PROJECT SUMMARY
Radial chromosomes (hereafter “radials”) have been part of the scientific consciousness for more than 50
years. Radials are operationally defined as fusions involving two or more nonhomologous chromosomes and
as having more than one centromere. In normal cells, radials are quite rare, and they almost never form
spontaneously. Radials have an important clinical relevance, however, since their appearance is a diagnostic
hallmark of the bone marrow failure, cancer predisposition and premature aging syndrome, Fanconi anemia
(FA). Interestingly though, even in FA cells, radial chromosomes do not appear spontaneously with high
frequency but usually need to be induced by exposure to clinically-relevant DNA crosslinking reagents, such as
mitomycin C (MMC) and cisplatin. Despite their diagnostic use for many decades, radial formation is still poorly
understood, and the literature is filled with conflicting reports about what is required mechanistically for their
production. In our first Specific Aim, we describe the construction of a human cell line that is defective for the
radiation sensitive 18 (RAD18) gene. RAD18 encodes an E3 ubiquitin ligase and is required for translesion
synthesis (TLS), an error-prone DNA damage tolerance pathway, which is a subpathway in the FA-mediated
DNA crosslink repair process. Consistent with this role, RAD18-null human cells generated high levels of MMC-
induced radials. Unexpectedly, the absence of DNA ligase III (LIGIII) and DNA polymerase theta (POLQ; both
key components of alternative non-homologous end joining; A-EJ)) dramatically reduced RAD18-dependent,
MMC-induced radial formation. Thus, the radials generated by the absence of RAD18 appear to require the A-
EJ pathway. We propose to confirm these observations and extend the generality of these studies by
determining the requirement for A-EJ in MMC-induced radial formation in Bloom syndrome (BLM) (another
cancer predisposition and premature aging syndrome) and FA-defective cells.
Amazingly, there is not a single report in the literature regarding the molecular makeup of a radial
chromosome fusion junction. This is unfortunate, since an analysis of fusion junctions for a wide variety of other
types of translocations (e.g., canonical reciprocal chromosomal translocations and telomere fusions) has provide
a wealth of data regarding the molecular signatures of such junctions as well as insight into the repair
mechanisms responsible for their formation. In our Specific Aim 2, we propose to rectify this egregious omission.
Specifically, we describe a protocol (Radial-Seq) for isolating, sequencing and bioinformatically analyzing radial
chromosome fusions. We expect this analysis to confirm the formation of such junctions by A-EJ and to greatly
extend our understanding of the molecular makeup of a radial fusion. In toto, these experiments should
mechanistically define the requirements for radial formation which should benefit basic researchers; in addition,
the assignment of A-EJ for their formation may help clinicians therapeutically hinder their formation.