State Letter

Summary by Sara Mingu

Sheung Chung Ng & Dirk Görlich

TL;DR: Nup98 is an important component of the nuclear pore complex permeability barrier. In vitro studies using a simplified Nup98 version reveal LCST-type phase separation. They also provide estimations for the enthalpy, entropy and free energy contribution per repeat to the phase separation and inter-repeat cohesion.

The nuclear pore complex permeability barrier is a condensed phase assembled from cohesive FG-repeats of intrinsically disordered, FG-rich nucleoporins such as Nup98. In vitro, both Nup98 itself and an engineered, simplified version of it (prf.GLFG(52x12)) were found to phase separate, forming droplets with similar biophysical properties which recapitulate the permeability barrier functions. Models describing NPC permeability barrier functionality have pointed to the importance of the cohesive interactions between the hydrophobic FG domains, especially in Nup98, in the formation and function of the NPC permeability barrier.

To determine the strength of FG interactions, the authors first study the partition of various prf.GLFG variants with different repeat numbers into various host FG phases. The results showed that each additional FG repeat contributed a specific free energy increment to the partitioning. The cohesion increased with increasing salt concentration, supporting the notion that inter-FG cohesion is driven by hydrophobic interactions, and that the cohesive potential is evolutionarily conserved. To gain more information, the authors established a phase diagram for prf.GLFG(52x12) phase separation behavior, as well as measuring the phase transition temperatures for different conditions. They also used an alternative approach and measured the saturation concentrations after centrifugation (condense phase was pelleted). In all cases, and also for various other Nup98 variants, they consistently observed lower critical saturation temperature (LCST)-type behavior, where the saturation concentration decreased with increasing temperatures. Comparison of phase separation of prf.GLFG variants of different repeat numbers showed that saturation concentrations increase exponentially with FG repeat number. Finally, the authors explored the effect of inter-FG cohesion with permeability barrier properties, such as transport selectivity. They found that FG cohesion could particularly impact the partition (and transport) of FG-repellent cargo bound to a transport receptor.

A single FG interaction is very weak, and can be already dissociated by thermal motion. This implies that there is high local mobility and that a large fraction of FG motifs are 'free' at any given time and available for interactions with cargo. On the other hand, due to the large number of FG motifs present in the NPC, s stable barrier can be formed against FG-phobic species. The LCST phase separation behavior of prf.GLFG studied in this work resembles micelle formation from non-ionic detergent monomers. They are both driven by the hydrophobic effect and favored by increasing salt concentration. It is likely that FG phases, like some micelles, are hydrophobic in nature. For the NPC, this could mean a dense, hydrophobic 'core' composed of the highly cohesive Nup98 at the center, and more hydrophilic 'cap' composed of less cohesive nucleoporins (Nup153, Nup159, etc) at the cytoplasmic and nuclear sides.

Link to paper

Dirk Görlich is a Professor and Director at the Max Planck Institute for Multidisciplinary Sciences

Summary by Simone Alessio del Grosso

Anisha Shakya, Seonyoung Park, Neha Rana, John T King

TL; DR: In vivo, H1 condenses into liquid like droplets and co-localizes with heterochromatin. In vitro, H1 is able to condense large segments of DNA and nucleosomes over a broad range of salt concentration. Of the core histones, only H2A shows phase separating behavior, while the others precipitate. ATP, among other nucleotides, was shown to aid the formation of spherical, liquid like polynucleosomal condensates.

The link between liquid-liquid phase separation (LLPS) and silenced heterochromatin is still widely unexplored. Therefore Shakya et al. systematically investigated linker histone H1’s as well as the histone core proteins’ phase separating behavior with DNA and nucleosomes. Through two color confocal microscopy, they show that eGFP-tagged H1 condenses into liquid like puncta in nuclei of HeLa cells during interphase and co-localizes with the chromatin marker HP1ɑ. In vitro, droplet formation is readily seen with a 150-bp long DNA. Additionally, using fluorescence correlation spectroscopy (FCS) and a probe, they observe smaller diffusion times the longer the DNA is, rationalizing this with an increase in mesh size. Among the core histone proteins, only H2A showed phase separation with the same 150 bp long DNA, while H2B, H3 and H4 precipitated. Similarity of the histone core proteins suggests additional drivers for LLPS like charge and sequence patterning might be important. Also, the authors point out H2A’s C terminal tail, known to make contacts with nucleosomal and linker DNA, bind to H1 and to be subject to PTMs. By means of FRET-labelled mononucleosomes (5’ Cy3-DNA, Cy5-H2A) they observe droplet formation after both addition of H1 and H2A. Comparable ratios of Cy5/Cy3 emission intensity between nucleosomes in buffer and nucleosomes in the H1/H2A droplets, lead them to conclude that the structural integrity of the nucleosome is maintained.

Phase diagrams were measured for condensates consisting of H1, H2A and HP1a and differently sized DNA and nucleosomes. H1 forms droplets across a wide range of salt concentration (150-500 mM NaCl), with droplets being most pronounced at 150 mM. Similar trends are observed for both H2A and HP1ɑ. Experiments conducted with polynucleosomes show a similar broad salt concentration range for LLPS of H1 (150-400 mM). At 150 mM NaCl, spherical droplets are only observed with mononucleosomes, irrespective of the protein tested. Polynucleosomes all form irregularly shaped condensates. Above 300 mM NaCl, all samples exhibit spherical droplets. Fluorescence after photobleaching revealed that the labelled proteins were mobile in spherical condensates and arrested in irregular ones.

As a result, the observation that liquid like condensates are able to form in HeLa cells seemed to stand in contrast to the fact that only mononucleosomes form spherical liquid like droplets. To check their hypothesis that additional factors aid condensation of large polynucleosomes in vivo, experiments with heptanucleosomes and H1 with addition of ATP (or other nucleotides) were performed. In fact, this led to observation of faster recovery after photobleaching and spherical shapes. The authors conclude by putting their findings into the larger context of LLPS and heterochromatin formation and argue that partitioning of small molecules and proteins into chromatin droplet is crucial for LLPS in vivo.

Link to paper

John King leads a group at the Institute for Basic Science, Center for Soft and Living Matter (S. Korea), which focuses on understanding how molecular and macromolecular structure and dynamics manifest in the bulk-level behavior of soft materials.

Summary by Dr. Monica Rieth

Vahid Sheikhhassani | Barbara Scalvini | Julian Ng | Laurens W. H. J. Heling | Yosri Ayache | Tom M. J. Evers | Eva Estébanez-Perpiña | Iain J. McEwan | Alireza Mashaghi

TL;DR: A novel approach to study molecular dynamics and topological conformations of the long intrinsically disordered N-terminal domain of the human androgen receptor is described. Results from the work suggest a regulatory DNA binding site is situated at the C terminal end of the long-disordered N-terminal domain that attenuates binding to a separate DNA binding site on the protein.

The human androgen receptor is an important nuclear transcription factor that interacts with hormones such as testosterone. It belongs to a larger family of steroid hormone nuclear receptors and binds to DNA in a ligand-dependent manner, regulating transcription. The receptor is expressed in numerous cell types and tissues. The N-terminal transactivation domain (NTD; residues 1-538) of the receptor is variable while the DNA-binding domain (DBD) is highly conserved and structured. The NTD is very large and dynamic making it a challenging target for structure/conformational elucidation. In this paper, the authors investigate the topological dynamics and structure of the intrinsically-disordered N-terminal domain (NTD), which includes the structurally distinct N-terminal region (NR, residues 1-224) and C-terminal region (CR, residues 225-538). A long-term goal is to use the structural information extracted from these studies to develop small molecule inhibitors that target the NTD.

Predicting the topology of intrinsically disordered proteins is difficult because of their flexibility and sampling of a large conformational space. The authors use a combination of molecular dynamics (MD) simulations and circuit topology (CT) analysis to probe the dynamics of the NTD from the human androgen receptor. From this analysis, it was revealed that flexible-disordered regions of both the NR and CR compete for binding to the DBD on the receptor, suggesting a possible mechanism of regulation. The analyses uncovered a small molecule inhibitor binding site within the CR that serves to regulate this activity.

Authors began by using I-TASSER to predict a 3D structure of the long N-terminal domain, AR-NTD (1-538). The structure was refined using the MD simulation program, GROMACS. A series of additional computational and bioinformatics programs were used to verify the accuracy of the predicted model structure. Next, dynamics were performed using GROMACS beginning with the minimized energy-predicted structure, refined from the I-TASSER pipeline.

Molecular docking experiments (ClusPro) were employed to study binding interactions of the NTD with the AR ligand binding domain (LBD) and DBD. Binding of the inhibitor, EPI-001, to the CR was also investigated to understand the mechanism of inhibition. CT analysis was performed to map the residue contact points of the NR and CR to better understand the evolution of any transiently folded sub-structures. CT analysis shows the NR is truly dynamic and conformational “folding” is random, while the CR less so. The latter is capable of assuming multiple sub-structures indicating less dynamic flexibility.

Link to paper

Alireza Mashaghi, Medical Systems Biophysics and Bioengineering, Leiden Academic Centre for Drug Research, Faculty of Science, Leiden University, Leiden, The Netherlands