Organelle membrane-associated proteins recruit cGAS via phase separation to facilitate its membrane localization

Reviewer #1 (Public review):

Summary:

This manuscript by the Yin group presents interesting findings that organelle-tethered intrinsically disordered "MEMCA" scaffolds, as exemplified by ZDHHC18 at the Golgi and MARCH8 at endosomes, enhance the engagement of cGAS with organelle-proximal condensates, thereby sequestering cGAS from cytosolic DNA sensing and negatively regulating innate immunity.

Strengths:

These findings suggest a previously unrecognized mechanism by which Golgi/endosomal IDR scaffolds modulate cGAS activity, with implications for antiviral defense and tumor immunology. The study is conceptually intriguing and potentially impactful.

Weaknesses:

While the manuscript addresses a novel aspect of cGAS regulation, additional mechanistic insights and targeted validations are needed to ensure robustness:

(1) How do ZDHHC18/MARCH8 enhance cGAS engagement? Do they act as bridges to form a ternary, membrane-tethered cGAS-DNA-MEMCA complex, or alter cGAS condensate properties allosterically?

(2) Is organelle cGAS capture selective? For instance, can other palmitoyltransferases/E3 ligases be substituted for ZDHHC18/MARCH8?

(3) Why does membrane association suppress cGAS enzymic activity, as dsDNA still resides in cGAS condensation?

Reviewer #2 (Public review):

Summary:

The authors found that cGAS, a DNA sensor, relocalizes to organelle membranes (ER, Golgi, endosomes) upon DNA stimulation, revealing spatial regulation of its activity. ZDHHC18 and MARCH8 recruit cGAS to Golgi/endosomes via intrinsically disordered regions (IDRs), driving phase-separated condensates. This sequestration of cGAS-dsDNA complexes suppresses innate immune signaling, uncovering a novel regulatory mechanism.

Strengths:

The work overall is very interesting. The authors provided molecular and biochemical evidence.

Weaknesses:

Overall, the work is very interesting. However, the quality of some of the data does need to be improved, and more experiments need to be performed.

The following points need to be addressed:

(1) In Figure S7, no direct binding between cGAS and MARCH8 or ZD18 IDR is observed, and the interaction only occurs after DNA stimulation. However, Figure 5 shows cGAS recruitment to ZD18 or MARCH8 IDR droplets, suggesting direct interactions. This apparent discrepancy should be clarified.

(2) The authors propose that recruiting cGAS to organelle membranes reduces its activity, as demonstrated by the FKBP experiment. However, ZD18 and MARCH8 also post-translationally modify cGAS. Do both mechanisms contribute to this effect, and can the authors test this?

(3) To demonstrate the functional importance of MEMCA, the authors should test IFN production or STING activation in cells.

(4) Does the IDR of MARCH8 or ZD18 influence the interaction between cGAS and DNA?

(5) Which region of cGAS does the IDR of MARCH8 or ZD18 interact with: the cGAS-CD or the cGAS-N-terminus?

(6) The in vitro LLPS experiments with cGAS, DNA, and ZD18/MARCH8 should be conducted under physiological conditions.

Reviewer #3 (Public review):

Summary:

In this study by Shi et al., the authors evaluate if cGAS is recruited to the membranes of intracellular organelles. Using a combination of biochemical fractionation and imaging techniques, the authors propose that upon recognition of DNA, cGAS translocates to various subcellular locations, including the golgi, endoplasmic reticulum, and endosomes. Mechanistically, the authors propose that upon localizing to the Golgi or endosome, cGAS binding to MARCH8 and ZDHHC18 prevents cGAS activity by incorporating cGAS and dsDNA into biomolecular condensates. However, in its current form, the study does not directly address this question.

Strengths:

The question of evaluating cGAS sub-cellular localization as a mechanism for controlling activity is interesting, and there is some evidence that cGAS is localized to sub-cellular organelle membranes.

Weaknesses:

(1) The well-established nuclear localization of cGAS is not adequately addressed in the cell lines used and is inconsistent with the findings.

(2) Previous studies have shown that ZDHHC18 and MARCH8 control cGAS activity, which detracts somewhat from the novelty.

(3) A lot of inconsistency in the cell lines and artificial expression systems used across the study.

(4) A key element missing is showing that in the absence of ZDHHC18 or MARCH8, the loss of endogenous cGAS localization to the various sub-cellular organelles increases cGAMP synthesis and downstream STING activation in primary cells. There is an over-reliance on artificial expression systems. An important experiment to validate the hypothesis would be to evaluate endogenous cGAS localization in MARCH8- and ZDHHC18-deficient primary cells. Further, there should be evaluation of endogenous STING responses in MARCH8- and ZDHHC18-deficient primary cells in tandem with the localization studies.

(5) There are a large number of grammatical errors throughout the manuscript which should be addressed.

Author response:

Below we outline our provisional responses to the major points raised in the public reviews, and our planned revisions:

(1) Mechanistic model of how ZDHHC18/MARCH8 engage the cGAS–DNA condensate (Reviewer #1 & #2

We will add a dedicated subsection and a working-model figure describing our current view: IDRs of ZDHHC18 (Golgi) and MARCH8 (endosomes) engage pre-formed cGAS–DNA condensates at organelle membranes, and thereby tune cGAS activity through PTMs. We will explicitly discuss bridge-like versus allosteric modes by perform additional LLPS experiment (e.g. FRAP assay) to detect any IDR-driven changes in condensate properties, and explain how these scenarios fit our data.

(2) Selectivity beyond ZDHHC18/MARCH8 (Reviewer #1)

We will expand the text to explain existing evidence indicating that, in addition to ZDHHC18 or MARCH8, other post-translational modification (PTM) enzymes and/or membrane-associated scaffolds may also modulate cGAS. We will summarize our current datasets that support this possibility and outline how this selectivity relates to organelle identity.

(3) Why membrane association suppresses cGAS activity (Reviewer #1)

We will provide a concise mechanistic rationale—integrating our published work—to explain how membrane-proximal sequestration can limit cGAS catalysis despite cGAS–DNA coexistence within condensates. Specifically, we will discuss (i) IDR-dependent changes in condensate properties, and (ii) PTMs by ZDHHC18/MARCH8 that allosterically reduce catalytic efficiency; we will clearly cross-reference our prior publications that bear on these points.

(4) Reconciling Fig. S7 (DNA-dependent binding) with Fig. 5 (recruitment to IDR droplets) (Reviewer #2)

We will add text to clarify experimental context and readouts to prove that there is no real contradiction between Fig. S7 and Fig. 5. In the experiment shown in Fig. 5, PEG (a macromolecular crowding agent) was added to the system, which facilitates the formation of IDR phase-separated droplets. Under these conditions, cGAS partitions into the IDR condensates, leading to the observed recruitment. In contrast, Fig. S7 examines the direct physical interaction between cGAS and the IDRs using biochemical pull-down assays and shows that no direct interaction occurs in the absence of DNA. These two results reflect different experimental contexts and are therefore not mutually exclusive.

(5) Planned additional tests to address specificity and mechanism (Reviewer #2)

DNA pull-down: to test whether IDRs alter cGAS–DNA affinity, we will compare cGAS binding to DNA with/without MEMCA IDRs (and with charged-residue mutants).

Domain mapping: to determine which region of cGAS engages MEMCA IDRs, we will map binding using cGAS N-terminus/core-domain truncations and key surface mutants.

Physiological in vitro LLPS: we will repeat cGAS–DNA–IDR LLPS assays under physiological buffer conditions and report partition coefficients, FRAP, and phase diagrams to ensure physiological relevance.

(6) Image clarity and data presentation (Reviewer #2):

We will improve image resolution, add zoomed-in insets with organelle markers, and provide more significant Cy5-ISD signal.

(7) Nuclear localization of cGAS and system considerations (Reviewer #3)

We will explicitly document the nuclear signal of cGAS observed in our confocal experiments, detail the cell lines and expression systems used. We will also clarify cGAS nuclear localization in the cell lines used.

(8) Endogenous validation and cell line consistency (Reviewer #3):

We will perform experiments in primary cells (knockout macrophages) to address the concern of relying on overexpression.

(9) Language and grammar (Reviewer #3):

We will thoroughly revise the manuscript for grammar and clarity.

Together, these planned revisions will strengthen the mechanistic basis of our findings and provide direct evidence for the physiological role of organelle-tethered IDRs in regulating cGAS activity.