Search

NODAL/Activin signaling orchestrates key processes during embryonic development via SMAD2. How SMAD2 activates programs of gene expression that are modulated over time however, is not known. Here we delineate the sequence of events that occur from SMAD2 binding to transcriptional activation, and the mechanisms underlying them. NODAL/Activin signaling induces dramatic chromatin landscape changes, and a dynamic transcriptional network regulated by SMAD2, acting via multiple mechanisms. Crucially we have discovered two modes of SMAD2 binding. SMAD2 can bind pre-acetylated nucleosome-depleted sites. However, it also binds to unacetylated, closed chromatin, independently of pioneer factors, where it induces nucleosome displacement and histone acetylation. For a subset of genes, this requires SMARCA4. We find that long term modulation of the transcriptional responses requires continued NODAL/Activin signaling. Thus SMAD2 binding does not linearly equate with transcriptional kinetics, and our data suggest that SMAD2 recruits multiple co-factors during sustained signaling to shape the downstream transcriptional program. DOI: http://dx.doi.org/10.7554/eLife.22474.001, eLife digest To allow a complex animal to develop from a small bundle of cells, the cells need to be able to communicate with each other to coordinate their activities. Furthermore, this communication needs to continue in adulthood to keep the body in balance and to prevent diseases such as cancer. The cells communicate by releasing signals that influence the behavior of their neighbors by activating proteins called transcription factors. These proteins then change the activity of particular genes in the nucleus by binding to specific places on a structure called chromatin (the structure in which the genes are packaged). One group of signaling molecules is known as the transforming growth factor beta superfamily, which is crucial for embryos to develop correctly. Failure to control these signals can also promote the growth of tumors. However, it is not clear how the detection of these signals at the surface of the cell leads to changes in the activity of genes inside the nucleus. Two transforming growth factor beta signals called Activin and NODAL cause a transcription factor known as SMAD2 to move into the nucleus where it can alter gene activity. Here Coda, Gaarenstroom et al. investigated how SMAD2 transmits the Activin/Nodal signal in mouse cancer cells. The experiments showed that SMAD2 can change the activities of genes in multiple ways. SMAD2 can bind to places in the chromatin that are either easy to access (which typically contain genes that are already “switched on”) as well as areas that are difficult to access (which generally contain genes that are “switched off”). As a result, SMAD2 increases the activity of genes that were already active, but also switches on on genes that were previously inactive. Coda, Gaarenstroom et al. also found evidence that SMAD2 remained bound to chromatin after long periods of Activin/NODAL signaling. For some genes, this resulted in high gene activity, but in other cases this decreased the gene’s activity. Therefore, future experiments will investigate which other proteins help SMAD2 to change gene activity at later times. DOI: http://dx.doi.org/10.7554/eLife.22474.002

Mitotic chromosomes were one of the first cell biological structures to be described, yet their molecular architecture remains poorly understood. We have devised a simple biophysical model of a 300 kb-long nucleosome chain, the size of a budding yeast chromosome, constrained by interactions between binding sites of the chromosomal condensin complex, a key component of interphase and mitotic chromosomes. Comparisons of computational and experimental (4C) interaction maps, and other biophysical features, allow us to predict a mode of condensin action. Stochastic condensin-mediated pairwise interactions along the nucleosome chain generate native-like chromosome features and recapitulate chromosome compaction and individualization during mitotic condensation. Higher order interactions between condensin binding sites explain the data less well. Our results suggest that basic assumptions about chromatin behavior go a long way to explain chromosome architecture and are able to generate a molecular model of what the inside of a chromosome is likely to look like. DOI: http://dx.doi.org/10.7554/eLife.05565.001, eLife digest The genetic material of living things is made up of long strands of DNA. Human cells contain about two meters of DNA split between 46 chromosomes. These chromosomes carry all the instructions to build a human body. To fit all of this information inside each human cell, the DNA is wrapped around hundreds of thousands of proteins such that the chromosomes each resemble a string of beads. Most of the time the chromosomes in a cell are only loosely arranged. But, when a cell prepares to divide into two new cells, its chromosomes become more compacted. This allows the DNA to withstand the physical forces involved when the copies of the chromosomes are pulled into the two daughter cells, and it makes it easier for the cell to handle its genetic material. If a chromosome breaks during cell division, it can result in diseases such as cancer. Several proteins—collectively called condensins—work to compact (or condense) the chromosomes. These proteins are found in a wide range of species, but it remains poorly understood how they cause chromosomes to become more compact. Due to the technical limitations of current imaging methods, it has not been possible to directly visualize the path of the DNA strand within a compacted chromosome. However, Cheng et al. have now overcome this limitation by combining experimental analyses and computational simulations. Cheng et al. used computer modeling to simulate a piece of chromosome that was about the same size as a chromosome from a single-celled microorganism called budding yeast. This model could accurately recreate the behavior of chromosomes as observed in non-dividing cells—and revealed that these chromosomes are in a relaxed state. Cheng et al. then modeled what happens when condensins are introduced. As expected, the chromosomes became more compacted and the model's behavior was then validated using further experiments. This predicted that condensin complexes, bound to regions along the chromosome's length, interact to form pairs that continually separate and form new pairs with other condensins; and that these ‘dynamic pairwise’ interactions compact the chromosome. The current model describes a relatively small chromosome and, in the future, extending the model to larger chromosomes could shed insight. DOI: http://dx.doi.org/10.7554/eLife.05565.002