Cellular mechanisms of DNA replication initiation
Life depends on the faithful transmission of genetic information from parent to progeny. Cellular life stores genetic content in the form of DNA and cells contain machinery that copies and disseminates duplicated DNA to daughter cells during cell division. Although the molecules which ready DNA for replication are well known, how their action is coordinated in cellular space and time remain a mystery. Further, replication factors have diverged markedly in the metazoan tree of life bringing into question the precise conservation of events across evolution. We are using in vivo and in vitro methods to mechanistically dissect the early steps of DNA replication in a variety of metazoan model organisms. The information we glean from cellular studies directly guides our efforts to develop advanced reconstitution systems to further dissect the molecular underpinnings of this fundamental process.
ORC in fly embryos (transgene) and human cells (knock-in).
Specialization of initiator disordered domains in protein phase separation
The metazoan DNA replication licensing factors (ORC, Cdc6 and Cdt1) contain long intrinsically disordered regions (IDRs) that facilitate chromatin association in cells and drive DNA-dependent self-assembly and phase separation in vitro. Biomolecular condensates that form upon initiator-DNA binding selectively recruit replication factors but exclude non-partner phase separating proteins. Interestingly, licensing factor IDRs form a sequence class all their own and we have discovered a functionally diverse family of chromatin-associated factors with IDRs that have similar sequence features to licensing factors. We seek to understand how this set of sequences facilitates DNA-dependent phase separation, how their propensity to phase separate impacts the various chromatin-contextualized processes with which they're involved, and we aim to uncover the molecular basis for the inherent sorting capacity of replication condensates.
The molecular language of intrinsically disordered regions
A large fraction of protein sequences are intrinsically disordered (approximately 40% of the proteome). Unlike globular domains, it is unclear if protein intrinsically disordered regions (IDRs) can be comprehensively classified and we therefore cannot predict IDR function from sequence alone. Our lab develops bioinformatic algorithms to identify statistically non-random sequence features of IDRs and we have discovered a sequence-spanning level of organization in these enigmatic sequences. These computational tools are being used to comprehensively classify disordered sequence types. Beyond developing a predictive understanding of IDRs, we are also interested in the language of inter-IDR interactions, such as those which drive phase separation. We are therefore investigating the molecular determinants of inter-IDR interactions in an unbiased and proteome-wide fashion with the ultimate goal of developing algorithms for a priori prediction of inter-IDR interaction and condensate sorting.
Local compositional bias bestows IDRs with a modular architecture.