Forged by DXZ4, FIRRE, and ICCE: How Tandem Repeats Shape the Active and Inactive X Chromosome

Abstract

Recent efforts in mapping spatial genome organization have revealed three evocative and conserved structural features of the inactive X in female mammals. First, the chromosomal conformation of the inactive X reveals a loss of topologically associated domains (TADs) present on the active X. Second, the macrosatellite DXZ4 emerges as a singular boundary that suppresses physical interactions between two large TADdepleted "megadomains." Third, DXZ4 reaches across several megabases to form "superloops" with two other X-linked tandem repeats, FIRRE and ICCE, which also loop to each other. Although all three structural features are conserved across rodents and primates, deletion of mouse and human orthologs of DXZ4 and FIRRE from the inactive X have revealed limited impact on X chromosome inactivation (XCI) and escape in vitro. In contrast, loss of Xist or SMCHD1 have been shown to impair TAD erasure and gene silencing on the inactive X. In this perspective, we summarize these results in the context of new research describing disruption of X-linked tandem repeats in vivo, and discuss their possible molecular roles through the lens of evolutionary conservation and clinical genetics. As a null hypothesis, we consider whether the conservation of some structural features on the inactive X may reflect selection for X-linked tandem repeats on account of necessary cis-and trans-regulatory roles they may play on the active X, rather than the inactive X. Additional hypotheses invoking a role for X-linked tandem repeats on X reactivation, for example in the germline or totipotency, remain to be assessed in multiple developmental models spanning mammalian evolution.

Chromosome Compartments on the Inactive X Guide TAD Formation Independently of Transcription during X-Reactivation

“…Previous studies have shown that the active and inactive X chromosome have strikingly different 3D conformations (Bansal et al, 2019;Deng et al, 2015;Giorgetti et al, 2016;Minajigi et al, 2015;Nora et al, 2012;Rao et al, 2014;Splinter et al, 2011). In particular, while the active mouse X chromosome has an autosome-like structure, exhibiting active A and inactive B compartments and topologically associating domains (TADs), the inactive X chromosome is thought to lack compartmentalization and to solely exhibit TADs of attenuated strength .…”

Contiguous Erosion of the Inactive X in Human Pluripotency Concludes With Global DNA Hypomethylation

“…2C,G). This evolutionarily conserved tandem repeat forms long-range CTCF-mediated chromatin loops with two other tandem repeats DXZ4 and ICCE on the Xi, and transcribes a long non-coding RNA (ncRNA) primarily from the Xa 52,53 . The three FIRRE DMPs gaining DNAme reside inside the tandem repeat, characterized by closing (“local repeat”) ATAC peaks, while probes near its major promoter decrease in DNAme (Fig.…”

“…2,3). FIRRE has been shown to form conserved Xi-specific long-range interactions with the third female-specific YY1 cluster, the macrosatellite DXZ4 which functions as a topological boundary on the Xi (reviewed in 53 ). Both of these non-coding tandem repeat loci maintain focal euchromatin on the otherwise escapeepoor long arm, and reside at the center ( FIRRE ) or periphery ( DXZ4 ) of the earliest reactivation clusters (Fig.…”

Effects of forced cohesin eviction and retention on X-inactivation and autosomes

“…The superloops are not present on the Xa. Current models postulate that superloops form through loop extrusion of cohesin (Bansal et al, 2019;Bonora et al, 2018), but currently unknown is whether cohesin depletion is required to form the Xi superstructure and deposit the H3K27me3 Polycomb mark. Does cohesin binding antagonize the establishment of the Xi?…”

“…The Firre locus also normally retains its cohesins on the Xi (Fig.3E, RAD21 track) (Bonora et al, 2018;Froberg et al, 2018;Giorgetti et al, 2016). It may thereby form a superloop with Dxz4 through the retained cohesins, potentially through a cohesin-mediated loop extrusion mechanism across a 30Mb distance (Bansal et al, 2019;Bonora et al, 2018). Intriguingly, our allelic Hi-C map revealed a cohesin-dependent 'stripe' extending from Dxz4 (Fig.3D, inset).…”