Common fragile sites (CFSs) are particularly vulnerable regions of the genome that become visible as breaks, gaps, or constrictions on metaphase chromosomes when cells are under replicative stress. Impairment in DNA replication, late replication timing, enrichment of A/T nucleotides that tend to form secondary structures, the paucity of active or inducible replication origins, the generation of R-loops, and the collision between replication and transcription machineries on particularly long genes are some of the reported characteristics of CFSs that may contribute to their tissue-specific fragility. Here, we validated the induction of two CFSs previously found in the human fetal lung fibroblast line, Medical Research Council cell strain 5 (MRC-5), in another cell line derived from the same fetal tissue, Institute for Medical Research-90 cells (IMR-90). After induction of CFSs through aphidicolin, we confirmed the expression of the CFS 1p31.1 on chromosome 1 and CFS 3q13.3 on chromosome 3 in both fetal lines. Interestingly, these sites were found to not be fragile in lymphocytes, suggesting a role for epigenetic or transcriptional programs for this tissue specificity. Both these sites contained late-replicating genes NEGR1 (neuronal growth regulator 1) at 1p31.1 and LSAMP (limbic system-associated membrane protein) at 3q13.3, which are much longer, 0.880 and 1.4 Mb, respectively, than the average gene length. Given the established connection between long genes and CFS, we compiled information from the literature on all previously identified CFSs expressed in fibroblasts and lymphocytes in response to aphidicolin, including the size of the genes contained in each fragile region. Our comprehensive analysis confirmed that the genes found within CFSs are longer than the average human gene; interestingly, the two longest genes in the human genome are found within CFSs: Contactin Associated Protein 2 gene (CNTNAP2) in a lymphocytes’ CFS, and Duchenne muscular dystrophy gene (DMD) in a CFS expressed in both lymphocytes and fibroblasts. This indicates that the presence of very long genes is a unifying feature of all CFSs. We also obtained replication profiles of the 1p31.1 and 3q13.3 sites under both perturbed and unperturbed conditions using a combination of fluorescent in situ hybridization (FISH) and immunofluorescence against bromodeoxyuridine (BrdU) on interphase nuclei. Our analysis of the replication dynamics of these CFSs showed that, compared to lymphocytes where these regions are non-fragile, fibroblasts display incomplete replication of the fragile alleles, even in the absence of exogenous replication stress. Our data point to the existence of intrinsic features, in addition to the presence of long genes, which affect DNA replication of the CFSs in fibroblasts, thus promoting chromosomal instability in a tissue-specific manner.

To complete mitosis, the bridge that links the two daughter cells needs to be cleaved. This step is carried out by the endosomal sorting complex required for transport (ESCRT) machinery. AKTIP, a protein discovered to be associated with telomeres and the nuclear membrane in interphase cells, shares sequence similarities with the ESCRT I component TSG101. Here we present evidence that during mitosis AKTIP is part of the ESCRT machinery at the midbody. AKTIP interacts with the ESCRT I subunit VPS28 and forms a circular supra-structure at the midbody, in close proximity with TSG101 and VPS28 and adjacent to the members of the ESCRT III module CHMP2A, CHMP4B and IST1. Mechanistically, the recruitment of AKTIP is dependent on MKLP1 and independent of CEP55. AKTIP and TSG101 are needed together for the recruitment of the ESCRT III subunit CHMP4B and in parallel for the recruitment of IST1. Alone, the reduction of AKTIP impinges on IST1 and causes multinucleation. Our data altogether reveal that AKTIP is a component of the ESCRT I module and functions in the recruitment of ESCRT III components required for abscission.

Membrane-enclosed organelle compartmentalization is not the only way by which cell processes are spatially organized. Phase separation is emerging as a new driver in the organization of membrane-less compartments and biological processes. Liquid–liquid phase separation has been indicated as a new way to control the kinetics of molecular reactions and is based on weak multivalent interactions affecting the stoichiometry of the molecules involved. In the nucleus, liquid–liquid phase separation may represent an ancestral means of controlling genomic activity by forming discrete chromatin regions, regulating transcriptional activity, contributing to the assembly of DNA damage response foci, and controlling the organization of chromosomes. Liquid–liquid phase separation also contributes to chromatin function through its role in the reorganization of the nuclear periphery in the post-mitotic phase. Herein, we describe the basic principles regulating liquid–liquid phase separation, analyze examples of phase separation occurring in the nucleus, and dedicate attention to the implication of liquid–liquid phase separation in the reorganization of the nuclear periphery by the endosomal sorting complexes required for transport (ESCRT) machinery. Although some caution is warranted, current scientific knowledge allows for the hypothesis that many factors and processes in the cell are yet to be discovered which are functionally associated with phase separation.

During open mitosis, reassembly of the nuclear envelope requires the coordinated recruitment of the ESCRT machinery, initiated by the chromatin-associated factor BAF1 and the nuclear-envelope-associated factor LEM2. Because telomeres are enriched at the reforming envelope, we investigated whether ESCRT factors contribute to telomere integrity. Reduction in the pivotal nuclear ESCRT factor CHMP7 caused DNA damage, heterochromatin disorganization, and telomere defects, including sister telomere associations and telomere free ends. Extending this analysis, we found that additional ESCRT components, including TSG101, VPS28, CHMP4B, and the ESCRT-associated factor AKTIP/Ft1, also contribute to telomere integrity, although with different strengths. Genetic interaction analyses suggest that CHMP7 converges in a common pathway with CHMP4B and AKTIP/Ft1, while it functions in parallel routes to TNKS1, a telomere-specific regulator of the shelterin TRF1. More genetic analyses indicated that BAF1 and LEM2 contribute to safeguarding of telomeres during nuclear envelope reassembly. Because defects in nuclear envelope dynamics and chromatin–membrane coupling are hallmarks of disorders associated with nuclear deformation and fragility, including aging and cancer, our findings contribute a new angle into these conditions and suggest potential targets for selectively modulating telomere maintenance pathways.