Sometimes a driver has a foot on the brake but keeps it poised for a quick shift to the gas pedal, as when waiting for a light to change. When parking the car, however, it's better to keep the foot off the gas – and pull the emergency brake to ensure that the vehicle won't move.
A few years ago, while working in London, Ana Pombo and her colleagues discovered a parallel mechanism that cells use to control the production of RNA and protein molecules. In some cases, cells were both silencing genes and keeping them poised for action. Now, her lab at the MDC has discovered that silencing and poising come hand-in-hand throughout differentiation. In a study based on neurons, they show that cells use poising over their entire lifespans – even for some genes that are normally parked and never used. The work appears in the current issue of.
Silencing and poising go hand-in-hand
In gene activation, the feet and pedals are molecules that cells attach to DNA sequences to start or stop the transcription of RNA molecules. There are two basic types of brakes. One is used mainly to silence genes over the long term; if they are likely to be needed, a more temporary method is used. The Pombo team was studying embryonic stem cells when they discovered cases in which cells were temporarily silencing genes but simultaneously setting the stage for activation. In other words, the brake was on, but the molecular foot was near the gas.
"That would make sense in embryonic stem cells," Pombo says. "They're about to specialize, which will require a number of silenced genes to become active very quickly. Poising is an efficient way to eliminate some steps and provide the cell a faster reaction to the differentiation signals." The finding raised new questions: did the combination of silencing and poising only happen in stem cells, or was it also used later during specialization?
Lab members Carmelo Ferrai and Elena Torlai Triglia, lead authors on the paper, tackled these questions across the entire genome using neuronal differentiation, in collaboration with colleagues in London, Singapore, Naples, and Altuna Akalin of the MDC Bioinformatics Platform group. They found that poising always accompanies the more temporary type of gene silencing throughout the process of neuronal specialization – even for genes that usually never become active.
The molecules of poising
Permanent gene silencing – the emergency brake – usually involves blanketing DNA itself with chemical tags called methyl groups. This generally draws DNA into a very tight form so that molecules including RNA polymerase II (RNAPII) can't attach themselves to nearby sequences to begin the process of transcribing genes into RNA.
More temporary silencing involves a cluster of proteins called the Polycomb repressor complexes, or PRC. They also apply chemical tags – but their targets are the long tails of histone proteins, associated with DNA, rather than the DNA sequence itself. This process also tightens nearby strands of DNA, but not quite as much. There's still enough room for RNAPII, whose appearance usually marks the activation of a gene.
The new study proves that the PRC brake is always accompanied by RNAPII, though at these genes RNAPII doesn't make the mature RNAs that lead to the production of proteins. Instead, poised forms of RNAPII seem to transcribe very short RNA molecules that are probably broken down quickly.
A brake that could be released – but isn't
Activating a gene requires pressing a "button" on the surface of RNAPII – again through the attachment of a chemical tag. This apparently causes PRC to release its grip and loosen the packing of nearby DNA. That allows other molecules to enter the region, attach themselves to RNAPII, and build a larger machine to transcribe the poised gene.
For some genes this never happens; they remain inactive all the way through the final stages of neuronal differentiation. "This mainly affected a certain type of gene," Carmelo Ferrai says. "Surprisingly it was often genes that drive cell specialization – not of neurons, but other cell types, such as those involved in the development of cardiac tissue."
It's not clear why neurons apply a mechanism that typically silences genes temporarily to genes that will never become active. One possibility, Pombo says, is that this might not strictly be true: the genes may have some function in emergencies, for example when nerves have been damaged and need to adapt.
What would happen to cells in which the genes overcome silencing?
The functions of the genes near poised RNAPII suggest yet another possibility. Normally cells develop until they reach a stage of "terminal differentiation," a point at which their types are fixed and they no longer reproduce.
"But there are indications that some cells retain a very limited capacity to switch types," Elena Torlai Triglia says. "If that happens in the wrong context, the cell might become confused about its stage of development and identity – a phenomenon often seen in cancer."
One question the lab will pursue is what would happen to cells in which the genes overcome silencing. Another will be to see if the pattern is specific to neurons, or if it applies to many other types of cells. For now, the discovery of widespread poising has sharpened scientists' view of the type of gene silencing managed by PRCs and their many functions in cells.