Guest Blog Series 2026

By Andres Hernandez Maduro

Chromosomal DNA strands are long. In each of your body cells, there is approximately two metres worth of DNA packed tightly in a chamber (the nucleus) around five micrometres in diameter. They are also quite sticky, and if thrown into a random soup of molecules will tend to aggregate into clumps. Thus, it is the curiosity of many-a-scientist that cellular DNA is well-organised and shows a remarkably consistent three-dimensional structure – as if a ball of yarn were to form exactly the same weaves, knots and loops every time you got a cat to play with it.

My name is Andres, and it is probably not surprising that I work in the DNA Topology lab at Oxford. As a DPhil student, I study how DNA in eukaryotic cells not only loops and folds in predictable patterns, but also how nuclear proteins regulate chromatin structure and keep chromosomes where they should be over the cell cycle. One key player in this process is cohesin, a proteinaceous ring that is able to grab onto DNA and progressively extrude it into loops, similar to a rope sliding through a belay device. Once generated, DNA loops allow genes and their regulatory sequences to come close. They also provide chromosomes with structural rigidity so that they can be more easily segregated during cell division. A question, now: given the precise chromatin structure we see in cells, how can the position and size of DNA loops be regulated?

In mammals, the answer is the occurrence of physical boundaries. As cohesin extrudes loops, a protein called CCCTC-binding factor (CTCF) is placed in neighbouring regions of the DNA in a directionally convergent manner. These stop further loop extrusion by stalling cohesin from moving any further, acting in the above scenario as the brakes locking the belay. Recently, a collaborative study between groups in Harvard and MIT combined protein analyses with chromatin landscape assays to elucidate how CTCF establishes this blockage.

Doubling the scaffold

Convergent CTCF proteins have long been known to pair up (i.e. dimerise) when establishing boundaries for DNA loops. However, how CTCF interacts with the surrounding nucleosomes – units of DNA wrapped around scaffold proteins to create chromatin – has remained unclear. Rather than tackling the issue from the top-down, Valeriano et al (Biorxiv, 2026) began by designing and purifying a construct made of only a single nucleosome with the 42 bp DNA binding sequence for CTCF stretching out. They then incubated this with purified CTCF proteins and analysed their interactions.

Strikingly, the results indicated that CTCF proteins did not form stable dimers by themselves, not even in combination with isolated DNA strands. Instead, nucleosomes were necessary to mediate this interaction, with the presence and orientation of the CTCF binding sequence greatly enhancing their binding affinity. To better understand how nucleosome-CTCF assemblies come to be, the team used cryo-electron microscopy (cryo-EM) to scrutinise several millions of particles and create a structural model at the atomic scale. Looking at the model, we see that dimerisation occurred not only between CTCF proteins – but also between the nucleosomes themselves.

Further analysis of a dinucleosome construct revealed that CTCF is able to dramatically enhance the multimerisation of adjacent nucleosomes, and does so without a CTCF-histone interface. Instead, it appears that holding DNA strands together creates the proximity required to catalyse nucleosome stacking. How this stacking is expounded on by multiple sequential CTCF binding sites on chromosomal DNA will be a point of future study.

Stacking troubles

So now we have a model – but what happens when you perturb it in vivo? To test the physiological relevance of CTCF dimerisation, the team engineered mouse embryonic stem cell (mESC) lines that harboured point mutations at the interface of CTCF that allows its self-interaction. Unsurprisingly, the more mutations that were introduced, the fewer CTCF proteins that were bound to DNA and the more slowly that cells grew. Attempting to induce the differentiation of mutant mESCs into primordial germ cell-like cells (PGCLCs; biologists have great names for things, I know) also didn’t work, suggesting that CTCF dimers are needed for development.

One factor influenced by faulty CTCF proteins is the DNA loops aforementioned. As the authors found out by measuring DNA-DNA contacts, cohesin-mediated loop extrusion was largely impaired without CTCF dimerisation. The same high-resolution genomic contact maps were used to infer patterns of nucleosome positioning around CTCF sites. Notably, while nucleosomes retained an ordered arrangement even in mutant CTCF lines, their signature reads were reduced and corresponded to a model in which nucleosome stacking is mediated by CTCF dimerisation in vivo.

Valeriano et al have made strides into a field that has been murky for decades, yet work remains. In particular, even higher-resolution 3D genomics techniques will be required to elucidate how chromatin is remodelled to create the space needed to accommodate not only cohesin, but the CTCF proteins brought together to stack nucleosomes and anchor loops in place. Furthermore, how these anchors are removed during cell division in a timely fashion is still not precisely understood. For now, we keep researching.

---