The lysosomal LAMTOR-Rag complex functions as a checkpoint for antiviral interferon production
Zeming Feng, Lulu Wang, Shujun Chen, Sihan Cao, Miao Lei, Xiuzhen Yang, Kaixiong Ma, Shi Yu, Huina Hu, Kaixuan Zheng, Xin Xu, Qi Zheng, Shaobo Wang, Wenxiang Hu, Chun-Yan Lim
TL;DR
The lysosomal LAMTOR-Rag complex acts as a nutrient-sensitive checkpoint for antiviral interferon production, linking metabolism and immunity.
Contribution
The study identifies a lysosome-specific checkpoint for IFN-β production that operates independently of mTORC1.
Findings
LAMTOR-Rag complex is essential for IFN-β production across multiple PRR signaling pathways.
Rag GTPase regulates IRF-5/7 and p38 MAPK recruitment to lysosomes, stabilizing Ifnb1 mRNA.
Disruption of the LAMTOR-Rag-FLCN-p38 axis impairs antiviral responses in vitro and in vivo.
Abstract
Lysosomes are emerging as important signaling hubs for antiviral defense, yet how they enable type I interferon (IFN-β) production is unclear. Here, we identify an evolutionarily repurposed lysosomal pathway, centered on the LAMTOR-Rag GTPase complex, that governs IFN-β production through dual transcriptional and post-transcriptional regulation. Genetic ablation of LAMTOR or Rag GTPases in macrophages abolishes IFN-β responses despite intact pattern recognition receptor (PRR) signaling, uncovering a lysosome-specific checkpoint essential for antiviral immunity. Mechanistically, Rag GTPase activity controls IRF expression to prime IFN transcription, while upon PRR stimulation, the tumor suppressor FLCN recruits p38 MAPK to lysosomes, where Rag-dependent p38 phosphorylation stabilizes Ifnb1 mRNA. Nutrient availability dynamically modulates Rag nucleotide states and thereby its activation,…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 18
Figure 19- —http://dx.doi.org/10.13039/501100001809MOST | National Natural Science Foundation of China (NSFC)
- —R&D Program of Guangzhou National Laboratory
- —Overseas Experts Supporting Programs in China
- —National High-level Talents Program of China
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsinterferon and immune responses · Cell death mechanisms and regulation · Autophagy in Disease and Therapy
Introduction
Type I interferons (IFNs), originally identified for their ability to ‘interfere’ with viral replication, are induced and secreted by host cells in response to viral infection (Pestka, 2007; McNab et al, 2015; Ivashkiv and Donlin, 2014). The emergence of IFNs as early as jawed vertebrates, such as cartilaginous sharks and bony fish, represents a seminal event in the acquisition of host innate immunity against infectious viruses during vertebrate evolution (Secombes and Zou, 2017). In particular, major IFN isoforms, including IFN-α and IFN-β, trigger a signaling cascade by binding to IFN-α/β receptors (IFNARs) on the cell surface, culminating in a potent antiviral transcription program involving hundreds of interferon-stimulated genes (ISGs) that can disrupt every step of viral replication (Ivashkiv and Donlin, 2014; Schoggins, 2019).
Recognition of viral infections by host cells requires the sensing of pathogen-associated molecular patterns (PAMPs) via cytosolic [RIG-I-like receptors (RLRs)] or membrane-bound [Toll-like receptors (TLRs)] pattern recognition receptors (PRRs) (Ivashkiv and Donlin, 2014; McNab et al, 2015; Janeway, 1989). Viral PAMPs are often nucleic acid structures of the viral genome or viral replication intermediates, all of which are distinct molecular signatures absent from host cells (Iwasaki, 2012). Upon engagement, PRR signaling triggers the nuclear translocation of several key interferon regulatory factors (IRFs), particularly IRF-3 and IRF-7, which coordinate on the IFN-α/β enhancer the assembly of an enhanceosome consisting of other transcription factors (TFs), including nuclear factor kappa B (NF-κB) and AP-1 (comprising ATF-2/c-Jun heterodimers). This enhanceosome further recruits coactivators, the chromatin-remodeling complex, and the RNA Pol II transcriptional machinery to the IFN-α/β promoter, driving IFN expression (Panne et al, 2007; Honda and Taniguchi, 2006).
As a consequence of virus-immune coevolution, IFN production can be elicited in almost every cell type because viruses have evolved to infect all of them, whereas IFNARs are ubiquitously expressed in almost all cells, allowing them to acquire an antiviral state (McNab et al, 2015; Hoffmann et al, 2015). Hence, the IFN program, as a pivotal component of innate immunity, is intrinsically hardwired into the host genome and is subject to mechanistic regulation from viral recognition to signal propagation, transcriptional activation to IFN secretion, and autocrine-paracrine signaling to the ISG response.
Lysosomes are intracellular catabolic endpoints responsible for the degradation of macromolecules via hydrolytic activity while also acting as a critical signaling hub that integrates nutrient availability with cellular metabolic responses (Perera and Zoncu, 2016; Lim and Zoncu, 2016). Central to this regulatory function is the evolutionarily conserved mTORC1 pathway that coordinates cell growth, autophagy, and metabolism in response to nutrient fluctuations. The activation of mTORC1 is spatially regulated at the lysosomal limiting membrane, where a dedicated scaffolding machinery consisting of the LAMTOR (late endosomal/lysosomal adapter and MAPK and mTOR activator) complex, also known as the Ragulator, and its cognate binding partners, the heterodimeric Rag GTPases, mediates the recruitment of mTORC1 in response to nutrient signals (Liu and Sabatini, 2020; Goul et al, 2023).
In addition to their role in nutrient sensing, lysosomes have a primordial function in cell-autonomous immunity against viruses and other pathogens (Deretic, 2021). By leveraging their degradative and detoxifying capabilities, lysosomes facilitate the engulfment and destruction of viral particles or debris delivered via the endocytic, autophagic, and phagocytic pathways, thereby restricting the initial infection and viral spread within the host cell. Additionally, lysosomes contribute to TLR-mediated IFN signaling by harboring multiple TLRs, including TLR3, 7/8, and 9, in which the acidic lumen enables proteolytic processing and PAMP recognition, thereby ensuring complete receptor activation (Lim and Zoncu, 2016; Fitzgerald and Kagan, 2020). Dysregulation of these lysosomal functions can exacerbate infections and inflammatory responses, with severe consequences for cellular health and host immunity (Fitzgerald and Kagan, 2020; McNab et al, 2015).
While the lysosomal membrane has emerged as a key platform for orchestrating innate immune responses (Heinz et al, 2020; Evavold et al, 2021; Zheng et al, 2021), the direct mechanistic link between lysosomes and IFN induction remains unclear. Given the resource-intensive nature of the IFN program, we reasoned that its effective execution requires precise coordination between lysosomal nutrient-sensing and nuclear gene expression programs. In this study, we identified the lysosomal LAMTOR-Rag GTPase complex as a central regulator that dynamically modulates PRR-induced IFN production, thereby shaping the cell-wide innate immunity. Through distinct nucleotide-loading states, this signaling complex fine-tunes the expression profile of IRFs essential for IFN induction while concurrently activating p38 MAPK to stabilize Ifnb1 mRNA. This dual mechanism integrates transcriptional and post-transcriptional controls to ensure robust IFN responses without altering PRR signaling or relying on the mTORC1 pathway. Our findings reveal a lysosome-specific signaling paradigm with broad implications for understanding the organelle-mediated regulation of innate immunity and provide new directions for the development of host-directed antiviral strategies.
Results
A screen for lysosome signaling factors with antiviral activity
To investigate the molecular link between lysosomes and IFN induction, we aimed to identify lysosome-resident signaling factors involved in antiviral defense. We analyzed a published “master list” of lysosome-associated proteins and identified 50 candidates (~8.2%) annotated with signaling functions in Gene Ontology (GO) and Reactome databases (Shin et al, 2022) (Fig. 1A; Dataset EV1). To assess their antiviral roles, we conducted a targeted shRNA screen in RAW264.7 mouse macrophages, which robustly mount the innate immune response. Using a GFP-reporter Influenza A virus (IAV PR8-NS1-GFP) (Zhao et al, 2022), we measured viral infectivity via GFP fluorescence upon gene knockdown (Fig. 1A). This screen revealed that depletion of the LAMTOR/Ragulator complex (composed of Lamtor1-5 subunits) or its binding partners, Rag GTPases, enhanced IAV infectivity (Fig. 1B). While these proteins are known regulators of mTORC1 signaling and pyroptosis (Zheng et al, 2021; Evavold et al, 2021; Liu and Sabatini, 2020; Goul et al, 2023), their role in antiviral activity is unprecedented. Notably, knockdown of SLC15A4 (a TLR7-9 immune regulator) and TLR3 (a viral nucleic acid sensor) also increased infectivity, validating the robustness of our screening (Kim et al, 2008b; Heinz et al, 2020) (Fig. 1B). We confirmed Lamtor subunit knockdown by immunoblotting and observed elevated GFP fluorescence in depleted cells post-IAV infection using live-cell imaging (Fig. 1C,D). Consistent with this, qRT-PCR showed increased viral mRNA (NS1 and HA) levels in LAMTOR-deficient cells (Fig. EV1A), implicating the LAMTOR complex in antiviral resistance. Given that IFN induction is central to antiviral defense, we focused on the role of LAMTOR-Rag in the IFN pathway.Figure 1. The lysosomal LAMTOR is antiviral and triggers the IFN signature via PRRs.(A) Schematic workflow for targeted shRNA screening in RAW264.7 macrophages to identify lysosomal antiviral factors by assaying IAV viral infectivity using NS1-GFP fluorescence as a readout. (B) Ranking of the target genes based on the fold change of normalized fluorescence from cells expressing the indicated shRNAs in comparison to control cells expressing shRNA against luciferase. Data were mean ± SD, n = 3 biological replicates per knockdown, two independent shRNAs were evaluated for each gene. (C) Immunoblot analysis of the indicated Lamtor proteins in cells depleted of LAMTOR components using shRNA. Actin was used as a loading control. (D) Representative micrographs of RAW macrophages depleted of LAMTOR components following infection with IAV PR8-NS1-GFP at an MOI of 0.25 for 24 h. Scale bars: 100 µm. A minimum of four independent fields of view was obtained for each sample. (E) Schematic diagram depicting the importance of a lipidation site within the N-terminal region of Lamtor1 in controlling the LAMTOR lysosomal localization. 3A, non-lysosomal G2A/C3A/C4A mutant Lamtor1, WT wildtype. (F) Immunoblot analysis of the indicated proteins in Lamtor1 knockout (KO) mouse RAW264.7 macrophage lines reconstituted with vector control or FLAG-tagged human WT and 3A mutant LAMTOR1 accordingly. Actin was used as a loading control. (G) Time-course analysis of IFN-β secretion from Lamtor1 KO, WT, and 3A cells after stimulation with Poly(I:C) (30 µg ml^−1^), R848 (0.5 µg ml^−1^), CpG-B (1 µM), LPS (0.5 µg ml^−1^), or cGAMP (5 µg ml^−1^), as measured by ELISA. Data were mean ± SD, n = 3 biological replicates per group for each time point; two-way repeated measures ANOVA followed by Dunnett’s multiple comparisons test, adjusted P value as indicated. (H) Innate immune signaling in WT and 3A cells in response to PRR stimulation. Cells were lysed and analyzed for the levels of the indicated proteins and phosphorylation status of S6K1 (T389), TFEB (S122), TBK1 (S172), IRF-3 (S396), and STING (S365). (I) qRT-PCR analysis of Ifnb1 expression in Lamtor1 KO, WT, and 3A cells after 14 h of stimulation with PRR agonists. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s multiple comparisons test, adjusted P value as indicated. Mock-treatment controls were shared across PRR stimulation conditions. (J) Principal component analysis (PCA) of the RNA-seq transcriptome of LAMTOR1 WT and 3A cells in the presence or absence of the indicated PRR agonists. n = 3 biological replicates per group, differentially expressed genes (DEGs) with a false discovery rate (FDR) < 0.05, by DESeq2. (K,** L**) Venn diagram (K) and heatmap (L) of the transcriptome revealing the overlap of DEGs in IFN response between LAMTOR1 WT and 3 A cells across different inducers of innate immunity. (M–O) mRNA expression levels of Irf7 (M), Cxcl10 (N), and Mx2 (O) in WT and 3A cells after 14 h of stimulation with PRR agonists. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. (P) Immunoblot analysis of ISGs in WT and 3A cells in response to PRR stimulation. Representative results for A–I, M–P from two independent experiments. Source data are available online for this figure.
Loss of LAMTOR abolishes IFN-β gene transcription
The pentameric LAMTOR complex attaches to lysosomal membranes exclusively through N-terminal lipidation of Lamtor1; disruption of this modification abolishes lysosomal localization (Sancak et al, 2010; Nada et al, 2009) (Fig. 1E). To probe its functional role, we generated Lamtor1 knockout (KO) RAW264.7 macrophages using CRISPR-Cas9 and reconstituted them with lentiviral 3×FLAG-tagged human LAMTOR1 (wildtype or a non-lysosomal 3A mutant). Immunoblotting confirmed expression in these cell lines (Fig. 1F). Since FLAG immunofluorescence was ineffective in RAW cells, we validated the diffuse cytoplasmic distribution of the 3A mutant in stable HeLa cells, consistent with its loss of lysosomal anchoring (Fig. EV1B).
To examine whether LAMTOR broadly regulates innate immunity, we screened several PRR agonists, including polyinosinic-polycytidylic acid (Poly(I:C)) for TLR3, resiquimod (R848) for TLR7/8, CpG-B for TLR9, lipopolysaccharide (LPS) for TLR4, and cGAMP for stimulator of interferon genes (STING). Consistent with its antiviral role (Figs. 1D and EV1A), Lamtor1 depletion in macrophages impaired IFN-β secretion in a time-dependent manner across all PRR stimulation. This defect was rescued by WT LAMTOR1 but not by the 3A mutant, underscoring the requirement for lysosomal localization (Fig. 1G). We then examined whether LAMTOR modulates PRR signaling pathways. While 3A-expressing cells retained normal PRR expression and intact downstream signaling, including phosphorylation of TANK-binding kinase 1 (TBK1), STING, and IRF-3 (Figs. EV1C and 1H), they displayed a pronounced loss of Ifnb1 and Il1b mRNA induction, but not Tnfa, regardless of the PRR stimulus (Figs. 1I and EV1D,E). This dissociation between intact signaling and disrupted transcription reveals a previously unrecognized role for lysosomal LAMTOR in regulating nuclear gene expression, independent of canonical PRR pathway activation.
Pattern recognition receptors require LAMTOR to instruct the IFN signature
To investigate the transcriptional response to LAMTOR loss, we performed bulk RNA-seq on WT and 3A cells with or without PRR stimulation. Principal component analysis (PCA) revealed a clear shift in global gene expression after 14-h PRR agonist treatment, along with a significant divergence between 3A and WT cells even under mock conditions (Fig. 1J). Gene set enrichment analysis (GSEA) showed that WT, but not 3A, cells exhibited strong enrichment for innate immune pathways, including activation of innate immune response, defense response to virus, and cellular response to interferon-beta, consistent with impaired IFN-β induction in LAMTOR-deficient cells (Fig. EV2A–C). Differential expression analysis (FDR <0.05, fold change ≥2) identified 75 differentially expressed genes (DEGs) functionally linked to innate immunity, viral infection, and antiviral IFN responses, as confirmed by GO, KEGG, and Reactome analyses (Figs. 1K and EV2D–F). Strikingly, all of these DEGs were interferon-stimulated genes (ISGs) annotated in the Interferome database (Rusinova et al, 2013), with PRR-responsive expression strictly dependent on LAMTOR (Fig. 1L). Consistent with these findings, PRR-challenged 3 A cells exhibited blunted ISG responses, as evidenced by the reduced mRNA levels of Irf7, Cxcl10, and Mx2 (Fig. 1M–O) and diminished protein expression of ISG15, MDA-5, IFI44, and MOV-10 (Fig. 1P). Notably, while IFNs and IFN-regulated cytokines were downregulated, C-C motif chemokines and TNF ligand superfamily members were significantly upregulated in 3A cells, indicating that cytokine induction was not globally impaired (Fig. EV2G). Together, these results suggest that LAMTOR, via lysosomal localization, governs cell-autonomous innate immune response by selectively regulating IFN induction.
LAMTOR-Rag partnership licenses an antiviral state
LAMTOR anchors heterodimeric Rag GTPases to the lysosomal membrane, enabling mTORC1 activation in response to nutrients (Sancak et al, 2010; Kim et al, 2008a). Given this canonical function, we investigated whether Rag GTPases, mTORC1 activity, or both, contribute to the innate immune response. Our shRNA screen identified Rag GTPases, including RagA, RagB, and RagC paralogs, as critical hits, in addition to Lamtor family proteins (Fig. 1B; Dataset EV1). To determine whether the LAMTOR-Rag interaction is required for PRR-induced immunity, we leveraged prior structural insights, which revealed two key Lamtor1 motifs (NIV: N64/I66/V68; LVV: L154/V155/V156) essential for Rag GTPase binding (de Araújo et al, 2017) (Fig. EV3A–C). We generated alanine-substituted mutants (NIV and LVV) and reconstituted them in Lamtor1 KO cells. FLAG immunoprecipitation and immunofluorescence confirmed that these mutations abolished LAMTOR-Rag interactions without disrupting LAMTOR assembly or lysosomal localization (Fig. EV3D,E). In contrast, the 3A mutant impaired lysosomal targeting, but not complex integrity (Figs. EV1B and EV3E). Critically, cells expressing NIV or LVV mutants exhibited blunted IFN-β induction at both the mRNA and protein levels (Fig. EV3F,G). Thus, both physical LAMTOR-Rag interactions and lysosomal localization are crucial for PRR-induced immunity.
Consistent with this, CRISPR deletion of either Rraga or Rragc in cells abrogated PRR responses, phenocopying Lamtor1 loss, as evidenced by suppressed Ifnb1 mRNA expression and IFN-β secretion compared with parental RAW cells (Fig. 2A,B). Similar to Lamtor1 KO cells, Rraga KO and Rragc KO cells exhibited markedly reduced ISG protein levels upon PRR stimulation (Fig. 2C–E). Importantly, the loss of these genes in macrophages did not alter fundamental lysosomal features, including the maintenance of acidic luminal pH, ruling out lysosomal dysfunction as a confounding factor (Ma et al, 2017) (Fig. EV3H). These findings demonstrate that LAMTOR, in partnership with Rag GTPases, mediates the PRR-dependent IFN program.Figure 2LAMTOR-Rag mediates the IFN program in macrophages and in vivo.(A) Ifnb1 mRNA expression in parental, Lamtor1 KO, Rraga KO, and Rragc KO cells after stimulation with PRR agonists. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. (B) IFN-β secretion from parental, Lamtor1 KO, Rraga KO, and Rragc KO cells after stimulation with PRR agonists, as measured by ELISA. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls in (A, B) were shared across PRR stimulations, except for cGAMP stimulation, which used separate mock controls to account for reagent lot variability. (C–E), Immunoblot analysis of the indicated ISGs in parental, Lamtor1 KO (C), Rraga KO (D), and Rragc KO (E) cells in response to PRR stimulation. (F) Schematic diagram of the Rragc-floxed allele and generation of Rragc^Lyz2^ mice. Exons 2–3 of Rragc are flanked by two loxP sites. (G) Genotyping (top) and immunoblotting (bottom) showing the myeloid-specific deletion of RagC in Rragc^Lyz2^ mice. Brain tissue was used as a negative control. BMDM, bone marrow-derived macrophage. (H) Survival rate (top) and body weight change (bottom) were monitored following PR8 infection in Rragc^f/f^ and Rragc^Lyz2^ mice. Data were mean ± SD of body weight measurements (bottom), n = 10 mice per group. (I) Viral titers in bronchoalveolar lavage fluid were quantified using the TCID_50_ assay 3 d after infection. Data were mean ± SD, n = 5 mice per group. (J) IFN-β levels in bronchoalveolar lavage fluid were measured 3 d after infection. Data were mean ± SD, n = 5 mice per group. (K) Representative hematoxylin and eosin staining of lung tissue sections from PR8-infected mice. n = 5 mice per group. Representative results for A–E, G–K from two independent experiments. Source data are available online for this figure.
To assess the physiological relevance of LAMTOR-Rag in host defense against viral infection, we generated myeloid-specific Rragc KO (Rragc^Lyz2^) mice by crossing Rragc-floxed mice (Rragc^flox/flox^) with Lyz2-Cre mice (Fig. 2F). Genotyping PCR and immunoblotting confirmed the efficient deletion of RagC in bone marrow-derived macrophages (BMDMs) from Rragc^Lyz2^ mice (Fig. 2G). We infected 8-week-old Rragc^Lyz2^ mice and littermate Rragc^f/f^ controls with H1N1 (PR8) via intranasal inoculation. Strikingly, Rragc^Lyz2^ mice exhibited significantly higher mortality than the controls (Fig. 2H). Correlating with this phenotype, bronchoalveolar lavage fluid from infected Rragc^Lyz2^ mice contained elevated viral loads and reduced IFN-β levels, indicating impaired antiviral immunity (Fig. 2I,J). Furthermore, histopathological analysis revealed exacerbated acute lung injury and immune cell infiltration in Rragc^Lyz2^ mice (Fig. 2K). Together, these findings establish a critical in vivo role for LAMTOR-Rag in innate immune protection against viral infection.
The mechanism by which LAMTOR-Rag induces IFN is separable from mTORC1
Given the established function of the LAMTOR-Rag complex in regulating mTORC1 signaling, we first investigated whether downstream mTORC1 kinase activity contributes to IFN induction. While LAMTOR-Rag deficiency modestly reduced phosphorylation of canonical mTORC1 substrates (S6K1, 4E-BP1, and S6) in resting cells (Fig. EV4A), pharmacological inhibition revealed a more complex role for mTORC1 in PRR-induced innate immunity. Pretreatment with either rapamycin (an allosteric mTORC1 inhibitor) or Torin 1 (an ATP-competitive mTORC1/2 inhibitor) prior to PRR stimulation demonstrated that while mTORC1 inhibition selectively attenuated TLR3-induced Ifnb1 gene expression (Fig. EV4B), it broadly suppressed IFN-β secretion across all PRR stimuli (Fig. EV4C). Both inhibitors produced similar phenotypic effects and effectively blocked mTORC1 activity, with Torin 1 additionally inhibiting phosphorylation of the noncanonical lysosomal mTORC1 substrate TFEB (Fig. EV4D). These results indicate that mTORC1 primarily regulates IFN-β production at the translational level (Liu and Sabatini, 2020), consistent with its established role in global protein synthesis and secretion (Narita et al, 2011; Nüchel et al, 2021; Kaeser-Pebernard et al, 2022), while exerting PRR-specific effects on transcription.
To address potential off-target effects of prolonged pharmacological inhibition, we employed CRISPR-mediated Rptor deletion. As Raptor is essential for mTORC1 assembly and activity (Liu and Sabatini, 2020; Goul et al, 2023), its loss significantly reduced mTORC1 activity (Fig. EV4E). Mirroring the pharmacological inhibition, Rptor KO cells failed to respond to TLR3, selectively impairing Poly(I:C)-induced Ifnb1 mRNA expression and IFN-β secretion (Fig. EV4F,G). Interestingly, Raptor depletion enhanced IFN-β production upon TLR4, TLR7/8, and TLR9 activation. In contrast, cGAMP treatment reduced Ifnb1 mRNA levels in Rptor KO cells, although protein levels were less affected. Despite these PRR-specific differences, Rptor KO cells maintained intact ISG protein expression (Fig. EV4H). Unlike the context-dependent effects of Raptor depletion, Lamtor1 deficiency uniformly abolished IFN responses across all stimulation conditions (Fig. EV4F,G). Considering the metabolic demands associated with IFN induction, we tested whether constitutive mTORC1 activation (via lysosome-targeted Raptor-Rheb15) could rescue IFN defects in Lamtor1-deficient cells (Sancak et al, 2010) (Fig. EV4I–K). However, elevated mTORC1 activity failed to restore IFN responses, demonstrating that LAMTOR’s role in IFN induction is mTORC1-independent.
The transcription factors TFEB and TFE3, which regulate lysosomal biogenesis, are well-established noncanonical substrates of mTORC1. Their phosphorylation, which strictly depends on the LAMTOR-Rag axis, results in cytosolic retention and functional inactivation (Cui et al, 2023; Napolitano et al, 2020; Fernandes et al, 2024). Consistent with this, in RAW macrophages, ablation of Lamtor1 lysosomal anchoring or LAMTOR-Rag deficiency abolished TFEB phosphorylation under both steady-state and PRR-stimulated conditions (Figs. 1H and EV5A–C). A concurrent reduction in TFEB/TFE3 protein levels was also observed, which we attributed to their nuclear accumulation and subsequent incomplete extraction with the lysis buffer used in this study (Figs. 1H and EV5A–C).
Based on this established regulatory mechanism, we hypothesized that TFEB and TFE3 might suppress IFN transcriptional activation following LAMTOR-Rag ablation. Contrary to this expectation, genetic ablation of Tfe3 completely abolished IFN induction (Fig. EV5D–F), a phenotype that was recapitulated when Tfeb was deleted in both parental and Lamtor1 KO cells (Fig. EV5G–I). Thus, these data indicate an unexpected positive and essential role for TFEB and TFE3 in IFN regulation, independent of the LAMTOR-Rag–mTORC1 axis.
In summary, our findings demonstrate that LAMTOR-Rag drives IFN induction via a mechanism distinct from mTORC1 and its lysosomal substrates, TFEB and TFE3. Furthermore, mTORC1 activation downstream of LAMTOR-Rag is insufficient to dictate the PRR-induced IFN signature.
LAMTOR-Rag is an integral part of cellular innate immunity
Given the profound defects in IFN induction upon LAMTOR-Rag loss, we compared the cells disrupted for MyD88 or UNC93B1 using CRISPR. MyD88 is an essential adapter for TLR4, TLR7/8, and TLR9 signaling (Fitzgerald and Kagan, 2020), while UNC93B1 mediates trafficking of TLR3, TLR7/8, and TLR9 to endolysosomes for nucleic acid sensing (Kim et al, 2008b; Majer et al, 2019) (Fig. EV6A). Unlike Lamtor1 loss, which entirely abolished IFN responses, MyD88 or UNC93B1 depletion had distinct PRR-dependent effects (Fig. EV6B–E). Myd88 deletion abrogated TLR4/7/8/9 responses, but increased sensitivity to TLR3 and STING pathways, as shown by elevated Ifnb1 induction and IFN-β release upon Poly(I:C) and cGAMP treatment (Fig. EV6B–D). Conversely, Unc93b1 ablation impaired endolysosomal TLRs, while enhancing cGAMP responses (Fig. EV6B,C,E). We further validated the role of LAMTOR in THP-1 cells, a human monocytic line with intact TLR4/7/8 signaling. Lamtor1 silencing with three distinct shRNAs via lentiviral transduction significantly impaired IFNB1 induction, IFN-β secretion, and ISG transcription (ISG15, RSAD2, and IFITM1) upon R848/LPS stimulation (Fig. EV6F–K).
In addition to the endomembrane-bound PRRs, MAVS, a mitochondrial antiviral signaling hub, is activated by cytosolic dsRNA sensors RIG-I and MDA-5 (McNab et al, 2015; Hoffmann et al, 2015; Iwasaki, 2012). MAVS oligomers recruit TBK1 to phosphorylate IRF-3, driving its nuclear translocation and IFN induction (Honda and Taniguchi, 2006) (Fig. EV7A). To test the role of LAMTOR in MAVS-mediated IFN response, we transfected RAW cells with Poly(I:C) to specifically activate MAVS. Both Lamtor1 KO and 3A cells showed reduced IFN-β production and impaired ISG expression (Fig. EV7B,C). Similar defects were observed with Rraga or Rragc deletion (Fig. EV7D,E). Unlike PRRs, which are mainly restricted to the immune cells, MAVS functions ubiquitously (Fitzgerald and Kagan, 2020). In non-immune HEK293T cells, Lamtor1 knockdown blocked ISRE-driven reporter activity (Fig. EV7F,G) and disrupted MAVS-induced IFN responses (Fig. EV7H–K). Concordantly, LAMTOR1 KO enhanced IAV infectivity as quantified by NS1-GFP flow cytometry (Fig. EV7L,M). Together, these results establish the lysosomal LAMTOR-Rag complex as a central regulator of IFN program and cell-intrinsic immunity.
Nucleotide states of RagA/C regulate the IFN program
Rag GTPases function as obligate heterodimers (RagA/B with RagC/D), where coordinated nucleotide-loading states dictate their activity in mTORC1 lysosomal recruitment. We hypothesized that immune cells, particularly macrophages, exploit these nucleotide-dependent conformational changes to modulate the IFN program. To test this, we reconstituted Rraga KO cells with FLAG-tagged wild-type (WT), GTP-bound (Q66L), or GDP-bound (T21L) RagA (Figs. 3A,B and EV8A,B). While RagA WT restored IFN responses (Ifnb1 induction, IFN-β secretion, ISG expression), GTP-bound RagA further enhanced these effects, and GDP-bound RagA failed to rescue the phenotype. Conversely, Rragc KO cells expressing GTP-bound RagC (Q119L) exhibited complete IFN deficiency, whereas WT or GDP-bound RagC (S74N) restored PRR sensitivity, with RagC-GDP cells showing augmented responses (Figs. 3C,D and EV8C,D). These findings demonstrate that distinct Rag nucleotide states exert opposing effects on the IFN program, consistent with their reciprocal regulation of TFEB phosphorylation, further reinforcing the TFEB/TFE3-independent mechanism (Fig. EV8E–H).Figure 3. The nucleotide-loading state of Rag modulates IRF expression to regulate IFN induction.(A) Ifnb1 mRNA expression in Rraga KO cells reconstituted with FLAG-tagged WT, GTP-bound (Q66L), or GDP-bound (T21L) forms of mouse RagA after stimulation with PRR agonists. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls were shared across PRR stimulations, except for cGAMP. (B) IFN-β secretion from cells reconstituted with the WT, GTP-bound, and GDP-bound forms of mouse RagA after stimulation with PRR agonists, as measured by ELISA. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls were shared across PRR stimulation conditions. (C) Ifnb1 mRNA expression in Rragc KO cells reconstituted with FLAG-tagged WT, GTP-bound (Q119L), or GDP-bound (S74N) forms of mouse RagC after stimulation with PRR agonists. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. (D) IFN-β secretion from cells reconstituted with the WT, GTP-bound, and GDP-bound forms of mouse RagC after stimulation with PRR agonists, as measured by ELISA. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls in (C, D) were shared across PRR stimulations, except for cGAMP stimulation, which used separate mock controls to account for reagent lot variability. (E) Heatmap comparing the gene expression profiles of IRF family members between LAMTOR1 WT and 3A cells in response to PRR stimulation. (F) Immunoblot analysis of the indicated TFs in parental, Lamtor1 KO, WT, and 3A cells. (G) Immunoblot analysis of the indicated TFs in parental, Lamtor1 KO, Rraga KO, and Rragc KO cells. (H,** I**) Immunoblot analysis of the indicated TFs in Rraga KO (H) or Rragc KO cells (I) reconstituted with recombinant cDNAs encoding the corresponding WT, GTP-bound, or GDP-bound variants. Representative results for (A–D, F–I) from two independent experiments. Source data are available online for this figure.
The proper expression, activation, and recruitment of IRFs are essential for IFN-β induction, and their disruption likely underlies the suppressed IFN program in LAMTOR-Rag-deficient cells. RNA-seq analysis of WT and 3 A cells revealed a shared differential TF expression pattern after PRR stimulation. Specifically, 3 A cells showed markedly reduced mRNA levels of Irf1, Irf5, Irf7, Stat1, Stat2, and Irf9, which are key positive regulators of IFN and ISG responses (Figs. 3E and EV8I). Consistent with this transcriptional profile, the loss of LAMTOR-Rag universally diminished IRF protein levels, with IRF-5 and IRF-7 being the most affected, as validated in 3 A, Lamtor1 KO, Rraga KO, and Rragc KO cells (Fig. 3F,G). Notably, Stat1/2 protein levels remained stable despite transcript reduction, and exogenous IFN-β induced normal Stat1 phosphorylation in LAMTOR-Rag-deficient cells (Fig. EV8J), demonstrating cell-autonomous control of IFN induction by lysosomal LAMTOR-Rag signaling.
Given the prominent role of Rag GTPases in the IFN program (Figs. 2A–K and 3A–D), we hypothesized that the nucleotide-loading states of Rag heterodimers might control the expression of IFN-inducing IRFs. Indeed, reintroducing either WT RagA or GTP-bound RagA into Rraga KO cells restored IRF-5 and IRF-7 protein levels (Fig. 3H). Mirroring this, WT RagC or GDP-bound RagC rescued IRF expression in Rragc KO cells (Fig. 3I). In contrast, GDP-bound RagA and GTP-bound RagC suppressed IRFs in a dominant-negative manner (Fig. 3H,I). Together, these data indicate that (1) impaired IFN induction and ISG responses in LAMTOR-Rag-deficient cells stem from IRF depletion and (2) the nucleotide-loading configuration of Rag heterodimers is essential for maintaining IRF levels, thereby driving the IFN program.
Because Rag nucleotide states reflect the cellular nutrient status, we investigated whether nutrient availability governs PRR-induced IFN expression. Indeed, LAMTOR-Rag-deficient macrophages showed impaired amino acid-dependent mTORC1 activation, as evidenced by reduced S6K1 phosphorylation upon refeeding (Fig. EV8K). Remarkably, amino acid deprivation abolished PRR-induced Ifnb1 transcription—except for cGAMP stimulation, which paradoxically increased Ifnb1 levels (Fig. EV8L). This exception reflected STING pathway hyperactivation during starvation, as shown by elevated TBK1, STING, and IRF-3 phosphorylation levels (Fig. EV8M). Despite the enhanced cGAMP responses, all PRR pathways required amino acids for IFN-β secretion and ISG production (Fig. EV8N,O). Thus, Rag heterodimers coordinate nutrient sensing with IRF expression to dynamically regulate IFN responses in macrophages.
p38 MAPK acts downstream of LAMTOR-Rag for IFN induction
As the most potent transcriptional activator of IFN-β among IRF family members (Fitzgerald and Kagan, 2020; Honda and Taniguchi, 2006), IRF-3 protein expression remained stable despite LAMTOR-Rag deficiency or alterations in Rag GTPase nucleotide state. This stability stood in stark contrast to the marked reduction observed for other IRFs (Fig. 3F–I). Furthermore, although Lamtor1 deletion impaired IFN induction and the ISG response (Fig. EV7B–E), it did not affect MAVS-induced IRF-3 phosphorylation, dimerization, or nuclear translocation, as confirmed by nuclear-cytoplasmic fractionation (Fig. 4A,B). Exogenous expression of both IRF-5 and IRF-7 in Lamtor1 KO cells produced only a partial rescue of IFN responses (Fig. 4C–E). These findings indicate that canonical PRR-mediated IRF activation is necessary but insufficient for robust IFN production, revealing an essential requirement for a concurrent lysosome-dependent signaling pathway to fully execute the IFN program.Figure 4. Loss of LAMTOR-Rag function abolishes p38 MAPK activation.(A) Immunoblot analysis of IRF-3 activation in Lamtor1 KO cells transfected with Poly(I:C). Cells were lysed and analyzed for levels of the indicated proteins, phosphorylation status of IRF-3 (S396), and dimerization of IRF-3. (B) Nuclear translocation of IRF-3 in Lamtor1 KO cells transfected with Poly(I:C) as analyzed by nuclear-cytoplasmic fractionation. Tubulin was used as a cytosolic control, whereas PARP-1 was used as a nuclear control. (C) Ectopic expression of IRF5 and IRF7 in Lamtor1 KO cells. (D, E) Induction of Ifnb1 transcript (D) and IFN-β release (E) from parental, Lamtor1 KO, and Lamtor1 KO cells stably expressing IRF-5 and IRF-7 after PRR stimulation. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls in (D, E) were shared across PRR stimulations. (F) Schematic overview of the ATAC-seq workflow for chromatin accessibility analysis in WT and 3A cells in response to cGAMP stimulation, with n = 3 biological replicates per group. (G) Chromatin accessibility and functional enrichment of differentially accessible regions. Heatmaps display normalized ATAC-seq signal intensity across a 10 kb region (±5 kb) centered on genomic loci within the indicated clusters (left). GO term analysis of biological processes significantly associated with the ‘up’ and ‘stable’ clusters in 3A cells (right). Enrichment was calculated using rGREAT with a hypergeometric test (FDR <0.05). The top ten enriched terms are shown. (H) Coordinated transcriptional and chromatin remodeling upon LAMTOR loss. Genes concurrently identified by RNA-seq and ATAC-seq are shown in the Venn diagram (left). The biological processes most significantly associated with the upregulated 3A-specific gene cluster are listed (right). GO enrichment was calculated via hypergeometric distribution using Clusterfiler (FDR <0.05). (I) HOMER motif TF analysis of the indicated gene clusters. (J) ATAC-seq analysis highlighting accessible chromatin regions of Ifnb1 in WT and 3A cells under mock- and cGAMP-treated conditions. n = 3 biological replicates per group. (K) Schematic representation illustrating communal changes in chromatin accessibility of the indicated genes in the absence of LAMTOR. (L) Immunoblot analysis of phosphorylation status of ERK1/2 (T202/Y204), JNK (T183/Y185), and p38 (T180/Y182) in Lamtor1 KO, Rraga KO, and Rragc KO cells. (M) Immunoblot analysis of the phosphorylation status of the indicated MAPKs, ATF-2 (T71), and c-Jun (S73) in Rraga KO cells and Rragc KO cells reconstituted with WT or its variants. Representative results for (A–E and L, M) from two independent experiments. Source data are available online for this figure.
To investigate the altered TF profile caused by LAMTOR loss, we performed ATAC-seq in WT and 3 A cells with or without cGAMP stimulation (Buenrostro et al, 2015) (Figs. 4F and EV9A). While cGAMP treatment robustly increased chromatin accessibility in WT cells, this response was severely blunted in 3A cells (Fig. 4G). Integrated RNA-seq/ATAC-seq analysis revealed 3A-specific gene signatures linked to open chromatin regions (Fig. 4H). GO enrichment analysis identified significant regulation of MAPK activity (Fig. 4G,H), and HOMER motif scanning of 3A-specific accessible regions predicted predominant AP-1 family binding (Fig. 4I), consistent with the known MAPK regulation of AP-1 (Canovas and Nebreda, 2021). Visual inspection of the key loci (Ifnb1, Irf5, Irf7, and Mda5) revealed that LAMTOR loss altered the baseline chromatin accessibility, which became unresponsive to cGAMP stimulation (Figs. 4J,K and EV9B–D). These findings suggest that LAMTOR deficiency may impair IFN induction by disrupting MAPK signaling and downstream AP-1 activation.
Supporting this hypothesis, depletion of Lamtor1, RagA, or RagC specifically abolished the phosphorylation of p38 MAPK—but not that of other MAPK family members, JNK and ERK1/2—under both basal and PRR-stimulated conditions (Figs. 4L and EV9E–G). Moreover, the active Rag state (RagA-GTP/RagC-GDP), which promotes the IFN program, enhanced p38 phosphorylation compared to the inactive state (RagA-GDP/RagC-GTP) (Fig. 4M). Notably, this regulation was specific to p38, as phosphorylation of its canonical downstream targets, ATF-2 and c-Jun, remained unaltered (Fig. 4M). Among the four p38 MAPK isoforms, our RNA-seq data identified Mapk14 (p38α) and Mapk11 (p38β) as the predominant isoforms expressed in murine macrophages. To corroborate the critical role of p38 in IFN induction, we generated Mapk14 KO and Mapk11 KO cells and found that they were defective in PRR-stimulated IFN responses, while maintaining intact sensitivity to exogenous IFN-β, thereby recapitulating the phenotype of LAMTOR-Rag deficiency (Fig. EV10A–D). Importantly, canonical p38 activators, including osmotic stress (sorbitol) and oxidative stress (sodium arsenite), failed to trigger IFN responses in parental cells as well as in cells lacking Lamtor1 or MAPK14/11 (Fig. EV10E,F). Collectively, these data establish that p38 MAPK activation is necessary for PRR signaling to drive IFN production.
Rag nucleotide states are modulated by GATOR1 (acting as a RagA/B GAP) and the folliculin complex (FLCN-FNIP1/2 as a RagC/D GAP) (Perera and Zoncu, 2016; Goul et al, 2023). Since the active Rag conformation requires GTP-bound RagA and GDP-bound RagC, we predicted that disrupting these regulators would have opposing effects on IFN induction. Indeed, Flcn KO cells showed reduced p38 MAPK phosphorylation and abolished Ifnb1 transcription and IFN-β secretion, phenocopying the Rragc deletion (Fig. 5A–C). Conversely, depletion of the GATOR1 component DEPDC5 enhanced p38 phosphorylation and amplified IFN production, consistent with RagA-GTP accumulation (Fig. 5A–C). Most importantly, neither RagA-GTP nor RagC-GDP expression rescued PRR-induced IFN responses in Mapk14/Mapk11 DKO cells (Fig. 5D–F), positioning p38 as the essential downstream effector of Rag GTPases. Collectively, these findings support a model wherein p38 MAPK phosphorylation is dynamically regulated by Rag GTPase nucleotide states, which in turn are governed by the antagonistic actions of their cognate GAPs (Fig. 5G).Figure 5. Rag nucleotide-loading state controls p38 MAPK activation necessary for IFN induction.(A) Immunoblot analysis of p38 MAPK phosphorylation in Rragc KO, Flcn KO, and Depdc5 KO cells compared with parental cells. (B) Ifnb1 mRNA expression in parental, Flcn KO, and Depdc5 KO cells after PRR stimulation. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls were shared across PRR stimulations, except for cGAMP. (C) IFN-β secretion from parental, Flcn KO, and Depdc5 KO cells after PRR stimulation, as measured by ELISA. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls were shared across PRR stimulations, except for cGAMP and Poly(I:C)/FuGENE. (D) Ectopic expression of FLAG-tagged RagA-GTP (Q66L) or RagC-GDP (S74N) in Mapk14/11 DKO cells. (E) Ifnb1 mRNA expression in Mapk14/11 DKO cells expressing FLAG-tagged RagA-GTP or RagC-GDP after PRR stimulation. Data are mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. (F) IFN-β secretion from Mapk14/11 DKO cells expressing FLAG-tagged RagA-GTP or RagC-GDP after PRR stimulation. Data were mean ± SD, n = 3 biological replicates per group; two-way ANOVA followed by Tukey’s test, adjusted P value as indicated. Mock-treatment controls in (E, F) were shared across PRR stimulations, except for cGAMP. (G) Schematic showing epistasis across the GATOR1 and FLCN-FNIP1/2 complex, RagA/C GTPases, and p38 MAPK. Representative results for (A–F) from two independent experiments. Source data are available online for this figure.
p38 MAPK is recruited to lysosomes by FLCN to stabilize Ifnb1 mRNA
To determine whether the catalytic activity of p38 MAPK is essential for IFN regulation, we engineered cells stably expressing MAPK14 WT, a kinase-dead mutant (D168A), or a constitutively active mutant (D176A/F327S) (Fig. EV10G). Dominant-negative inhibition by the D168A mutant robustly suppressed IFN production at both the mRNA and protein levels, whereas the constitutively active mutant did not enhance IFN responses beyond WT levels (Fig. EV10H,I). These data imply that p38 kinase activity is necessary but not limiting for IFN induction. Since p38 regulates IFN independently of its canonical nuclear targets (ATF-2/c-Jun), we investigated its subcellular localization. Nuclear-cytoplasmic fractionation revealed predominant cytoplasmic p38 localization, with no detectable nuclear accumulation following PRR stimulation, irrespective of LAMTOR status (Fig. 6A). Given that p38 operates downstream of Rag GTPases, we postulated its association with lysosomal membranes. To test this, we isolated native phagolysosomes from macrophages using polymer-based magnetic beads, following an established protocol (Majer et al, 2019) (Fig. 6B). Immunoblot analysis of MAPK family members showed that only p38 MAPK and ERK1/2 were detectable in purified phagolysosomes (Fig. 6C). Notably, Lamtor1 depletion, which severely disrupts lysosomal Rag GTPase localization, caused a modest but specific decrease in phagolysosomal p38 MAPK levels, without affecting ERK1/2 (Fig. 6C). Furthermore, PRR stimulation significantly increased lysosomal p38 abundance without altering mTOR or Raptor levels (Fig. EV10J).Figure 6p38 MAPK is required to regulate Ifnb1 mRNA stability.(A) Nuclear-cytoplasmic fractionation assays to determine the p38 MAPK localization in response to PRR stimulation. (B) Schematic workflow for the isolation of native phagolysosomes. (C) Immunoblot analysis of phagolysosomes isolated from parental and Lamtor1 KO cells. WCL whole cell lysate. (D) Immunoblot analysis of phagolysosomes isolated from parental, Flcn KO, and Depdc5 KO cells. The asterisk (*) indicates a non-specific band detected by the FLCN antibody. (E) Interaction between mBaoJin-MAPK14 and FLAG-tagged mouse FLCN or the indicated variants, as assessed by FLAG immunoprecipitation in HEK293T cells. (F) Representative micrographs showing co-localization of mBaoJin-MAPK14 with LysoView-positive lysosomes. Scale bars: 10 µm. (G) Measurements of Ifnb1 mRNA decay using Roadblock-qPCR in parental, Lamtor1 KO, and Mapk14/11 DKO cells. Data were mean ± SD, n = 3 biological replicates per group. (H) Measurements of Ifnb1 mRNA decay using Roadblock-qPCR in parental, Depdc5 KO, Mapk14/11 DKO, and Depdc5/Mapk14/Mapk11 TKO cells. Data were mean ± SD, n = 3 biological replicates per group. (I) Representative micrographs of RAW macrophages of the indicated genotypes following infection with IAV PR8-NS1-GFP at an MOI of 0.25 for 24 h. Scale bars: 100 µm. A minimum of four independent fields of view was obtained for each sample. (J) Schematic representation of the dual mechanism by which the lysosomal LAMTOR-Rag complex regulates IFN production. Representative results for (A–I) from two independent experiments. Source data are available online for this figure.
To determine whether the nucleotide-loading state of Rag GTPases regulates the lysosomal recruitment of p38, we analyzed Flcn KO and Depdc5 KO cells (Fig. 6D). DEPDC5 deficiency, which induces constitutive Rag GTPase and mTORC1 activity, significantly enhanced the phagolysosomal recruitment of both Raptor and p38. Conversely, FLCN depletion completely abolished p38 localization to phagolysosomes (Fig. 6D). Co-IP assays in HEK293T cells demonstrated a direct interaction between MAPK14 and FLCN that requires the Longin and DENN domains (Lawrence et al, 2019) (Fig. 6E). Consistent with Rag GTPase regulation, live-cell imaging in RAW macrophages showed that mBaoJin-MAPK14 co-localized with LysoView-labeled lysosomes under nutrient-replete conditions, and this co-localization was abolished upon amino acid withdrawal (Fig. EV10K). Furthermore, mBaoJin-MAPK14 failed to co-localize with lysosomes in Flcn KO cells, in contrast to its punctate localization in Depdc5 KO and control cells (Fig. 6F). Together, these results establish FLCN as a critical regulator that couples the nucleotide state of Rag GTPases to the lysosomal recruitment and subsequent phosphorylation of p38 MAPK.
Given the short half-life of antiviral Ifnb1 mRNA and its reliance on RNA-binding proteins for stability and translation (Piccirillo et al, 2014), we hypothesized that extranuclear p38 MAPK might regulate its post-transcriptional fate. We employed Roadblock-qPCR to analyze Ifnb1 mRNA decay kinetics (Fig. 6G). After 14-h Poly(I:C) transfection, cells were pulse-labeled with 4-thiouridine (4sU) at varying time points. Nascent 4sU-containing transcripts were then selectively depleted via N-ethylmaleimide (NEM) modification during cDNA synthesis, allowing quantification of pre-existing (non-4sU-labeled) Ifnb1 mRNA decay by qPCR (Watson and Thoreen, 2022).
Remarkably, Ifnb1 mRNA degradation was significantly accelerated in both Lamtor1 KO and Mapk14/Mapk11 DKO cells compared to parental cells (Fig. 6G). In contrast, DEPDC5 deficiency substantially prolonged Ifnb1 mRNA half-life, an effect attributable to both enhanced transcription and elevated p38 phosphorylation that was fully reversed by Mapk14/Mapk11 deletion (Fig. 6H). Accordingly, p38 MAPK ablation in Depdc5 KO cells counteracted the heightened IFN responses (Fig. EV11A–C). These data strongly support a key post-transcriptional role for p38 MAPK in stabilizing Ifnb1 mRNA downstream of the LAMTOR-Rag complex. Crucially, overexpression of the constitutively active p38 mutant in Lamtor1 KO, Rraga KO, and Rragc KO cells failed to rescue IFN deficiency, underscoring the parallel requirement for transcriptional activation alongside this post-transcriptional regulation (Fig. EV11D–F). Finally, impaired IFN production in Lamtor1 KO, Flcn KO, and Mapk14/11 DKO cells enhanced viral susceptibility, whereas the potent antiviral state in DEPDC5-null cells was strictly p38-dependent (Fig. 6I).
Discussion
Our study establishes lysosomes as critical signaling hubs for IFN induction, revealing that the LAMTOR-Rag GTPase complex orchestrates IFN-β production through two synergistic mechanisms: (1) transcriptional priming of IRFs (e.g., IRF-5/7) and (2) p38 MAPK-mediated stabilization of Ifnb1 mRNA (Fig. 6J). Disruption of this axis abrogates IFN-β production across multiple PRR pathways in macrophages and impairs antiviral defense in mice, positioning lysosomes as noncanonical regulators of innate immunity. While lysosomes are classically viewed as degradative organelles, their emerging role in immune signaling has been attributed to TLR housing and autophagy-mediated pathogen clearance (Deretic, 2021; Fitzgerald and Kagan, 2020). Our work expands this paradigm by identifying the LAMTOR-Rag complex as a bona fide signaling module that bridges lysosomal membranes to IFN induction, independent of mTORC1 signaling or canonical PRR activation. The strict dependence of IFN-β production on lysosome-localized LAMTOR underscores spatial regulation as a defining feature of this pathway. Notably, the dissociation between intact PRR signaling (TBK1/IRF3 phosphorylation) and impaired IFN responses in LAMTOR-deficient cells suggests that lysosomes serve as an indispensable checkpoint to ensure coordinated transcriptional and post-transcriptional control.
Beyond their canonical roles in nutrient sensing and mTORC1 recruitment (Liu and Sabatini, 2020; Goul et al, 2023; Sancak et al, 2010; Kim et al, 2008a), Rag GTPases emerge as immune regulators whose nucleotide-loading states dictate IFN output via three mechanisms. First, Rag nucleotide cycling governs the expression of IFN-inducing transcription factors IRF-5 and IRF-7, thereby priming transcriptional activation of the IFN program. Second, it dynamically regulates p38 MAPK activation, which is critical for post-transcriptional stabilization of Ifnb1 mRNA. Third, nutrient availability signals fine-tune Rag nucleotide loading (Kim et al, 2008a; Sancak et al, 2010), ensuring that resource-intensive IFN production is synchronized with the cellular anabolic capacity. Consequently, the active Rag state (RagA-GTP/RagC-GDP) promotes IFN induction, whereas the inactive Rag state (RagA-GDP/RagC-GTP) suppresses it. This exquisite role highlights the context-specific functions for Rag GTPases, akin to their recently described involvement in pyroptosis (Evavold et al, 2021; Zheng et al, 2021). Furthermore, LAMTOR components have been implicated in lysosomal MEK1-ERK signaling (Schaeffer et al, 1998; Nada et al, 2009; Teis et al, 2006; Wunderlich et al, 2001). Together, these findings suggest that the evolutionarily conserved LAMTOR-Rag complex acts as a pleiotropic signaling module with inherent adaptability to regulate diverse physiological processes.
Within the broader context of immunity, Lamtor1 has emerged as a critical regulator of M2 macrophage polarization (Kimura et al, 2016). Myeloid-specific deletion in mice leads to elevated proinflammatory cytokines, such as TNF-α and IL-6, during sepsis. While lysosomal transcription factors TFEB and TFE3 have been implicated in cytokine regulation (Pastore et al, 2016; Hayama et al, 2018), our findings reveal their distinct role in IFN responses: rather than suppressing IFN production downstream of the LAMTOR-Rag axis, they act as essential transcriptional activators of IFN genes. This suggests that PRR signaling may regulate TFEB/TFE3 nuclear translocation and activity independently of canonical mTORC1-mediated nutrient sensing (Napolitano et al, 2020; Cui et al, 2023; Fernandes et al, 2024). Future studies should explore the interplay between TFEB/TFE3 and IRFs, particularly their potential co-regulation through shared regulatory elements (e.g., CLEAR/ISRE elements) in IFN and ISG promoters.
Notably, IRF-5 and IRF-7 expression is tightly coupled to the LAMTOR-Rag axis, with their levels dictated by Rag nucleotide-loading states. We propose that nutrient-dependent Rag cycling establishes a preemptive priming mechanism—analogous to NLRP3 pre-activation in inflammasome formation (Barnett et al, 2023)—ensuring rapid antiviral responses upon pathogen detection. Although the mechanistic details require further investigation, this coordinated IRF regulation likely optimizes the balance between metabolic demands and antiviral defense. Our work further clarifies the role of mTORC1 in IFN induction, which primarily operates through global translation and protein secretion rather than direct activation of IFN transcription, except in TLR3 signaling (Narita et al, 2011; Nüchel et al, 2021; Kaeser-Pebernard et al, 2022). Together with emerging literature (Heinz et al, 2020; Kobayashi et al, 2014; Harvey et al, 2025; Evavold et al, 2021), our findings reinforce the interplay between metabolic rewiring (e.g., amino acid sensing) and immune signaling, further positioning lysosomes as central hubs for integrating these pathways (Settembre and Perera, 2024; Deretic, 2021).
Unlike mitogenic MAPKs, p38 kinases are stress-responsive effectors coordinating cellular adaptation to various environmental and genotoxic insults (Canovas and Nebreda, 2021). Their activation outcomes depend on stimulus-specific spatiotemporal organization. For instance, during DNA damage, p38 translocates to the nucleus and phosphorylates transcription factors like ATF-2 and c-Jun, triggering protective gene expression to promote DNA repair and cell survival (Canovas and Nebreda, 2021). In contrast, p38 remains predominantly extranuclear during innate immune pathway activation, highlighting context-dependent regulation. Here, we identify a previously unrecognized lysosome-localized p38 subset whose regulation diverges from canonical three-tiered MAPK signaling cascades. Specifically, under PRR signaling, FLCN recruits p38 to phagolysosomal membranes, where Rag GTPase nucleotide cycling drives localized activation. We propose that Rag nucleotide states recruit upstream regulators to spatiotemporally control p38 activity at lysosomes, though the precise mechanisms warrant future investigation.
Antiviral IFN and ISG mRNAs are inherently unstable, undergoing rapid decay to prevent excessive inflammation (Piccirillo et al, 2014). Genetic ablation of p38 MAPK accelerates Ifnb1 mRNA degradation and abolishes PRR-induced IFN-β production, even under active Rag signaling, underscoring the critical role of mRNA stability regulation in IFN immunity. Our study reveals that p38 MAPK acts as a downstream effector of the LAMTOR-Rag pathway, linking lysosomal signaling to transcript longevity and providing new insights into the post-transcriptional control of IFN responses.
How does p38 MAPK post-transcriptionally regulate the IFN program? Previous studies demonstrate that RNA-binding proteins HuR/ELAVL1 and Tristetraprolin (TTP/ZFP36) reciprocally regulate the half-life of IFN and ISG mRNAs by binding to AU-rich elements (AREs) in their 3’-untranslated regions (Herdy et al, 2015; Rothamel et al, 2021). Notably, the p38 substrate MAPK-activated protein kinase 2 (MAPKAPK-2) phosphorylates TTP, blocking its ability to recruit the CCR4-NOT deadenylase complex and thereby stabilizing target mRNAs (Canovas and Nebreda, 2021; Clement et al, 2011). Conversely, during DNA damage, p38 phosphorylates HuR, promoting its cytoplasmic accumulation and enhancing transcript stabilization to enforce cell-cycle arrest (Lafarga et al, 2009). These findings suggest that p38 may control IFN mRNA stability through opposing effects on TTP and HuR, though whether such regulation occurs at lysosomes remains to be investigated.
Although lysosomes are ancient organelles conserved from yeast to humans, their role in IFN regulation likely represents a key evolutionary innovation, coinciding with the emergence of a fully functional interferon system in jawed vertebrates (Secombes and Zou, 2017). The repurposing of the LAMTOR-Rag complex for IFN regulation exemplifies gene co-option, allowing higher organisms to integrate sophisticated innate immune responses with metabolic capacity. The severe impairment of antiviral defense in mice with myeloid-specific Rragc deletion underscores the physiological relevance of this pathway. Given the lysosome’s sensitivity to metabolic stress, our findings may explain how nutrient fluctuations modulate immune outcomes during infection. Tight transcriptional and post-transcriptional control may prevent aberrant IFN production under nutrient-limited conditions, ensuring metabolically sustainable immune responses. Therapeutically, targeting Rag GTPase cycling or p38 activity could provide new strategies to enhance IFN responses in immunocompromised hosts or mitigate excessive inflammation. Conversely, viruses may exploit this pathway, as hinted by the increased infectivity observed in LAMTOR-deficient cells. Future studies should explore viral evasion mechanisms and the potential role of lysosomal signaling in IFN-mediated autoimmunity.
In summary, we identify the LAMTOR-Rag–p38 axis as a lysosome-centered mechanism linking organelle signaling to innate immunity. By integrating nutrient sensing with immune activation, this pathway ensures precise control of IFN responses, opening new avenues to modulate antiviral defense and inflammation.
Methods
Reagents and tools tableReagent/resourceReference or sourceIdentifier or catalog number Experimental models Influenza A virus IAV (strain H1N1; PR8)(Zhao et al, 2022)Influenza A virus IAV (PR8-NS1-GFP)(Zhao et al, 2022)M. musculus strainsConditional floxed Rragc (Rragc^fl^); C57BL/6 JCyagen BiosciencesS-CKO-11702Lyz2-Cre; C57BL/6 JGemPharmatechT003822Rragc^flox/flox^; Cre- (Rragc^f/f^); C57BL/6 JThis studyRragc^flox/flox^; Cre + (Rrag^Lyz2^); C57BL/6 JThis studyPrimary culturesRragc^f/f^ BMDM; female; C57BL/6 JThis studyRragc^Lyz2^ BMDM; female; C56BL/6 JThis studyCell linesRAW264.7ATCCTIB-71RAW264.7 Lamtor1 KOThis studyRAW264.7 Lamtor1 KO + hLAMTOR1 WTThis studyRAW264.7 Lamtor1 KO + hLAMTOR1 3 AThis studyRAW264.7 Lamtor1 KO + hLAMTOR1 NIVThis studyRAW264.7 Lamtor1 KO + hLAMTOR1 LVVThis studyRAW264.7 Lamtor1 KO + mIRF5 + mIRF7This studyRAW264.7 Lamtor1 KO + mRaptor-Rheb15This studyRAW264.7 Lamtor1 KO + hMAPK14 D176A/F327SThis studyRAW264.7 Rraga KOThis studyRAW264.7 Rraga KO + mRagA WTThis studyRAW264.7 Rraga KO + mRagA Q66LThis studyRAW264.7 Rraga KO + mRagA T21LThis studyRAW264.7 Rraga KO + hMAPK14 D176A/F327SThis studyRAW264.7 Rragc KOThis studyRAW264.7 Rragc KO + mRagC WTThis studyRAW264.7 Rragc KO + mRagC Q119LThis studyRAW264.7 Rragc KO + mRagC S74NThis studyRAW264.7 Rragc KO + hMAPK14 D176A/F327SThis studyRAW264.7 Rptor KOThis studyRAW264.7 Flcn KOThis studyRAW264.7 Depdc5 KOThis studyRAW264.7 Tfe3 KO #1This studyRAW264.7 Tfe3 KO #2This studyRAW264.7 Tfeb KOThis studyRAW264.7 Lamtor1/Tfeb KOThis studyRAW264.7 Myd88 KOThis studyRAW264.7 Unc93b1 KOThis studyRAW264.7 Mapk14 KOThis studyRAW264.7 Mapk11 KOThis studyRAW264.7 Mapk14/11 DKOThis studyRAW264.7 expressing hMAPK14 WTThis studyRAW264.7 expressing hMAPK14 D168AThis studyRAW264.7 expressing hMAPK14 D176A/F327SThis studyRAW264.7 Mapk14/11 DKO + mRagA Q66LThis studyRAW264.7 Mapk14/11 DKO + mRagC S74NThis studyRAW264.7 Depdc5/Mapk14/11 TKOThis studyTHP.1ATCCTIB-202THP-1 shLuciferaseThis studyTHP-1 shLAMTOR #1This studyTHP-1 shLAMTOR #2This studyTHP-1 shLAMTOR #3This studyHEK293TATCCCRL-3216HEK293T LAMTOR1 KOThis study Recombinant DNA pLentiCRISPRv2Addgene#52961psPAX2Addgene#12260pMD2.GAddgene#12259Details of all plasmidsThis studyTable EV1 Antibodies phospho-S172-TBK1 (IB 1:1000)Cell Signaling Technology#5483 (D52C2)TBK1 (IB 1:1000)Cell Signaling Technology#3504 (D1B4)phospho-S365-STING (IB 1:1000)Cell Signaling Technology#13647 (D2P2F)STING (IB 1:1000)Cell Signaling Technology#72971 (D8F4W)phospho-S396-IRF-3 (IB 1:1000)Cell Signaling Technology#4947 (4D4G)IRF-3 (IB 1:1000)Cell Signaling Technology#4302 (D83B9)phospho-T389-S6K1 (IB 1:1000)Cell Signaling Technology#9234 (108D2)S6K1 (IB 1:1000)Cell Signaling Technology#2708 (49D7)phospho-S65-4E-BP1 (IB 1:1000)Cell Signaling Technology#94514E-BP1 (IB 1:1000)Cell Signaling Technology#9644 (53H11)phospho-S235/236 S6 (IB 1:1000)Cell Signaling Technology#4858 (D57.2.2E)S6 (IB 1:1000)Cell Signaling Technology#2217 (5G10)phospho-T202/Y204-ERK1/2 (IB 1:1000)Cell Signaling Technology#9101ERK1/2 (IB 1:1000)Cell Signaling Technology#4695 (137F5)phospho-T183/Y185-JNK (IB 1:1000)Cell Signaling Technology#4668 (81E11)JNK (IB 1:1000)Cell Signaling Technology#9252phospho-T180/Y182-p38 (IB 1:1000)Cell Signaling Technology#4511 (D3F9)p38 (IB 1:1000)Cell Signaling Technology#8690 (D13E1)phospho-T71-ATF-2 (IB 1:1000)Cell Signaling Technology#24329ATF-2 (IB 1:1000)Cell Signaling Technology#35031 (D4L2X)phospho-S73-c-Jun (IB 1:1000)Cell Signaling Technology#3270 (D47G9)c-Jun (IB 1:1000)Cell Signaling Technology#9165 (60A8)Raptor (IB 1:1000)Cell Signaling Technology#2280 (24C12)RagA (IB 1:1000)Cell Signaling Technology#4357 (D8B5)RagC (IB 1:1000)Cell Signaling Technology#3360Lamtor1 (IB 1:1000)Cell Signaling Technology#8975 (D11H6)Lamtor2 (IB 1:1000)Cell Signaling Technology#8145 (D7C10)Lamtor3 (IB 1:1000)Cell Signaling Technology#8168 (D38G5)Lamtor4 (IB 1:1000)Cell Signaling Technology#12284 (D6A4V)Lamtor5 (IB 1:1000)Cell Signaling Technology#14633 (D4V4S)ISG15 (IB 1:1000)Cell Signaling Technology#2743MDA-5 (IB 1:1000)Cell Signaling Technology#5321 (D74E4)MyD88 (IB 1:1000)Cell Signaling Technology#4283 (D80F5)TLR3 (IB 1:1000)Cell Signaling Technology#6961 (D10F10)TLR4 (IB 1:1000)Cell Signaling Technology#14358 (D8L5W)TLR7 (IB 1:1000)Cell Signaling Technology#82658 (E4J3Z)MAVS (IB 1:1000)Cell Signaling Technology#83000 (E8Z7M)IRF-1 (IB 1:1000)Cell Signaling Technology#8478 (D5E4)IRF-4 (IB 1:1000)Cell Signaling Technology#62834 (E8H3S)IRF-5 (IB 1:1000)Cell Signaling Technology#96527 (E9I4Z)IRF-7 (IB 1:1000)Cell Signaling Technology#72073 (D8V1J)IRF-9 (IB 1:1000)Cell Signaling Technology#28845 (D9I5H)NF-κB p65 (IB 1:1000)Cell Signaling Technology#8242 (D14E12)Stat1 (IB 1:1000)Cell Signaling Technology#14994 (D1K9Y)Stat2 (IB 1:1000)Cell Signaling Technology#72604 (D9J7L)ATF-3 (IB 1:1000)Cell Signaling Technology#33593 (D2Y5W)VDAC (IB 1:1000)Cell Signaling Technology#4661 (D73D12)Golgin-97 (IB 1:1000)Cell Signaling Technology#13192 (D8P2K)PEX5 (IB 1:1000)Cell Signaling Technology#83020 (D7V4D)Calreticulin (IB 1:1000)Cell Signaling Technology#12238 (D3E6)α-tubulin (IB 1:1000)Cell Signaling Technology#2125 (11H10)PARP-1 (IB 1:1000)Cell Signaling Technology#9532 (46D11)phospho-S122-TFEB (IB 1:1000)Cell Signaling Technology#87932 (E9M5M)TFEB (IB 1:1000)Cell Signaling Technology#83010 (E5P9M)TFE3 (IB 1:1000)Cell Signaling Technology#14779FLAG/DYKDDDDK Tag (IB 1:1000)Cell Signaling Technology#14793 (D6W5B)FLAG/DYKDDDDK Tag (IF 1:400)Cell Signaling Technology#14793 (D6W5B)Actin (IB 1:1000)Cell Signaling Technology#4970 (13E5)MOV-10 (IB 1:1000)Proteintech Group10370-1-APCPEB1 (IB 1:1000)Proteintech Group13274-1-APIFI44 (IB 1:1000)SolarbioK006369PNPC1 (IB 1:1000)Abcam#ab134113LAMP2 (IF 1:400)Santa Cruz Biotechnologysc-18822TFEB (IB 1:1000)Bethyl LabsA303-673A-TTLR9 (IB 1:1000)Thermo Fisher Scientific14-9093-82 (M9.D6)Goat anti-mouse IgG-HRP conjugate (1:4000)Thermo Fisher ScientificG21040Alexa Fluor-568 goat anti-mouse IgG (1:400)Thermo Fisher ScientificA-11004Alexa Fluor-488 goat anti-rabbit IgG (1:400)Thermo Fisher ScientificA-11034Anti-FLAG M2 affinity gel (IP 1:50)Sigma-AldrichA2220Goat anti-rabbit IgG-HRP conjugate (1:4000)Vector LaboratoriesPI-1000-1Goat anti-rat IgG-HRP conjugate (1:4000)ElabscienceE-AB-1041 Oligonucleotides and other sequence-based reagents Details of all qRT-PCR primersThis studyTable EV2 Chemicals, enzymes and other reagents Poly(I:C)InvivoGentlrl-picR848InvivoGentlrl-r848CpG-BInvivoGentlrl-1668LPSInvivoGentlrl-eblpscGAMPInvivoGentlrl-nacga23-1RapamycinTargetmolT1537Torin 1TargetmolT6045SorbitolBeyotimeST1290Sodium ArseniteSigma-AldrichS7400Pierce EDTA-free protease inhibitor tabletsThermo Fisher ScientificA32965FuGENE HD transfection reagentFugent LLCHD-1000PolybreneMilliporeTR-1003-GPuromycinCalbiochem540411Mouse IFN-β ELISA kitMulti SciencesEK2236Human IFN-β ELISA kitMulti SciencesEK1236Dual-luciferase reporter assay kitVazymeDL101-01Nucleocytoplasmic fractionation kitBeyotimeP0027LysoView 550Biotium#70083DMEMThermo Fisher Scientific11965DMEM, no amino acids, no glucoseUS BiologicalD9800-27RPMI 1640Thermo Fisher Scientific118750.05% trypsin-EDTAThermo Fisher Scientific25300penicillin-streptomycin (10,000 U/mL)Thermo Fisher Scientific15140OptiMEM-IThermo Fisher Scientific31985070calcium- and magnesium-free DPBSThermo Fisher Scientific14190High-sensitive enhanced chemiluminescence detection kitVazymeE412-01Universal antibody diluentNCM BiotechWB500DMycoBlue Mycoplasma Detector kitVazymeD101Lenti-X ConcentratorClontech631231Quick Start Bradford 1× Dye ReagentBio-Rad#5000205ProMag® Magnetic MicrospheresPolysciences25029LysoSensor^TM^ Yellow/Blue DND-160Thermo Fisher ScientificL7545FastPure Cell/Tissue Total RNA Isolation kitVazymeRC112-01HiScript III All-in-one RT SuperMix Perfect for qPCR kitVazymeR333-01Taq Pro Universal SYBR qPCR Master MixVazymeQ712-034-thiouridine (4sU)Sigma-AldrichT4509N-EthylmaleimideSigma-Aldrich04259Normal donkey serumJackson ImmunoResearch017-000-121TRIzol reagentThermo Fisher Scientific15596026NEBNext Ultra RNA Library Prep Kit for IlluminaNew England Biolabs#7530TruePrep DNA Library Prep KitVazymeTD501MinElute PCR Purification KitQIAGEN28004TruePrep Index Kit V2 for IlluminaVazymeTD202 Software Prism 10GraphPad Software, Inc.ExcelMicrosoftAdobe IllustratorAdobeZEN liteZEISSAndor FusionOxford InstrumentsImaris for BC43Oxford InstrumentsIntegrative Genomics Viewer https://igv.org Benchling https://www.benchling.com OmicSmart https://www.omicsmart.com
Other CellDrop Automated Cell CounterDeNovixInvitrogen iBright CL1500Thermo Fisher ScientificeBlot L1 Fast Wet Transfer SystemGenScriptApplied Biosystems QuantStudio 1 Real-Time PCR systemThermo Fisher ScientificSynergy H1 Multimode Microplate ReaderBioTek InstrumentsAndor BC43 Benchtop Spinning Disk Confocal MicroscopeOxford Instruments-AndorZEISS Axio Vert.A1 FL-Inverted LED Fluorescence MicroscopeZeissBD FACSAria Fusion flow cytometerBD Biosciences
Mouse studies
Conditional floxed Rragc (Rragc^fl^, S-CKO-11702) C57BL/6 J mice were purchased from Cyagen Biosciences, whereas Lyz2-Cre (T003822) strain was acquired from GemPharmatech. Rragc^flox/flox^ mice were crossed with a Lyz2-Cre background to generate Rragc^Lyz2^ mice. Rragc^Lyz2^ mice were born at a normal Mendelian ratio. The mice were bred and housed under specific pathogen-free conditions in the animal facility of RUIYE Laboratories (RYEth-20230317191, Guangzhou). All animal experiments complied with the relevant ethical regulations and were approved by the Institutional Animal Care and Use Committee of Guangzhou National Laboratory. Unless otherwise stated, mice of the indicated genotypes were used for viral infection or BMDM isolation at 8–10 weeks of age.
Cell culture and cell lines
RAW264.7, HEK293T, THP-1, MDCK, and HeLa cells were obtained from the American Type Culture Collection (ATCC). RAW264.7, MDCK, and HEK293T cells were cultured in DMEM supplemented with 10% (v/v) inactivated fetal bovine serum (GIBCO, Thermo Fisher Scientific), penicillin (100 U ml^−1^), streptomycin (100 µg ml^−1^), 1 mM sodium pyruvate, and 2 mM L-glutamine. THP-1 and HeLa cells were cultured in RPMI 1640 supplemented with the aforementioned reagents. To maintain THP-1 cells, 50 µM of 2-mercaptoethanol was added to the culture medium. Primary bone marrow-derived macrophages (BMDMs) were generated from the femurs of the indicated mice (C57BL/6 J background, female, 8 weeks old) and were differentiated for seven days in complete DMEM supplemented with macrophage colony-stimulating factor-1 (20 ng ml^−1^). Cells were cultured at 37 °C in a 5% CO_2_ incubator. All the cell lines used in this study were free of mycoplasma contamination and were regularly tested using the MycoBlue Mycoplasma Detector kit (D101) from Vazyme. For amino acid starvation and refeeding experiments, the cell medium was replaced with amino acid-free DMEM (US Biological D9800-27) supplemented with 25 mM glucose and 2% dialyzed FBS for the indicated period and/or re-substituted with DMEM (Thermo Fisher Scientific 11965) containing 2% dialyzed FBS.
Plasmid constructs
cDNA encoding human LAMTOR1 (UniProt Q6IAA8) was generated by reverse transcription polymerase chain reaction (RT-PCR) using RNA extracted from HEK293T cells. A 3×FLAG tag was added to the C-terminus of LAMTOR1 and subcloned into the pLJM1 lentiviral vector using the Vazyme ClonExpress II One-Step Cloning Kit (C112-01). cDNA for human MAPK14 (UniProt Q16539) was generated similarly and inserted into the pRK5-based expression vector or the pLJM1 lentiviral vector, carrying a 3×FLAG or mBaoJin tag as specified in the experiments. cDNAs for mouse Rraga (UniProt Q80X95) and Rragc (UniProt Q99K70) were acquired by RT-PCR from RNA obtained from RAW264.7 cells. A FLAG tag was appended to the N-terminus of Rraga or Rragc and subcloned into the pLJM1 vector, as described above. Site-directed mutagenesis of LAMTOR1, Rraga, Rragc, and MAPK14 was performed using QuikChange II XL (Stratagene) according to the manufacturer’s instructions. The following plasmids, including pLentiCRISPRv2 (#52961), psPAX2 (#12260), and pMD2.G (#12259), were from Addgene. Sequence information of the plasmids used in this study is provided in Table EV1.
Generation of CRISPR knockout cells
CRISPR-Cas9-mediated knockout cell line generation was performed using pLentiCRISPRv2-GFP-puro, a construct modified from pLentiCRISPRv2-puro (Addgene #52961). Small guide RNAs (sgRNAs) were selected using the CRISPR design tool of Benchling or CHOPCHOP v3. At least two–three independent sgRNA sequences were tested for each gene. Oligonucleotides for sgRNAs were synthesized, annealed in vitro, and subcloned into BsmBI-digested pLentiCRISPRv2-GFP-puro. Approximately 1,000,000 cells were seeded in six-well plates. The next day, cells were transfected with pLentiCRISPRv2-GFP-puro containing the respective sgRNAs using the FuGENE transfection reagent. Twenty-four hours after transfection, the GFP-positive cells were sorted using a BD FACSAria Fusion flow cytometer. Single-cell-derived knockout clones were isolated, expanded, and screened using genotyping and immunoblotting. In the early phase of this study, we noted that non-targeting control RAW264.7 cell clones displayed significant discrepancies in response to PRR stimulation. Hence, parental RAW cells were used as a reference control when analyzing the gene knockout phenotypes. In all gene knockout and rescue experiments, the knockout line reconstituted with wild-type cDNA was used as the reference group for multiple comparisons.
Lentivirus packaging and transduction
HEK293T cells were transfected with lentiviral vectors and packaging plasmids, psPAX2 (Addgene #12260) and pMD2.G (Addgene #12259), using polyethylenimine (PEI) or calcium phosphate. Twenty-four hours later, the medium was changed; 48 h after transfection, the cell supernatants were collected, cleared at 1000 × g for 5 min, and filtered through 0.45-μm PES membrane filters. Lentiviruses were further concentrated using Lenti-X Concentrator (Clontech 631231). Target cells were infected with lentiviruses in the presence of 8 μg ml^−1^ polybrene (Millipore TR-1003-G). Twenty-four hours after infection, cells were given fresh medium and selected with 2 µg ml^−1^ puromycin (Calbiochem 540411). Protein expression was confirmed by immunoblotting and/or fluorescence microscopy.
Cell lysis, immunoblot analysis, and immunoprecipitation
Cells were washed with ice-cold PBS and lysed in lysis buffer (40 mM HEPES, pH 7.4, 1% Triton X-100, 10 mM β-glycerol phosphate, 10 mM sodium pyrophosphate, and 4 mM EDTA supplemented with EDTA-free protease inhibitor cocktail). After incubation on a rotator, the cell lysates were cleared of insoluble material by centrifugation at 17,000 × g for 10 min at 4 °C. The protein concentration of each sample was quantified using the Quick Start Bradford 1× Dye Reagent (#5000205) from Bio-Rad. Proteins were denatured in Laemmli SDS sample buffer at 70 °C for 10 min, separated by 10, 12, or 4–20% Bis-Tris SDS–PAGE (GenScript SurePAGE precast gels), and transferred onto Immobilon PVDF membranes (Millipore) in an eBlot L1 transfer system (GenScript). Membranes were blocked in TBS-Tween with 5% milk for 1 h at room temperature and incubated overnight at 4 °C with the indicated antibodies. Membranes were washed thrice with TBS-Tween buffer and incubated for 1 h with secondary HRP-conjugated antibodies in TBS-Tween with 5% milk. After extensive washing with TBS-Tween, membranes were developed using a highly sensitive enhanced chemiluminescence detection solution (Vazyme) on an Invitrogen iBright CL1500 imaging system (Thermo Fisher Scientific).
For FLAG immunoprecipitation, 20 µL of a well-mixed 50% slurry of anti-FLAG M2 Affinity Gel (Sigma-Aldrich A2220) was added to each lysate (1 mg ml^−1^) and incubated at 4 °C for 2 h under rotation. Subsequently, the beads were washed twice with lysis buffer and thrice with lysis buffer supplemented with 500 mM NaCl. Protein complexes were eluted from the beads using Laemmli SDS sample buffer at 70 °C for 10 min, resolved by SDS–PAGE, and analyzed by immunoblotting as described above.
Quantitative RT-PCR analysis
Total RNA was extracted from cells using a Vazyme FastPure Cell/Tissue Total RNA Isolation kit (RC112-01) with DNase treatment and quantified using a TGem Plus Spectrophotometer (TIANGEN Biotech). Following the manufacturer’s instructions, cDNA was prepared from 1 µg of RNA by RT-PCR using a Vazyme HiScript III All-in-one RT SuperMix Perfect for qPCR kit (R333-01). Quantitative RT-PCR to assess the indicated mRNA expression was performed on an Applied Biosystems QuantStudio 1 Real-Time PCR system with Vazyme Taq Pro Universal SYBR qPCR Master Mix (Q712-03). Fold changes in expression were calculated by the ΔΔC_T_ method using mouse Gapdh or human ACTB as endogenous controls for mRNA expression. Mock-treatment controls were shared across PRR stimulation conditions from the same experimental batch, except for cGAMP stimulation in some cases, which used separate mock controls to account for reagent lot variability. All qRT-PCR primers are listed in Table EV2.
Roadblock-qPCR
To measure Ifnb1 mRNA stability, Roadblock-qPCR was conducted as previously described (Watson and Thoreen, 2022). Approximately 600,000 cells of the indicated genotypes were seeded into 12-well plate. The following day, the cells were stimulated with 0.5 µg ml^−1^ Poly(I:C) via FuGENE transfection for 14 h. Post-stimulation, cells were grown in complete media supplemented with 400 µM 4sU (Sigma-Aldrich T4509) and harvested at 2-h intervals for 8 h with three biological replicates per collection time point. Cells cultured in complete media without 4sU were used for measurement of the initial time point. Briefly, RNA was harvested as described above and treated with DNase I. For each condition or time point, 1 µg of RNA was placed in a clean 0.2 mL PCR tube, brought to a final volume of 11 µL with RNase-free water, and incubated at 65 °C for 5 min in a thermocycler. Subsequently, 10 µL NEM reaction buffer (125 mM Tris-HCl, pH 8.0, 2.5 mM EDTA) and 4-µL N-ethylmaleimide (Sigma-Aldrich 04259, 50 µg/µL in 100% ethanol) were added to the RNA in the PCR tubes, resulting in a total volume of 25 µL. The tubes were then incubated at 42 °C for 1.5 h in a thermocycler. RNA samples treated with NEM were purified by adding 3 µL of 3 M Sodium Acetate Solution, pH 5.2 (Thermo Fisher Scientific R1181), and 84 µl of isopropanol to the tubes. The mixtures were thoroughly combined and incubated at −20 °C overnight in a refrigerator. Subsequently, the samples underwent centrifugation at 13,000 × g at 4 °C for 10 min to discard the flow-through. The purified RNA samples were then washed twice with 75% ethanol, air-dried for 10 min, reconstituted with RNase-free water, and subsequently converted to cDNA. Quantitative PCR was then performed with primers targeting mouse Ifnb1 (Forward: CCAGCTCCAAGAAAGGACGA; Reverse: TGGATGGCAAAGGCAGTGTA) and mouse 18S rRNA (Forward: CGGAAAATAGCCTTCGCCATCAC; Reverse: ATCACTCGCTCCACCTCATCCT). qRT-PCR results were analyzed with a single exponential decay model to estimate the Ifnb1 mRNA half-life, using Prism 10.2.3 (GraphPad Software, Inc.).
Phagolysosome isolation
Native phagolysosomes were isolated using a protocol developed in the laboratory of G. Barton (Majer et al, 2019). Briefly, cells in a 150-mm culture dish were fed with 40 µL ProMag® Magnetic Microspheres (Polysciences 25029) for 4 h, followed by extensive washing with PBS. Cells were harvested in 10 mL sucrose homogenization buffer (250 mM sucrose, 3 mM imidazole, pH 7.4) and collected by centrifugation. Cells were resuspended in 2 mL buffer supplemented with protease inhibitor cocktail and 1 mM PMSF and disrupted in a Dounce glass homogenizer. The disrupted cells were gently rocked for 10 min on ice to liberate phagolysosomes. Microspheres were collected using a DynaMag-2 magnet (Invitrogen) and washed four times with buffer. Phagolysosome samples were denatured in 2× SDS buffer at 37 °C for 10 min, resolved by SDS–PAGE, and analyzed by immunoblotting.
Luciferase reporter assay
HEK293T cells expressing the indicated shRNAs (300,000 cells per well in a 12-well plate) were transfected with an ISRE firefly luciferase reporter plasmid and Renilla luciferase reporter plasmid using PEI. Cells were stimulated with Poly(I:C) via FuGENE transfection for 24 h, lysed, and analyzed for ISRE reporter activity using a dual-luciferase reporter assay kit (Vazyme DL101-01). Luciferase activity was measured using a Synergy H1 Multimode Microplate Reader (BioTek Instruments). Data were normalized to Renilla expression and presented as fold change in luciferase activity compared to unstimulated controls. Note that while cells expressing shRNA against luciferase served as controls, the corresponding shRNA targeting sequence was absent in the ISRE firefly luciferase reporter construct.
Enzyme-linked immunosorbent assay (ELISA)
Cells were quantified using a DeNovix CellDrop Automated Cell Counter and subsequently seeded into tissue culture-treated multiwell plates (RAW264.7: 250,000 cells per well in a 24-well plate; THP-1: 1,000,000 cells per well in a 24-well plate; HEK293T: 300,000 cells per well in a 12-well plate). The cells were stimulated with designated agonists for a duration of 14 h. Post-stimulation, cells were utilized for RNA extraction while cell culture supernatants were collected for analysis of IFN-β secretion via ELISA, following the manufacturer’s protocols. The ELISA kits for mouse IFN-β (EK2236) and human IFN-β (EK1236) were obtained from Multi Sciences. IFN-β concentrations were determined by extrapolation from a recombinant IFN-β standard curve. For experimental consistency, mock-treatment control data were shared across different PRR stimulation conditions when derived from the same batch. For experiments involving cGAMP stimulation, separate mock controls were used in some cases to account for batch-to-batch reagent variability. It should be noted that identical values in some replicates, primarily under mock conditions, resulted from the microplate reader’s detection limit at low signals, as confirmed with the manufacturer. All data were reported as measured.
Immunofluorescence
To stain the fixed cells, 200,000 HeLa cells were seeded on fibronectin-coated 18-mm round glass coverslips in 12-well plates and allowed to attach overnight. The cells were fixed for 15 min with 4% paraformaldehyde in PBS at room temperature and rinsed twice with PBS. The cells were then permeabilized with ice-cold methanol for 10 min at −20 °C, washed thrice with PBS, and blocked with 5% normal donkey serum (Jackson ImmunoResearch 017-000-121) at room temperature for 1 h. Cells were stained overnight at 4 °C with the indicated primary antibodies (rabbit anti-FLAG, 1:400; mouse anti-LAMP2, 1:400) in blocking solution. The cells were washed thrice with PBS and stained for 1 h with fluorophore-conjugated secondary antibodies (1:400) at room temperature. The cells were washed thrice with PBS and counterstained with Hoechst 33342 (1:1000) in PBS. The coverslips were mounted onto microscope slides using Fluoromount-G (SouthernBiotech 0100-01).
Fluorescence microscopy
Spinning disk confocal imaging was performed on an Andor Benchtop Confocal Microscope (Oxford Instruments-Andor BC43) equipped with a 4.1 MP (6.5-μm pixel, 16-bit) monochrome camera and controlled by Andor Fusion acquisition software (version 1.1.1.17). Images were acquired using a 100 × Plan Apo 1.45 NA objective (Nikon). Blue, green, red, and far-red fluorescence was excited using 405, 488, 561, and 638 nm lasers, respectively. All acquired images were processed using Imaris for BC43 and prepared for figure assembly using Adobe Illustrator 2023 version 27.2. For widefield fluorescence microscopy, images were captured using an LD A-Plan 10 × objective (ZEISS) on a ZEISS Axio Vert.A1 FL-Inverted LED Fluorescence Microscope equipped with a 2.83 MP (4.54-μm pixel, 14-bit) Axiocam 503 color camera. All images were processed using the ZEISS ZEN 2.3 lite software and exported to Adobe Illustrator 2023 version 27.2 for figure assembly.
Endolysosomal pH measurement
Endolysosomal pH was measured using LysoSensor^TM^ Yellow/Blue DND-160 (Thermo Fisher Scientific L7545) with a microplate reader-based method as previously described (Ma et al, 2017). Cells of different genotypes were passaged and seeded into a 96-well black/clear-bottom microplate at 30,000 cells per well. Both generations of the pH calibration curve and collection of experimental data were performed simultaneously in a single 96-well plate. The following day, the medium was changed, and cells were incubated with 1 µM LysoSensor in RAW cell culture medium at 37 °C for 5 min, then rinsed twice with PBS, and left with 100 µL of PBS. Cells were incubated in triplicate with 100 µL of each pH buffer for various time intervals to generate the pH calibration curve. LysoSensor fluorescence was measured immediately using a Synergy H1 Multimode Microplate Reader (BioTek Instruments), with excitation wavelengths of 329 and 384 nm, and emission wavelengths of 440 and 540 nm. For the experimental groups, cells were incubated for the same time periods used in generating the pH calibration curve at 37 °C in 100 µL of pre-warmed, normal cell culture medium containing 1 µM LysoSensor, then rinsed twice with PBS, and left with 100 µL of PBS for readout. The fluorescence intensity ratio of emission readings at 440 nm to emission readings at 540 nm was calculated, and the pH of cells was measured accordingly using the pH calibration curve generated in parallel with the experiment.
Virus infection
The Influenza A virus IAV (H1N1 strain; PR8 or PR8-NS1-GFP) used in this study was provided by S Wang (Zhao et al, 2022). Cells were infected with the virus in the culture medium at the indicated multiplicity of infection (MOI). IAV titers were determined using a standard plaque assay with MDCK cells. For the IAV replication assay, HEK293T cells of the respective genotypes were infected with IAV PR8-NS1-GFP at the indicated MOI in serum-free DMEM for 2 h, after which the cells were washed with PBS and cultured in serum-free DMEM supplemented with 0.7 µg ml^−1^ N-tosyl-L-phenylalanine chloromethyl ketone (TPCK)-treated trypsin and 0.2% (v/v) BSA. PR8-NS1-GFP replication was assessed by determining the percentage of GFP-positive cells by flow cytometry. For in vivo studies, 8-week-old female littermate mice were infected with 50 PFU of IAV-H1N1 (PR8) via intranasal administration. The infected mice were closely monitored for survival and body weight loss for 16 days. Another batch of mice was sacrificed 3 days after infection, and bronchoalveolar lavage fluid (BALF) samples were collected for analysis. BALF samples were analyzed for IFN-β levels by ELISA and viral titers by the TCID_50_ assay. The lung tissue fractions were fixed with formalin and processed for histological analysis. All procedures involving viral strains were conducted in an Animal Biosafety Level 2 (ABSL2) facility at the Guangzhou National Laboratory, in accordance with both institutional biosafety committee protocols and the ARRIVE 2.0 guidelines.
Targeted shRNA screen
Fifty genes encoding proteins involved in signal transduction were selected from the lysosomal proteomic dataset as shRNA targets according to their functional annotation using GO terms and Reactome. Two independent shRNAs were designed for each gene, inserted into the pLKO.1-TRC cloning vector and prepared for lentiviral production. RAW264.7 macrophages were plated at 10,000 cells per well in black-walled, clear-bottom 96-well microplates, allowed to adhere for 4 h, and transduced with lentiviral particles, with three technical replicates per shRNA. After 48 h, cells were washed with PBS and supplemented with fresh medium. The next day, cells were infected with IAV PR8-NS1-GFP at an MOI of 0.25 for 24 h. The cells were washed once with PBS, incubated with 5 μg ml^−1^ Hoechst 33342 for 5 min, and washed thrice with PBS. Fluorescence was measured using a Synergy H1 Multimode Microplate Reader (BioTek Instruments). The GFP intensity was divided by Hoechst 33342 to derive the normalized fluorescence for each well, eliminating the variation caused by the well-to-well cell density. Averaged data are expressed as the fold change in fluorescence of IAV-infected cells expressing shRNA against luciferase.
RNA-sequencing
RAW macrophages of the indicated genotypes were seeded into 12-well plates (500,000 cells per well). The following day, cells were supplemented with fresh medium and stimulated with the indicated PRR agonists for 14 h. Cells were harvested and centrifuged at 500 × g for 5 min at 4 °C to remove the supernatant. Total RNA was extracted from the cell pellets using TRIzol reagent (Thermo Fisher Scientific 15596026) following the manufacturer’s instructions and quantified using Qubit 2.0 (Thermo Fisher Scientific). RNA quality per sample was assessed by on-chip electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies). Libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs #7530) and sequenced on a NovaSeq 6000 (Illumina) with a maximum read depth of approximately 40 billion paired-end reads per run by Gene Denovo Biotechnology Co., Ltd. (Guangzhou, China). High-quality clean reads from sequenced libraries were acquired after filtering by fastp (version 0.18.0) and trimming by Bowtie2 (version 2.2.8). The paired-end clean reads were mapped to the mouse reference genome (Ensembl release 106) using HISAT 2.2.4. The mapped reads of each sample were assembled using StringTie v1.3.1, and gene expression values were estimated with Fragments Per Kilobase of transcript per million mapped reads (FPKM) using RSEM software. Differential expression analysis of the RNA-seq dataset was conducted with DESeq2 (v.1.18.1). Genes were identified as differentially expressed with a false discovery rate (FDR) cut-off of 0.05 and an absolute fold change ≥2. Plots for enrichment analysis, including GO terms, KEGG, and Reactome, were generated using OmicSmart, a real-time interactive online platform for bioinformatic analysis (https://www.omicsmart.com).
ATAC-sequencing
Samples were prepared as described in the RNA-seq section, except that the cell pellets were resuspended in freezing medium containing 90% FBS and 10% DMSO, and aliquoted into cryovials. The samples were kept in a gradient cooling box for 10 min at 4 °C, 30 min at 20 °C, overnight at −80 °C, and then sent to the Gene Denovo Biotechnology Co. Ltd. (Guangzhou, China) for library construction and genomic sequencing. Libraries were prepared using the TruePrep DNA Library Prep Kit (Vazyme TD501) according to a protocol previously described (Buenrostro et al, 2015). Frozen cell samples were thawed quickly in a 37 °C water bath and centrifuged at 500 × g for 5 min at 4 °C. Cell pellets were resuspended by gentle pipetting for 3 min in cold lysis buffer containing 10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl_2_, and 0.1% IGEPAL CA-630. The resuspended cell pellets were centrifuged at 500 × g for 5 min at 4 °C to generate the crude nuclei. After removing the supernatant, nuclei were subjected to a transposition reaction for 30 min at 37 °C with gentle mixing. Transposed DNA was purified using a QIAGEN MinElute PCR Purification Kit (QIAGEN 28004). PCR amplification was performed using the N5XX and N7XX barcode oligos provided in the TruePrep Index Kit V2 for Illumina (Vazyme TD202). Amplified libraries were cleaned using AMPure XP beads (Beckman Coulter, Inc.), analyzed using an Agilent 2100 Bioanalyzer (Agilent Technologies), and sequenced using a NovaSeq 6000 (Illumina). ATAC-seq raw reads were trimmed to remove low-quality reads and adapters using trim-galore (version 0.6.7). Next, trimmed reads were aligned to the mouse reference genome (GRCm39) using Bowtie2 (version 2.4.5) with the “--very-sensitive -X 1000” option; reads mapped to the mitochondria were ignored. SAMtools (version 1.15.1) was applied to extract properly mapped reads, and duplicated reads were removed by sambamba (version 0.8.2). The Reads Per Kilobase per million mapped reads (RPKM) normalization method of deepTools bamCoverage (version 3.5.1) was used to transform alignment bam files into read coverage files in the bigWig format. Peak calling was performed using MACS2 (version 2.2.7.1) with the parameter “--SPMR --nomodel–q 0.001 --cut-off”. Differentially accessible peaks were identified using the DiffBind R package (version 3.8.4), peaks with FDR of <0.05, fold change ≥2 were identified as significantly differentially accessible regions, and stable regions were identified with a fold change <1.5. Motif enrichment analysis was performed using the HOMER (version 4.11) function “findMotifsGenome.pl” with default options. The Integrative Genomics Viewer (IGV, version 2.10.0) was used for data visualization.
Statistical analysis
Statistical analyses were performed using Prism 10.2.3 (GraphPad Software, Inc.). Data were expressed as means ± standard deviation (SD). Comparisons between two groups were performed using an unpaired two-tailed Student’s t-test. Comparisons of more than two groups and grouped data were performed using one- or two-way ANOVA, where applicable and corrected for multiple comparisons between groups or to a reference group using Tukey’s or Dunnett’s test. In multiple testing scenarios, adjusted P values greater than 0.0001 are reported as exact values; those less than 0.0001 are reported as P < 0.0001. The survival rate of mice after IAV infection was analyzed using the Kaplan–Meier method followed by the Gehan–Breslow–Wilcoxon test. Unless otherwise specified, n represents the number of individual biological replicates and is represented in the graphs as one dot per sample. All representative results were obtained from at least two to three independent experiments, except for the RNA-seq and ATAC-seq experiments. The experiments were neither randomized nor blinded, and the sample size was not predetermined by any statistical means. At least three samples were used for each experimental group and under each condition.
Supplementary information
Table EV1 Table EV2 Peer Review File Dataset EV1 Source data Fig. 1 Source data Fig. 2 Source data Fig. 3 Source data Fig. 4 Source data Fig. 5 Source data Fig. 6 Figure EV8A and B Source Data Expanded View Figures