PREreview de PARP1 Exhibits an Enzymatically Inactive Chromatin Binding Mode

Overall assessment

In this manuscript, Fiorenza et al. biochemically and biophysically characterize the binding of PARP1 to a minimal tri-nucleosome (NLE-Tri) substrate lacking free DNA ends. PARP1 is a key enzyme in the DNA damage response. When it binds to damaged DNA, PARP1 undergoes a conformational change that triggers its enzymatic activity. In the absence of DNA damage, PARP1 binds and compacts chromatin. How PARP1 interacts with undamaged DNA/chromatin while staying inactive is a critical question in the field. Additionally, PARP inhibitors (PARPi) are a class of drugs used to target cancers with deficiency in DNA repair, and any off-target effects of how such drugs affect normal PARP1 function in the absence of DNA damage are not well understood.

The biophysical data in this study strongly support the conclusion that PARP1 binds and compacts intact chromatin in a way that is distinct from its binding to damaged DNA, providing a mechanism for how PARP1 remains inactive in the absence of DNA damage while still engaging chromatin. Furthermore, the described binding mode is shown to be unaffected by clinically approved and in-trial PARPi, advancing the field’s understanding of how such therapies might affect normal cell function.

The precise mechanism of how PARP1 engages the linker DNA could benefit from some additional experiments and/or discussion in the text, summarized below in the major and minor points.

Major Points

1. Lines 167-168; 328-330: The authors state that “one PARP1 molecule can bind both linkers simultaneously” and that this is the mechanism of chromatin compaction by PARP1. However, it is unclear mechanistically how PARP1 engages on both linkers: do all the Zn and BRCT domains coordinate both DNA molecules? A previous publication from this group (Rudolph and Muthurajan et al., 2021; doi: 10.1016/j.molcel.2021.11.014) determined a cryo-EM structure of the PARP1 BRCT domain bound to a mononucleosome, showing that the BRCT domain can directly interact with the nucleosomal DNA. It is then possible that a single PARP1 could be coordinating one linker and the distal nucleosome to compact the tri-nucleosome, rather than both linkers.

   • Adding an assay (such as DNaseI footprinting) to determine the specific binding sites could make the claim that both linkers are bound by a single PARP1 more convincing. Alternatively, referencing affinities of the individual domains for linker DNA versus nucleosomal DNA could be addressed in the text to support the conclusion that there is preferential interaction with the linker DNA.

   • Inclusion of a non-linker-ended dinucleosome substrate for comparison into the experiments may help distinguish different modes of linker DNA binding by comparing affinities and compaction between one versus two linkers.

   • Most of the key experiments in this manuscript are performed with a 2-2.5x excess of PARP1 over NLE-Tri, which by mass photometry and EMSA (50 nM trinucleosome) show 1:1 stoichiometry. However, the MNase protection experiment supporting linker DNA binding is performed with 4x more of each component (200 nM trinucleosome). Demonstrating that the preferred 1:1 stoichiometry does not change with increasing concentration could better support the claim that a single molecule of PARP1 indeed protects both linkers of a trinucleosome.

Minor Points

1. Line 24: The abstract states that there is “cooperative” binding of the Zn1, Zn2, Zn3, and BRCT domains, but the manuscript does not demonstrate cooperativity with the single-domain deletion mutants used. Furthermore, the four consecutive DNA-binding domains may increase avidity by tethering multiple DNA-binding domains rather than binding cooperatively. Measuring Kd values for combinations of isolated domains (Zn1-Zn2, Zn3-BRCT, etc.) in comparison to N-PARP could resolve this ambiguity and provide a more rigorous biophysical characterization of the system. Alternatively, the authors could consider removing the word “cooperative” and instead qualitatively describe the binding in the text to avoid this confusion.

2. Figure 2B: In Figure 1D-E, the fold-excess of PARP1 was used rather than the absolute concentration. Because this experiment is similarly used to identify a ratio of trinucleosome to PARP1, the authors could consider changing the X-axis to fold-excess for consistency and clarity so that the reader can better connect the different experiments in Figures 1 and 2.

3. Figure 4D-E: PARPi were shown to have no effect on chromatin binding and compaction. PARPi are known to bind the catalytic domain, which the authors determined does not have a role in the studied mechanism of chromatin binding (Fig. S6, where N-PARP is sufficient to bind and compact). Briefly stating the binding site for PARPi in the text could link the two findings and provide a stronger mechanistic basis for the model presented.

4. Does PARP1 compete with linker histone for the same binding sites? Even if this is not directly tested in this study, it could be a helpful point to address in the discussion to help contextualize the newly characterized binding mode in normal chromatin environments.

5. The full-length PARP1 was expressed and purified from insect cells (lines 423-424), whereas the PARP1 mutants were expressed and purified from E. coli (lines 448-449), raising the question of whether the bacterially-expressed protein behaves similarly to the insect-expressed protein. The authors could consider presenting or referencing data that the two expression systems produce PARP1 with similar activity profiles to resolve this question.

Disclosure statement: I have anonymously reviewed this preprint as part of the HHMI Transparent and Accountable Peer Review training program. I selected this preprint based on my training in protein-nucleic acid biochemistry/biophysics and structural biology. This report was written with the support of my research mentor. Claude Sonnet 5.5 was used to provide suggestions on clarity and tone, but AI tools were not used to write or generate any contents of the report.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.