Transcription-driven genome organization: a model for chromosome structure and the regulation of gene expression tested through simulations

TL;DR

This paper proposes a physical model where genome organization and gene regulation are driven by clustering of transcription machinery, explaining the formation of contact domains and the influence of 3D proximity on gene activity, supported by simulations.

Contribution

It introduces a comprehensive physical model linking genome folding to transcription regulation, emphasizing the role of transcription factories and loops in gene activity.

Findings

Clusters of active RNA polymerases are key architectural features.

Contact domains and compartments reflect loops and clusters.

Simulations support the influence of 3D organization on gene expression.

Abstract

Current models for the folding of the human genome see a hierarchy stretching down from chromosome territories, through A/B compartments and TADs (topologically-associating domains), to contact domains stabilized by cohesin and CTCF. However, molecular mechanisms underlying this folding, and the way folding affects transcriptional activity, remain obscure. Here we review physical principles driving proteins bound to long polymers into clusters surrounded by loops, and present a parsimonious yet comprehensive model for the way the organization determines function. We argue that clusters of active RNA polymerases and their transcription factors are major architectural features; then, contact domains, TADs, and compartments just reflect one or more loops and clusters. We suggest tethering a gene close to a cluster containing appropriate factors -- a transcription factory -- increases the firing frequency, and offer solutions to many current puzzles concerning the actions of enhancers, super-enhancers, boundaries, and eQTLs (expression quantitative trait loci). As a result, the activity of any gene is directly influenced by the activity of other transcription units around it in 3D space, and this is supported by Brownian-dynamics simulations of transcription factors binding to cognate sites on long polymers.