TADs or no TADS: Lessons From Single-cell Imaging of Chromosome Architecture

Highlights

• Eukaryotic genomes are organized in a hierarchical multiscale fashion.

• We focus on the contribution of novel imaging technologies to chromatin organization.

• There is a high degree of stochasticity in genome folding in single cells.

• TADs are functional units but not necessarily stable physical entities.

• The stochastic nature of folding is likely essential to regulate chromatin transactions.

Abstract

Eukaryotic genomes are folded in a hierarchical organization that reflects and possibly regulates their function. Genomewide studies revealed a new level of organization at the kilobase-to-megabase scale termed “topological associating domains” (TADs). TADs are characterized as stable units of chromosome organization that restrict the action of regulatory sequences within one “functional unit.” Consequently, TADs are expected to appear as physical entities in most cells. Very recent single-cell studies have shown a notable variability in genome architecture at this scale, raising concerns about this model. Furthermore, the direct and simultaneous observation of genome architecture and transcriptional output showed the lack of stable interactions between regulatory sequences in transcribing cells. These findings are consistent with a large body of evidence suggesting that genome organization is highly heterogeneous at different scales. In this review, we discuss the main strategies employed to image chromatin organization, present the latest state-of-the-art developments, and propose an interpretation reconciling population-based findings with direct single-cell chromatin organization observations. All in all, we propose that TADs are made of multiple, low-frequency, low-affinity interactions that increase the probability, but are not deterministic, of regulatory interactions.

Introduction

Storing large amounts of information in small spaces while being able to readily access it rapidly and efficiently remains a current technological challenge of man-made storage devices. Eukaryotic cells face a similar challenge as they need to pack around 2 m of linear DNA into a micrometer-sized nucleus. Yet, despite this remarkable degree of compaction, the information contained in the DNA sequence is replicated and translated in a very short amount of time and with great accuracy to ensure inheritance to daughter cells and continuous normal cellular functioning.

At the nucleus, the linear genome is organized in a hierarchical multiscale fashion ranging from a few kilobases to megabases that is, in most cases, reflected by the three-dimensional (3D) folding of chromatin (Fig. 1). Additionally, the minimal structure of chromatin, the nucleosome, represented by DNA wrapped around specific proteins (histones), can carry different covalent chemical modifications that will determine how DNA will interact with transcriptional and reparation machineries and thus chromatin state (for an extended review, see Ref. [1]). Compact and repressed chromatin (heterochromatin) domains are usually found at the nucleus periphery and segregated from active and large (eu-)chromatin domains [2] where chromatin is thought to be actively transcribed [3]. On a larger scale, it is now well accepted that individual chromosomes are organized into territories (CT) [4,5] and rarely intermingle with each other. However, the internal organization of chromosomes remained largely unknown until recently. Over the past decade, several breakthroughs in high-throughput sequencing enabled the mapping of pairwise interactions between genetic loci with genomic specificity and genomewide coverage [[6], [7], [8], [9]] unveiling novel hierarchical levels of eukaryotic chromatin organization. Topologically associated domains (TADs) have been defined as regions displaying enriched interactions with neighboring DNA and appear in Hi-C matrices as squares on the main diagonal [[10], [11], [12]]. Thus, the definition of TADs describes a novel feature of Hi-C maps and does not involve any specific factor (e.g., cohesin) or mechanism (e.g., loop extrusion) [13]. Finally, active/repressed TADs often associate with each other in the nucleus to form active (A) or inactive (B) compartments (Fig. 1) [6,[10], [11], [12],14,15].

Hierarchical organization critically impacts nuclear activities such as transcription, replication as well as cellular events such as cell-cycle regulation, key cell fate decisions, and embryonic development, yet the mechanisms and the exact role of each structural element, particularly at the TAD level, are still a matter of debate. Particularly, from genomewide methods, two contrasting hypotheses regarding the role of TADs and their structural interpretation have arisen (see next section). Fluorescent microscopy is in a privileged place to reveal the nuclear organization in single cells and settle this apparent contradiction. The aim of this review is to present an overview of microscopy findings over the last decade that lay the bases of present discoveries and the current state-of-the-art on high-throughput–high-content imaging technologies. We will put into context and discuss the results from these findings to present a general picture of genome organization and a new interpretation that reconciles both models.