Sophisticated Conversations between Chromatin and Chromatin Remodelers, and Dissonances in Cancer

Abstract

The establishment and maintenance of genome packaging into chromatin contribute to define specific cellular identity and function. Dynamic regulation of chromatin organization and nucleosome positioning are critical to all DNA transactions-in particular, the regulation of gene expression-and involve the cooperative action of sequence-specific DNA-binding factors, histone modifying enzymes, and remodelers. Remodelers are molecular machines that generate various chromatin landscapes, adjust nucleosome positioning, and alter DNA accessibility by using ATP binding and hydrolysis to perform DNA translocation, which is highly regulated through sophisticated structural and functional conversations with nucleosomes. In this review, I first present the functional and structural diversity of remodelers, while emphasizing the basic mechanism of DNA translocation, the common regulatory aspects, and the hand-in-hand progressive increase in complexity of the regulatory conversations between remodelers and nucleosomes that accompanies the increase in challenges of remodeling processes. Next, I examine how, through nucleosome positioning, remodelers guide the regulation of gene expression. Finally, I explore various aspects of how alterations/mutations in remodelers introduce dissonance into the conversations between remodelers and nucleosomes, modify chromatin organization, and contribute to oncogenesis.

Keywords: BAF; CHD; INO80; ISWI; SWI/SNF; cancer; chromatin remodeling; nucleosome positioning; promoter; transcription regulation.

Figure 1

Chromatin processes, nucleosome spacing and phasing, and remodelers’ contribution. (A) Contribution of remodelers to various chromatin processes. Using their ATPase subunit, remodelers can contribute to (i) chromatin opening (light green background) by ejecting nucleosomes, irregularly spacing nucleosomes through sliding, evicting dimers, or altering DNA wrapping; (ii) chromatin assembly (light pink background) by maturating deposition or regularly spacing nucleosome; and (iii) nucleosome editing (light purple background) by modifying octamer composition through installation or removal of histone variants (blue partial cylinder). (B) Contribution of remodelers to nucleosome spacing and phasing. (1) Within a genome population, the deposition of histone octamers results in random nucleosome positioning with a mix of blockage (red) and exposure (green) of specific cognate sites for DNA-binding proteins. (2) Upon remodeling by an assembly remodeler, without an obstacle, nucleosomes are regularly distributed along the DNA, resulting in regular spacing, but without phasing, as access to particular cognate sites remains heterogeneous across the population. (3) Alternatively, upon remodeling by an assembly remodeler, in the presence of an obstacle, e.g., a DNA sequence or a bound barrier factor, nucleosomes are regularly distributed along the DNA, resulting in regular spacing and phasing with homogeneous access to particular cognate sites across the population. (4) From a population of regularly spaced and phased nucleosome arrays, upon remodeling by an opening remodeler, nucleosomes can remain phased but lack regular spacing, leading to a homogeneity in the exposure of binding sites across the genome population.

Figure 2

ATPase-based classification of remodelers, model for DNA translocation, and mechanistic and regulatory contacts between a remodeler and a nucleosome. (A) ATPase-based classification of remodelers. All remodelers contain a SWI2/SNF2-related ATPase subunit characterized by an ATPase/translocase domain (ATD), which is split into two RecA-like lobes, lobe 1 (yellow) and lobe 2 (orange), and sufficient to perform DNA translocation. Remodelers can be separated into four subfamilies based on the conserved domains flanking the ATPase domain and the length of the insertions between the lobes (gray). The INO80 subfamily is the only subfamily bearing a long insertion. Types and position of domains further define each subfamily. In SWI/SNF, an HSA helix, a post-HSA, an α2 helix, a SuppH-containing Protrusion 1 helix (P1), and a Brace that together form a structural hub bridging the lobes; a SnAC domain along with a basic-patch (BP); AT-hooks; and a bromodomain. In ISWI, an AutoN region, NegC region, APB domain, and HSS (HAND–SANT–SLIDE) module. In CHD, possible PHD fingers, tandem chromodomains, a region similar to NegC, and a DBD (DNA-binding domain). In INO80, an HSA helix and a post-HSA. Possible remodeling outcomes from each subfamily are color-coded in the background of each ATPase as in Figure 1A. (B) Model for DNA translocation. The lobes of the ATPase domain (depicted as in (A)) undergo an ATP binding- and hydrolysis-dependent conformational cycle that correlates with alternating high affinity for DNA, driving DNA translocation. Lobes are colored (yellow for lobe 1; orange for lobe 2) when they have a high affinity for DNA and depicted in gray when they have a low affinity for DNA. Only the tracking strand of DNA, along which the lobes move in a 3′–5′ direction, is represented. The movements of the lobes are visualized by colored arrows, and the DNA translocation is depicted with back arrows. The precise step in which the inorganic phosphate (Pi) is released is unknown. Model inspired from Reference [13] and modified to incorporate results and observations from References [14,15], leading to an updated model in which lobe 1 is the stationary lobe. (C) Mechanistic and regulatory contacts between remodelers and nucleosomes. Beyond the lobes (here depicted at SHL2 from the dyad), depending on the subfamily, multiple mechanistic and regulatory contacts can be established between a remodeler (brown shape) and a nucleosome (cylinder with wrapped DNA in black; dyad axis depicted). The ATPase itself or additional subunits interact with (1) the second DNA gyre; (2) the extranucleosomal DNA on the entry or exit side; (3) the proximal and/or distal histone acidic patches (HAP); and (4) the histone tails, for example, H4 tail depicted here (red). Direction of potential translocation is indicated by blue arrows.

Figure 3

Cryo-EM structures of nucleosome-bound remodelers. Remodelers from the (A) ISWI subfamily are yeast Isw1 (6JYL), and human SNF2H (6NE3); (B) CHD subfamily are yeast Chd1 (5O9G) and human CHD4 (6RYR); (C) SWI/SNF subfamily are yeast Snf2 (5X0Y), yeast RSC (6KW3), and human BAF (6LTJ); and (D) INO80 subfamily are yeast Swr1 (6GEJ) and C. thermophilum Ino80 (6FML) bound to a nucleosome are depicted in the same orientation resulting from a structural alignment of their respective histone octamers before being separated into panels. Corresponding features are depicted with the same color codes. The ATPase subunits are depicted in spheres, with lobe 1 in yellow, lobe 2 in orange (as in Figure 2), and the rest of the subunit in green. The other subunits, if any, are depicted as colored cartoons. The dyad base-pair of each nucleosome is depicted in red. A few remarkable features have been named and indicated by arrows. All the remodelers from the various subfamilies engage the nucleosome in a similar or an inverted manner.

Figure 4

Hand-in-hand progressive increase in complexity of the regulatory conversations between remodelers and nucleosomes with the increase in challenges of remodeling processes. On the way (arrows) to successfully achieve their respective remodeling outcomes (top), remodelers from each subfamily need to overcome challenges (left orange gradient) associated with the processes they carry out (right blue gradient) by using shared and specific subunits and domains (black text along the arrows) corresponding to distinct nucleosomal features (blue text along the arrows). There is a positive correlation: the more complex the process is, the more challenges need to be overcome, and the higher is the need for complex regulation and multiple contacts with the nucleosome, and thus the more structurally and compositionally complex the remodelers are. D–H interaction stands for DNA–histone interaction.

Figure 5

Promoter architectures and nucleosome occupancy impact how remodelers are recruited. For adequate regulation, most genes in yeast blend features from two types of promoter architectures. (A) Open promoters, in their repressed state, present an NDR without (left) or with (right) a fragile nucleosome (FN) adjacent to the TSS (black arrow), containing poly-A and G/C-rich motifs, along with activator/GRF-binding sites, and flanked by well-positioned H2A.Z-containing nucleosomes, leading to regular spacing and phasing in a genome population. Open promoters are common at constitutively active genes, and their activation necessitates minor (if any) remodeling (+1 nucleosome sliding or FN ejection) through direct remodeler recruitment by an activator or a GRF. (B) Closed promoters, in their repressed state, contain a continuum of canonical nucleosomes covering their TSS, leading to a genome population lacking phasing. Closed promoters are common at highly regulated genes, and their activation necessitates the binding of a pioneer transcription factor (PTF) in the nucleosomal context, followed by the recruitment of a remodeler, which opens the promoter by ejecting nucleosomes, rendering the activator-binding site accessible.