Every cell in our body that contains a nucleus stores 2 meters of DNA in a volume that is only 1/100th of millimeter across (or 1/10th of the width of human hair). This is achieved through the non-random folding of the DNA. Correct folding of DNA is crucial for processes such as DNA replication, DNA repair and gene transcription. Failures in these processes can lead to diseases such as cancer or developmental abnormalities (i.e. congenital defects). Therefore a proper understanding behind the mechanisms that drive the folding of DNA inside the nucleus is important to understand how regulatory processes can go wrong.
A large group of proteins is crucial to the correct folding of the DNA inside the nucleus. However, many of these proteins are essential for the survival of the cell. Therefore, standard loss-of-function analyses (i.e. knock-out or knock-down) does not work for these proteins. Studying genome organization is further complicated by the fact that the genome organization is highly dynamic. To circumvent this problem, we have implemented so-called acute protein depletion methods. Using these methods we are able to achieve near-complete depletion of a specific protein with sub-hour time resolution. This allows us to chart rapid differences in the organization of the genome.
The Role of DNA Loops in Expression
These tools have allowed us to understand how a dynamic motor protein complex called cohesin can regulate genes by making dynamic loops in the DNA. By changing the dynamics of the motor, genes involved in maintaining the identity of the cell were misregulated. This led to the differentiation of stem cells.
We identified a new DNA structure that we call "fountains". These are regions where DNA loops spread out from a central point. Fountains appear in areas with many enhancers (DNA sequences that boost gene activity). We found that genes near fountains need cohesin to be active, and fountains exist in many different animals including mice, frogs, zebrafish, and worms.
Correcting the Scientific Literature
We created an acute degradation line for the protein, ZNF143, which had been associated with DNA loops for a decade. However, our experiments were not able to show a function for this protein in the formation of DNA loops. After careful investigation, we discovered that a commonly used research antibody (used to detect proteins) was detecting CTCF in addition to ZNF143. This mistake had confused the field for years. We found that ZNF143's real job is helping turn on genes that make proteins for mitochondria.
How 3D DNA Organization Affects Development
Studies into proteins that regulate DNA loops showed that their loss has limited effect on gene activity. But this is puzzling because mutations in these proteins cause serious birth defects (like Cornelia de Lange syndrome, caused by mutations in cohesin or its helper protein NIPBL).
To investigate this, we used "gastruloids"—lab-grown structures made from stem cells that develop like early embryos. We mapped which DNA regions were accessible in developing gastruloids and found they closely resembled real embryos.
When we removed CTCF or NIPBL during gastruloid development, something surprising happened: the cells could still differentiate into different cell types, but the overall 3D shape and structure of the gastruloid did not form correctly. This shows that cell differentiation (becoming specialized cells) and morphogenesis (building 3D body structures) are separate processes. We also discovered that CTCF has two different jobs: early in development it directly activates genes, but later it's needed for making DNA loops.
