Functional chimeric mRNAs encode proteins in mammalian immunity

The diversity of proteins produced by cells is typically defined by the repertoire of individual genes encoded in the genome. This diversity can be amplified through pre-mRNA cis-splicing, where varied pairings of exons within a given transcript can be differentially fused. It is unclear whether mechanisms beyond cis-splicing exist to diversify the protein-coding capacity of mammalian cells. In unicellular and invertebrate organisms, the process of trans-splicing, whereby exons from distinct pre-mRNAs are fused to create hybrid proteins, has been reported. In trypanosomes and nematodes, a common pre-mRNA is trans-spliced to various other mRNAs to promote transcript stability and translation³; by contrast, in Drosophila, this process generates functionally diverse transcripts and proteins⁴. In healthy mammals, few examples of trans-splicing exist, and functional chimeric mRNAs (chRNAs) are typically associated with oncogenic transformation of cells where genomic translocations fuse disparate genes at the DNA level^5,6. Expression of some chRNAs known to be produced by DNA translocation have also been described in non-malignant tissue and are proposed to form through trans-splicing^7,8,9. However, it remains to be elucidated whether, in healthy mammalian cells, widespread fusion of mRNA from distinct genes can produce chRNAs that encode functional proteins. However, one can predict that such fusion events would expand the number of physiological protein-encoding mRNAs far beyond what is currently appreciated in our genomes.

Despite advances and widespread adoption of RNA sequencing (RNA-seq) methodologies, endogenously expressed chRNAs in mammals have largely evaded discovery because of technical limitations. cDNA synthesis-based approaches are used extensively and rely on viral reverse transcriptase enzymes to convert RNA to cDNA. However, viral reverse transcriptase enzymes can generate artificial fusion transcripts through template switching¹⁰. Furthermore, as mammalian chRNAs are not annotated in reference transcriptomes, candidate chRNAs multimap to different parts of the genome during alignment, resulting in their routine removal during standard RNA-seq analyses. Short-read sequencing also limits chRNA detection, as read fragments are assigned as chimeric only if they span the junction at which one gene meets the other; otherwise, these reads are assigned to the parent genes of the chRNA¹¹. Moreover, additional challenges arise from difficulties in accurately resolving multimapping reads that display short junction sequences¹² and/or map to multiple similar loci¹³. Notwithstanding these caveats, conventional RNA-seq datasets have consistently identified putative chRNAs in healthy human tissue^14,15. Whether these chRNAs arise artificially during library preparation, or whether they are physiologically relevant, remains an open question.

To identify chRNAs while avoiding artifacts from cDNA synthesis, we performed Oxford Nanopore PromethION direct RNA-seq analysis of polyadenylated RNA from steady-state, tissue-reparative and inflammatory mouse bone-marrow-derived macrophages (BMDMs; Fig. 1a). Ten biological replicates yielded 52.9 million passing reads (Extended Data Fig. 1a and Supplementary Table 1), of which over 90% mapped to the mouse genome (Extended Data Fig. 1b). Read identity (Extended Data Fig. 1c), average Phred scores (Extended Data Fig. 1d) and read lengths (Extended Data Fig. 1e) were consistent with high-quality published datasets, and marker-gene expression confirmed macrophage polarization states (Extended Data Fig. 1f).

The computational tools LongGF, JAFFAL and Genion^16,17,18 were benchmarked to detect BCR-ABL1 in K562 cells¹⁹ (Supplementary Table 2) and used to identify chRNAs containing annotated splice donor and acceptor sites from two distinct genes. This identified 30,390 candidates across activation states (Fig. 1b, Extended Data Fig. 1g and Supplementary Tables 3 and 4). These chRNAs had a median length of 1,681 nucleotides, compared with 2,190 nucleotides for their parent transcripts (Extended Data Fig. 1h). Parent genes were distributed throughout the genome (Fig. 1b), with around 88% of chRNA species arising from interchromosomal pairings (Fig. 1c) and about 4.5% and 7.5% from proximal (<10⁶ bases) and distal (≥10⁶ bases) intrachromosomal loci, respectively (Fig. 1c and Extended Data Fig. 1i), suggesting contributions from both readthrough transcription via cis-splicing and trans-splicing from distal loci.

Limited sequencing depth and modest agreement among the long-read detection tools (Extended Data Fig. 1g and Supplementary Table 1) precluded robust quantification, prompting orthogonal validation. We generated high-depth Illumina RNA-seq data (Supplementary Table 5) and analysed these data using six fusion-detection algorithms^{20,21,22,23,24}, producing an implausibly large candidate set (Supplementary Table 6) consistent with false positives generated through reverse-transcription and PCR template switching¹⁰. Nevertheless, comparison with the direct RNA catalogue identified over 250 chRNAs with matching junctions (Supplementary Table 7), including candidates associated with differential macrophage polarization (Fig. 1d). Although direct RNA-seq analysis predominantly detected interchromosomal chRNAs, co-detected chRNA species were largely intrachromosomal (Fig. 1e and Extended Data Fig. 1j), with proximal species detected more frequently and supported by more reads than distal or interchromosomal species (Extended Data Fig. 1k–m), consistent with the prevalence of local and readthrough transcription^25,26.

As short reads rarely span sufficient sequence on both sides of a chRNA junction, this approach can also generate false negatives for specific isoforms¹¹. We therefore designed tandem NanoString nCounter probes that require simultaneous hybridization across each RNA junction (Fig. 1a). Over 500 probe sets were designed to target chRNAs detected by all three long-read tools (Supplementary Table 8), alongside scrambled controls with the expectation of false negatives. More than 100 high-quality probes detected constitutive or polarization-regulated chRNAs (Fig. 1f and Supplementary Table 7). Some chRNA species were validated by both Illumina and NanoString, whereas others were unique to either platform (Fig. 1d,f and Supplementary Table 7); notably, NanoString validated a greater proportion of interchromosomal species (Fig. 1e,g).

PCR and Sanger sequencing independently validated chRNA junctions (Extended Data Fig. 2a–o and Supplementary Table 9), including a Gsdmd-Tmem106a amplicon spanning Gsdmd exons 1–2 fused to Tmem106a exons 6–9 (Fig. 1h–i). Gsdmd-Tmem106a was among the most inflammation-induced NanoString candidates (Fig. 1f), and both parent genes have established roles in lipopolysaccharide (LPS)-induced inflammation^2,27. The same exon–exon junction was detected in C57BL/6J, BALB/cJ and wild-derived mice (Fig. 1j–k). Analysis using quantitative PCR with reverse transcription (RT–qPCR) further confirmed LPS-inducible expression, peaking at 6 h in parallel with the parent genes (Fig. 1l–n). The junction and LPS responsiveness of Cd274-Lacc1 were similarly confirmed (Extended Data Fig. 2p–s). In vivo, Gsdmd-Tmem106a was detected and induced in lung macrophages after influenza A infection, brain macrophages/microglia during Escherichia coli meningitis and peritoneal macrophages after LPS administration (Fig. 1o–q).

Applying this pipeline to steady-state and inflammatory human monocyte-derived macrophages (Extended Data Fig. 3a–g and Supplementary Table 10), we identified over 900 chRNAs (Extended Data Fig. 3h–i and Supplementary Table 11), including inter- and intrachromosomal species (Extended Data Fig. 3j) with properties resembling those of mouse chRNAs (Extended Data Fig. 3k–l). qPCR further confirmed their detection and differential regulation by LPS (Extended Data Fig. 3m–o).

Despite lower sequencing depth in the human dataset (Extended Data Fig. 3b and Supplementary Table 10), cross-species analysis identified 33 chRNAs sharing parent genes between mice and humans (Extended Data Fig. 4a and Supplementary Table 12). Although we did not identify a GSDMD-TMEM106A chRNA in these cells, approximately 8 displayed at least 50% mRNA sequence homology and 28 encoded predicted proteins with at least 50% conservation (Extended Data Fig. 4a,b and Supplementary Table 12), accompanied by closely matched domain architectures (Extended Data Fig. 4c). These included HDAC8-CITED1 and TBC1D22B-RNF8, which retained functional domains from both parent genes (Extended Data Fig. 4d–e) and were validated by Sanger sequencing in mouse and human macrophages (Extended Data Fig. 4f–g).

Thus, integrating direct RNA-seq with short-read RNA-seq, NanoString, PCR and expression analyses provides a generalizable framework for identifying, validating and prioritizing endogenous chRNAs for functional studies in immunity and inflammation.

To distinguish chromosomal translocation from RNA-level fusion, we performed long-range genomic PCR analysis of DNA from untreated and LPS-stimulated BMDMs using primers positioned in the junction exons, including Gsdmd exon 2 and Tmem106a exon 6. Neither Gsdmd-Tmem106a nor Cd274-Lacc1 produced the expected kilobase-scale genomic amplicon under either condition (Extended Data Fig. 5a,b), arguing against detectable chromosomal translocations and supporting RNA-level fusion through trans-splicing.

As our chRNA catalogue was defined using annotated splice sites, we tested whether canonical splicing machinery contributes to chRNA formation. Inhibition of the SF3B spliceosome complex with pladienolide B^28,29 or knockdown of U1 small nuclear RNA abolished LPS-induced Gsdmd-Tmem106a expression (Fig. 2a,b). Inhibition of RNA polymerase II with actinomycin D³⁰ similarly eliminated Gsdmd-Tmem106a and Cd274-Lacc1, as well as their parent transcripts (Fig. 2c–e and Extended Data Fig. 5c–e), supporting the co-transcriptional joining of newly synthesized RNAs.

We next examined whether chRNA formation is facilitated by spatial proximity between parent genes. Genome-wide in situ Hi-C analysis revealed that interchromosomal parent–gene pairs identified by direct RNA-seq (Extended Data Fig. 1g) were significantly enriched for DNA–DNA interactions relative to randomized exon pairs (Fig. 2f). As CTCF regulates genome organization³¹, we depleted Ctcf using small interfering RNA (siRNA; Fig. 2g) and performed Hi-C after 6 h of LPS stimulation. LPS enhanced contacts between chRNA parent genes, whereas Ctcf depletion abolished this increase, demonstrating that these interactions are dynamically regulated and CTCF dependent (Fig. 2h,i).

Functionally, Ctcf depletion reduced LPS-induced Gsdmd-Tmem106a without affecting Gsdmd or Tmem106a expression (Fig. 2j), and similarly reduced Cd274-Lacc1 without altering either parent gene (Extended Data Fig. 5f). Finally, junction-specific primers (Fig. 2k) detected only the Gsdmd exon 2 and Tmem106a exon 6 junction isoform (Fig. 2l), demonstrating that fusion is highly selective rather than a random consequence of proximity. Together, these findings show that inflammation promotes CTCF-dependent interactions between chRNA parent genes, enabling specific trans-splicing and chRNA formation in macrophages.

To test chRNA protein-coding potential, we focused on Gsdmd-Tmem106a, which retains the Gsdmd 5′ untranslated region and canonical start codon. The Tmem106a-derived sequence was predicted to be translated out of frame, producing a novel C terminus that is absent in the annotated mouse proteome (Fig. 3a).

We generated Gsdmd-Tmem106a-MYC knock-in mice (G-T^MYC mice) carrying a C-terminal MYC tag within the chimeric ORF located in exon 6 of Tmem106a before the predicted chimeric stop codon (Fig. 3b and Extended Data Fig. 6a,b). As the Tmem106a-derived sequence would be translated out of frame, the tag is exclusive to GSDMD–TMEM106A. MYC immunoblotting detected an LPS-inducible protein in G-T^MYC BMDMs but not in wild-type BMDMs (Fig. 3c). A junction-specific siRNA selectively depleted Gsdmd-Tmem106a (Fig. 3d), while an independent siRNA targeting Gsdmd exon 2 (encoded on chromosome 15) was used to test whether this Gsdmd exon is required to produce the chromosome-11-encoded MYC tag. Both reduced the MYC signal, confirming its chimeric origin (Fig. 3e).

A custom antibody against the novel C-terminal peptide similarly detected an LPS-inducible protein diminished by both siRNAs (Fig. 3f). The detected endogenous GSDMD–TMEM106A migrated above its predicted molecular mass, whereas overexpressed GSDMD–TMEM106A with a C-terminal HA tag produced both the expected approximately 15 kDa protein and higher-molecular-mass species (Fig. 3g), suggesting multiple modified states.

Finally, we generated Gsdmd-Tmem106a^Stop (G-T^Stop) mice by introducing a single G-to-T substitution into Tmem106a exon 6 (Extended Data Fig. 6c). This created a stop codon after the third Tmem106a-derived amino acid without altering canonical GSDMD or TMEM106A translation (Fig. 3h). GSDMD–TMEM106A was absent from G-T^Stop BMDMs but robustly induced by LPS in wild-type cells (Fig. 3i), establishing that Gsdmd-Tmem106a encodes a stable, LPS-inducible protein.

To define GSDMD–TMEM106A function, we examined its contribution to inflammasome activity. Canonical GSDMD is cleaved by inflammatory caspases, releasing GSDMD-NT to oligomerize and form membrane pores that enable secretion of caspase-1 (CASP-1)-processed IL-1β and IL-18 (refs. ^1,2,32). After siRNA-mediated Gsdmd-Tmem106a knockdown (Fig. 3d), LPS- and nigericin-induced IL-1β and IL-18 release was significantly reduced (Fig. 4a and Extended Data Fig. 7a). LDH release after prolonged nigericin treatment was also reduced, although less markedly than early cytokine secretion (Fig. 4b), whereas non-inflammasome-dependent IL-6 secretion was unaffected (Extended Data Fig. 7b).

BMDMs from G-T^Stop mice similarly displayed reduced early IL-1β release but normal IL-6 secretion (Fig. 4c,d), with a modest reduction in late-stage cell lysis (Fig. 4e). Thus, both acute depletion and genetic loss of GSDMD–TMEM106A selectively impair inflammasome-dependent cytokine release and pyroptosis.

As IL-1β promotes antibacterial immunity^33,34 and GSDMD protects against Salmonella enterica subspecies enterica serovar Typhimurium (hereafter, S. Typhimurium)³⁵, we challenged G-T^Stop and wild-type mice with S. Typhimurium. G-T^Stop mice exhibited significantly greater bacterial burdens in the spleen and liver compared with their co-housed littermates (Fig. 4f,g). Conversely, in lethal LPS-induced sepsis, in which GSDMD contributes to disease severity³², G-T^Stop mice were protected from early hypothermia (Fig. 4h) and, whereas the majority of G-T^Stop mice survived for 96 h, all of the wild-type controls succumbed within 48 h (Fig. 4i).

Gain-of-function experiments produced reciprocal effects. Lentiviral GSDMD–TMEM106A overexpression increased IL-1β release after NLRP3 activation (Extended Data Fig. 7c). Similarly, transfection with Gsdmd-Tmem106a mRNA enhanced IL-1β secretion and accelerated LDH release, whereas mRNA encoding only the Gsdmd-derived region (Gsdmd-Tmem106a^ΔCT (G-T^ΔCT-encoding mRNA)) did not (Extended Data Fig. 7d–e), demonstrating that the Tmem106a-derived sequence is required. Although absolute IL-1β release differed between overexpression and loss-of-function approaches, the relative effects were consistent across models (Extended Data Fig. 7f).

We next administered LNP-encapsulated Gsdmd-Tmem106a, Gsdmd-Tmem106a^ΔCT or non-translating control mRNA 18 h before S. Typhimurium challenge. Gsdmd-Tmem106a administration significantly reduced liver bacterial burden at 24 h compared with both controls (Fig. 4j), demonstrating that exogenous GSDMD–TMEM106A enhances antibacterial defence. However, in LPS-induced sepsis, Gsdmd-Tmem106a administration accelerated hypothermia within 3 h compared with Gsdmd-Tmem106a^ΔCT and control mRNAs (Fig. 4k). As inflammasome cytokines and GSDMD activity promote sepsis lethality³², we measured serum cytokines 6 h after LPS. Gsdmd-Tmem106a specifically increased IL-1β (Extended Data Fig. 7g) without altering IL-6, IFNγ, TNF, CCL2 or IL-12p70 (Extended Data Fig. 7h–l). Nearly all Gsdmd-Tmem106a-treated mice subsequently reached the humane end points within 72 h, whereas Gsdmd-Tmem106a^ΔCT and control-treated mice were protected (Fig. 4l).

Finally, because Gsdmd^−/− mice resist lethal LPS sepsis³², we tested whether Gsdmd-Tmem106a requires endogenous GSDMD. Gsdmd-Tmem106a-induced sepsis lethality was completely abolished in Gsdmd^−/− mice (Fig. 4l), and Gsdmd-Tmem106a overexpression did not induce IL-1β release in Gsdmd^−/− immortalized BMDMs (iBMDMs) (Extended Data Fig. 7m). Thus, GSDMD–TMEM106A functions through canonical GSDMD to amplify IL-1β release, enhancing antibacterial defence while exacerbating sepsis immunopathology.

To position GSDMD–TMEM106A within the inflammasome pathway, we first examined priming and proteolytic activation. Neither siRNA-mediated depletion nor genetic loss in G-T^Stop macrophages altered LPS-induced Il1b, Il6, Nlrp3 or Gsdmd expression (Fig. 3d and Extended Data Fig. 8a–g). Gsdmd-Tmem106a loss also did not affect GSDMD or CASP-1 cleavage after LPS and nigericin treatment (Fig. 5a and Extended Data Fig. 8h), placing its activity downstream of inflammasome assembly and GSDMD cleavage.

As GSDMD–TMEM106A requires canonical GSDMD (Fig. 4l and Extended Data Fig. 7m), we tested whether the proteins function cooperatively. In Gsdmd^−/− iBMDMs expressing a fixed amount of GSDMD, increasing GSDMD–TMEM106A-encoding mRNA dose-dependently enhanced IL-1β release (Extended Data Fig. 7n). Although Gsdmd-Tmem106a represents less than 10% of parental Gsdmd expression (Extended Data Fig. 7o), supplementing GSDMD-encoding mRNA with 10% GSDMD–TMEM106A-encoding mRNA nearly doubled IL-1β release, whereas an equivalent increase in GSDMD had little effect (Fig. 5b). GSDMD–TMEM106A^ΔCT-encoding mRNA did not enhance secretion, demonstrating that the TMEM106A-derived C terminus is required for synergy with GSDMD-NT.

As GSDMD–TMEM106A lacks the inhibitory GSDMD C terminus, we investigated whether it associates with cellular membranes before inflammasome activation. Recombinant FLAG-tagged GSDMD–TMEM106A (Extended Data Fig. 9a–d) modestly bound to the inner-leaflet lipids phosphatidic acid and phosphatidylserine^36,37 as well as the mitochondrial lipid cardiolipin¹ (Extended Data Fig. 9e). Epitope-tagged GSDMD–TMEM106A localized to the plasma membrane without inflammasome activation (Extended Data Fig. 9f,g) and accumulated at membrane protrusions during pyroptosis, when cells become rounded and swollen³⁸ (Fig. 5c). Subcellular fractionation further showed that endogenous GSDMD–TMEM106A translocates to membranes during LPS priming (Fig. 5d).

Mutation of the GSDMD-derived membrane-anchor residues Phe50 and Trp51 (ref. ³⁹) in GSDMD–TMEM106A-encoding mRNA caused the GSDMD–TMEM106A^F50G/W51G–HA protein to remain cytoplasmic (Extended Data Fig. 9h). By contrast, the non-functional GSDMD–TMEM106A^ΔCT–HA protein still localized to membranes, showing that membrane targeting depends solely on the GSDMD-derived region (Extended Data Fig. 9i). Only full-length, membrane-localized GSDMD–TMEM106A enhanced IL-1β release and cell lysis (Fig. 5e and Extended Data Fig. 9j), indicating that membrane localization is necessary for activity.

To test whether membrane-localized GSDMD–TMEM106A interacts with GSDMD-NT, we performed co-immunoprecipitation after NLRP3 activation. Endogenous GSDMD-NT co-precipitated with GSDMD–TMEM106A–HA but not cytoplasmic GSDMD–TMEM106A^F50G/W51G–HA or the non-translating control (Fig. 5f), demonstrating a physical interaction restricted to the plasma membrane.

AlphaFold 3 predicted co-folding of GSDMD–TMEM106A with the GSDMD-NT pore, positioning its TMEM106A-derived region within the pore (Extended Data Fig. 10a). This region contains four residues conserved in GSDMD-NT acidic patches that are required for IL-1β release³⁹ (Extended Data Fig. 10b). Substitution of these residues (E72A, E74A, D89A and D92A) altered predicted co-folding (Extended Data Fig. 10c) and abolished GSDMD–TMEM106A-mediated enhancement of IL-1β release (Extended Data Fig. 10d), establishing their functional importance.

Endogenous GSDMD–TMEM106A loss slowed GSDMD-NT pore formation, as measured by uptake of propidium iodide (PI) (Fig. 5g–h, Extended Data Fig. 10e and Supplementary Videos 1 and 2), whereas overexpression accelerated it (Extended Data Fig. 10f). In cells with doxycycline-inducible GSDMD-NT⁴⁰, Gsdmd-Tmem106a, but not Gsdmd-Tmem106a^ΔCT or control mRNA, increased the rate of PI uptake (Fig. 5i,j), further demonstrating activity downstream of inflammasome activation. Finally, in cell-free liposome-permeabilization assays using purified proteins (Extended Data Fig. 9a–d), recombinant GSDMD–TMEM106A–FLAG, but not recombinant GSDMD–TMEM106A^ΔCT–FLAG, directly enhanced GSDMD-NT-mediated pore formation (Fig. 5k).

Together, these findings establish GSDMD–TMEM106A as a membrane-localized cofactor that physically and functionally cooperates with GSDMD-NT to accelerate pore formation, amplify early cytokine release and enhance subsequent pyroptosis.

Mammalian cells contain considerable genomic dark matter of functional importance. Over the past several decades, the constricting central tenets of what defines a eukaryotic protein-coding gene have slowly relaxed in lockstep with technological advancement, enabling the identification of novel protein-coding entities. Non-AUG start codons driving translation initiation, evidence of bicistronic translation and small ORF-encoded microproteins have now all been described in mammalian cells in diverse cellular processes ranging from immunity to muscle performance^41,42,43. Indeed, in addition to classical linear splicing of exons in a gene, backsplicing events have been identified that create circular mRNA with protein coding potential^44,45,46. Here we demonstrate that distinct genes distantly located throughout the genome can create chRNAs during normal physiological responses.

A central finding of this study is that chRNA biogenesis is a rapid, stimulus-responsive process rather than a consequence of genomic instability or splicing noise. Pro-inflammatory signals robustly induce chRNAs in macrophages, with formation occurring at annotated exon boundaries and requiring the canonical spliceosome. For a subset of chRNAs, parent genes are brought into proximity through interchromosomal DNA contacts, which are enhanced within 6 h of inflammation and depend on the genome-organizing protein CTCF³¹. chRNA formation is also highly precise: Gsdmd and Tmem106a, for example, consistently join at Gsdmd exon 2 and Tmem106a exon 6. The determinants of this specificity remain unclear but may involve repetitive elements analogous to those involved in circular RNA biogenesis⁴⁷, specialized trans-splicing or RNA-binding factors, nuclear condensates, and chromatin or RNA modifications. Notably, tumour-associated fusion transcripts such as JAZF1-JJAZ1 (ref. ⁷) and PAX3-FOXO1 (ref. ⁹) can also arise through physiological trans-splicing without chromosomal rearrangement. Together with the proposed RNA-before-DNA model⁴⁸ and evidence that three-dimensional genome organization influences translocation-partner selection^49,50, our findings raise the possibility that chRNAs mark loci that recurrently interact in nuclear space and may become susceptible to translocation during genotoxic DNA repair. Determining whether our chRNA catalogue is enriched for recurrent translocation partners, and directly visualizing how chromatin proximity is coordinated with pre-mRNA joining, will be important future directions.

Our findings support a model in which LPS-induced GSDMD–TMEM106A localizes to the plasma membrane before inflammasome-mediated GSDMD cleavage, creating a prepositioned landing zone that accelerates recruitment and assembly of liberated GSDMD-NT. Consistent with this model, GSDMD–TMEM106A physically interacts with cleaved GSDMD-NT at the membrane, and AlphaFold 3 predicts that its TMEM106A-derived region cooperatively folds with the membrane-inserting region of the GSDMD-NT oligomer. As GSDMD pores are thought to assemble from membrane-associated monomers⁴⁰, this interaction may stabilize only after GSDMD-NT oligomerization, potentially explaining its absence with the cytoplasmic GSDMD–TMEM106A^F50G/W51G–HA mutant. Analogous to Drosophila Myd88 and mammalian PIP2-binding adaptors that concentrate immune effectors at specific membrane domains⁵¹, binding of GSDMD–TMEM106A to phosphatidic acid and phosphatidylserine may enhance the spatial precision and kinetics of pore formation. Accordingly, its loss attenuates—but does not abolish—cytokine release and pyroptosis, indicating that it amplifies rather than enables canonical pore formation. Its additional binding to cardiolipin raises the possibility that it contributes to GSDMD-mediated mitochondrial damage⁵². Defining its interaction domains, accessory factors and pore architecture will require structural and biophysical approaches, particularly cryo-electron microscopy³⁹. Notably, endogenous GSDMD–TMEM106A migrates predominantly above its predicted molecular mass, whereas HA-tagged overexpression produces both higher- and expected-molecular-mass species, suggesting distinct modified, oligomeric and monomeric pools. As the retained GSDMD region contains known sites of phosphorylation⁵³, oxidation^40,54, succination⁵⁵ and ubiquitination^56,57, defining these biochemical states and their effects on GSDMD–TMEM106A activity represents an important future direction.

Our study reveals a hidden layer of mammalian gene regulation in which spatial genome organization and regulated splicing unite distant genes to generate protein-coding chRNAs. Their potential conservation in mouse and human macrophages suggests that transcript fusion is an underappreciated mechanism for expanding transcriptomic and proteomic diversity. By encoding proteins with novel domain architectures and functions, chRNAs may also broaden the pharmacologically targetable proteome. Their physiological prevalence, including potential fusions involving multiple parent genes and cell type- or context-specific networks, remains largely unclear. Our study provides a generalizable framework for chRNA discovery, validation and manipulation, demonstrating that chRNA-specific interference, mRNA delivery and gene editing can modulate their activity in cells and in vivo. Together, these findings establish chRNAs as a therapeutically actionable layer of mammalian gene regulation that expands the functional proteome and opens a landscape for biological discovery and therapeutic intervention.

Long-read direct RNA-seq

RNA libraries were prepared using the Direct RNA Sequencing kit (SQK-RNA002). Sequencing was performed on the Oxford Nanopore PromethION P24 sequencer using the FLO-PRO002 flow cell. Sequencing was performed using MinKNOW v.22.03.4 and MinKNOW Core v.5.0.0 with a pore scan frequency of 1.5 h, active channel selection set to on and reserved pores set to on. Libraries were run for 72 h. Basecalling was performed using Bream v.7.0.9 and Guppy v.6.0.7. Data processing was performed on an Ubuntu 20.04 system. Sequencing was performed by the BioMicro Center at Massachusetts Institute of Technology.

Computational analysis of long-read direct RNA-seq data

Individual passing FASTQ files from the same sample were merged to generate a single FASTQ file per sample. Except in the case of JAFFAL¹⁷ v.2.2 and v.2.3, all mouse data analyses were performed using reference files from GENCODE⁵⁸ release M28 (GRCm39), and all human data analyses were performed using reference files from GENCODE Release 43 (GRCh38.p13). For JAFFAL, matched reference files were downloaded from the UCSC genome browser⁵⁹ and prepared using the instructions provided on the JAFFAL wiki (https://github.com/Oshlack/JAFFA/wiki). Genomic alignments for NanoPlot⁶⁰ v.1.42.0 were performed using Minimap2⁶¹ v.2.24 with the parameters -ax splice -uf -k14 and aligned reads were processed and analysed using SAMtools⁶² v.1.15. Gene quantification was performed by aligning FASTQ files to the reference transcriptome with Minimap2 using the parameters -ax splice -uf -k14 and then quantifying these alignments with Salmon⁶³ v.1.10.2 using the command salmon quant --ont -t <in.fasta> -lU --numBootstraps 30 -a <in.bam> -o <Results>. Salmon data files were imported into R using tximport⁶⁴ v.1.38.2 with counts scaled using the lengthScaledTPM method. Differential gene expression analysis was performed using DESeq2⁶⁵ v.1.50.2 with cooksCutoff=TRUE using the resulting Salmon quant files. log₂(transcripts per million (TPM) + 1) summary gene quantification data were derived directly from Salmon TPM abundance estimates for genes with expression > 0 in at least one sample.

We created a dedicated pipeline for chimeric RNA analysis called TYPHON, which is available at GitHub (https://github.com/erenada/TYPHON) and reimplements the manually computed chimeric RNA analysis performed in this study using Python modules as part of an integrated, modular workflow. TYPHON provides the ability for users to easily install and run the computational tools and commands used for chimeric RNA analysis in this paper. Please see the TYPHON GitHub page for more details and instructions on running TYPHON.

Chimeric RNA identification using LongGF¹⁶ v.0.1.2, JAFFAL¹⁷ and Genion¹⁸ v.1.2.3. Genomic alignments for LongGF were performed with Minimap2 using the parameters -ax splice -uf -k14–secondary=no -G 50k and processed using SAMtools. LongGF was run on name-sorted .bam files using the command LongGF <bam_file_in> <gtf_file_in> 100 50 100 2 0 1 0 > <Results.log>. JAFFAL was run with MIN_LOW_SPANNING_READS=1 in the make_final_table.R script using the previously generated JAFFAL-specific reference files. As Genion expects reference files to be formatted in the format of Ensembl⁶⁶ reference files, GENCODE reference files were reformatted to be compatible with Genion using a reference script from 10x Genomics⁶⁷ (https://support.10xgenomics.com/single-cell-gene-expression/software/release-notes/build), the command sed ‘s/^chrM/MT/;s/^chrX/X/;s/^chrY/Y/;s/^chr//’ and pygtftk⁶⁸ v.1.6.2 using the command gtftk convert_ensembl -i <input_gtf_file> -o <output_modified_gtf_file>. For running Genion, we patched the tool to be able to output the read IDs for all chimeric RNA detected. This was done by modifying line 1037 of the annotate.cpp file (located in ./genion/src from: “bool full_debug_output = false;” to “bool full_debug_output = true;” and line 1195 from “outfile ≪ fusion_id ≪ “\t” ≪ cand.second.forward.size() ≪ “\t”” to “outfile ≪ fusion_id ≪ “\t” ≪ read_id ≪ “\t” ≪ cand.second.forward.size() ≪ “\t””.). In TYPHON, we enabled full debug output and modified the Genion annotation step to emit one line per supporting read with the Read_ID appended as the final column. Genion was then compiled with this modified annotate.cpp file according to the author’s instructions, and this modified Genion installation was used to run Genion for all chimeric RNA analysis except the analysis of K562 cell data for BCR-ABL1 benchmarking^19,69, which was run using the standard Genion installation without any modifications. The selfalign.tsv file required for running Genion was generated by aligning the reference transcriptome fasta file against itself using Minimap2 with the parameters -X -c and the resulting selfalign.paf file was processed to tsv format using pafr v.0.0.2 and tidyverse v.2.0.0. Input .paf files for running each sample through Genion were generated from the same alignment files previously used for running LongGF using the Minimap2 command paftools.js sam2paf <input_sam_file> > <output_paf_file>. Genion was run using the parameters --min-support 1 and with a blank genomicSuperDups.txt file, as an updated version of the genomicSuperDups.txt file for GRCm39 was not available at the time of analysis (see the Genion GitHub page (https://github.com/vpc-ccg/genion) for further details). To keep mouse and human data analyses consistent, a blank genomicSuperDups.txt file was also used for running Genion on human long-read direct RNA-seq data.

For compiling results from LongGF, JAFFAL and Genion, all chimeras in the following results files were considered: the log results files from LongGF, the summary results files from JAFFAL, and the result.tsv files from Genion. For BCR-ABL1 benchmarking with Genion, results.tsv and .fail files were considered. Results from LongGF, JAFFAL and Genion were combined and made distinct by Read ID using tidyverse. Chimeric RNA fasta files were prepared from alignment files using SAMtools and SeqKit⁷⁰ v.2.4.0. Chimeric RNA reads were filtered using BLAST+ (blast)⁷¹ v.2.13.0 by generating blast reference files using the reference transcriptome and makeblastdb with the parameters -in <reference.fasta.fa> -parse_seqids -blastdb_version 5 -title <reference_title> -dbtype nucl -out <db_name> command, and then running blastn with the parameters -query <fasta_containing_chimeric_RNA_sequences> -db <db_name> -outfmt 6 -out <output_results.txt> on a single fasta file containing chimeric RNA reads. Processing of blast results and generation of exon-repaired chimeric RNA reads was performed using SeqKit, BEDOPS⁷² v.2.4.41, BEDTools⁷³ v.2.31.0, Clustalo⁷⁴ v.1.2.4, Clustalw⁷⁵ v.2.1, Biopython⁷⁶ v.1.81, Seqtk v.1.4, tidyverse and data.table v.1.15.2. In brief, chimeric RNAs that did not map to at least one isoform of each chimeric RNA parent gene were removed. Mapping to isoforms containing retained introns was disallowed. BLAST mappings were ranked by descending bit.score and the BLAST mapping with the highest bit.score was selected for each chimeric RNA parent gene. Blast mappings for each chimeric RNA were ordered according to BLAST mapping order in the original chimeric RNA read. The genomic coordinates for every exon present in each selected chimeric RNA isoform were imported from the reference GTF file using BEDTools and the breakpoint exons for each chimeric RNA transcript were estimated by calculating cumulative exon length on a per-transcript basis and comparing against the length of each chimeric RNA segment provided by BLAST. This method enabled the calculation of which exons were expected to be present in each repaired chimeric RNA transcript. Selected exon coordinates were then ordered and added to bed files. Corresponding exonic sequences were extracted from the reference genome and fasta files containing the exon-repaired reads for each chimeric RNA were generated.

Conservation analysis between chimeric RNA present in mice and humans was performed by conducting predicted chimeric protein analysis, constructing BLAST and BLASTp references using human chimeric RNA sequences and predicted chimeric protein sequences, and then running BLAST or BLASTp on mouse chimeric RNA transcripts and predicted chimeric protein sequences against the computed human chimeric RNA and chimeric protein databases. Chimeric RNA peptide prediction was performed using orfipy⁷⁷ v.0.0.4 using the parameters <chimeric_RNA_sequence_fasta.fa> --min 90 --max 1000000000 --procs 1 --strand f --start ATG --stop TAA,TAG,TGA --table 1 --outdir <out_directory> --pep <Predicted_peptide_fasta.fa>. The BLASTp reference construction command was makeblastdb with the parameters -in <reference_proteome.fasta> -parse_seqids -blastdb_version 5 -title <reference_title> -dbtype prot -out <db_name>. Nichenetr⁷⁸ v.2.2.0 was used to convert human to mouse gene symbols using the convert_human_to_mouse_symbols R function, and mouse and human transcript BLAST matches were required to be derived from parent gene orthologues in both species. BLASTp was run using the parameters -query <in_fasta> -db <db_name> -outfmt 6 -out <blast_result.txt>. The resulting BLAST and BLASTp results were ordered by decreasing bit.score (high to low) and alignment length (longer to shorter) and the top hit for each mouse chimeric RNA Read ID was selected. dplyr v.1.1.4, stringr v.1.5.1, data.table and rtracklayer⁷⁹ v.1.64.0 were used for processing of BLAST data. For applications downstream of exon repair, long-read RNA-seq-derived chimeric RNA chimera IDs were corrected to appear in biological order as determined using BLAST+ and exon repair, with coordinates and other information also reordered appropriately.

Short-read RNA-seq

RNA libraries were prepared using the KAPA mRNA HyperPrep Kit (Roche). Total-RNA samples were quantified using the Agilent 4200 TapeStation instrument, with the corresponding Agilent TapeStation RNA assay. The resulting RNA-integrity number (RIN) scores and concentrations were considered when qualifying samples to proceed. An average RIN score of >9 was recorded for all samples. Samples were normalized to 200 ng of input in 50 μl (4 ng μl⁻¹), and the mRNA was captured using oligo-dT beads as part of the KAPA mRNA HyperPrep workflow. cDNA synthesis, adapter ligation and amplification were conducted subsequently as part of the same workflow. After amplification, residual primers were eluted away using KAPA Pure Beads in a 0.63× SPRI-based cleanup. The resulting purified libraries were run on the Agilent 4200 Tapestation instrument, with the corresponding Agilent High Sensitivity D1000 ScreenTape assay to visualize the libraries and check that the size and concentrations of the libraries matched the expected product. qPCR using the KAPA Library Quantification kit, which uses primers complementary to the sequencing flowcell oligos, was run to confirm the functional concentration. Molarity values obtained from this assay were used to normalize all samples in equimolar ratio for one final pool. The pool was denatured and loaded onto the Illumina NovaSeq6000 instrument, with an S4 300-cycle kit to obtain paired-end 150 bp reads. The pool was loaded at 1.2 pM, with 5% PhiX spiked in as a sequencing control. Basecall files were demultiplexed through the Harvard BPF Genomics Core’s pipeline, and the resulting FASTQ files were used in subsequent analysis.

Computational analysis of short-read RNA-seq data

FASTQ files were trimmed using fastp v.0.23.2 with the default parameters and trimmed FASTQ files were used as input for short-read chimeric RNA detection unless otherwise specified. GENCODE M28 reference files were used unless otherwise specified. Chimeric RNA were detected in short-read RNA-seq data using n = 6 different computational tools, Arriba, FusionCatcher, STAR-Fusion, STAR-SEQR, JAFFA and Pizzly^{20,21,22,23,24}. Arriba v.2.4.0 was run by first running star v.2.7.11b with the parameters --outSAMattributes Standard --outSAMtype BAM Unsorted --outSAMunmapped Within --outFilterMultimapNmax 50 --peOverlapNbasesMin 10 --alignSplicedMateMapLminOverLmate 0.5 --alignSJstitchMismatchNmax 5 -1 5 5 --chimSegmentMin 10 --chimOutType WithinBAM HardClip --chimJunctionOverhangMin 10 --chimScoreDropMax 30 --chimScoreJunctionNonGTAG 0 --chimScoreSeparation 1 --chimSegmentReadGapMax 3 --chimMultimapNmax 50, and then running Arriba with the default settings and a blank blacklist.txt file. Reference files for running Arriba were generated using STAR --runMode genomeGenerate --genomeFastaFiles <input_fasta> --sjdbOverhang 150. Results from Arriba were processed to remove ambiguous intergenic chimeric RNA reads. The reference for FusionCatcher v.1.33 was built on 13 March 2025 using Ensembl 113 mouse reference files using fusioncatcher-build -g mus_musculus. FusionCatcher was run on untrimmed FASTQ files using the default settings other than --limitSjdbInsertNsj 3000000 --limitOutSJcollapsed 3000000. STAR-Fusion references were generated by downloading the Mouse_GRCm39_M31_CTAT_lib_Nov092022.source.tar.gz archive from https://data.broadinstitute.org/Trinity/CTAT_RESOURCE_LIB/ and following the ‘Building a Custom Genome Resource Library for Fusion Detection’ instructions from https://github.com/TrinityCTAT/ctat-genome-lib-builder/wiki to build the reference with GENCODE M28 reference files. STAR-Fusion v.1.12.0 was run with the parameters --no_annotation_filter --min_FFPM 0 --min_sum_frags 1.

STAR-SEQR was run using STAR-SEQR v.0.6.7 by first generating reference files using STAR --runMode genomeGenerate --genomeFastaFiles --sjdbOverhang 150 --genomeSAsparseD 1 with GENCODE M28 reference files. Individual FASTQ from different lanes for the same sample were concatenated into single R1 and R2 files and the parameters -m 1 -vv --keep_mito were used. JAFFA v.2.3 was run using the same reference files as used for long-read RNA-seq analysis on sample-level concatenated single R1 and R2 FASTQ files using the parameters <JAFFA-version-2.3/tools/bin/bpipe> run <JAFFA-version-2.3/JAFFA_direct.groovy>. pizzly v.0.37.3 was run on untrimmed fastq files by first running kallisto⁸⁰ v.0.48.0. The kallisto reference was generated by modifying the GENCODE M28 transcriptome reference using zcat </path/transcripts.fa.gz> | tr '|' ' ' | gzip -1> </path/transcripts.fixed.fa.gz> and building the reference using kallisto index -i </path/index.idx> </path/transcripts.fixed.fa.gz>. Kallisto was then run using kallisto quant with the --fusion parameter and pizzly was run using pizzly with the parameters -k 31 --align-score 2.

For calculation of junction-level overlap between chimeric RNA from long-read and short-read RNA-seq data, coordinates provided by Arriba, FusionCatcher, STAR-Fusion, STAR-SEQR and JAFFA were used and compared against long-read coordinates derived from exon-repair (that is, breakpoint exon coordinates). For FusionCatcher, coordinates were first translated from Ensembl 113 to GENCODE M28 coordinates by mapping the FusionCatcher-provided gene symbol to the corresponding GENCODE M28 gene symbol using Mouse Genome Informatics (MGI) (https://www.informatics.jax.org/) as an intermediate with the scCustomize⁸¹ v.3.2.0 Updated_MGI_Symbols function. Then, using exon breakpoints, the closest (in terms of absolute nucleotide distance) matching exon in the Ensembl reference was calculated against GENCODE coordinates to determine GENCODE M28 coordinates for each chimeric RNA found using FusionCatcher. These inferred GENCODE M28 coordinates were then used for junction-level overlap calculation. For Pizzly, which provides transcriptomic information, a strategy similar to that used for exon-repair was used to infer coordinate breakpoints for junction-level overlap. The transcript breakpoints in the annotated Pizzly chimeric RNA parent transcripts were compared against matched cumulative exon length metrics for GENCODE M28 transcripts and breakpoint exons were inferred based on minimum absolute nucleotide distance. An absolute nucleotide distance of ≤10 nucleotides between the short-read RNA-seq chimeric RNA gene A and gene B breakpoints and the long-read gene A and gene B breakpoints was required for a chimeric RNA to be considered to have the same breakpoint junction. For breakpoint junction calculation, long-read RNA-seq-derived chimeric RNA chimera IDs were corrected to appear in biological order as determined using BLAST+ and exon repair, with coordinates and other information also reordered appropriately. The chimera IDs from chimeric RNA derived from short-read RNA-seq were not reordered, as these chimeric RNA were not processed with BLAST/exon-repair.

FFPM-like metrics for each short-read RNA-seq tool were calculated using the broad strategy of FFPM_metric = (split_reads1 + split_reads2 + discordant_mates)/(total_mapped_reads/1,000,000) on a per tool basis (as non-identical methodologies and outputs are used/provided by each chimeric RNA detection tool). Notably, the detection of FFPM-like metrics for fusion genes, as well as any downstream differential analysis, is difficult and estimative by nature, and was used in this study for the purpose of selecting potentially interesting downstream candidates in tandem with long-read RNA-seq co-identification; metrics derived using this methodology should be considered an estimate and should not be interpreted in the same manner as standard gene expression and differential analysis metrics. For Arriba, the sum of split_reads1 + split_reads2 + discordant_mates was calculated on a chimera ID level with total mapped reads being derived from STAR-generated bam files using samtools view -@ 30 -c -F 1028. For FusionCatcher, Spanning_unique_reads was used in place of (split_reads1 + split_reads2 + discordant_mates) and the number for total mapped reads was derived from the ‘# reads with at least one reported alignment’ in the info.txt files provided by FusionCatcher. Metrics for STAR-Fusion were calculated using JunctionReadCount + SpanningFragCount in place of (split_reads1 + split_reads2 + discordant_mates) using the same total mapped reads counts as used for Arriba. For STAR-SEQR, the total mapped reads counts were extracted from the STAR-SEQR Log.final.out log files as the ‘Uniquely mapped reads number’ number, and metrics were calculated using NREAD_SPANS + NREAD_JXNLEFT + NREAD_JXNRIGHT in place of (split_reads1 + split_reads2 + discordant_mates). For Pizzly, the total mapped reads counts were extracted from the run_info.json files n_processed numbers provided by Pizzly, and metrics were calculated using total_paircount + total_splitcount instead of (split_reads1 + split_reads2 + discordant_mates). Finally, for JAFFA, total mapped reads counts were calculated using the result of (seqkit stats -T -j 4 <fastq_file> | tail -n +2 | cut -f4)*2 with the <sample>_filtered_reads.fastq.1.gz file for each sample as the input, with spanning.pairs + spanning.reads used in place of (split_reads1 + split_reads2 + discordant_mates). All FFPM-like metric results were filtered to only apply to short-read chimeric RNA with breakpoint junctions overlapping long-read RNA-seq-derived chimeric RNA breakpoint junctions. FFPM calculations were performed by taking individually derived FFPM-like metrics for each tool and calculating log_ffpm <- log₂(total_filter + 0.01). Differential testing was performed using limma⁸². Calculation of relative enrichment between interchromosomal and distal/proximal intrachromosomal chRNA was performed using the summed counts originally derived for ffpm-like metrics. Counts were combined for individual read IDs corresponding to the same chRNA ID or for the same chRNA ID and junction in the case of density plots. Significance testing was performed using the R t.test function with the default settings.

NanoString nCounter analysis

NanoString nCounter analysis was conducted on purified, extracted RNA by the Genetics and Genomics Core at Boston Children’s Hospital in accordance with standard NanoString protocols using a NanoString nCounter Max System. Analysis of NanoString RCC files from BMDMs was performed using the NanoTube⁸³ v.1.6.0 package using nSolver normalization, a background proportion of 0.33 and a background threshold of 2. The normalized count matrix of passing chimeric RNA expression data was converted to log₂ format using the formula log₂(x + 0.5), where x is defined as the normalized count matrix. For BMDM data, chimeric RNAs that were expressed below the background, defined as a value of −1 in the log₂ normalized matrix, in less than 75% of replicates of any one polarization condition were removed from the analysis. Differential expression analysis was performed using the runLimmaAnalysis function. For analysis of Gsdmd-Tmem106a expression in vivo, the nSolver analytical pipeline was followed using standard settings using nSolver version ≤4.0.

Hi-C processing

The Hi-C libraries were processed as previously described with changes in parameters⁸⁴. Libraries were aligned with the presplit_map.py script in the diffHic package (v.1.38.0)⁸⁵. Reads were split into 5′ and 3′ segments if they contained the MboI ligation signature (GATCGATC), in cutadapt (v.4.8)⁸⁶ with the default parameters. Reads were aligned to the mm39 build of the mouse genome with bowtie2 (v.2.5.3)⁸⁷ in single-end mode. The FixMateInformation command from the Picard suite v.1.117 (https://broadinstitute.github.io/picard/) was applied to synchronize mate information for each read pair. Duplicates were marked with MarkDuplicates from Picard and sorted by name with SAMtools. Each BAM file was processed to identify alignments for each read to a specific MboI restriction fragment with the preparePairs function in diffHic. Duplicate reads and those with mapping quality scores below 10 were discarded. Read pairs were determined to be dangling ends and removed if the pairs of inward-facing reads or outward-facing reads on the same chromosome were separated by less than 1,000 bp for inward-facing and outward-facing reads. Read pairs with fragment sizes above 1,000 bp were removed. Biological replicates were summed with the mergePairs function. HDF files of read pairs were converted to the Juicer format .hic with juicer tools (v.1.22.01)⁸⁸.

APA

Aggregate peak analysis (APA) was performed using the R package GENOVA⁸⁹ v.1.0.1. Files were loaded using the load_contacts function with resolution 500 kb and balancing equal TRUE. A total of 1,047 interchromosomal chRNA fusions found in BMDMs and co-detected by LongGF, JAFFAL and Genion were used for APA. Two controls were created for comparisons. For randomized chRNA fusion partners, 10,000 random combinations of the interchromosomal fusion regions were created, intrachromosomal combinations were removed. For random exons, 10,000 exon combinations were created by sampling from all exons in the genome; the inbuilt mm39 annotation from Rsubread⁹⁰ version ≤2.26.0 was used to define exons and intrachromosomal combinations were removed. APA plots were created using the APA function with size_bin=11 and dis_thres = c(500000, Inf) and all plotted on the same scale. APA was quantified with the quantify function and the fold-change value used for each interaction (the ratio of the foreground and background). Fold-change values in cases in which the background was 0 were removed for all samples. P values were calculated using a paired t-test of the log₂(0.5 + fold change) values.

Mice

All mouse experiments were carried out in compliance with Harvard Medical School Animal Care and Use Committee protocols. For most experiments unless otherwise specified in the text, experiments were conducted using male and female, 8–12-week-old C57BL/6J (000664) mice from Jackson Laboratory in an unblinded manner. BALB/cJ (000651) and wild-derived WSB/EiJ (001145) mice were also procured from Jackson Labs. Gsdmd-knockout mice were gifted by I. Chiu.

To generate G-T^MYC mice, a ssDNA donor oligonucleotide containing a flexible linker sequence followed by a MYC epitope tag sequence flanked with homology arms was provided to facilitate homology-directed repair insertion of the sequence into the C terminus of the Gsdmd-Tmem106a open reading frame. In brief, a guide RNA (GAGGGACGTTACCTGCTCAC) and ssDNA Ultramer (IDT) donor template (GGCCCATTGGCCAGGGTGGTTCTGGTGGTGGTTCTGGTGGTGGTTCTGGTGAACAAAAACTCATCTCAGAAGAGGATCTGTGAGCAGGTAACGTCCCTCTTCCTGGCCAGCCCTCTACACAGGCATGTAATGTGGTGACGGAGAGGGGCCAGCCTGAGGT) (bold font represents the knock-in cassette; non-bold represents homology arms) were designed targeting the stop codon locus of Gsdmd-Tmem106a in exon 6 of Tmem106a. The sequence of the knock-in mouse was confirmed using Sanger sequencing of the genomic loci. The forward primer (TTTCTGGTGCATCTGAAGACA) and reverse primer (GAGACCAAAGGGCACTCACT) were designed to amplify the mutation site. Knock-in alleles were identified as 496 bp PCR products and wild-type alleles were identified as 430 bp PCR products by gel electrophoresis.

To generate G-T^Stop mice, a ssDNA donor oligonucleotide containing a point mutation with homology arms was provided to facilitate homology-directed repair insertion of the sequence into exon 6 of Tmem106a. Two guide RNAs (AGTGACACAGCTGACGGCCG; AGATGTTCAGGACATTCTGG) and ssDNA Ultramer (IDT) donor template (TCTTTTTTAAAAAAGGTCTGAAAAAAATTTTAAATTATGCTTTTCCTTGCTGTCAACTCCAGAATGTCCTTAACATCTTCAACAGCAACTTCTATCCCATCACAGTGACACAGCTGACGGCCGAAGTGCTCCACCAGGCCTCT) were designed targeting exon 6 of Tmem106a. The sequence of knock-in mice was confirmed using Sanger sequencing of the genomic loci. The forward primer (ACACCCTCTTTCTGGTGCAT) and reverse primer (TACCTCAGCAGCCCTTCAGT) were designed to amplify the mutation site.

Mouse BMDM generation and polarization

BMDMs were generated from progenitor cells isolated from the femurs and tibias of C57BL/6J (obtained from the Jackson Laboratory), G-T^Stop or G-T^MYC mice and maintained in macrophage-colony stimulating factor (M-CSF) (50 ng ml⁻¹) for 7 days. Medium containing M-CSF was refreshed after 4 days. To induce an inflammatory phenotype, cells were stimulated with LPS (100 ng ml⁻¹) (serotype O55:B5; Enzo, ALX-581-013-L001) and recombinant mouse IFNγ (20 ng ml⁻¹) (PeproTech, 315-05) for 24 h. To induce a tissue reparative phenotype, cells were stimulated with recombinant mouse IL-4 (20 ng ml⁻¹) (PeproTech, 214-14) and mouse recombinant IL-13 (20 ng ml⁻¹) (PeproTech, 210-13) for 24 h.

Human MDM generation and polarization

MDMs were generated from CD14⁺ cells. For Oxford Nanopore Sequencing, anonymous peripheral blood samples were collected from the Mississippi Valley Regional Blood Center as waste cellular products and CD14⁺ cells were isolated using the MojoSort Human CD14 Selection Kit (BioLegend, 480026). For PCR and Sanger sequencing, anonymous cryopreserved CD14⁺ monocytes were purchased from Lonza (2W-400C). All monocytes were stimulated with M-CSF (50 ng ml⁻¹) (BioLegend, 574806) and granulocyte-macrophage-colony stimulating factor (GM-CSF) (50 ng ml⁻¹) (BioLegend, 572904) in RPMI 1640 containing 10% FBS for 7 days. To induce an inflammatory phenotype, cells were stimulated with LPS (100 ng ml⁻¹) (serotype O55:B5; Enzo, ALX-581-013-L001) and recombinant IFNγ (20 ng ml⁻¹) (BioLegend, 570204) for 24 h.

DNA and RNA extraction, Sanger sequencing and PCR

RNA and DNA were extracted from BMDMs using Trizol according to the manufacturer’s protocol. cDNA synthesis for PCR and RT–qPCR was performed using Maxima Reverse Transcriptase (Thermo Fisher Scientific, EP0741). PCR and RT–qPCR for chRNA were performed using a forward primer in gene A and a reverse primer in gene B. The primers used for PCR and Sanger sequencing of chRNAs that were co-identified by all three long-read RNA-seq tools and have a parent gene with an annotated function in immunity are provided in Supplementary Table 9. PCR and RT–qPCR for CD44-FAM168B was performed using a forward primer against CD44 (CCCATCCCAGACGAAGACAG) and a reverse primer against FAM168B (GTACACGGCAGTCTGGTAGG). PCR and RT–qPCR for ITGAX-CD52 was performed using a forward primer against ITGAX (GGGATGCCGCCAAAATTCTC) and a reverse primer against CD52 (GCTGAGACGTGTCACCTCAA). PCR and Sanger sequencing for HDAC8-CITED1 in human cDNA and Hdac8-Cited1 in mouse cDNA were performed using a forward primer in HDAC8 (human: TTATGACTGCCCAGCCACTG; mouse: TTGCGACGGAAATTTGACCG) and a reverse primer in CITED1 (human: TCCCGAGGAACTAGTGGGAG; mouse: TGCCTTGCGATCCTTCACTC). PCR and Sanger sequencing for TBC1D22B-RNF8 in human cDNA and Tbc1d22b-Rnf8 in mouse cDNA were performed using a forward primer in TBC1D22B (human: CGTCAACTTCTCTCCAGCCA; mouse: TGACATTCCGAGGACGAACC) and a reverse primer in RNF8 (human: TCCGACAAATGGGGCATTCT; mouse: CCTTCACGTCTGAGCTTAGGTC). PCR was performed using a ProFlex 3×32-well PCR system from Applied Biosystems. For long-range genomic PCRs, Phusion high fidelity DNA polymerase was used (New England Biolabs, M0530S) and the extension time was increased to amplify products up to 10 kb. Sanger sequencing was performed on a 1% agarose-gel-purified PCR product using a forward primer. Gel purification was conducted using the Monarch Spin DNA Gel Extraction kit (New England Biolabs, T1120L). RT–qPCR was performed using Sybr Green (Thermo Fisher Scientific, S7567) on the QuantStudio 5 Real-Time PCR System and all gene expression was assayed in duplicate.

LPS injection for NanoString peritoneal macrophage isolation

C57BL/6J mice (male, aged 8 weeks) from Jackson Laboratory were treated intraperitoneally with 100 μl PBS or LPS (E. coli strain O111:B4; InvivoGen, tlrl-3pelps) in PBS (5 mg kg⁻¹) for 16 h. Peritoneal lavage was collected in PBS containing 3% FBS and F4/80⁺ cells were isolated using MACS magnetic bead enrichment with anti-F4/80 MicroBeads (Miltenyi Biotec, 130-110-443) according to the manufacturer’s protocol. F4/80⁺ cells from three mice within the same treatment group were pooled in Trizol and RNA was extracted for NanoString nCounter analysis.

Intracisternal infection and brain macrophage isolation

C57BL/6J mice (male, aged 8 weeks) from Jackson Laboratory were subjected to intracisternal injection of 5 μl PBS or 1.2 × 10³ CFU E. coli (strain C5; ATCC, 700973) in 5 μl PBS. Then, 18 h after treatment, the mice were euthanized and perfused for collection and homogenization of the hindbrain. Isolated cells were resuspended in HBSS and an enrichment for myeloid cells was performed by adding 90% Percoll and centrifuging at 500g for 60 min at 4 °C. CD11b⁺ cells were isolated using MACS magnetic bead enrichment with anti-CD11b MicroBeads (Miltenyi Biotec, 130-126-725) according to the manufacturer’s protocol. Cells were collected in Trizol and RNA was extracted for NanoString nCounter analysis.

Intranasal infection and pulmonary macrophage isolation

C57BL/6J mice (female, aged 8 weeks) from Jackson Laboratory were infected intranasally with 30 μl PBS or influenza A (strain PR8) for 7 days. In brief, after infection, mice were euthanized and the perfused lung was collected. Lung tissue was digested in 2 mg ml⁻¹ collagenase IV solution in PBS with DNase I (100 μg ml⁻¹) at 37 °C shaking for 30 min. Homogenized and filtered lung tissue was subjected to a 33%/66% Percoll gradient and centrifuged at 3,000 rpm for 30 min at room temperature. Immune cells in the interface were isolated for FACS. Live CD45⁺F4/80⁺CD64⁺ single cells were sorted and stored in Trizol for RNA extraction.

Pladienolide B splice inhibitor treatment

BMDMs were plated 1.5 × 10⁶ cells per well in a 6-well plate and co-treated with 100 ng ml⁻¹ LPS (Serotype O55:B5; Enzo, ALX-581-013-L001) and 100 nM pladienolide B (PlaB) (Santa Cruz Biotechnology, sc-391691) or 0.1% DMSO (Sigma-Aldrich, D2438) for 6 h. Cells were collected in 1 ml of Trizol per well for RNA extraction.

Actinomycin D transcription inhibitor treatment

BMDMs were plated 1.5 × 10⁶ cells per well in a 6-well plate and co-treated with 100 ng ml⁻¹ LPS (Serotype O55:B5; Enzo, ALX-581-013-L001) and 5 μg ml⁻¹ actinomycin D (Gibco, 11805017) or 0.1% DMSO (Sigma-Aldrich, D2438) for 3 h. Cells were collected in 1 ml of Trizol per well for RNA extraction.

siRNA transfection for knockdown

Two siRNA were designed and pooled to target the junction sequence of Gsdmd-Tmem106a (CAGAACCAGAAUGUCCUGAACAU; AGAACCAGAAUGUCCUGAACA). Scrambled target sequences were generated for scrambled control siRNA design (AUAGCGAUACCAACAUAUCGACG; GAUUAACACGAACUAGCGAAC). siRNAs targeting Ctcf (CGAUCAGUUUCUGUCCUGAAAG), Gsdmd (GUGGUCAAGAAUGUGAUCAAGG) and Rnu1a1 (GGAGAUACCAUGAUCACGAAGGUGG) were additionally designed and synthesized by IDT. Negative control siRNA (siCtrl) was purchased from IDT (51-01-14-03). BMDMs were plated at a density of 2 × 10⁵ cells per well in a 24-well-plate for cytokine secretion and LDH assays, and 1 × 10⁶ cells per well in a 6-well-plate for all RT–qPCR and western blot analyses. Once cells adhered (over 12 h after plating), siRNA was transfected in Lipofectamine RNAiMAX (Thermo Fisher Scientific, 13778100) according to the manufacturer’s protocol. Assays examining the effects of siG-T and siGsdmd were performed 48 h after transfection. Assays examining the effects of siCtcf and siRnu1a1 were conducted 24 h after transfection.

Endogenous GSDMD–TMEM106A western blots

BMDMs were generated from homozygous G-T^MYC and wild-type littermate control mice. Cells were plated at 8 × 10⁶ and left untreated or stimulated with LPS (100 ng ml⁻¹) (serotype O55:B5; Enzo, ALX-581-013-L001) for 6 or 24 h. After stimulation, whole-cell lysates were collected using NP40 lysis buffer containing 10% PMSF and 1× protease inhibitor and subjected to western blot using anti-MYC antibody (9B11) (Cell Signaling Technology, 2276S) (1:1,000) and anti-mouse HRP-linked secondary antibody (Cell Signaling Technology, 7076S) (1:3,000). Anti-β-Actin antibody (13E5) (Cell Signaling Technology, 4970, 1:1,000) was used as the loading control for whole-cell lysates.

Production and purification of a custom antibody against GSDMD–TMEM106A were performed by Thermo Fisher Scientific. In brief, rabbits were immunized with peptide antigen CGRPGHQQPPPAHRPIGQ, including an initial injection and three booster injections over the course of 56 days. Antibody was affinity purified from serum and titred by indirect ELISA against our target peptide. For detection of endogenous GSDMD–TMEM106A in wild-type and G-T^Stop mice, BMDMs were left untreated or were treated with LPS (100 ng ml⁻¹) (serotype O55:B5; Enzo, ALX-581-013-L001) for 6 h. Whole-cell lysates were collected using NP40 lysis buffer containing 10% PMSF and 1× protease inhibitor and processed for western blotting using anti-GSDMD–TMEM106A custom antibody (1:500) and anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). For detection of endogenous GSDMD–TMEM106A in different cellular fractions of wild-type and G-T^Stop mice, 5 × 10⁶ BMDMs were collected after no stimulation, after 6 hr LPS (10 ng ml⁻¹) treatment alone, or after LPS prime and nigericin treatment (10 μM) for 15 min. Subcellular fractionation was conducted using the Thermo Fisher Scientific Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Fisher Scientific, 78840) according to the manufacturer’s instructions. Western blotting was performed using anti-GSDMD–TMEM106A custom antibody (1:500) and anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). Anti-α-tubulin antibody (DM1A, Cell Signaling Technology, 3873, 1:1,000) was used as subcellular fractionation control for cytoplasmic protein isolation. Anti-Na,K-ATPase antibody (Cell Signaling Technology, 3010, 1:1,000) was used as subcellular fractionation control for membrane protein isolation.

All western blots were conducted using 4–12% NuPAGE Bis-Tris gels (Thermo Fisher Scientific, NP0321BOX) according to the manufacturer’s protocol for the XCell SureLock Mini-Cell system and transferred using the XCell II Blot Module (Thermo Fisher Scientific, EI9051).

mRNA transfection for overexpression

Codon optimized mRNA expressing the amino acid sequence of interest was designed and synthesized by Moderna. Gsdmd-Tmem106a mRNA was designed to encode the wild-type GSDMD–TMEM106A sequence, while Gsdmd-Tmem106a^ΔCT mRNA encodes the same truncated sequence expressed by G-T^Stop mice, whereby a stop codon is placed just three amino acids into the Tmem106a-encoded C terminus. mRNA was transfected into BMDMs using MessengerMax lipofectamine (Thermo Fisher Scientific, LMRNA001) for 24 h before the assay. Single mRNA transfections for measuring IL-1β secretion were performed in 24-well plates in which cells were plated at 2E5 cells-per-well in 500 μl medium and treated with 500 ng mRNA, according to the MessengerMax lipofectamine protocol. Transfections of mRNA encoding GSDMD and GSDMD–TMEM106A or GSDMD–TMEM106A^ΔCT encoding mRNA together were performed in 96-well plates in which cells were plated at 4 × 10⁴ cells per well in 100 μl medium and treated with 120 ng mRNA. In experiments using different doses of GSDMD–TMEM106A, non-translating control mRNA was used to normalize total mRNA input across samples.

Subcellular fractionation was conducted using a Thermo Fisher Scientific Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Fisher Scientific, 78840) according to the manufacturer’s instructions. Western blotting for HA was performed using anti-HA antibody (C29F4) (Cell Signaling Technology, 3724S, 1:1,000) and anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). Anti-α-tubulin antibody (DM1A) (Cell Signaling Technology, 3873, 1:1,000) was used as the subcellular fractionation control for cytoplasmic protein isolation. Anti-Na,K-ATPase antibody (Cell Signaling Technology, 3010, 1:1,000) was used as subcellular fractionation control for membrane protein isolation.

BMDMs transfected with mRNA expressing GSDMD–TMEM106A–HA (G–T–HA mRNA) or a non-translating control mRNA were observed for GSDMD–TMEM106A localization under a Nikon Ti inverted microscope using plan apo ×60 oil-immersion objective lens. Wheat germ agglutinin (WGA) with Alexa fluor 488 conjugate (Thermo Fisher Scientific, W11261) was used to stain the plasma membrane before cell fixation, using 2% PFA for 10 min, and cell permeabilization, using a PBS solution containing 2% BSA and 0.1% Triton X-100. Blocking was conducted using 2% BSA for 1 h at room temperature. HA-tagged GSDMD–TMEM106A was marked with an anti-HA antibody (16B12) conjugated to PE (1:500) (BioLegend, 901518) and visualized using a 561 nm laser. Membrane stain was visualized using a 488 nm laser.

Pyroptosis assay, including cytokine measurement, GSDMD and CASP-1 western blotting, LDH and PI uptake

IL-1β, IL-18 and IL-6 were measured in cell-free supernatants. R&D Systems ELISA kits were used according to the manufacturer’s instructions for IL-18 (R&D Systems, DY7625-05) and IL-6 (R&D Systems, DY406-05), and BioLegend kits were used for IL-1β measurements (BioLegend, 432616). IL-1β and IL-18 were measured in the supernatant after 6 h LPS (10 ng ml⁻¹ for siRNA and G-T^Stop experiments; 100 ng ml⁻¹ for lentivirus and mRNA transfection experiments; serotype O55:B5; Enzo, ALX-581-013-L001) prime alone and after addition of nigericin (10 μM) (InvivoGen, Tlrl-nig-5) for 30 min to activate the NLRP3 inflammasome. IL-6 secretion was measured after a 6 h LPS treatment alone. Absorbance was read using the Synergy HTX plate reader at 450 nm. The samples were diluted appropriately to fall within the standard range of 2,000 to 31.3 pg ml⁻¹ for IL-1β ELISAs and 250 to 7.8 pg ml⁻¹ for IL-6 ELISAs. Gsdmd-knockout iBMDMs were gifted from the laboratory of J. Kagan and were not tested for mycoplasma contamination.

Western blots for GSDMD and CASP-1 were performed on BMDM cell lysates left at steady state, after 6 h LPS (10 ng ml⁻¹) alone, or after LPS prime and nigericin treatment (10 μM) for 15 min. Whole-cell lysates were collected using NP40 lysis buffer containing 10% PMSF and 1× protease inhibitor. Western blots for GSDMD and CASP-1 were performed using anti-GSDMD antibody (E9S1X, Cell Signaling Technology, 39754, 1:1,000) and mixed anti-CASP-1 (E9R2D, Cell Signaling Technology, 83383) and anti-cleaved CASP-1 (E2G2I, Cell Signaling Technology, 89332) antibodies (1:1,000), respectively, and anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). Anti-β-actin antibody (13E5, Cell Signaling Technology, 4970, 1:1,000) was used as the loading control for whole-cell lysates.

LDH was measured in cell-free supernatants after 6 h LPS prime (10 ng ml⁻¹ for siRNA and G-T^Stop experiments; 100 ng ml⁻¹ for lentivirus and mRNA transfection experiments; serotype O55:B5; Enzo ALX-581-013-L001) and 2 h nigericin treatment (10 μM; InvivoGen, Tlrl-nig-5). Promega CytoTox 96 Non-Radioactive Cytotoxicity Assay kits were used according to the manufacturer’s instructions (Promega, G1780). Maximum LDH was determined by replacing the medium with fresh medium containing 1% Triton X-100 for 15 min. The absorbance was read using the Synergy HTX plate reader at 490 nm.

For the measurement of PI uptake, cells were plated in black 96-well plates with optically clear bottoms. After mRNA transfection for GSDMD–TMEM106A overexpression, PI uptake was assessed after 6 h LPS prime (100 ng ml⁻¹) and 1 h nigericin treatment (10 μM). Diluted PI solution (1:500, Sigma-Aldrich, P4864-10ML) was added to the cell culture medium 30 min before the timepoint. The plate was centrifuged at 500g for 5 min and population PI staining was assayed using the Synergy HTX plate reader, conducting bottom reading of fluorescence with an excitation wavelength of 530 nm and emission wavelength of 617 nm. Maximum PI was determined after lysis of cells with 1% Triton X-100. For analysis of PI uptake in G–T^Stop and wild-type BMDMs, diluted PI solution (1:500) was added to the cell culture medium after LPS priming at the time of nigericin treatment (5 μM) and the plate was imaged using an Incucyte Live-Cell Analysis System with 5% CO₂ and 37 °C incubation with phase and red channel images taken every 15 min for 5 h. The counts of red fluorescent cells in a well were quantified to determine the total PI uptake. Maximum PI uptake is reported as the percentage of the number of red fluorescent cells at a time over the number of red fluorescent cells in the same well after 5 h of nigericin treatment. For live imaging of PI uptake by G-T^Stop and wild-type BMDMs, after LPS priming, diluted PI solution (1:500) was added to the cell culture medium at the time of nigericin treatment (5 μM). The plate was imaged using the Nikon Ti inverted microscope with the Physik Instrumente Piezo Z motor using a ×20 objective lens under 5% CO₂ and 37 °C incubation; bright-field and 561 nm laser channel images taken every 2.5 min for 2 h.

LNP-encapsulated mRNA treatment

Wild-type littermate mice were randomly assigned to mRNA treatment groups and treated intravenously through tail vein injection with 100 μl of lipid nanoparticle (LNP)-encapsulated mRNA (1 mg kg⁻¹) for 18 h before LPS sepsis or S. Typhimurium infection.

LPS sepsis

For serum cytokine analysis, female mice treated with LNP-encapsulated mRNA were injected intraperitoneally with 100 μl LPS (E. coli strain O111:B4; InvivoGen, tlrl-3pelps) in PBS (3 mg kg⁻¹) and monitored every 3 h for change in body temperature using a thermometer equipped with a rectal probe. Then, at 6 h after LPS injection, mice were bled retro-orbitally for serum cytokine analysis by Eve Technologies using the Mouse Cytokine Proinflammatory Focused 10-Plex Discovery Assay Array (Eve Technologies, MDF10). For assessment of survival, female mice treated with LNP-encapsulated mRNA were injected intraperitoneally with 100 μl LPS (E. coli strain O111:B4; InvivoGen, tlrl-3pelps) in PBS (6 mg kg⁻¹). G-T^Stop and wild-type littermate control males were injected intraperitoneally with 100 μl LPS (E. coli strain O111:B4; InvivoGen, tlrl-3pelps) in PBS (15 mg kg⁻¹) and the body temperature was recorded every 12 h for 24 h. All mice were regularly monitored for humane end-point survival, defined as having lost greater than 25% of their body temperature, severely hunched posturing and a severe lack of normal ambulation. Sample sizes were selected based on previous experiments.

S. Typhimurium Infection

Salmonella was grown from glycerol stock overnight in Luria broth (LB) with ampicillin (100 μg ml⁻¹) then diluted to obtain an optical density at 600 nm (OD₆₀₀) of 0.9 (10⁹ CFU ml⁻¹). Mice were injected intraperitoneally with 100 μl stationary-phase S. Typhimurium in PBS (10² for G-T^Stop experiments) (10⁵ CFU for in vivo LNP experiments). Then, 24 h after infection, tissue was collected and homogenized in 2 ml of PBS. Tissue homogenate supernatants were plated onto LB agar with ampicillin (100 μg ml⁻¹) for CFU calculations. Sample sizes were based on previous experiments.

Co-immunoprecipitation

Wild-type BMDMs (10⁷) were transfected with mRNA encoding GSDMD–TMEM106A–HA, GSDMD–TMEM106A^F50G/W51G−HA or a non-translating control. Then, 24 h later, cells were stimulated with LPS (100 ng ml⁻¹) for 4 h, followed by treatment with nigericin (10 μM) and glycine (10 mM) for 30 min. Cells were lysed by adding 1 ml of Pierce IP lysis buffer (Thermo Fisher Scientific, 87787) to the dish for 30 min with gentle rocking at 4 °C. Cleared lysates were added to 25 μl of prewashed Pierce Anti-HA magnetic beads (Thermo Fisher Scientific, 88836) and incubated overnight at 4 °C for immunoprecipitation of HA-tagged protein. After thorough washing, protein was eluted from anti-HA magnetic beads by adding NuPAGE LDS sample buffer and NuPAGE sample reducing agent (Thermo Fisher Scientific, NP0007 and NP0004) and heating at 70 °C for 10 min. Immunoblotting was conducted using an antibody against cleaved GSDMD (Cell Signaling Technology, 10137S, 1:1,000), and against the HA-tag (C29F4) (Cell Signaling Technology, 3724S, 1:1,000) with anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). Immunoblotting of input control lysates was performed using an anti-GSDMD antibody (E9S1X, Cell Signaling Technology, 39754, 1:1,000), β-actin antibody (13E5, Cell Signaling Technology, 4970, 1:1,000) and an HA-tag antibody (C29F4, Cell Signaling Technology, 3724S, 1:1,000).

Lentiviral transduction

Plasmids and lentiviral packaging were designed and performed by Vector Builder. For immunofluorescence microscopy, BMDMs were plated 2 × 10⁵ cells per well in a black 24-well-plate optically clear bottom pretreated with poly-l-lysine. For inflammasome activation and IL-1β-release assays, BMDMs were plated 2 × 10⁵ cells per well in standard tissue-culture-treated 24-well-plates. For subcellular protein fractionation BMDMs were plated at 1.5 × 10⁶ cells per well in standard tissue-culture-treated 6-well plates. In all experiments cells were transduced at a multiplicity of infection of 10 in medium containing polybrene (8 μg ml⁻¹). Fresh medium was added 3 h after transduction and cells were assayed 72 h later.

BMDMs transduced with lentivirus expressing GSDMD–TMEM106A–HA (Lenti(G-T-HA)) or an empty vector control were observed for GSDMD–TMEM106A localization under a Nikon Ti inverted microscope with the Physik Instrumente Piezo Z motor using a plan apo ×100 oil-immersion objective lens. WGA with Alexa fluor 488 conjugate (Thermo Fisher Scientific, W11261) was used to stain the plasma membrane before cell fixation, using 4% PFA for 10 min, and cell permeabilization, using a PBS solution containing 2% BSA and 0.1% Triton X-100. Blocking was conducted using 2% BSA for 1 h at room temperature. HA-tagged GSDMD–TMEM106A was marked with an anti-HA antibody (16B12) conjugated to PE (1:500, BioLegend, 901518) and visualized using the 561 nm laser. Membrane staining was visualized using a 488 nm laser.

Subcellular fractionation was conducted using the Thermo Fisher Scientific Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Fisher Scientific, 78840) according to the manufacturer’s instructions. Western blotting for HA was performed using anti-HA antibody (C29F4, Cell Signaling Technology, 3724S, 1:1,000) and anti-rabbit HRP-linked secondary antibody (Cell Signaling Technology, 7074S, 1:3,000). Anti-α-tubulin antibody (DM1A, Cell Signaling Technology, 3873, 1:1,000) was used as the subcellular fractionation control for cytoplasmic protein isolation. Anti-Na,K-ATPase antibody (Cell Signaling Technology, 3010, 1:1,000) was used as subcellular fractionation control for membrane protein isolation.

Doxycycline-inducible GSDMD-NT experiments

iBMDMs stably expressing doxycycline-inducible GSDMD-NT⁴⁰ were gifted from the laboratory of J. Kagan and were not tested for mycoplasma contamination. Cells were plated at 3 × 10⁴ cells per well in a black 96-well plate with optically clear bottoms and maintained in medium containing puromycin and G418. Once cells adhered (over 12 h after plating), mRNA was transfected for GSDMD–TMEM106A, GSDMD–TMEM106A^ΔCT or a non-translating control for 24 h. To induce GSDMD-NT expression, cells were treated with doxycycline hyclate (Sigma-Aldrich, D9891-1G) in fresh medium (0.25 µg ml⁻¹). Diluted PI solution (1:300, Sigma-Aldrich, P4864-10ML) was added to the cell culture medium at the time of doxycycline treatment and the plate was imaged using the Incucyte Live-Cell Analysis System under 5% CO₂ and 37 °C incubation; phase and red channel images were taken every hour for 24 h. The count of red fluorescent cells in a well was quantified to determine PI uptake. Maximum PI uptake is reported as the percentage of the number of red fluorescent cells at a time over the number of red fluorescent cells in the same well after 24 h of doxycycline treatment.

Recombinant protein constructs

Full-length GSDMD was cloned into the pDB.His.SUMO vector using Gibson Assembly Master Mix (New England Biolabs, E2611L). GSDMD–TMEM106A–FLAG (rG–T–FLAG) or rGSDMD–TMEM106A^ΔCT−FLAG were cloned into the pDB.His.MBP-3C vector, downstream of the N-terminal His₆–maltose-binding protein (MBP) tag and human rhinovirus 3C protease site, using the Gibson Assembly Master Mix.

Expression and purification of proteins in bacteria

pDB-His-SUMO-GSDMD, pDB.His.MBP-3C-rG-T-FLAG and pDB.His.MBP-3C-rG-T^ΔCT-FLAG constructs were transformed into E. coli BL21 (DE3) cells (Agilent Technologies, 230280). Transformants were plated and incubated overnight at 37 °C. Single colonies were inoculated into LB medium containing 50 μg ml⁻¹ kanamycin and cultured at 37 °C. At an OD₆₀₀ of 0.6, protein expression was induced with 500 μM isopropyl β-d-1-thiogalactopyranoside, and cells were grown for 16 h at 18 °C before collection.

For His-SUMO-GSDMD, cells were pelleted by centrifugation at 4,000g for 30 min and resuspended in buffer A (40 mM HEPES, pH 7.0, 150 mM NaCl) supplemented with 5 mM imidazole. Cells were lysed by sonication, and His–SUMO-tagged GSDMD was captured on Ni-NTA resin and eluted with buffer A containing 300 mM imidazole. The His–SUMO tag was removed by ULP1 protease at 4 °C for 4 h. Cleaved tag was separated using Ni-NTA, and the flow-through containing GSDMD was further purified on the Superdex 200 Increase 10/300 GL size-exclusion column (Cytiva) equilibrated in buffer A. Peak fractions were assessed by SDS–PAGE and snap-frozen for downstream assays.

For pDB-His-MBP-3C-rG-T-FLAG or pDB.His.MBP-3C-rG-T^ΔCT-FLAG, cells were pelleted by centrifugation at 4,000g for 30 min and resuspended in buffer A. Cells were lysed by sonication (2 s on, 8 s off; 5 min total on time; 40% amplitude) on ice, and clarified by centrifugation at 18,000g for 1 h. The supernatant was incubated with anti-FLAG G1 affinity resin (GenScript, L00432-25) for 2 h at 4 °C with gentle rotation. After extensive washing, proteins were eluted with buffer A containing 150 ng µl⁻¹ 3×FLAG peptide (Sigma-Aldrich, F4799-25MG), analysed for purity and snap-frozen for subsequent assays.

Expression and purification of proteins in mammalian cells

pcDNA3.1-GSDMD-Flag constructs were transfected into Expi293 cells that were maintained in 1,000 ml Expi293 expression medium, fed with 6 mM KCl and grown to 2.5 × 10⁶ cells per ml, using polyethylenimine (PEI, Polysciences). At 12 h after transfection, the culture was fed with 10 mM sodium butyrate and 10 ml of a 45% d-(+)-glucose solution. The cells were then grown for an additional 2 days before being collected by centrifugation at 4,000g for 30 min.

The cell pellet was resuspended in buffer A and lysed by sonication (2 s on, 8 s off, for a total on time of 3.5 min at 40% power). The lysates were clarified by ultracentrifugation at 40,000 rpm for 1 h. The supernatant was collected and incubated with Flag G1 affinity resin (GenScript, L00432-25) for 2 h at 4 °C with gentle rotation. After washing, the protein was eluted using buffer A with 150 ng μl⁻¹ 3× Flag peptide (Sigma-Aldrich, F4799-25MG). The eluted GSDMD protein was snap-frozen for other assays. To obtain highly palmitoylated GSDMD, ROT (Sigma-Aldrich, R8875-1G) at 10 μM was used to treat Expi293 cells for 4 h before collection.

Liposome leakage assay

Liposomes were prepared as previously described^39,91. In brief, cardiolipin (CL), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (PE) and phosphatidylcholine (PC) (Avanti Polar Lipids) were mixed at a mass ratio of 5:8:4. The solvent was evaporated under a gentle stream of N₂ gas, and the dried lipid film was resuspended in 200 μl buffer B (20 mM HEPES, pH 7.4, 150 mM NaCl, 50 mM sodium citrate and 15 mM TbCl₃). The suspension was extruded 45 times through 100 nm Whatman Nuclepore Track-Etched membranes to obtain homogeneous liposomes. The filtered suspension was purified through a size-exclusion column (Superose 6, 10/300 GL) in buffer C (20 mM HEPES, 150 mM NaCl) to remove TbCl₃ outside liposomes. Fractions close to void were pooled to produce a stock of PC/PE/CL liposomes.

Bacteria or mammalian expressed full-length GSDMD was precleaved by CASP-1 for 1 h. His–MBP–3C–G–T–Flag or His–MBP–3C–G–T^ΔCT−Flag was precleaved by 3C for 30 min. Catalytically active CASP-1 was expressed and purified as previously described⁹¹. In brief, the p20 and p10 subunits (without tags) were individually expressed in E. coli BL21 (DE3) as inclusion bodies. The p20–p10 complex was assembled by denaturation and refolding, and further purified by HiTrap SP cation-exchange chromatography (GE Healthcare Life Sciences).

For leakage assays, liposomes were diluted four times in buffer D (20 mM HEPES, 150 mM NaCl and 15 μM dipicolinic acid). Leakage of Tb³⁺ from liposomes was monitored by the increase in fluorescence after binding to dipicolinic acid in buffer D. Pre-cleaved proteins were added to 384-well plates (Corning, 3820) containing PC/PE/CL liposomes. The fluorescence intensity was recorded at 545 nm (excitation 276 nm) immediately after mixing and monitored for 2 h on a SYNERGY microplate reader (Biotek). The baseline emission before protein addition was defined as F₀. After 2 h, 0.1% Triton X-100 was added to achieve complete Tb³⁺ release, and the mean of the top three fluorescence readings was defined as F₁₀₀. The percentage of Tb³⁺ release at each timepoint is defined as follows: Tb³⁺ release (%) = (F − F₀) × 100/(F₁₀₀ − F₀).

Protein–lipid binding assay

Lipid-binding assay for rG–T–Flag was performed using membrane lipid strips (Echelon Biosciences, P-6002) according to the manufacturer’s instructions. In brief, lipid strip membranes were blocked with 3% BSA in PBS containing 0.1% Tween-20 (PBST) for 1 h at room temperature. The strips were then incubated for 1 h with pre-cleaved rG–T–Flag (His–MBP–3C–rG–T–Flag was pre-cleaved with 3C protease for 30 min to remove the His–MBP–3C tag) at 2 μg ml⁻¹ in 3% BSA/PBST. After three washes with PBST (5 min each), bound proteins were detected using anti-Flag M2-HRP antibody (1:1,000). All steps were performed at room temperature.

Protein structure prediction and visualization

Prediction of three-dimensional protein structures was performed using AlphaFold 3⁹² with AlphaFold Server (https://alphafoldserver.com/) using a random seed value of 9999. Protein structure images were created using Mol*⁹³ and the RCSB PDB platform⁹⁴ (https://www.rcsb.org/).

Schematic design

Schematics for Figs. 1a, 2k and 3a and Extended Data Figs. 3a and 9a were created using BioRender. All other schematics were created using PowerPoint or Inkscape software.

Data analysis

All sequencing data analysis was performed on a Linux GNU Pop!_OS LTS 22.04 64-bit system or Ubuntu 26.04 LTS 64-bit system. Python analysis was performed using Anaconda Distribution version ≤26.1.1 and Python programming language version ≤3.13.9. Analysis in R was performed using R version ≤4.5.2 running in R Studio version ≤2026.05.0 Build 218. Graphs generated in R were produced using the ggplot2 suite⁹⁵ version ≤4.0.3 and ComplexHeatmap⁹⁶ v.2.26.1. Microsoft Excel (v.2607, build 16.0.20228.20110) 64-bit and LibreOffice Calc 26.2.4.2 (X86_64) were used for table generation and data organization. Microsoft Windows 11 Pro 24H2 OS and MacOS BigSur v.11.6 were used for additional data analysis. AlphaFold 3 folding was run using AlphaFold Server (https://alphafoldserver.com/welcome). Mol* and the RCSB PDB platform for AlphaFold 3 result visualization were accessed at https://www.rcsb.org/3d-view.

Except where otherwise indicated, statistical significance for experiments with multiple groups and multiple independent variables was determined using two-way ANOVA with Sidak’s multiple-comparison test. One-way ANOVA with Tukey’s multiple-comparison test was used to determine statistical significance for experiments with multiple groups and a single independent variable. Unpaired t-tests were performed to determine significance between two variables. All non-R/Python adjusted P values were calculated using GraphPad Prism v.10.0.2. GraphPad Prism was run on MacOS BigSur v.11.6.

Ethics

All mouse experiments were performed in accordance with protocols approved by the Harvard Medical School Institutional Animal Care and Use Committee.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Long-read direct RNA-seq data have been deposited in the Gene Expression Omnibus (GEO) under accession code GSE267147 (mouse) and GSE277057 (human). Short-read RNA-seq data have been deposited in the GEO under accession code GSE324139. Hi-C sequencing data have been deposited in the GEO under accession code GSE324391. K562 cell data were downloaded from GEO accession GSE123191 (ref. ⁶⁹). GENCODE reference files for GENCODE M28 release (GRCm39) and GENCODE 43 release (GRCh38.p13) are available online (https://www.gencodegenes.org/). UCSC reference files are available online (https://genome.ucsc.edu/) and can be accessed through the Table Browser (https://genome.ucsc.edu/cgi-bin/hgTables). Ensembl reference files are available online (https://www.ensembl.org/info/data/ftp/index.html?redirect=no) and can be downloaded through the ftp site (https://ftp.ensembl.org/pub/). Source data are provided with this paper.

Custom code for the TYPHON wrapper script is available at GitHub (https://github.com/erenada/TYPHON).

1. Liu, X. et al. Inflammasome-activated gasdermin D causes pyroptosis by forming membrane pores. Nature 535, 153–158 (2016).

   Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

2. Shi, J. et al. Cleavage of GSDMD by inflammatory caspases determines pyroptotic cell death. Nature 526, 660–665 (2015).

   Article  ADS  CAS  PubMed  Google Scholar 

3. Michaeli, S. Trans-splicing in trypanosomes: machinery and its impact on the parasite transcriptome. Fut. Microbiol. 6, 459–474 (2011).

   CAS  Google Scholar 

4. Dorn, R., Reuter, G. & Loewendorf, A. Transgene analysis proves mRNA trans-splicing at the complex mod(mdg4) locus in Drosophila. Proc. Natl Acad. Sci. USA 98, 9724–9729 (2001).

   ADS  CAS  PubMed  PubMed Central  Google Scholar 

5. Daley, G. Q., Van Etten, R. A. & Baltimore, D. Induction of chronic myelogenous leukemia in mice by the P210bcr/abl gene of the Philadelphia chromosome. Science 247, 824–830 (1990).

   ADS  CAS  PubMed  Google Scholar 

6. Nowell, P. C. & Hungerford, D. A. Chromosome studies on normal and leukemic human leukocytes. J. Natl Cancer Inst. 25, 85–109 (1960).

7. Li, H., Wang, J., Mor, G. & Sklar, J. A neoplastic gene fusion mimics trans-splicing of RNAs in normal human cells. Science 321, 1357–1362 (2008).

   ADS  CAS  PubMed  Google Scholar 

8. Bracht, J. R. et al. Chromosome fusions triggered by noncoding RNA. RNA Biol. 14, 620–631 (2017).

   PubMed  Google Scholar 

9. Yuan, H. et al. A chimeric RNA characteristic of rhabdomyosarcoma in normal myogenesis process. Cancer Discov. 3, 1394–1403 (2013).

   CAS  PubMed  Google Scholar 

10. Cocquet, J., Chong, A., Zhang, G. & Veitia, R. A. Reverse transcriptase template switching and false alternative transcripts. Genomics 88, 127–131 (2006).

    CAS  PubMed  Google Scholar 

11. Heyer, E. E. et al. Diagnosis of fusion genes using targeted RNA sequencing. Nat. Commun. https://doi.org/10.1038/s41467-019-09374-9 (2019).

12. Latysheva, N. S. & Babu, M. M. Discovering and understanding oncogenic gene fusions through data intensive computational approaches. Nucleic Acids Res. 44, 4487–4503 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

13. Almeida da Paz, M., Warger, S. & Taher, L. Disregarding multimappers leads to biases in the functional assessment of NGS data. BMC Genom. https://doi.org/10.1186/s12864-024-10344-9 (2024).

14. Chen, C., Haddox, S., Tang, Y., Qin, F. & Li, H. Landscape of chimeric RNAs in non-cancerous cells. Genes https://doi.org/10.3390/genes12040466 (2021).

15. Singh, S. et al. The landscape of chimeric RNAs in non-diseased tissues and cells. Nucleic Acids Res. 48, 1764–1778 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

16. Liu, Q. et al. LongGF: computational algorithm and software tool for fast and accurate detection of gene fusions by long-read transcriptome sequencing. BMC Genom. 21, 793 (2020).

    CAS  Google Scholar 

17. Davidson, N. M. et al. JAFFAL: detecting fusion genes with long-read transcriptome sequencing. Genome Biol. 23, 10 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

18. Karaoglanoglu, F., Chauve, C. & Hach, F. Genion, an accurate tool to detect gene fusion from long transcriptomics reads. BMC Genom. 23, 129 (2022).

    Google Scholar 

19. Drexler, H. L. et al. Revealing nascent RNA processing dynamics with nano-COP. Nat. Protoc. 16, 1343–1375 https://doi.org/10.1038/s41596-020-00469-y (2021).

20. Uhrig, S. et al. Accurate and efficient detection of gene fusions from RNA sequencing data. Genome Res. https://doi.org/10.1101/gr.257246.119 (2021).

21. Nicorici, D. et al. FusionCatcher—a tool for finding somatic fusion genes in paired-end RNA-sequencing data. Preprint at bioRxiv https://doi.org/10.1101/011650 (2014).

22. Haas, B. J. et al. Accuracy assessment of fusion transcript detection via read-mapping and de novo fusion transcript assembly-based methods. Genome Biol. 20, 213 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

23. Melsted, P. et al. Fusion detection and quantification by pseudoalignment. Preprint at bioRxiv https://doi.org/10.1101/166322 (2017).

24. Davidson, N. M., Majewski, I. J. & Oshlack, A. JAFFA: high sensitivity transcriptome-focused fusion gene detection. Genome Med. https://doi.org/10.1186/S13073-015-0167-X (2015).

25. Caldas, P. et al. Transcription readthrough is prevalent in healthy human tissues and associated with inherent genomic features. Commun. Biol. 7, 100 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

26. Ribeiro, D. M., Ziyani, C. & Delaneau, O. Shared regulation and functional relevance of local gene co-expression revealed by single cell analysis. Commun. Biol. 5, 876 (2022).

    Article  Google Scholar 

27. Zhang, X. et al. Inactivation of TMEM106A promotes lipopolysaccharide-induced inflammation via the MAPK and NF-κB signaling pathways in macrophages. Clin. Exp. Immunol. 203, 125–136 (2021).

    CAS  PubMed  Google Scholar 

28. Kotake, Y. et al. Splicing factor SF3b as a target of the antitumor natural product pladienolide. Nat. Chem. Biol. 3, 570–575 (2007).

    Article  CAS  PubMed  Google Scholar 

29. Cretu, C. et al. Structural basis of splicing modulation by antitumor macrolide compounds. Mol. Cell 70, 265–273 (2018).

    CAS  PubMed  Google Scholar 

30. Perry, R. P. & Kelley, D. E. Inhibition of RNA synthesis by actinomycin D: characteristic dose-response of different RNA species. J. Cell. Physiol. 76, 127–139 (1970).

    CAS  PubMed  Google Scholar 

31. Ling, J. Q. et al. CTCF mediates interchromosomal colocalization between Igf2/H19 and Wsb1/Nf1. Science 312, 269–272 (2006).

    ADS  CAS  PubMed  Google Scholar 

32. Kayagaki, N. et al. Caspase-11 cleaves gasdermin D for non-canonical inflammasome signalling. Nature 526, 666–671 (2015).

    Article  ADS  CAS  PubMed  Google Scholar 

33. Netea, M. G. et al. IL-1β processing in host defense: beyond the inflammasomes. PLoS Pathog. 6, e1000661 (2010).

34. Schultz, M. J. et al. Role of interleukin-1 in the pulmonary immune response during Pseudomonas aeruginosa pneumonia. Am. J. Physiol. Lung Cell. Mol. Physiol. 282, L285–L290 (2002).

    CAS  PubMed  Google Scholar 

35. Chen, K. W. et al. Noncanonical inflammasome signaling elicits gasdermin D-dependent neutrophil extracellular traps. Sci. Immunol. 3, eaar6676 (2018).

    PubMed  Google Scholar 

36. Leventis, P. A. & Grinstein, S. The distribution and function of phosphatidylserine in cellular membranes. Annu. Rev. Biophys. 39, 407–427 (2010).

    CAS  PubMed  Google Scholar 

37. Van Meer, G., Voelker, D. R. & Feigenson, G. W. Membrane lipids: where they are and how they behave. Nat. Rev. Mol. Cell Biol. 9, 112–124 (2008).

    Article  PubMed  PubMed Central  Google Scholar 

38. Yu, P. et al. Pyroptosis: mechanisms and diseases. Signal Transduct. Target. Ther. 6, 128 (2021).

    PubMed  PubMed Central  Google Scholar 

39. Xia, S. et al. Gasdermin D pore structure reveals preferential release of mature interleukin-1. Nature 593, 607–611 (2021).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

40. Evavold, C. L. et al. Control of gasdermin D oligomerization and pyroptosis by the Ragulator-Rag-mTORC1 pathway. Cell 184, 4495–4511 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

41. Jackson, R. et al. The translation of non-canonical open reading frames controls mucosal immunity. Nature 564, 434–438 (2018).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

42. Brubaker, S. W., Gauthier, A. E., Mills, E. W., Ingolia, N. T. & Kagan, J. C. A bicistronic MAVS transcript highlights a class of truncated variants in antiviral immunity. Cell https://doi.org/10.1016/j.cell.2014.01.021 (2014).

43. Anderson, D. M. et al. A micropeptide encoded by a putative long noncoding RNA regulates muscle performance. Cell 160, 595–606 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

44. Wang, Y. & Wang, Z. Efficient backsplicing produces translatable circular mRNAs. RNA https://doi.org/10.1261/rna.048272.114 (2015).

45. Memczak, S. et al. Circular RNAs are a large class of animal RNAs with regulatory potency. Nature https://doi.org/10.1038/nature11928 (2013).

46. Chen, C.-K. et al. Structured elements drive extensive circular RNA translation. Mol. Cell 81, 4300–4318 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

47. Jeck, W. R. et al. Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA 19, 141–157 (2013).

    CAS  PubMed  Google Scholar 

48. Rowley, J. D. & Blumenthal, T. The cart before the horse. Science 321, 1302–1304 (2008).

    CAS  PubMed  Google Scholar 

49. Engreitz, J. M., Agarwala, V. & Mirny, L. A. Three-dimensional genome architecture influences partner selection for chromosomal translocations in human disease. PLoS ONE 7, e44196 (2012).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

50. Lin, C. et al. Nuclear receptor-induced chromosomal proximity and DNA breaks underlie specific translocations in cancer. Cell 139, 1069–1083 (2009).

    MathSciNet  CAS  PubMed  PubMed Central  Google Scholar 

51. Marek, L. R. & Kagan, J. C. Phosphoinositide Binding by the Toll Adaptor dMyD88 controls antibacterial responses in Drosophila. Immunity 36, 612–622 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

52. Miao, R. et al. Gasdermin D permeabilization of mitochondrial inner and outer membranes accelerates and enhances pyroptosis. Immunity 56, 2523–2541 (2023).

    CAS  PubMed  PubMed Central  Google Scholar 

53. Chu, X. et al. Gasdermin D-mediated pyroptosis is regulated by AMPK-mediated phosphorylation in tumor cells. Cell Death Dis. 14, 469 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

54. Devant, P. et al. Gasdermin D pore-forming activity is redox-sensitive. Cell Rep. https://doi.org/10.1016/j.celrep.2023.112008 (2023).

55. Humphries, F. et al. Succination inactivates gasdermin D and blocks pyroptosis. Science https://doi.org/10.1126/science.abb9818 (2020).

56. Shi, Y. et al. E3 ubiquitin ligase SYVN1 is a key positive regulator for GSDMD-mediated pyroptosis. Cell Death Dis. https://doi.org/10.1038/s41419-022-04553-x (2022).

57. Chu, X. et al. Ubiquitination of gasdermin D N-terminal domain directs its membrane translocation and pore formation during pyroptosis. Cell Death Dis. 16, 181 (2025).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

58. Frankish, A. et al. GENCODE: reference annotation for the human and mouse genomes in 2023. Nucleic Acids Res. 51, D942–D949 (2023).

    CAS  PubMed  PubMed Central  Google Scholar 

59. Nassar, L. R. et al. The UCSC Genome Browser database: 2023 update. Nucleic Acids Res. 51, D1188–D1195 (2023).

    CAS  PubMed  PubMed Central  Google Scholar 

60. De Coster, W., D’Hert, S., Schultz, D. T., Cruts, M. & Van Broeckhoven, C. NanoPack: visualizing and processing long-read sequencing data. Bioinformatics 34, 2666–2669 (2018).

    PubMed  PubMed Central  Google Scholar 

61. Li, H. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 34, 3094–3100 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

62. Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).

    PubMed  PubMed Central  Google Scholar 

63. Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods 14, 417–419 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

64. Soneson, C., Love, M. I. & Robinson, M. D. Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. F1000Research https://doi.org/10.12688/F1000RESEARCH.7563.2 (2016).

65. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. https://doi.org/10.1186/s13059-014-0550-8 (2014).

66. Dyer, S. C. et al. Ensembl 2025. Nucleic Acids Res. 53, D948–D957 (2025).

    CAS  PubMed  PubMed Central  Google Scholar 

67. Zheng, G. X. Y. et al. Massively parallel digital transcriptional profiling of single cells. Nat. Commun. https://doi.org/10.1038/ncomms14049 (2017).

68. Lopez, F. et al. Explore, edit and leverage genomic annotations using Python GTF toolkit. Bioinformatics 35, 3487–3488 https://doi.org/10.1093/bioinformatics/btz116 (2019).

69. Drexler, H. L., Choquet, K. & Churchman, L. S. Splicing kinetics and coordination revealed by direct nascent RNA sequencing through nanopores. Mol. Cell 77, 985–998 (2020).

    CAS  PubMed  Google Scholar 

70. Shen, W., Le, S., Li, Y. & Hu, F. SeqKit: a cross-platform and ultrafast toolkit for FASTA/Q file manipulation. PLoS ONE https://doi.org/10.1371/journal.pone.0163962 (2016).

71. Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinform. https://doi.org/10.1186/1471-2105-10-421 (2009).

72. Neph, S. et al. BEDOPS: high-performance genomic feature operations. Bioinformatics 28, 1919–1920 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

73. Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).

    CAS  PubMed  PubMed Central  Google Scholar 

74. Sievers, F. et al. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol. Syst. Biol. 7, 539 https://doi.org/10.1038/msb.2011.75 (2011).

75. Larkin, M. A. et al. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947–2948 (2007).

    CAS  PubMed  Google Scholar 

76. Cock, P. J. A. et al. Biopython: freely available Python tools for computational molecular biology and bioinformatics. Bioinformatics 25, 1422–1423 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

77. Singh, U. & Wurtele, E. S. orfipy: a fast and flexible tool for extracting ORFs. Bioinformatics 37, 3019–3020 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

78. Browaeys, R., Saelens, W. & Saeys, Y. NicheNet: modeling intercellular communication by linking ligands to target genes. Nat. Methods 17, 159–162 (2020).

    Article  CAS  PubMed  Google Scholar 

79. Lawrence, M., Gentleman, R. & Carey, V. rtracklayer: an R package for interfacing with genome browsers. Bioinformatics 25, 1841–1842 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

80. Bray, N. L., Pimentel, H., Melsted, P. & Pachter, L. Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 34, 525–527 https://doi.org/10.1038/nbt.3519 (2016).

81. Marsh, S. E. scCustomize: custom visualizations & functions for streamlined analyses of single cell sequencing. Zenodo https://doi.org/10.5281/zenodo.5706430 (2021).

82. Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).

    PubMed  PubMed Central  Google Scholar 

83. Class, C. A., Lukan, C. J., Bristow, C. A. & Do, K. A. Easy NanoString nCounter data analysis with the NanoTube. Bioinformatics https://doi.org/10.1093/bioinformatics/btac762 (2023).

84. Johanson, T. M. et al. Transcription-factor-mediated supervision of global genome architecture maintains B cell identity. Nat. Immunol. 19, 1257–1264 (2018).

    Article  CAS  PubMed  Google Scholar 

85. Lun, A. T. L. & Smyth, G. K. diffHic: a Bioconductor package to detect differential genomic interactions in Hi-C data. BMC Bioinform. 16, 258 (2015).

    Google Scholar 

86. Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet 17, 10–11 (2011).

    Google Scholar 

87. Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

88. Durand, N. C. et al. Juicer provides a one-click system for analyzing loop-resolution Hi-C experiments. Cell Syst. 3, 95–98 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

89. Van Der Weide, R. H. et al. Hi-C analyses with GENOVA: a case study with cohesin variants. NAR Genom. Bioinform. 3, lqab040 (2021).

    PubMed  PubMed Central  Google Scholar 

90. Liao, Y., Smyth, G. K. & Shi, W. The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Res. https://doi.org/10.1093/nar/gkz114 (2019).

91. Du, G. et al. ROS-dependent S-palmitoylation activates cleaved and intact gasdermin D. Nature 630, 437–446 (2024).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

92. Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

93. Sehnal, D. et al. Mol∗Viewer: modern web app for 3D visualization and analysis of large biomolecular structures. Nucleic Acids Res. 49, W431–W437 (2021).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

94. Berman, H. M. et al. The Protein Data Bank. Nucleic Acids Res. 28, 235–242 https://doi.org/10.1093/nar/28.1.235 (2000).

95. Wickham, H. ggplot2: Elegant Graphics for Data Analysis Second Edition (Springer, 2015).

96. Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016).

    CAS  PubMed  Google Scholar 

Download references

Author notes

1. Gang Du

   Present address: Institute of Precision Medicine, the First Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong, China

2. Liang Shan

   Present address: Institute of Human Immunology, Shenzhen Medical Academy of Research and Translation, Shenzhen, Guangdong, China

3. These authors contributed equally: Olivia Venezia, Harry Kane

Authors and Affiliations

1. Department of Immunology, Harvard Medical School, Boston, MA, USA

   Olivia Venezia, Harry Kane, Hongshen Wang, Pathricia Tilstam, Samuel W. Kazer, Georgia Gunner, Daisy A. Hoagland, Eren Ada, Ademi Zhakyp, Sandhya Kumar, Coco Duizer, Owen Searle, Chris Geo Provido, Joia Capocchi, Bahawar Sharif Dhillon, Ruth A. Franklin, Isaac M. Chiu & Ruaidhrí Jackson

2. Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, USA

   Gang Du, Shilong Yang, Liam B. Healy & Hao Wu

3. Program in Cellular and Molecular Medicine, Boston Children’s Hospital, Boston, MA, USA

   Gang Du, Shilong Yang, Liam B. Healy, Judy Lieberman & Hao Wu

4. The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia

   Hannah D. Coughlan, Timothy M. Johanson & Rhys S. Allan

5. Department of Medical Biology, The University of Melbourne, Parkville, Victoria, Australia

   Hannah D. Coughlan, Timothy M. Johanson & Rhys S. Allan

6. Broad Institute of MIT and Harvard, Cambridge, MA, USA

   Samuel W. Kazer & Jose Ordovas-Montanes

7. Division of Gastroenterology, Boston Children’s Hospital and Harvard Medical School, Boston, MA, USA

   Samuel W. Kazer & Jonathan C. Kagan

8. Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, USA

   Samuel W. Kazer & Jose Ordovas-Montanes

9. Department of Stem Cell and Regenerative Biology, Harvard University, Cambridge, MA, USA

   Daisy A. Hoagland & Ruth A. Franklin

10. Moderna, Cambridge, MA, USA

    Megha G. Basavappa, Kanchana Ravichandran & Kate L. Jeffrey

11. Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Leonel Joannas & Jorge Henao-Mejia

12. Institute for Immunology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

    Leonel Joannas & Jorge Henao-Mejia

13. Division of Infectious Diseases, Department of Medicine, Washington University School of Medicine, St Louis, MO, USA

    Qiankun Wang & Liang Shan

14. Department of Chemistry, Yale University, New Haven, CT, USA

    Sarah Slavoff

15. Division of Protective Immunity, Department of Pathology and Laboratory Medicine, Children’s Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA, USA

    Jorge Henao-Mejia

16. Harvard Stem Cell Institute, Harvard University, Cambridge, MA, USA

    Ruth A. Franklin & Jose Ordovas-Montanes

17. The Gene Lay Institute of Immunology and Inflammation, Brigham and Women’s Hospital, Mass General Hospital and Harvard Medical School, Boston, MA, USA

    Isaac M. Chiu

18. Division of Gastroenterology, Hepatology and Nutrition, Boston Children’s Hospital, Boston, MA, USA

    Jose Ordovas-Montanes

19. Program in Immunology, Harvard Medical School, Boston, MA, USA

    Jose Ordovas-Montanes

20. Department of Pediatrics, Harvard Medical School, Boston, MA, USA

    Judy Lieberman

Contributions

R.J. conceived and initiated the project. O.V. led the wet laboratory experiments and H.K. led the computational aspects of the project, with initial help from S.W.K. and J.O.-M. G.D., S.Y., L.B.H., J.L. and H. Wu produced recombinant proteins and performed lipid-binding assays and liposome-leakage assays. H.D.C., T.M.J. and R.S.A. generated and analysed genome-wide in situ Hi-C data. H. Wang conducted key experiments relating to GSDMD–TMEM106A protein localization and P.T. generated RNA used for ONT analysis. G.G. and I.M.C. conducted the intracisternal infection and brain myeloid cell experiments and provided resources for Incucyte studies. D.A.H. and R.A.F. generated the lung macrophage data. M.G.B., K.R. and K.L.J. generated mRNA for transfection and LNP studies and provided critical experimental advice and instruction. E.A. provided computational assistance for producing TYPHON. A.Z. assisted in making the injection mix and formulation for the creation of G-T^Stop mice. S.K. and C.D. assisted in chRNA PCRs and Sanger sequencing. O.S., C.G.P. and J.C. aided in data analysis and presentation. B.S.D. provided the Gsdmd-knockout mice. L.J. and J.H.-M. designed and generated the G-T^MYC mice. Q.W. and L.S. provided human macrophage RNA for chRNA identification. S.S. provided advice and discussion regarding chRNA protein translation and modification. J.C.K., J.L. and H. Wu provided key reagents, advice and discussion on experiments pertaining to GSDMD. O.V., H.K. and R.J. wrote the manuscript with contributions from all of the authors in the interpretation of the data and conclusions of the findings. R.J. supervised and led the study.

Corresponding author

Correspondence to Ruaidhrí Jackson.

Competing interests

K.L.J., M.G.B. and K.R. are currently employees of Moderna, and hold shares and/or stock options in the company. J.C.K. consults and holds equity in Corner Therapeutics, Larkspur Biosciences, MindImmune Therapeutics and Neumora Therapeutics. J.O.-M. reports compensation for consulting services with Tessel Biosciences and Radera Biotherapeutics. None of these relationships have impacted this study. The other authors declare no competing interests.

Peer review information

Nature thanks the anonymous reviewers for their contribution to the peer review of this work.

Reporting Summary (download PDF )

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions