Molecular determinants of phase separation for Drosophila DNA replication licensing factors

  1. Matthew W Parker
  2. Jonchee A Kao
  3. Alvin Huang
  4. James M Berger
  5. Michael R Botchan

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This paper studies the role of phase separation in replication initiation, with a focus on Cdt1. Sorting out the relative roles of phase separation and other mechanisms will require a detailed dissection of the amino acids driving phase separation, which can then be used to probe the role of phase separation in cells. Here the authors perform extensive and comprehensive analyses that are well done and that set the scene for a full dissection of the role of condensation in replication initiation inside cells.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Liquid–liquid phase separation (LLPS) of intrinsically disordered regions (IDRs) in proteins can drive the formation of membraneless compartments in cells. Phase-separated structures enrich for specific partner proteins and exclude others. Previously, we showed that the IDRs of metazoan DNA replication initiators drive DNA-dependent phase separation in vitro and chromosome binding in vivo, and that initiator condensates selectively recruit replication-specific partner proteins (Parker et al., 2019). How initiator IDRs facilitate LLPS and maintain compositional specificity is unknown. Here, using Drosophila melanogaster ( Dm ) Cdt1 as a model initiation factor, we show that phase separation results from a synergy between electrostatic DNA-bridging interactions and hydrophobic inter-IDR contacts. Both sets of interactions depend on sequence composition (but not sequence order), are resistant to 1,6-hexanediol, and do not depend on aromaticity. These findings demonstrate that distinct sets of interactions drive condensate formation and specificity across different phase-separating systems and advance efforts to predict IDR LLPS propensity and partner selection a priori.

Article activity feed

  1. Version published to 10.7554/elife.70535 on eLife
  2. eLife

    Evaluation Summary:

    This paper studies the role of phase separation in replication initiation, with a focus on Cdt1. Sorting out the relative roles of phase separation and other mechanisms will require a detailed dissection of the amino acids driving phase separation, which can then be used to probe the role of phase separation in cells. Here the authors perform extensive and comprehensive analyses that are well done and that set the scene for a full dissection of the role of condensation in replication initiation inside cells.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    This study by Parker et al is an in-depth analysis of the chemical "rules" that underlie liquid-liquid phase separation (LLPS) by the Drosophila melanogaster (D.m.) Cdt1 replication initiation protein. Given that Orc1 and Cdc6 replication initiation proteins have similar internal disordered regions (IDR) that promote phase separation, the findings of the current report are likely to generalize to the initiator class of IDRs that induce LLPS.

    The author's provide an excellent intellectual background on IDR composition and determinants for LLDS in many different systems. This allows the authors to effectively communicate to readers the comparison of the initiator IDR with a wide variety of IDRs that induce LLPS in other types of proteins. The analysis suggests that the initiator class of IDR is quite distinct from some of the most commonly studied proteins that lead to IDR-induced LLPS.

    A strength of the study is that it uses several techniques to analyze inherent LLPS formation in the D.m. Cdt1 replication initiator protein. This includes uses of hydrophobic disruption agents, salt, use of various polyanions as scaffolds, and use of PEG as a crowding agent. The authors mutate the Cdt1 IDR sequence in many ways, following a well explained logistical reasoning for identification of the determinants within the IDR that are required for LLPS formation. Thus they explain their studies of the IDR for aromatic residues, for short repeats, for linear and branched hydrophobic residues. The authors then perform extensive mutagenesis to test their analysis of the IRD sequence. The mutagenesis includes deletions of large segments of the IDR, complete total scrambling of the IDR sequence, removal of hydrophobic residues, replacement of hydrophobic residues. In each case, the mutants are assayed for relative efficiency of LLPS formation using either PEG or DNA to induce LLPS.

    The authors conclude that the Cdt1 IDR amino acid composition, but not its sequence, is required for LLPS formation. This was a surprise. They found that aromatic residues were not required, unlike the case for some other types of proteins, but that the hydrophobic residues are required, regardless of sequence context. Furthermore, they determine that the hydrophobic residues mediate inter-IDR interactions and that these provide the main force behind LLPS formation. They find that interaction with a polyanion, like DNA (or numerous other types of polyanions), helps the phase separation process, but is not the main driving force of LLPS by Cdt1.

    Overall, this study is an important "next step" in the study of LLPS, and the authors convincingly identify the chemical determinants that are necessary for LLPS formation in Cdt1. The results are clear, and the data support the author's conclusions. Furthermore, the results with Cdt1 are likely to extend to the Orc1 and Cdc6 initiator proteins, which have IDRs that appear to have similar characteristics to Cdt1. The authors make the interesting proposal that this "like-recruits-like" IDR interaction in the LLPS may contribute to the basis by which these "initiator LLPS particles" specifically attract initiator proteins while excluding other types of proteins that undergo LLPS.

    The studies of this report may form a template for identification of the determinants within the IDRs in other types of proteins that undergo LLPS.

  4. eLife

    Reviewer #2 (Public Review):

    In this paper, the authors study the Drosophila replication initiation protein Cdt1 in vitro. This study builds on previous work by the Berger and Botchan labs in which the authors reported that Cdt1 as well as Cdc6 and ORC, other replication initiation factors, undergo liquid liquid phase separation (LLPS) when combined with a short stretch of DNA in vitro. This phase separation is dependent on the intrinsically disordered region (IDR) in the N-terminus of the Cdt1 protein. Importantly, the condensed phase recruits the Mcm helicases required complex that is required for replication initiation. In addition, the IDR seems to be regulated by cyclin dependent kinases, and dephosphorylation at the end of mitosis could allow the formation of initiator condensates on DNA to initiate recruitment of other replication factors (Parker et al., 2019).

    In the current manuscript, the authors address the questions of which molecular hallmarks of Cdt1 IDR control DNA-dependent phase separation. They find that positively charged amino acids play a crucial role while aromaticity - which was reported to drive phase separation of other proteins - is not important for Cdt1 phase separation. Indeed, other negatively charged polymers can phase separate with Cdt1 as well. Interestingly, the charge distribution in the IDR does not seem to be important for the phase separation process. However, because Cdt1 molecules with a more uniform charge distribution have a different elution profile in size exclusion chromatography, the charge distribution may serve another unknown function. In addition, the authors describe an essential role of hydrophobic amino acids (in particular Leucin, Isoleucin and Valine) in driving the self-interaction of Cdt1 molecules that is required for phase separation.

    In conclusion, the authors propose that phase separation of replication initiation factors on DNA requires both, the electrostatic interactions of the positively charged Cdt1 IDR with the DNA polymer, as well as the hydrophobic IDR-IDR interaction which mostly involve Leucin and Isoleucine.

    Overall, this is a good follow up work dissecting, in detail, the molecular interactions required for replication initiation factor phase separation. As in their previous work, the experiments are well performed and of good quality. The authors created a series of Cdt1 mutants that, in combination, elucidate which types of interactions contribute to phase separation. However, given that studying the effect of mutations on the phase separation behaviour of Cdt1 is the main focus of this work, the simple 'depletion assays' which the authors performed throughout the work do not give sufficient quantitative insight and phase diagrams should be prepared for wt and mutant proteins. In addition, the core assumption of this work is that all protein-dependent phase separation is driven by aromatic residues. While this was shown for several proteins, there are other examples which includes Michael Rosen's work on NCK and N-WASP (Li et al., Nature, 2012) or work on the nucleolar protein NPM1 (Mitrea et al. or Ferrolino et al., Nat Commun, 2018), to just name a few. Further, many examples have been described in which the interaction between a protein and a nucleic acid is driving phase separation (e.g. Shrinivas et al., Mol. Cell, 2019 or Guillén-Boixet et al./Sanders et al./Yang et al., Cell, 2020 or Boeynaems et al, PNAS, 2019), similarly to what the authors described for DNA and some of which has been shown to be driven by basic amino acids.

    Although it was not a prior authors goal of this study, we learn very little about the relevance of the presented work for replication initiation in vivo. Is the recruitment of other replication factors influenced by the balance of electrostatic and hydrophobic interactions, why are the mutants not tested for recruitment of other initiation factors? How does this work compare to older, established systems for replication initiation? While RNA is freely diffusing inside the cell, this is not equally true for the long DNA polymer. It is therefore unclear, how replication phase separation could work inside the cell.

    Other major points:

    Considering that this is an in vitro study of Cdt1 solely, and the main assays are a pelleting assay, as well as fixed time point fluorescent microscopy, differences among mutants are hard to judge. Phase diagrams of wildtype and mutant proteins should be prepared. As well saturation concentrations should be determined and microscopy images of the different phase separation assays should be provided. Biophysical parameters of condensates are not addressed at all and would give considerable insight into the role of the different amino acid interactions.

    The authors show that hexanediol does not dissolve Cdt1 droplets. In addition, they convincingly establish that phase separation of Cdt1 is driven by the interaction between basic amino acids and RNA as well as between interactions of hydrophobic residues but not aromatic amino acids. This gives interesting insight into the type of interactions that are disrupted by hexanediol. Given that this chemical is widely used in the literature to probe for phase separation, this is an important finding of their work and should be discussed further.

    The authors dissect in detail which molecular interactions are driving phase separation of Cdt1. A picture emerges in which Cdt1 is soluble in the aqueous solution due to charged (mostly negatively charged) amino acids. However, through binding to DNA those charges are neutralized, exposing almost exclusively hydrophobic amino acids to the aqueous environment of the cell. This is a likely explanation for how those hydrophobic interactions are driving phase separation in the presence of DNA. This model (or an alternative model provided by the authors) should be discussed in more detail, for which the authors are suggested to re do Fig. 6.

  5. eLife

    Reviewer #3 (Public Review):

    The central hypothesis of the manuscript is that initiator proteins have different characteristics that lead to their phase separation away from other proteins that undergo LLPS. Consistent with the authors' hypothesis, they find that ORC, Cdc6 and Cdt1 IDRs have distinct sequences and that LLPS by these proteins is resistant to 1.6 hexanediol, in contrast to many other LLPS-forming proteins. In the rest of the paper, the authors analyze the requirements for Cdt1 LLPS in detail. They find that the sequence composition of the IDRs associated with these initiation proteins are distinct from other proteins involved in LLPS. The studies provide strong evidence that the role of DNA in the induction of LLPS by Cdt1 is as a multivalent ion. Finally, the show that branched chain hydrophobic but not aromatic residues are important for Cdt1 LLPS formation.

    The strength of this manuscript are the detailed studies of the requirements for the Cdt1 IDR to undergo LLPS. The authors provide clear evidence as to the sequence determinants of Cdt1 LLPS. The authors use the identification of Cdt1-Cdt1 interactions in the absence of DNA to identify critical aspects of the sequence for self-interaction and then show that they are important for the DNA-dependent LLPS formation.

    The weaker part of the paper is evidence that these observations can be extended to the other 'initiator IDRs'. It would substantially improve the manuscript to test whether similar mutations in another initiator IDR (either ORC or Cdc6) have the same consequences. It would also be nice to determine whether the mutations that prevent Cdt1 from phase separating on its own would also be enough to prevent it from joining an ORC-DNA or Cdc6-DNA LLPS event.

    Overall, this manuscript describes a thorough analysis of the determinants of LLPS by DNA and Cdt1.

  6. Version published to 10.1101/2021.04.12.439485v1 on bioRxiv