A major part of eukaryotic nuclear DNA is organized into nucleosomes that are not distributed randomly, but occupy defined positions with respect to the DNA sequence. Such nucleosome positioning is fundamental to all genomic processes as it regulates the accessibility of DNA. For example, a transcription factor binding site may be incorporated in a nucleosome and therefore be less accessible, or it may reside in a nucleosome free region and be readily bound by its corresponding factor. In our first line of studies, we wish to understand what determines nucleosome positioning (for review see Lieleg et al., 2015, Chromosoma). In our second line, we study how positioned nucleosomes in one biological state become remodeled into an altered organization of a different state. This happens, for example, during promoter chromatin opening upon gene induction (for review see Korber and Barbaric, 2014, Nucleic Acids Res.). The recurring theme of both our lines of research is that ATP-dependent nucleosome remodeling enzymes play a key role. As model systems we use the unicellular yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe and combine their powerful genetic tools with biochemical techniques. Especially regarding the biochemical approach, we established the first genome-wide in vitro reconstitution system that enables us in a unique way to pinpoint the roles of individual factors in nucleosome positioning.
Elisa Oberbeckmann and Iris Langstein: Molecular Mechanism of chromatin remodeling enzymes and barrier factors for nucleosome positioning
In vivo studies can be very valuable for finding factors involved in nucleosome positioning as respective mutants may show altered nucleosome positioning patterns in vivo. However, with such mutant studies it is often very difficult if not impossible to distinguish direct from indirect, and specific from unspecific effects and to recognize what is sufficient. Therefore we developed over more than ten years an in vitro biochemistry approach. The in vitro reconstitution of nucleosomes onto DNA by biophysical methods in the absence of other factors, like salt gradient dialysis, usually does not yield the proper nucleosome positions as observed in vivo. Therefore, proper nucleosome positioning requires factors beyond the nucleosomes themselves, so called trans-factors. Indeed, we established a chromatin assembly system that includes trans-factors and allows the generation of properly positioned nucleosomes in vitro: Our former graduate students Christina Hertel, Franziska Ertel and Christian Wippo showed that nucleosomes reconstituted onto yeast DNA by salt gradient dialysis are moved to their proper positions upon incubation with a yeast whole cell extract (Korber and Hörz, 2004, J. Biol. Chem.; Hertel et al., 2005, Mol. Cell. Biol.; Wippo et al., 2009, Mol. Cell. Biol.; Wippo et al., 2011, EMBO J.; Wippo and Korber, 2012, Meth. Mol. Biol.). This process is ATP-dependent and Christian Wippo, in collaboration with the group of B. Franklin Pugh at the Pennsylvania State University, extended this approach to the genome level (Zhang et al., 2011 Science; Krietenstein et al., 2012, Methods in Enzymology; Fig. 1).
Figure 1. Genome-wide in vitro reconstitution of in vivo-like nucleosome positioning for the budding yeast genome. An S. cerevisiae plasmid library was assembled by salt gradient dialysis (SGD). The resulting nucleosome occupancy was mapped by MNase-anti-H3-ChIP-seq and displayed for more than 4000 genes after alignment at their transcriptional start sites (TSS, start of shaded arrow). Genes were grouped (clusters 1 to 5) according to similar nucleosome occupancy patterns around the TSS. Ex vivo chromatin („Native“) is shown for comparison. The same SGD chromatin was treated with yeast whole cell extract (WCE) with or without ATP. The proper generation of in vivo-like nucleosome depleted regions (blue region just upstream of the TSS) and regular arrays (yellow stripes) essentially depends on factors from the WCE and ATP. Figure modified from Zhang et al., 2011, Science.
This offers us a unique tool to study nucleosome positioning mechanisms at genome-scale by a biochemical approach. Together with the Pugh group and with the group of Craig L. Peterson (University of Massachusetts Medical School) our former graduate student Nils Krietenstein undertook a candidate approach and used mutant extracts and purified factors in order to identify the nucleosome positioning determinants. Recently, we succeeded in finding a minimal system of purified nucleosome remodeling enzymes and so called “barrier” factors (also called “general regulatory factors”, GRFs) that could recapitulate the generation of basic nucleosome positioning patterns around yeast promoters (Krietenstein et al., 2016, Cell; Fig. 2). This demonstrates a direct, specific and to a surprisingly large extent also sufficient role of ATP-dependent nucleosome remodeling enzymes and barriers in setting up a dynamic, self-organizing nucleosome positioning pattern.
Elisa studies now how individual remodeling enzymes, for example the INO80 complex, turn DNA sequence into nucleosome positioning. Iris investigates how barrier factors work.
Figure 2. The combination of purified nucleosome remodeling enzymes (like ISW2, ISW1a, INO80, or RSC), general regulatory factors (GRFs, = “barriers”), and DNA sequence features (like poly(dA:dT) elements or shape features) is sufficient to reconstitute in vitro (symbolized by Eppendorf tube) the basic nucleosome organization at most yeast promoters. This organization consists of a nucleosome free region (NFR) just upstream of the gene and well positioned nucleosomes (+1, +2, +3, ...) over the coding region. Graphical abstract created by Nils Krietenstein and taken from Krietenstein et al., 2016, Cell.
Maria Walker: In vitro reconstitution of nucleosome positioning mechanisms in S. pombe
During recent years the fission yeast Schizosaccharomyces pombe has become a favorite model organism in the chromatin field as many features, like the organization of heterochromatin and centromeres and the presence of the RNAi system, are more similar to multicellular eukaryotes than those of the more classic budding yeast model Saccharomyces cerevisiae (= baker’s yeast). Our former graduate student Alexandra Lantermann pioneered the genome-wide mapping of nucleosome positioning in S. pombe using high resolution tiling arrays (Lantermann et al., 2009, Methods). A comparison of this map with those published by others for S. cerevisiae strongly suggested that nucleosome positioning mechanisms are not universally conserved, but evolutionarily plastic (Lantermann et al., 2010, Nat. Struct. Mol. Biol.). Julia Pointner took up Alexandra’s work on S. pombe nucleosome positioning and found that the CHD1-type nucleosome remodeling enzymes Hrp1 and Hrp3 are necessary for the alignment of regular arrays over genic regions (Pointner et al., 2012, EMBO J.; Fig. 3). Such impaired arrays lead to substantial cryptic antisense transcription. This demonstrated that a conserved function, the generation of regular nucleosomal arrays over genic regions and the prevention of cryptic transcription, is carried out by different types of nucleosome remodeling enzymes in different yeasts. S. pombe uses CHD1-type remodelers only while S. cerevisiae employs also ISWI-type remodelers (Gkikopoulos et al., 2011, Science). In light of Nils’s findings we now attribute the evolutionary plasticity of nucleosome positioning mechanisms to the species-specific use of nucleosome remodeling enzymes.
Maria Walker took over the S. pombe work in our group and works on establishing the analogous in vitro reconstitution approach for S. cerevisiae. Her position is funded by the Bavarian Research Network for Molecular Biosystems (BioSysNet; http://biosysneteng.jimdo.com/) and in this context we collaborate with the group of Ulrich Gerland at the Technical University of Munich.
Figure 3. TSS-aligned overlay of nucleosome occupancy for S. pombe wild type and hrp1 hrp3 mutant. Genome-wide nucleosome occupancy was determined by MNase-chip (figure taken from Lieleg et al., 2014, Chromosoma; data taken from Pointner et al., 2012, EMBO J.).
Our former PhD student Corinna Lieleg initiated a collaboration with the group of Hendrik Dietz at the Technical University of Munich. The Dietz group specializes in nano-structures based on complementary DNA strand hybridizations (“DNA origamis”) and explores their manifold applications. Corinna designed, prepared and validated mononucleosomes with single strand handles that could be incorporated into DNA origamis via complementary strand hybridization. In particular, the Dietz group established a device that can be used as a molecular force spectrometer for analyzing molecular interactions (Fig. 6). After incorporating nucleosomes into this device the energy landscapes of stacking and unwrapping nucleosomes could be measured (Funke et al., 2016, Science Advances, Funke et al., 2016, ACS Nano Letters). Maria continues this collaboration.
Figure 4. DNA origami force spectrometer allows direct measurement of nucleosome-nucleosome interactions (double headed arrow). Nucleosomes were attached to DNA origami via single strand hybridization handles. Red spiral symbolizes torsional spring of DNA origami device that counteracts nucleosome-nucleosome interactions. Red and green spheres show positions of dyes that allow monitoring of opening angles via FRET. Figure taken from Funke et al., 2016, Science Advances.
Our group originated in 2005 from the group of Wolfram Hörz, who pioneered the S. cerevisiae PHO5 and PHO8 promoters as classical model systems for the role of chromatin in gene regulation (for review see Korber and Barbaric, 2014, Nucleic Acids Res.). These promoters harbor positioned nucleosomes that control access to the transactivator Pho4 and to the transcriptional machinery. Upon promoter induction, these nucleosomes become remodeled, which leads to extensive DNaseI hypersensitive sites (Fig. 5). The Hörz group elucidated many aspects of this remodeling, and we continue the study of the mechanisms of PHO promoter remodeling by in vitro and in vivo techniques (Hertel et al., 2005, Mol. Cell. Biol.; Korber et al., 2006, J. Biol. Chem.; Barbaric et al., 2007, J. Biol. Chem.; Wippo et al., 2009, Mol. Cell. Biol.; Ertel et al., 2010, Mol. Cell. Biol.). In this context we also continue the close collaboration with the group of Slobodan Barbaric at the University of Zagreb, Croatia. For example, we resolved the long standing question if the RSC remodeling complex, the only remodeler essential for viability in yeast, has a role in PHO promoter opening (Musladin et al., 2014, Nucl. Acids Res.). We found a surprising complexity of the remodeler network at the PHO5 promoter. Five different ATP dependent remodelers (RSC, SWI/SNF, INO80, Isw1, Chd1) cooperate to achieve wild type promoter opening kinetics.
Figure 5. Schematics of the chromatin organization at the S. cerevisiae PHO5 and PHO8 promoters in their repressed and induced state. In the repressed state, both promoter regions are organized into positioned nucleosomes (open and closed large circles) with short hypersensitive linker regions (sHS, arrows). The promoters are induced upon phosphate depletion of the cells, which activates the transactivator Pho4 (small circles denote Pho4 binding sites). This leads to the complete eviction or partial remodeling (stippled circles) of the positioned nucleosomes and generates an extensive hypersensitive site (eHS, bold horizontal line), which is a hallmark of the induced promoters. Promoter chromatin remodeling involves a surprisingly complex cofactor network like the remodeler ATPases Snf2, Ino80 and others, the histone acetyltransferases Gcn5 and Rtt109, the histone chaperone Asf1, and probably even more.
Sebastian Sommer: In vivo reconstitution of heterochromatin at yeast PHO promoters
Heterochromatin is loosely defined as transcriptionally repressive chromatin. While the promoter chromatin structure at PHO promoters is already repressive under non-inducing conditions, Sebastian explores if more repressive, maybe even uninducible chromatin can be engineered here.