Figures and data
cGAS is associated with organelle membranes.
(a) THP-1 cells were either untreated (Mock), stimulated with HT-DNA (2 μg/mL for 2 h), lysed in the presence of 1% Triton X-100, or subjected to both HT-DNA stimulation and Triton X-100 lysis, as indicated. Cytoplasmic extracts were subjected to OptiPrep density gradient centrifugation (10%–40%). Twenty-three fractions were collected from top (light) to bottom (dense) and analyzed by immunoblotting using antibodies against EEA1 (early endosome marker), GM130 (Golgi marker), Calnexin (endoplasmic reticulum marker), and cGAS. The yellow dashed triangle indicates that the distribution of the endoplasmic reticulum is affected by Triton X-100, while the yellow solid triangle indicates that cGAS remains present in the redistributed ER fractions. The green dashed triangle marks the redistribution of the Golgi apparatus upon Triton X-100 treatment, and the green solid triangle shows that cGAS is still retained in the altered Golgi fractions.
(b) Schematic diagram showing the experimental setup for detecting cGAS interaction and activation using the split sfGFP system. cGAS is tagged with G10-Flag and HA-G11, which in the presence of DNA, bring together GFP1−9, G10 and G11 fragments to reconstitute sfGFP, allowing fluorescence detection. Variants include cytosolic sfGFP, ERM-sfGFP, 2xFyve-sfGFP and GM130N for different subcellular localizations.
(c) Fluorescence microscopy images showing the localization of sfGFP-tagged cGAS in different cellular compartments (cytosol, endoplasmic reticulum (ER), Golgi apparatus and endosome) with and without HT-DNA. For the positive control, sfGFP1-9 fragment was fused to TRAF6, a protein known to spontaneously form aggregates upon overexpression, thus enabling successful assembly of the full sfGFP. The negative control consisted of the human serum albumin signal peptide (MKWVTFISLLFLFSSAYSRGVFRR) fused to the N-terminus of sfGFP1-9 and a KDEL sequence fused to the C-terminus, ensuring retention in the ER lumen. Scale bars = 5 µm.
(d) CLEM reveals the subcellular localization of cGAS to organelle membranes. Top-left: Correlative image combining light and electron microscopy, scale bar: 2 μm; Top-right: Confocal microscopy showing GFP-cGAS localization in HeLa cells stably expressing GFP-cGAS, following 2-hour transfection with HT-DNA (2 μg/ml), scale bar: 4 μm; Bottom-left: Transmission electron microscopy (TEM) image of the same cell, operating voltage: 100 kV, scale bar: 5 μm; Bottom-right: Enlarged view of the boxed region in the bottom-left panel, scale bar: 500 nm. The dashed white lines indicated the condensates of GFP-cGAS and HT-DNA. The red arrows indicate contacts between the endoplasmic reticulum (ER) and cGAS, the blue arrows indicate contacts between small vesicles and cGAS, and the yellow arrows indicate contacts between the Golgi apparatus and cGAS.
(e-f) Fraction of Golgi apparatus (e) or early endosome (f) of THP1 cells in the absence or presence of HT-DNA (2 μg/ml). Subcellular fractionation of the Golgi apparatus (e) and early endosomes (f) was performed using commercial organelle isolation kits (Invent Biotechnologies) according to the manufacturer’s instructions. Western blot analysis was performed to detect the presence of cGAS in the purified fractions. GM130 and EEA1 were used as markers for the Golgi apparatus and early endosomes, respectively. GAPDH served as a cytosolic fraction control.
MEMCA facilitates organelle membrane localization of cGAS via IDR.
(a) Representative confocal fluorescence microscopy images of HeLa cells stably expressing GFP-cGAS (green) and full-length mCherry-ZDHHC18FL (red), transfected with or without Cy5-labeled interferon stimulatory DNA (Cy5-ISD, 2 μg/mL, 1 h; magenta). In the absence of DNA stimulation (–Cy5-ISD), GFP-cGAS is diffusely distributed in the cytosol and nucleus, while mCherry-ZDHHC18FL displays punctate cytoplasmic localization with minimal overlap. Upon Cy5-ISD transfection (+Cy5-ISD), GFP-cGAS redistributes to cytoplasmic puncta that strongly co-localize with mCherry-ZDHHC18FL, suggesting condensate formation at DNA-associated compartments. Scale bars: 10 μm.
(b) Enlarged high-magnification views of boxed regions from panel (a), highlighting the spatial convergence of GFP-cGAS, mCherry-ZDHHC18FL, and Cy5-ISD. White arrows indicate DNA-containing condensates that are positive for both cGAS and ZDHHC18FL, indicating a tripartite interaction upon activation. Scale bars: 2 μm.
(c) Representative confocal images of HeLa cells stably expressing GFP-cGAS and the IDR deletion mutant mCherry-ZDHHC18ΔIDR. In unstimulated cells (–Cy5-ISD), both GFP-cGAS and ZDHHC18ΔIDR show diffuse distributions, with no apparent co-localization. Upon Cy5-ISD transfection, GFP-cGAS forms DNA-associated puncta, but ZDHHC18ΔIDR remains largely excluded from these structures, exhibiting weak or no co-localization. Scale bars: 10 μm.
(d) High-magnification views of boxed regions from panel (c) reveal that although GFP-cGAS is recruited to Cy5-ISD-positive condensates (magenta and green), these structures lack enrichment of ZDHHC18ΔIDR (red), suggesting that the IDR of ZDHHC18 is required for its DNA-induced co-localization with cGAS. Scale bars: 2 μm.
(e) Line profile analysis of fluorescence intensity in (b).
(f) Line profile analysis of fluorescence intensity in (d).
(g) Representative confocal fluorescence microscopy images of HeLa cells stably expressing GFP-cGAS (green) and full-length mCherry-MARCH8FL (red), transfected with or without Cy5-labeled interferon stimulatory DNA (Cy5-ISD, 2 μg/mL, 1 h; magenta). In the absence of DNA stimulation (–Cy5-ISD), GFP-cGAS is diffusely distributed in the cytosol and nucleus, while mCherry-MARCH8FL displays punctate cytoplasmic localization with minimal overlap. Upon Cy5-ISD transfection (+Cy5-ISD), GFP-cGAS redistributes to cytoplasmic puncta that strongly co-localize with mCherry-MARCH8FL, suggesting condensate formation at DNA-associated compartments. Scale bars: 10 μm.
(h) Enlarged high-magnification views of boxed regions from panel (g), highlighting the spatial convergence of GFP-cGAS, mCherry-MARCH8FL, and Cy5-ISD. White arrows indicate DNA-containing condensates that are positive for both cGAS and MARCH8FL, indicating a tripartite interaction upon activatio. Scale bars=2 µm.
(i) Representative confocal images of HeLa cells stably expressing GFP-cGAS and the IDR deletion mutant mCherry-MARCH8ΔIDR. In unstimulated cells (–Cy5-ISD), both GFP-cGAS and MARCH8ΔIDR show diffuse distributions, with no apparent co-localization. Upon Cy5-ISD transfection, GFP-cGAS forms DNA-associated puncta, but MARCH8ΔIDR remains largely excluded from these structures, exhibiting weak or no co-localization. Scale bars=10 µm.
(j) High-magnification views of boxed regions from panel (c) reveal that although GFP-cGAS is recruited to Cy5-ISD-positive condensates (magenta and green), these structures lack enrichment of MARCH8ΔIDR (red), suggesting that the IDR of MARCH8 is required for its DNA-induced co-localization with cGAS. Scale bars=2 µm.
(k) Line profile analysis of fluorescence intensity in (h).
(l) Line profile analysis of fluorescence intensity in (j).
(m and n) HeLa cells stably co-expressing eGFP-cGAS (green) and mCherry-ZDHHC18IDR (red; IDR-only fragment of ZDHHC18) (m) and mCherry-MARCH8IDR (red; IDR-only fragment of MARCH8) (n) were transfected with or without Cy5-labeled interferon stimulatory DNA (Cy5-ISD, 2 μg/mL, magenta) for 1 hour, followed by fixation and confocal microscopy. Each panel shows the merged image (left), eGFP-cGAS signal (middle), and mCherry signal (right). Scale bar: 10 μm.
MEMCAIDR forms biomolecular condensates in cells.
(a) Schematic diagram showing the structure of ZDHHC18 and MARCH8 proteins, highlighting their IDR and transmembrane domains (TM). The opto-droplets system utilizes Cry2-mCherry fused to the IDR of MARCH8 and ZDHHC18 to induce condensate formation upon blue light (488 nm) activation.
(b) Fluorescence microscopy images of cells expressing opto-MARCH8IDR with and without blue light activation. The light-induced condensate formation is shown in the presence of blue light. Light ON: 3 min cycles of 4s light (488 nm) and 10s resting. Representative fluorescence images are shown. Scale bars=5 µm.
(c) Fluorescence microscopy images of cells expressing opto-ZDHHC18IDR with and without blue light activation. The light-induced condensate formation is shown in the presence of blue light. Light ON: 3 min cycles of 4s light (488 nm) and 10s resting. Representative fluorescence images are shown. Scale bars=5 µm.
(d) Quantification of mean mCherry spot/cell for opto-MARCH8IDR before and after light activation, showing the kinetics of condensate formation.
(e) Quantification of mean mCherry spot/cell for opto-ZDHHC18IDR before and after light activation, showing the kinetics of condensate formation.
(f) After 5 minutes of light activation, blue light illumination was discontinued, and changes in the mCherry signal were subsequently monitored. Decay of mean mCherry spot/cell for opto-MARCH8IDR after cessation of blue light activation, indicating the dynamics of condensate dissolution.
(g) After 5 minutes of light activation, blue light illumination was discontinued, and changes in the mCherry signal were subsequently monitored. Decay of mean mCherry spot/cell for opto-ZDHHC18IDR after cessation of blue light activation, indicating the dynamics of condensate dissolution.
(h) Diagram illustrating the calculation of aspect ratio to determine the shape of the condensates.
(i) Three-dimensional analysis of opto-ZDHHC18IDR condensates showing the aspect ratio close to 1, indicating a spherical shape. Scale bar=5 µm.
(j) Three-dimensional analysis of opto-MARCH8IDR condensates showing the aspect ratio close to 1, indicating a spherical shape. Scale bar=5 µm.
(k) TIRF-SIM images of opto-ZDHHC18IDR condensates providing high-resolution details of the condensate structure. Scale bar=0.5 µm.
(l) TIRF-SIM images of opto-MARCH8IDR condensates providing high-resolution details of the condensate structure. Scale bar=0.5 µm.
(m) Schematic diagram of the opto-IDR assay setup using LOV-Turbo and Cry2-mCherry to investigate the formation and composition of MEMCAIDR condensates upon blue light activation.
(n and o) Volcano plots showing proteins enriched in proximity to MARCH8IDR (n) or ZDHHC18IDR (o) following opto-genetically induced condensate formation using the LOV-Turbo system. THP-1 cells stably expressing LOV-Turbo-Cry2-mCherry-MARCH8IDR or LOV-Turbo-Cry2-mCherry-ZDHHC18IDR constructs. Cells were exposed to 488 nm blue light for 30 minutes (“light-on”) or maintained in the dark (“light-off”) in the presence of 500 μM biotin, followed by streptavidin pulldown and LC-MS/MS analysis of biotinylated proteins. Proteins are plotted by their log₂ fold-change (light-on vs light-off) on the x-axis and the –log₁₀(p-value) from unpaired Student’s t-tests on the y-axis. Vertical dotted lines represent ±1 log₂ fold change; horizontal dotted lines denote a significance threshold of p = 0.05. Significantly enriched proteins (right quadrant) are highlighted in red and labeled.
MEMCAIDR undergoes liquid-liquid phase separation in vitro.
(a) Phase separation of eGFP-ZDHHC18IDR at varying concentrations (5, 10, 25, and 50 µM) and PEG-8000 concentrations (0, 5%, 10%, and 15%). Fluorescence microscopy images show the formation of liquid droplets indicating phase separation. Scale bars=100 µm.
(b) Phase separation of eGFP-MARCH8IDR at varying concentrations (5, 10, 25, and 50 µM) and PEG-8000 concentrations (0, 5%, 10%, and 15%). Fluorescence microscopy images show the formation of liquid droplets indicating phase separation. Scale bars=10 µm.
(c) Fluorescence recovery after photobleaching (FRAP) analysis of eGFP-MARCH8IDR (10µM) droplets with 10% PEG-8K. Images show the fluorescence recovery over time after photobleaching, indicating dynamic exchange within the droplets. Scale bar=0.2 µm.
(d) Fluorescence recovery after photobleaching (FRAP) analysis of eGFP-ZDHHC18IDR (10µM) droplets 10% PEG-8K. Images show the fluorescence recovery over time after photobleaching, indicating dynamic exchange within the droplets. Scale bar=0.2 µm.
(e) Quantification of FRAP recovery curves for eGFP-MARCH8IDR droplets, showing the mean fluorescence intensity over time post-bleaching. Time 0 indicates the end of photobleaching and the start of recovery. The mean ± SD are shown. N = 3 liquid droplets.
(f) Quantification of FRAP recovery curves for eGFP-ZDHHC18IDR droplets, showing the mean fluorescence intensity over time post-bleaching. Data represent the mean ± SEM from three independent experiments. Time 0 indicates the end of photobleaching and the start of recovery. The mean ± SD are shown. N = 3 liquid droplets.
(g) Reversibility of droplet formation for eGFP-ZDHHC18IDR (top row) and eGFP-MARCH8IDR (bottom row) upon dilution. Fluorescence microscopy images show droplet formation at 25 µM protein concentration and the reduction in size and number of droplets upon 1/2 dilution. Scale bars = 20 µm.
MEMCAIDR biomolecular condensates recruit cGAS and dsDNA.
(a) Fluorescence microscopy images showing the recruitment of eGFP-cGAS (green) (2 μM) to mCherry-ZDHHC18IDR (red) (10 μM) condensates with treatment of 16-HD or high salt (300 mM NaCl). The images demonstrate the effect of these treatments on the recruitment process. Scale bars=20 µm.
(b) Fluorescence microscopy images showing the recruitment of eGFP-cGAS (green) (2 μM) to mCherry-MARCH8IDR (red) (10 μM) condensates with treatment of 16-HD or high salt (300 mM NaCl). The images demonstrate the effect of these treatments on the recruitment process. Scale bars=20 µm.
(c) Fluorescence microscopy images showing the recruitment of Cy5-ISD (green) (0.5 μM) to mCherry-ZDHHC18IDR (red) (10 μM) condensates with treatment of 16-HD or high salt (300 mM NaCl). The images demonstrate the effect of these treatments on the recruitment process. Scale bars=20 µm.
(d) Fluorescence microscopy images showing the recruitment of Cy5-ISD (green) (0.5 μM) to mCherry-MARCH8IDR (red) (10 μM) condensates under different conditions: buffer, 16-HD, and high salt. The images demonstrate the effect of these treatments on the recruitment process. Scale bars=20 µm.
(e and f) Fluorescence microscopy images showing in vitro condensates formed by recombinant eGFP-cGAS (green, 2 μM) and mCherry-MARCH8IDR (red, 10 μM) (e) or mCherry-ZDHHC18IDR (red, 10 μM) (f) incubated in reaction buffer (20 mM Tris pH 7.5, 150 mM NaCl, 5 mM MgCl2) at room temperature for 30 minutes. Confocal images show distinct phase-separated droplets containing both fluorescent proteins. Arrows indicate representative condensates used for fluorescence intensity line profiling. Scale bars: 5 μm.
(g and h) Fluorescence intensity profiles along the white line in panel e (g) and f (h), plotted as a function of distance (μm). Fluorescence intensities were extracted using ImageJ.
(i and j) Fluorescence microscopy images showing the result of in vitro mixing of Cy5-labeled interferon stimulatory DNA (Cy5-ISD, magenta, 0.5 μM) with recombinant mCherry-MARCH8IDR (red, 10 μM) (i) or mCherry-ZDHHC18IDR (red, 10 μM) (j) in reaction buffer (20 mM Tris pH 7.5, 150 mM NaCl, 5 mM MgCl2). Samples were incubated at room temperature for 30 minutes prior to imaging by confocal microscopy. White arrows indicate representative puncta. Scale bars: 5 μm.
(k and l) Fluorescence intensity line-scan analysis along the white arrow shown in i (k) and j (l), plotting Cy5 (magenta) and mCherry (red) signal intensities across the selected condensates. Fluorescence intensities were extracted using ImageJ.
(m and o) Representative fluorescence microscopy images showing in vitro reconstitution of phase-separated condensates formed by mixing recombinant eGFP-cGAS (green, 2 μM), mCherry-MARCH8IDR (red, 10 μM) (m) or mCherry-ZDHHC18IDR (red, 10 μM) (o), and Cy5-labeled interferon stimulatory DNA (Cy5-ISD, magenta, 0.5 μM) in reaction buffer (20 mM Tris pH 7.5, 150 mM NaCl, 5 mM MgCl2). Following 30 minutes of incubation at room temperature, samples were imaged using confocal microscopy. The merged image shows clear overlap of all three components within the same condensates. White arrows indicate regions of interest used for subsequent line profile analysis. Scale bars: 5 μm.
(n and p) Quantification of fluorescence intensity along the white line indicated in (m) and (o). Line-scan analysis reveals co-enrichment of eGFP-cGAS (green), mCherry-MARCH8IDR (m) or ZDHHC18IDR (o) (red), and Cy5-ISD (magenta) within the same micron-scale condensates. Fluorescence intensities were extracted using ImageJ.
(q and r) HeLa cells stably expressing eGFP-cGAS (green) and mCherry-Cry2-MARCH8IDR (red) (q) or mCherry-Cry2-ZDHHC18IDR (red) (r) were subjected to light-inducible phase separation. Cells were maintained in the dark (“light off”, top row) or exposed to 488 nm blue light for 2 minutes (“light on”, bottom row), followed by confocal imaging. The white arrows indicate the cells subjected to light activation. Each panel shows merged (left), green (middle), and red (right) fluorescence channels. Scale bars=5 µm.
MEMCA biomolecular condensates regulates cGAS activity.
(a) Schematic representation of the experimental setup to study cGAS activity regulation via rapamycin-induced biomolecular condensates. Fusion proteins eGFP-cGAS-FRB and mCherry-FKBP12 or its variants (ERM-mCherry-FKBP12, mCherry-FKBP12-2xFYVE and GM130N-mCherry-FKBP12) were stably expressed in THP-1 cells via lenti-virus package. Addition of rapamycin induces dimerization, leading to the formation of biomolecular condensates, followed by HT-DNA (2 μg/mL) stimulation for 6 hours to measure cGAS activity.
(b-c) THP-1 cells or CRISPR-mediated knockout lines for ZDHHC18 (b) and MARCH8 (c) were stably transduced with eGFP-cGAS-FRB and various mCherry-FKBP12 fusion constructs (GM130N-mCherry-FKBP12 (b) and mCherry-FKBP12-2xFYVE (c)) to direct cGAS to specific organelles upon rapamycin treatment. Cells were treated with 100 nM rapamycin for 30 minutes to induce FRB-FKBP12 dimerization, followed by stimulation with 2 μg/mL HT-DNA for 6 hours. Intracellular cGAMP levels were quantified using a luciferase-based bioassay in THP-1 Lucia™ ISG cells (InvivoGen), and data are presented as relative fluorescence intensity. Data are presented as mean ± SEM; **** P<0.0001.
(d) Schematic of ZDHHC18 constructs used in (f): full-length ZDHHC18 and ΔIDR ZDHHC18 lacking the intrinsically disordered region.
(e) Schematic of MARCH8 constructs used in (g): full-length MARCH8 and ZD18IDR chimeric construct where MARCH8IDR is replaced with ZDHHC18IDR.
(f) Western blot analysis of ZDHHC18CRISPR and ZDHHC18 reconstituted cells showing the levels of palmitoylated (Palm) and input proteins in the presence and absence of HT-DNA.
(g) Cells were co-transfected with HA-ubiquitin, Flag-cGAS, and either MARCH8 variants. Immunoprecipitation (IP) using anti-Flag antibodies followed by immunoblotting (IB) with anti-HA antibodies was performed to detect ubiquitinated cGAS. Whole cell lysates (WCL) were also probed with anti-Flag and anti-HA antibodies to confirm expression levels.
(h and i) THP-1 Lucia™ ISG reporter cells were used to assess interferon pathway activity upon cytosolic DNA stimulation. CRISPR-Cas9-mediated knockout cell lines were generated for ZDHHC18 (h) or MARCH8 (i), and complemented with full-length, intrinsically disordered region–deleted (ΔIDR), or IDR-truncated constructs, as indicated. Cells were transfected with HT-DNA (2 μg/mL) or mock treated for 18 hours, and luciferase activity was measured as a readout of ISG pathway activation. Bar graphs show relative fluorescence intensity (mean ± SEM) from four biological replicates. Data are presented as mean ± SEM; **** P<0.0001. ** P<0.01.
