Genes are poised on the ‘starting blocks’, waiting to spring into action to build an embryo.


Mas G, Blanco E, Ballaré C, Sansó M, Spill YG, Hu D, Aoi Y, Le Dily F, Shilatifard A, Marti-Renom MA, Di Croce L.
“Promoter bivalency favors an open chromatin architecture in embryonic stem cells.”
Nat Genet, 50(10):1452-1462. doi: 10.1038/s41588-018-0218-5. Epub 2018 Sep 17.

It’s the day of the Big Race. All the cars are lined up on the starting grid with their engines revving and tyres gripping the tarmac, waiting for the signal to start. The lights change and most of the cars go shooting down the track in a blur of smoke and squealing tyres, while a few unfortunate competitors have stalled and will have to withdraw from the race.

The adrenaline-fuelled world of motor racing may seem a long way from the complex molecular processes that turn a single fertilised egg cell into an embryo, but they are no less thrilling.

In the first few days of development, a mammalian embryo like a human or mouse is nothing more than a small ball of stem cells, each with the potential to become any of the hundreds of types of cells that make up the body.

At this stage, life moves fast. Cells are multiplying rapidly and having to make quick decisions about whether to follow one fate or another. These choices are made by a set of genes that are controlled by bivalent promoters – two-way genetic control switches that are poised either to turn on in early development and rapidly drive high levels of gene activity, or to switch off and shut down the gene completely.

Researchers had previously discovered that there are opposing types of chemical ‘flags’, known as histone modifications, that are present on these two-way switches – one telling the gene to GO (be switched on), and the other acting as a repressive STOP sign. The active tags are put in place by a molecule called MLL2, while the silencing marks are put on by Polycomb proteins.

At the start of development these STOP and GO signals are perfectly balanced, keeping the gene poised and ready to quickly flip into the correct pattern of gene activity, either on or off, like a racing car on the starting line.

To find out more about the interplay between the two type of signals on the switches, Luciano Di Croce, a group leader at the CRG, teamed up with Marc A. Marti-Renom at the CNAG-CRG and Ali Shilatifard at Northwestern University in Chicago, publishing their findings in the journal Nature Genetics.

It’s obviously very difficult to study the first moments of development in the womb of a living animal, so the researchers turned to mouse embryonic stem cells growing in the lab. Under the right conditions, these cells will grow into tiny clumps known as embryoid bodies.

Although these clusters of cells aren’t exactly the same as a real embryo growing from a fertilised egg – and would never be capable of growing into a baby mouse – they capture some of the early decisions and changes in gene activity that happen during development, which are driven by genes controlled by bivalent switches.

Using genetic engineering techniques, the scientists removed MLL2 from embryonic stem cells. This effectively wiped out all the activating GO signals from the bivalent gene switches and left only the STOP signals, tipping the balance strongly in favour of genes being switched off.

As might be expected, many important developmental genes weren’t activated when they should have been, and the modified cells were no longer able to grow into embryoid bodies. Looking more closely at the location of these genes inside the nucleus, the researchers discovered that these genes had been relocated to regions that are usually associated with inactive genes.

“We found that changing the balance of histone modifications at these promoters had profound effects on gene activity and genome structure,” explains Di Croce.

“Genes that should normally be active were packed away in areas containing silent genes that are not normally needed in these cells – it’s the genetic equivalent of parking a car in the garage if you aren’t driving it any more,” adds Marti-Renom.

There was something else unusual about the genes with two-way switches in cells lacking MLL2, where the balance of signals had been tipped towards STOP. In normal embryonic stem cells, highly active genes form a loop so that their starts and ends are very close together. This means that the gene-reading machinery can quickly hop from the end back to the beginning to start again, like a racing car whizzing round and round a circuit.

But in cells without MLL2, the starts and ends of the genes were far apart, making it difficult to achieve very high levels of gene activity and revealing yet another way in which genes that should normally be poised for action are silenced.

Overall, the team’s findings start to illuminate the complex interplay between histone modifications and three-dimensional gene organisation at the very earliest stages of development, when cells are quickly making decisions about what to do in order to build an embryo. There are also implications for understanding what might have gone wrong when development goes awry, leading to miscarriage or birth defects, and in diseases involving disrupted gene activity such as cancer.

“We now know more about the role of histone modifications at bivalent promoters and why they are important for proper activation of the genes,” says Di Croce. “It’s clear that there needs to be a balance between active and repressive marks in order to maintain the looped conformation for quick activation, and we now understand what happens when that balance changes.”