Contribution context

Contribution & Prior-Work Reviewer

SCORE: 4 CONFIDENCE: 4

Summary

This manuscript reports the discovery of IPO9 as a previously unknown regulator of cGAS-STING signaling, identified through phenotypic screening and chemical proteomics. The work establishes that IPO9 modulates cGAS activity by disrupting inhibitory interactions between cGAS and free H2A-H2B histone dimers—a distinct mechanism from nucleosome-mediated inhibition. The contribution is real and substantive: IPO9's role in cGAS regulation is genuinely novel, the structural and biochemical evidence is solid, and the work opens a new regulatory axis. However, the novelty is somewhat narrower than framed, and key mechanistic claims rest on inferences that warrant explicit scrutiny.

Strengths

1. The authors correctly identify and validate IPO9 as a bona fide target of SR-218 through orthogonal methods (competitive labeling, genetic knockdown, recombinant protein assays) with clear SAR correlation, establishing genuine target engagement.

2. The cryo-EM structure of cGAS:H2A-H2B at 4.3 Å provides direct structural evidence that free histone dimers block DNA binding sites on cGAS while permitting dimerization, distinguishing this inhibitory mode from nucleosome-mediated sequestration.

3. The manuscript appropriately acknowledges mechanistic uncertainty, explicitly stating that "multiple mechanisms" may contribute to IPO9-mediated cGAS release and that RanGTP involvement cannot be excluded.

Weaknesses: Load-Bearing Claims

Claim 1: IPO9 is a "previously unknown regulator of cGAS-STING signaling."

The evidence supporting this claim is IPO9's ability to reverse H2A-H2B-mediated inhibition of cGAS in vitro (Fig. 3d) and the loss-of-function phenotype in cells (Fig. 3b,c). However, the manuscript does not establish that IPO9 naturally regulates cGAS in cells under physiological conditions. The in vitro rescue occurs at 500 nM IPO9 with an EC50 of 110 nM for H2A-H2B displacement (Fig. 3f), but neither the cellular concentration of IPO9 nor the abundance of free H2A-H2B dimers in the nucleus is reported. The cell-based knockdown experiments show that loss of IPO9 impairs dsDNA-dependent STING phosphorylation, but this could reflect a general defect in histone homeostasis or nucleosome assembly rather than a specific cGAS-regulatory function. The authors do not demonstrate that endogenous IPO9 and free H2A-H2B interact in cells, nor do they show that this interaction is rate-limiting for cGAS activation under any physiological stimulus. The claim would be stronger if the authors reported: (i) cellular IPO9 and free H2A-H2B concentrations; (ii) co-immunoprecipitation or proximity labeling data showing IPO9-H2A-H2B interaction in cells; or (iii) a condition (infection, DNA damage, etc.) where IPO9 knockdown specifically impairs cGAS activation relative to other innate pathways.

Claim 2: IPO9 disrupts cGAS:H2A-H2B interaction by direct displacement via its H18-19 loop.

The evidence is the cryo-EM structure showing minimal steric overlap between IPO9 and cGAS except at the H18-19 loop (Fig. 4d), combined with BLI data showing IPO9-dependent loss of cGAS:H2A-H2B binding signal (Fig. 3f). However, the BLI experiment measures binding kinetics on immobilized cGAS, not free cGAS in solution. The authors observe "a brief concentration-dependent increase in signal immediately after sensors were dipped in IPO9" (Extended Data Fig. 5f,g), which they interpret as a transient tripartite complex but could equally represent non-specific sensor effects or kinetic artifacts. Critically, the authors cannot distinguish between three mechanisms: (i) direct displacement of H2A-H2B from cGAS by IPO9 binding the histone acidic patch; (ii) indirect sequestration of free H2A-H2B away from cGAS; or (iii) sequestration of H2A-H2B from DNA (which they show IPO9 does, Extended Data Fig. 5d,h-i). The cryo-EM structure is at 4.3 Å resolution with preferred orientation requiring 30° tilt, precluding confident sidechain modeling. The authors state "resolution limits our ability to draw definitive conclusions about the interaction at an amino acid level" (Results section), yet they propose a specific mechanism involving the H18-19 loop. To distinguish these mechanisms, the authors should report: (i) BLI with soluble (not immobilized) cGAS; (ii) mutation of the IPO9 H18-19 loop and assessment of its ability to disrupt cGAS:H2A-H2B in vitro and rescue cGAS activity in cells; or (iii) direct measurement of ternary complex formation by size-exclusion chromatography or analytical ultracentrifugation.

Claim 3: Free H2A-H2B dimers inhibit cGAS with similar potency to nucleosomes.

The evidence is the malachite green assay showing comparable cGAMP inhibition by H2A-H2B and NCP at the concentrations tested (Fig. 3d). However, the assay uses 100 nM cGAS with 50 nM H2A-H2B or NCP—a 2:1 molar ratio that may not reflect physiological stoichiometry. The authors do not report the Kd values for cGAS:H2A-H2B or cGAS:NCP binding, making it impossible to assess whether the similar inhibition reflects similar affinity or simply saturation at the tested concentrations. The cryo-EM structure reveals a 2:2 cGAS:H2A-H2B stoichiometry, but the functional significance of this ratio is unclear. The authors should report: (i) dose-response curves for H2A-H2B and NCP inhibition of cGAS activity across a wider concentration range; (ii) Kd values from BLI or surface plasmon resonance; or (iii) quantification of free H2A-H2B dimer abundance in the nucleus relative to nucleosomes.

Weaknesses: Sweep

• The manuscript does not cite or discuss Cho et al. (2024, Nature 625:585–592) on MRE11-mediated release of cGAS from nucleosomes, which describes an alternative mechanism for cGAS liberation during tumorigenesis and should be positioned relative to the IPO9 axis.

• SR-218 shows poor metabolic stability (Extended Data Fig. 3a) and the authors acknowledge this precludes steady-state dosing, limiting the compound's utility as a tool and raising questions about whether the in vivo Trex1−/− experiment (Fig. 1g) achieved sufficient target engagement.

• The claim that IPO9 "moonlights" beyond nucleocytoplasmic transport (Discussion) is speculative; the authors provide no evidence that IPO9's cGAS-regulatory function is independent of its canonical RanGTP-dependent cargo release mechanism.

• The manuscript does not address whether SR-218 or IPO9 inhibition affects nucleosome assembly or chromatin structure, which could confound interpretation of the cGAS-specific effects.

• The in vivo pharmacodynamic marker (cGAMP in heart tissue) is indirect; the authors do not measure cGAS activity, STING phosphorylation, or downstream interferon responses in the Trex1−/− mice.

• The cryo-EM reconstruction uses a 30° tilt dataset to overcome preferred orientation, which may introduce anisotropic resolution and bias in the final model; the angular distribution and local resolution are not reported in the main text.

• The authors state that IPO9 "may actively remove histone dimers from the cGAS:H2A-H2B complex by displacing the acidic patch interaction with its H18-19 loop" (Results) but then note this "could only occur in the context of monomeric cGAS" (Extended Data Fig. 6c), leaving the functional relevance of the 2:2 dimer structure unresolved.

Questions

• What are the cellular concentrations of IPO9 and free H2A-H2B dimers, and do they support the in vitro EC50 values observed?

• Does mutation of the IPO9 H18-19 loop abolish its ability to disrupt cGAS:H2A-H2B binding in vitro and rescue cGAS activity in cells?

• What is the Kd for cGAS:H2A-H2B interaction, and how does it compare to cGAS:NCP?