Leonie's research lab at the IMP in Vienna, Austria is supported by the NET programme. Brona chats to her about sumo-wrestlers, boxers and the mysteries of molecular memory

B: How do cells remember who they are?

L: Well, I can tell you part of the story. So each cell expresses a particular handful of genes, and I work on how they remember to ‘express’ those genes

B: So cell identity has to do with which bits of the genome are switched on and which bits are switched off?

L: Yeah, so just to give you some numbers, the human genome has about 24 000 genes and flies have around 16 000

B: Considering their size [flies] that’s actually quite a lot of genes!

L: [laughs] Yeah, but any one cell will only be ‘expressing’ a few hundred genes within that cell

B: So in my liver cells, only like a few hundred genes of the 24 000 in my genome are switched on? But, they’re all in there…

L: Yeah, they’re all in there. Some of them are on in all cells, because there are genes that tell a cell to be a cell – housekeeping genes. Then there are just a handful of really specific genes that tell a liver cell to be a liver cell. But the big job in cell identity is actually keeping all the wrong genes switched off. It would be a disaster for the liver cell if it started expressing 90% of the genome

B: Hmm, it could get a bit schizophrenic then if the wrong genes get switched on.

So you work on flies? How do cells in flies remember what to be?

L: There’s the initial decision of the cell to become a brain cell or whatever…

B: Which happens in the developing embryo?

L: Yeah, then there’s a kind of maintenance phase where that ‘decision’ is remembered every cell division, and I work on this idea of maintenance and memory. If you’re a 3 year old kid and you have an experience in an art lesson and you become an artist later on, you don’t keep having the same experience, but it stays with you.

One thing that’s really important in this whole discussion is the idea of cascades or hierarchies of genes. So there may be a few hundred genes in your liver or brain, which help cells to be liver or brain cells

B: And in flies, is it a similar number of genes that are switched on in those cells?

L: Yeah. But to give a cell identity, you need to initiate the head gene in the hierarchy, and then all the other necessary genes get switched on. So one gene is the boss

B: So in a wing cell, the wing boss gets switched on and that rallies the troops needed to make a wing cell. So when wing cells divide, how do they remain true to their parents?

L: That’s something that we’re just finding out about. The particular regulators that I work on are Polycomb and Trithorax groups of proteins. I tend to think of Polycomb as a sumo wrestler and Trithorax is like a boxer, and they’re fighting eachother all the time. There’s only room for one of them to win in the boxing ring.

In flies, this boxing ring is represented by about 1% of the genome. They only fight over about 200 genes in the genome. Those are the boss genes, the ones at the top of the hierarchy that really determine the fate of cells.

So Polycomb and Trithorax fight over these genes. The outcome of the fight is decided in advance by what’s going on at the promoter of the gene…

B: And the promoter is like the on-off switch of the gene?

L: Yeah, so if the promoter is on, so is the gene that it controls. But, whatever switched that promoter on doesn’t stay there for the whole life of the fly. It’s like this experience you have when you’re 3, it’s not there anymore, but you remember it. So this is the mystery of Polycomb and Trithorax: they see what the experience was and reinforce it, reminding the cell every time it divides

B: So what do these proteins actually do? So these are groups of proteins…

L: Yeah, a family of sumo wrestlers

B: Only one of which will presumably be in any one cell at any one stage or…

L: No, not at all! They go around as a pack, a team of sumo wrestlers fighting a team of boxers. The tendency as far as we know is for the Polycomb group of proteins to win the fight. So the Trithorax group, the boxers have to work really hard to dislodge the Polycombs

B: So if the Polycombs, the sumo wrestlers win, what is the outcome for the gene?

L: The gene will be silenced

B: So you’re saying that most of the time, these genes will be silenced?

L: Yes, exactly. So it comes back to what I was saying about cell identity. Most of the genes need to be switched off.

B: So those sumo wrestler proteins go around keeping everything quiet?

L: Yep

B: Cor, that’s a lot of work, really isn’t it?

L: [laughs] Yeah, but it’s not so much work because they don’t do it for all 16 000 genes, they only do it for about 1%. They’re very clever, they just target the bosses. Then the bosses target the workers.

B: That’s super efficient!

L: The way they find the genes is by looking for little tags, so there’s a little boxing-ring sign on these genes. So it’s coded on top of the DNA, but can talk to both Polycomb and Trithorax

B: So these proteins aren’t reading the DNA sequence? They’re talking to a memory tag stuck on top of the DNA?

L: Yeah, that’s right. So it’s like a boxing-ring that would be rigged for either the sumo-wrestler or the boxer if you like. And we don’t exactly know what that tag is…

B: So this is a mystery.

L: Indeed. There are a lot of theories out there. Some people would disagree with me, some people would say they do know, but I think it’s still open.

So let’s go back to the idea of real cells rather than sumo wrestlers. What has to happen is that when the cell divides it has to duplicate everything. Then it has to separate the two new copies into two new cells. Making copies and fairly separating things between two new cells is a problem. We really don’t know what that physical tag is sitting on top of the DNA

B: So the idea is that this little memory mark on the DNA gets duplicated as well, when DNA replicates?

L: And carried over into new cells, yeah. There’s a huge amount of upheaval in this process. The chromosomes have to compact, physically a very traumatic event. So as the chromosomes get much more dense, many things bound to them jump off.

All gene activity closes down at that time, so you can’t really hypothesize that if the promoter is active, it stays active. All of the promoters get shut down during this stage of cell division. It’s like you’re going through a washing machine. Yet, coming out again you have to remember exactly how things were beforehand.

B: What a tremendous process! If there are 16 000 genes in the fly with only a couple of hundred genes switched on in a cell, new cells need to remember which ones to turn on again, and this is all down to little memory tags? Wow!

L: At the moment, there’s a huge debate about what the molecule is that serves as a memory tag…

B: So what do you think?

L: There are a lot of people that think it’s down to modifications of histones, proteins that bind to DNA. A lot of what’s driven epigenetic research in the last few years is that small modifications to these proteins can carry information.

My feeling is that that’s not enough. The amount of information you could have in a mark like that would not be enough for the Polycomb proteins. There is some evidence to support that the marks that correlate with activation wouldn’t be enough to cause activation

B: So you’re saying that some people believe that the mystery memory tag isn’t on the DNA itself, but on one of these histone proteins, or a bundle of them that DNA is wrapped around. But you don’t think that’s enough?

L: No I don’t think so, because that idea is based on correlation, that you often find these tags together with silencing or activation, but there’s no evidence that the tag can actually cause the silencing or activation.

That’s asking much more of the tag, so I think it’s more the consequence of silencing or activation that you get these modifications, which may reinforce either state, but there’s not enough evidence to conclude that if you have the tag, you’ll inevitably get silencing.

So what I think is that if you have one of these little DNA sequences [if you’re a boss gene] the default state is that Polycomb proteins sit there and silence the gene

B: So the sumo wrestlers have a tendency to just sit there and keep everything quiet, unless the boxers come along and say ‘get off that space, I’m having it’?

L: Yeah, and they really have to fight to do that. It’s not the default state. The thing is really rigged in favour of the sumo-wrestler.

I think traditionally the field has really worked on silencing, partly because the first genes that were discovered in the 1940s were the Polycomb group genes, which encode the sumo-wrestler proteins, so people have a much longer history of silencing. But I think we need to change our focus and ask ‘what makes the cell remember activation’? Remembering silence is easy

B: So how does this research translate into other animals and humans, in particular?

L: It’s becoming clear that there are many, many parallels. In 2006, people identified the target genes of Polycomb and Trithorax group proteins in mice and other organisms. Discovering these boss genes was a big breakthrough last year. So it really seems as through this strategy of targeting master regulators; silencing them if you’re Polycomb and activating them if you’re Trithorax, exists in other organisms