microRNAs, and each of them
has hundreds of messages that they downregulate. There are a couple cases
in which animal miRNAs direct the same type of mRNA cleavage as plant
miRNAs do, but for most targets they mediate a different type of mRNA
destabilization, or translational repression, or both. These messages code
for proteins involved in many different things. In fact, when you add it
all up, over a third of the human genes are conserved regulatory targets of
microRNAs. And many—most of them, in fact—are targeted by more
than one microRNA.
Is that number likely to change?
It will go up.
How high can it go?
To nearly one hundred percent. When I talked about the messages under
pressure to preserve their pairing to microRNA, I was talking about highly
conserved targets. There are additional microRNA targets that are
non-conserved. Their expression is dampened by microRNAs, but this
repression is either playing important roles only in particular species or
it has no role that is really under evolutionary selection.
So in addition to conserved targets, you have many, many non-conserved
targets, because microRNAs in mammals can recognize very short sequences in
the messenger RNAs. And these short sites are very common. As a result, you
have many, many messages that are downregulated by the microRNAs.
In addition, you have messages called "antitargets" because they are under
selective pressure to avoid targeting by some or all microRNAs. Putting it
all together, it is hard to escape the conclusion that microRNAs have a
very widespread impact on mRNA expression and evolution in animals.
Why do you think this small non-coding RNA story
took so long before it was recognized and elucidated?
add it all up, over a third of the
human genes are conserved
regulatory targets of
I think that if people had been looking for microRNAs, if they had known
what to look for, they would have been able to see them much earlier.
Everything was there 20 years ago to find microRNAs and to start learning
what they were doing. Of course it helps a lot to be working now when we
have all these genomes sequenced. But it would have been possible, if
people had looked for them. Perhaps nobody thought that there could be
interesting RNAs of this size range. Transfer RNAs—tRNA—were
already considered quite small. They’re about 76 nucleotides. And so
nobody was looking for anything smaller.
I can’t say first-hand what the block was, because I wasn’t in
the gene-regulation field. But if you look at how the first microRNA was
found, by Victor Ambros and his colleagues in 1993, they just followed the
genetics. They persevered. When you’re mapping a genetic mutation and
you don’t find a protein, some people are just going to say, "Oh,
this is too complex. I’ll just go and map a different mutation." To
find something this novel, you have to really stick to it, and that’s
what Victor’s lab did. When they didn’t see a protein, they
thought there still had to be something there and they continued to look
Another barrier is that these microRNAs, in many cases, don’t cause
huge changes in the output of the proteins, and so it’s harder to
find the consequences of losing the microRNA when it’s knocked out.
How do you see microRNA research evolving in the
next few years?
I think that virtually every gene, at some point in the growth and
development of the animal, is going to be found to be influenced by
microRNAs. So people working with particular genes, more and more, will be
considering microRNAs when they’re asking how a gene is regulated.
And, of course, the dysregulation of many genes is important in human
diseases. So microRNAs will be tied more and more to pretty much every
disease process in some way—perhaps sometimes in a small way,
sometimes in a more substantial way.
What are you working on now in your microRNA
We’ve been doing a lot of work trying to predict the targets of
microRNAs; fitting the microRNAs into gene regulatory networks. We’re
continuing to do that, and now we’re looking at what happens to the
proteins when you get rid of a microRNA. We want to know which proteins
change, and we’re using that information to learn more about how
targeting happens and about the biological consequences of the microRNA
A year ago we showed that there’s an oncogene, HMGA2, that loses its
microRNA regulation, and this appears to play a role as one of the key
steps leading to many human tumors. Others have shown that cells can have
too much of some microRNAs and that this can lead to tumors, or they can
have too little of other microRNAs and that can do it. What we showed is
that if you change one of the targets of a microRNA, so that the target can
no longer respond to the microRNA, you can also get tumorigenesis.
Following up on this finding, we want to learn what happens when other
particular targets are no longer regulated by microRNAs.
David P. Bartel
Howard Hughes Medical Institute
Whitehead Institute for Biomedical Research
Department of Biology
Cambridge, MA, USA
Dr. David P.
Bartel's most-cited paper with
1,370 cites to date: