Q: Can you describe how
you got started in homeotic gene research?
A: In the era when I was an undergraduate
student -- this was in the '70s -- we were getting the first glimpse
of what a gene actually looked like. That was very exciting. That was
real frontier stuff. There was an emerging sense that we could try to
finally figure out how genes control embryonic development. Back in
the 1910s and 1920s there was a strong sense that genes and gene
products played a very important role in the embryonic development of
at least certain animals. But from the 1920s to the 1970s, it was not
possible to physically isolate any specific gene. That opportunity
first came available, fortunately for me, at the time that I was a
student. So many of us thought, Wow, we can finally dig in and
identify these really mysterious genes that had been postulated in
the 1910s and '20s as playing an important role in embryonic
development. There really was a sense then of this opportunity, this
urgency that we can get them.
Q: How did the isolation of homeotic genes
start to gather speed?
A: By 1982, two major homeotic genes were
physically isolated -- UBX and Antennapedia ... We spent probably a
year or so trying to visualize the activity of the Antennapedia gene.
The method involved taking the isolated gene, the piece of DNA, and
seeing where in the embryo it was active. Now, genetic studies
suggested that the gene would be active in the thorax. An insect has
a head, a thorax, and an abdomen; the wings stick out of the thorax,
the middle part. The expectation is that Antennapedia would be
active, or expressed, in the developing thorax of the embryo. But who
knew? It was just an inference from genetic studies. It wasn't
obvious that the gene would only be active there. We knew it had to
be produced there, but it could have been produced other places, as
We painstakingly worked out this method, using
very boring kinds of genes to work out the technique, genes that were
activated really late in parts of the guts, which made enzymes to
help embryos and maggots digest food. They're very abundant kinds of
gene products, but these are the kinds of controls we had to do to
work out the method.
Now it was time to apply those techniques to this
mysterious homeotic gene, Antennapedia. ... These were very long days,
playing around with things just to get this technique to work.
Without Ernst [Hoffen, who Levine worked very closely with in the
Basel Laboratory], it would have never worked. He's a very calm Swiss
guy. He could have been a watchmaker. Very steady hands. You can't
imagine two guys more dissimilar than the two of us, but we
immediately became friends. If it relied on me, I would have given up
after a month or two because these were very tedious, painstaking
kinds of experiments with no immediate reward. ... No one had ever
seen where a gene like this was active in a developing embryo. And,
in fact, we did theoretical calculations that suggested we might
never see the Antennapedia gene if it's really rare.
But finally we did. Sure enough, the gene was
active in the middle part of the developing embryo. You had an embryo
here with no pattern. I mean, you could barely tell head from tail;
[it] didn't look anything like an adult fruit fly yet. But there was
a band of gene activity right in the middle of the embryo which was
destined to give rise to the thorax of the adult fly. That was a very
gratifying moment. It showed that this gene, which controls the
development of the thorax, is in fact active in that region of the
embryo long before those cells "know" that they're going to give rise
to a thorax. ... At the time that we saw this localized band of
activity, the embryo doesn't even know head from tail. It doesn't
know stomach from back. It's just a bag of cells, a so-called
blastula. So we knew that this was a very significant observation. We
were very excited.
Q: What insight did this give you into how
A: The insight was seeing exactly which
cells express the gene. We know that every cell in our body has the
same set of genes. Our skin cells have the same genes as our muscle
cells or our blood cells. But these cell types are totally different.
For example, red blood cells carry oxygen; skin or muscles don't. All
our cells have the same chromosomes, but different genes are active
in different cells.
Let me make this very clear: There's a gene
called globin. All of our cells have the globin genes. Our skin
cells, muscle cells, blood cells all have it. But the globin gene's
[active] in the blood cells. When active, the globin gene ultimately
produces a protein called hemoglobin, which carries oxygen. Well,
skin doesn't need to carry oxygen so it simply doesn't express the
globin gene. This is referred to as differential gene activity. All
our cells have all the same genes, but different cells express
different subsets of those genes. That's the key to development, and
probably the key to evolutionary change.
Q: Explain that. First of all, how is it the
key to development?
A: For example, there are homeotic genes
that control head development. They make head cells different from
tail cells. Well, imagine if that gene was expressed or active in the
tail! You'd have two heads. Clearly, it's very important in that
example that the head genes are only active in the part of the embryo
that will form the head, not in the part of the embryo that will form
the stomach, back, or the tail.
Q: And how is it responsible for evolutionary change?
A: We imagine that you can get change in body shape and
body plans by taking critical genes and getting them expressed in new
places. Imagine, let's say, a gene that's important for making limbs.
You want to create another insect, a bug that has another set of
limbs. And maybe that gene now gets activated in a part of the
developing embryo where it normally is quiet, where it's normally
Q: So at a certain point in embryonic development, these
homeotic genes are being kicked off. This means that different
segments of the body are beginning to be defined?
A: Right. It's so early, the cells are still plastic. They
don't know what they're going to do. We can prove that at this early
stage, they don't know if they're going to be a head cell, a stomach
cell, a tail cell. Seeing this early expression was very consistent
with the idea that the gene was making a thorax. Not just a
manifestation of the thorax, not just a marker for the thorax. It was
making the thorax because of its early time of activity.
Q: What was the next step in your research?
A: We saw where and when the Antennapedia was active.
Another lab did the same thing for UBX. ... We had the DNA for those
two genes. The two genes are located in totally different positions
on the chromosomes in the fly. Despite that, we noticed that both are
active and inactive in very similar patterns. It's true that
Antennapedia was expressed or active in the developing thorax, while
the UBX gene was specifically active in the developing abdomen of the
embryo. But even though they weren't exactly in the same cells, the
overall dynamics of the two gene activities were remarkably similar.
... Now, seeing similar patterns of gene activity doesn't mean that
the proteins that these genes ultimately produce are going to be
related. But, man, you could feel in your gut that the patterns were
so similar that they had to be related somehow. These were the first
two homeotic genes that were looked at, and there was a sense that
they had to be related.
That prompted Ernst Hoffen to reread Ed Lewis's 1978 paper more
carefully. I mean, we were just really struck by the overall
similarities, the overall dynamics, of these two different homeotic
genes. And sure enough, in Lewis's paper, he predicts that homeotic
genes are related to each other. ... He's basically explicitly
arguing that different homeotic genes arose in evolution from a
common ancestral gene. This means that they should share a common DNAsegment.
That's what it means to a molecular biologist, and we were molecular
biologists. So we wanted to see if there was a common tag in these
different homeotic genes. ...
This would mean that we could isolate other homeotic genes using
the shared segment between Antennapedia and UBX. Sounds trivial, but
this was more work, in fact, than developing the method to visualize
the gene activity in the first place.
When we started we only had the Antennapedia gene. We didn't have
access to the UBX gene. But we can ask theoretically, if we take this
gene, label it up, and screen a so-called library -- this means you
ask this labeled Antennapedia gene who will it recognize in a
collection of random pieces of DNA from the fruit fly. We want to
know if some of these new pieces of DNA that are recognized
correspond to new homeotic genes. Well, we wound up with a mess. We
took the whole Antennapedia gene and we labeled it in a special way
with a radioactive chemical. We threw it down on this plate that
contained random pieces of DNA. And there were just too many spots.
We knew that in principle, you could isolate new homeotic genes
starting with Antennapedia. There was just too much background noise.
New methods had to be developed. This is where Bill McGinnis came
Bill McGinnis, I'll tell you here and now, has the hands of an
angel. He is the most talented experimentalist I've ever seen. He
came into the lab six or nine months after I was there. Hoffen and I
had already done our thing to visualize where the Antennapedia gene
was active. We had already had a sense that we could perhaps isolate
new homeotic genes, starting with Antennapedia, but there was a noise
problem -- new methods had to be developed. And Bill McGinnis was
exactly the man for the job.
The technique he ultimately used sounds so trivial. But what he
did was to simply take the Antennapedia gene and cut it up into
pieces. That was his great insight: He just took different pieces. We
used the whole gene and there was noise on the whole gene. There were
other common DNA sequence tags on it that made it light up with
nonsense. But Bill was able to separate the nonsense from the good
stuff, and he found a piece of DNA from the Antennapedia gene that
specifically recognized only other homeotic genes when you throw them
on a plate that contained random pieces of DNA.
After it was discovered that Antennapedia and UBX share a common
DNA segment, things moved fast and furiously. It was now possible to
isolate basically all the homeotic genes in the fruit fly very
quickly, starting with the Antennapedia and UBX genes which had been
isolated through painstaking work. And so we had kind of this bag of
tricks. We had five or six genes that were sequentially active in the
developing fruit fly embryo. Some up in the head. Others back down in
the tail. And we had a sense that this set of genes was basically
distinguishing head from tail, [and] thorax from the head or the
Then the next big breakthrough was using the same DNA segment
shared between Antennapedia and UBX that was used to isolate the
other homeotic genes in fruit flies to see what it lights up in other
animals. Ed Lewis proposed that the homeotic genes might explain body
variations within arthropods. Arthropods include things like shrimp
and lobsters and insects. Maybe the homeotic genes could explain how
these related body plans nonetheless become different from each
other. He was even so bold to suggest that this bag of genes could
distinguish an earthworm, which contains a series of more or less
identical body segments, into an animal like a fruit fly or a
lobster, that has a series of segments that are all different from
each other. Nobody thought that this bag of genes would be conserved
in radically different animals like us, like vertebrates. And that
was just incredibly exciting.
This set of homeotic genes from fruit flies was used to basically
ask, Is there something related in other animals, starting with
earthworms? But then, for the heck of it, throwing on DNA from
vertebrates. And something lit up. And as people isolated these genes
and looked at their activities in developing, for example, mouse
embryos, it was clear that although the fruit fly embryo and the
mouse embryo couldn't look more different from each other if you just
watch them without the aid of molecular markers, if you put on these
probes, these DNA segments to ask where are the genes active, now you
see a common blueprint, a common blueprint that probably underlies
all animal life on this planet. That was very profound.
Q: Did this discovery change your understanding of the
mechanism of evolution?
A: It led to a change in all of our thinking. When I was a
student, there was a sense that all animals did things differently.
If a gene was conserved between different animals, or between plants
and animals, it was considered to be boring. How could it be
interesting, because all animals and plants are different from each
other? So, certainly, creating distinct forms of life couldn't depend
on a common set of molecules. And so to see that genes that are doing
such profound things in the fruit fly -- making head from tail,
stomach from back, thorax from abdomen -- are conserved, related in
other animals ... this was just not predicted by anybody. At least
nobody that I ever read. So this was very profound. It meant that
there could be a common blueprint for all animal life on this
Q How has it changed the way you look at animal life?
A: Seeing that there's this underlying, common molecular
logic in all life comforts me. I can envision when that bug over
there underwent embryonic development, and that opossum over there
underwent embryonic development, they did a lot of things in
There is a common embryonic blueprint deeply imbedded in most or
all animals on this planet. The embryo of an earthworm and the embryo
of a mouse really do share profound things. You wouldn't know it from
looking at them superficially, but they really do.
In the case of the discovery of common homeotic genes among all
animals, there was a strong sense in the '70s and the '80s that
embryonic development among different animals involved completely
different molecules, completely unrelated. This was such a strongly
held view. And so, yes, it came as a huge surprise not only to people
like my mother who says, "My God, an earthworm and a mouse? An
earthworm and me, sharing things in common?" But it came as a
surprise to other scientists that there was this profound
conservation of mechanism of building embryos among all these
different kinds of animals.