Molecular definition of the endogenous Toll-like receptor signalling pathways.

Material availability

All unique and newly generated materials (cell lines, plasmids, bacteria strains) are available from the corresponding author on request. Cell and bacteria culture, infections and treatments THP-1 cells (TIB202, ATCC) were maintained in RPMI 1640 medium (11875093, Thermo Fisher Scientific) supplemented with l-glutamine (25030081, Thermo Fisher Scientific), sodium pyruvate (11360070, Thermo Fisher Scientific), penicillin–streptomycin (15140122, Thermo Fisher Scientific) and 10% FBS (R&D systems), at 37 °C under a 5% CO 2 atmosphere. THP-1 cells were differentiated with 50 ng ml −1 phorbol 12-myristate 13-acetate (PMA, P1585, Sigma-Aldrich) for 3 days and rested for 2 days by replacing the differentiation medium with complete medium without PMA. iBMDMs and HEK293T cells (CRL-11268, ATCC) were cultured in DMEM supplemented with glutamine, sodium pyruvate, penicillin–streptomycin and 10% FBS, at 37 °C under a 5% CO 2 atmosphere. Induction of protein expression was performed with 200 ng ml −1 doxycycline (D9891, Sigma-Aldrich) overnight. Cells were treated with 100 ng ml −1 LPS O55:B5 (ALX-581–013, Enzo) or Pam3CSK4 (vac-pms, Invivogen) unless otherwise indicated. Other treatments to activate TLR signalling were performed with LTA-SA (tlrl-pslta, Invivogen), Pam2CSK4 (tlrl-pm2s, Invivogen), poly(I:C)(tlrl-pic, Invivogen), Fla-ST (tlrl-epstfla, Invivogen), R848 (vac-r848, Invivogen) or CpG-DNA (ODN1826; tlrl-1826, Invivogen). Cells were treated with 20 μM zimlovisertib/PF-06650833 (HY-19836, MedChemExpress) to inhibit IRAK4 kinase activity. L. monocytogenes strain 10403S was maintained on brain–heart infusion (BHI; 237500, Becton Dickinson) agar plates. To generate fluorescent bacteria for imaging, we cloned a Listeria codon-optimized mScarlet open reading frame (ORF), synthesized using gBlocks (Integrated DNA Technologies), into PstI-HF (R3140, NEB) and KpnI-HF (R3142, NEB)-digested pLOV plasmid (D.E.H. laboratory). Lm -mScarlet was generated by electroporating L. monocytogenes WT with pLOV-mScarlet Lm and selection with 7.5 μg ml −1 chloramphenicol (CHL; C0378, Sigma-Aldrich) on BHI agar plates. One day before infection, bacteria from a single colony were inoculated and grown overnight at 30 °C without shaking in 2 ml BHI broth supplemented with 10 μg ml −1 CHL. Bacteria from the overnight culture were collected by centrifugation at 1,000 g for 3 min, washed with PBS twice and resuspended in 1 ml plain DMEM. Cells were infected with L. monocytogenes at a multiplicity of infection of 10 (or 30 for live imaging), and infections were synchronized by centrifugation at 750 g for 10 min. Infected cells were washed three times with warm PBS at 1 h after infection, and fresh medium containing 40 μg ml −1 gentamicin (15750060, Thermo Fisher Scientific) was added. For secondary infection assays, iBMDMs were infected as described above and collected at 1 h after infection using PBS + 2.5 mM EDTA, washed five times with PBS and then added to the cells seeded for secondary infection with added gentamicin. The bacterial multiplicity of infection used for infections was confirmed by plating on BHI agar plates. An overview of all cell lines and bacteria strains is provided in Supplementary Table 1 .

Generation of MyD88–AGF knock-in cells

We used an adapted version of CRISPaint 47 to tag endogenous MyD88 protein with a multifunctional APEX2-mEGFP-3×Flag (AGF)-tag, which relies on transfection of the following three plasmids into the target cells: A single guide RNA (sgRNA) expression plasmid for targeting the MyD88 coding sequence on the genome. To generate plasmids expressing sgRNAs against human or mouse MyD88 , complementary DNA oligonucleotides encoding the crRNA sequences were annealed and ligated into BsmBI-v2 (R0739, NEB)-digested pMLM3636 (#43860, Addgene; a gift from Keith Joung) using Quick Ligation kit (M2200, NEB). sgRNA efficiency was confirmed by transfecting Cas9-expressing HEK293T or NIH3T3 cells using Lipofectamine 2000 (11668027, Thermo Fisher Scientific) followed by amplification of the Cas9 target loci from purified gDNA at 3 days post transfection using Q5 High-Fidelity DNA Polymerase (M0491, NEB) and analysing editing frequency with CRISPR ICE v3.0 (Synthego Performance Analysis, ICE Analysis. 2019. v.3.0). See Supplementary Table 2 for sgRNA sequences and genotyping primers. A tag donor plasmid encoding the AGF-tag. First, the GFP ORF on p EGFP -N1 (Clontech) was mutated using site-directed mutagenesis to obtain an ORF for a monomeric GFP A206K mutant ( mEGFP ). Next, mEGFP ORF was amplified from pmEGFP-N1, APEX2 ORF was amplified from V5-APEX2-SENP2 (129276, Addgene; a gift from A. Ting) 48 and the pCRISPaint-3xFlag-PuroR tag donor plasmid (1000000086, Addgene; a gift from V. Hornung) 47 was linearized by amplification using Q5 DNA polymerase and primers to produce 15 bp complementary overhangs. The three obtained amplicons were subsequently assembled into the final pCRISPaint-APEX2-mEGFP−3xFlag-PuroR plasmid using Gibson assembly (E2611, NEB). To reduce the amount of DNA integrated into the target cell genome, we opted to use replication-deficient minicircle DNA. To generate this minicircle tag-donor, the pCRISPaint-AGF-PuroR plasmid was digested with NheI-HF (R3131, NEB) and PspOMI (R0653, NEB), and the obtained tag-encoding insert was ligated into SpeI-HF (R3133, NEB) and PspOMI-digested pMC.BESPX-MCS1 (MN100A-1, System Biosciences). The obtained pMC.BESPX-AGF-PuroR plasmid was transformed into the ZYCY10P3S2T Escherichia coli minicircle DNA producer strain (MN900A-1, System Biosciences). Replication-deficient minicircle DNA was produced by inoculating a starter culture with bacteria from a single colony into 2 ml LB Miller broth (L3152, Sigma-Aldrich) containing 50 μg ml −1 kanamycin and growing them for approximately 12 h at 37 °C with shaking at 250 rpm. In the evening, 400 ml terrific broth containing 50 μg ml −1 kanamycin was inoculated with 100 μl of the starter culture and grown overnight at 37 °C with shaking at 250 rpm. Once the culture had reached an optical density at 600 nm (OD 600 ) of approximately 4–5, the culture pH was adjusted to 6.5 with 1 N NaOH. Subsequently, minicircle production was induced by adding induction mix (400 ml LB broth, 16 ml 1 N NaOH, 400 μl 20% l-arabinose (final concentration, 0.01%; A3256, Sigma-Aldrich) and by incubating at 32 °C with shaking at 250 rpm for 8 h. Minicircle DNA was purified from the bacteria using the Maxiprep kit (12163, Qiagen) according to the manufacturer’s instructions, but doubling the volume of P1, P2 and P3 buffers during the purification. Successful preparation of minicircle DNA was confirmed by agarose gel electrophoresis and whole-plasmid sequencing (Plasmidsaurus). A frame-selector plasmid for linearizing the tag-donor plasmid and for expressing Cas9; here we used pCAS9-mCherry-Frame+1 as provided in the CRISPaint Gene Tagging Kit (Addgene). Before electroporation, the three plasmids were further purified by reprecipitation with phenol–chloroform and resuspended in double-distilled H 2 O. WT THP-1 cells or iBMDMs were collected and prepared for electroporation by washing twice with PBS followed by resuspension of 1 × 10 5 cells in 10 μl electroporation buffer R and adding the plasmids: 200 ng pMLM3636-MyD88-sgRNA targeting plasmid, 50 ng pCAS9-mCherry-Frame+1 frame-selector plasmid and 600 ng minicircle AGF-PuroR tag donor plasmid (ratio 4:1:12). Electroporation was performed using the Neon Transfection System (Thermo Fisher Scientific) with the following parameters: 1,700 V, 20 ms width, 1 pulse. Immediately after electroporation, cells were placed into 24-well plates in 1.5 ml complete medium with 20% FCS and left to recover at 37 °C, 5% CO 2 atmosphere. The next day, the medium was replaced with selection medium containing 0.5 μg ml −1 puromycin (A1113803, Thermo Fisher Scientific) for THP-1 cells or 3 μg ml −1 puromycin for iBMDMs. Note that the puromycin concentration had to be adjusted empirically, as the expression level of the PuroR cassette depends on the expression level of the tagged gene. After 4 days, the medium was replaced again with complete medium without puromycin, and the cells expanded for about 10 days. Finally, the cells were sorted for homogenously high GFP fluorescence, to obtain cells with a homozygous knock-in of the AGF tag. MyD88 gene editing, tag addition and homozygosity of the cells was assessed by analytical PCR combining a forward primer for the 3′ end of the MyD88 coding sequence with a reverse primer for the MyD88 3′ UTR to check for presence of unedited alleles (WT probe) or with a reverse primer within the AGF-tag sequence to check for the presence of successfully edited alleles (AGF probe; oligonucleotide sequences are provided in Supplementary Table 2 ). The PCRs used purified gDNA from WT or edited cells and PCR amplicons were analysed by agarose gel electrophoreses and Sanger sequencing. Finally, the AGF knock-in cells were validated for expression of MyD88–AGF by immunoblotting. Lentiviral transductions Plasmid transfections were performed by using a 10× mix in OptiMEM I reduced serum medium (31985062, Thermo Fisher Scientific) containing DNA plasmids and Lipofectamine 2000 transfection reagent at a 1:2 ratio according to the manufacturer’s instructions. For lentiviral packaging, HEK293T cells were transfected with equimolar ratios of the transfer plasmid pMD2.G (Addgene, 12259) and psPAX2 (Addgene, 12260, both gifts from D. Trono). The medium was replaced after 16 h with complete medium containing 5 mM sodium butyrate (B5887, Sigma-Aldrich) and the cells were left to produce lentiviral particles for >24 h. The virus-containing supernatant was filtered through a 0.43 μm syringe filter and supplemented with 5 μM polybrene (TR1003, Sigma-Aldrich). The target cells were resuspended in 500 μl of the virus-containing medium and spinfected for 30 min at 1,000 g . After 1 h, 1 ml complete medium was added, and the cells left to rest. Selection of successfully transduced cells used 1 μg ml −1 puromycin for THP-1 cells or 7.5 μg ml −1 puromycin for iBMDMs. Generation of knockout cells with CRISPR–Cas9 sgRNA sequences targeting the genes of interest were designed using CRISPick 49 . Complementary DNA oligonucleotides encoding the crRNA sequences were annealed and cloned into BsmBI-v2-digested pLentiCRISPR-v2 (52961, Addgene; a gift from F. Zhang) 50 and transduced into target cells using lentiviral particles. After selection with puromycin, the sgRNA efficiency was assessed by Sanger sequencing of the gDNA surrounding the sgRNA-target site followed by CRISPR ICE analysis. Knockout pools with >60% editing scores were subcloned by serial dilution into 96-well plates. The obtained clones were screened for absence of the target proteins by immunoblotting. For each of the cell lines, >5 clones with a confirmed absence of target proteins were pooled before undergoing a second screen using CRISPR ICE analysis and immunoblotting. All crRNA sequences, editing efficiencies and genotyping primers can are provided in Supplementary Tables 3 and 4 .

Generation of cells with doxycycline-inducible mPlum–p65 expression

MyD88–AGF iBMDM target cells were transduced with pLenti–Tet–ctrl plasmid 51 using lentiviral transduction and selected with 15 μg ml −1 blasticidin S (A1113903, Thermo Fisher Scientific) to generate iBMDM MyD88–AGF + Tet cells expressing rtTA2 S -M2 and tTS doxycycline-responsive transactivators. To generate a doxycycline-inducible expression plasmid for mPlum–p65, the mPlum ORF was amplified from pmPlum-N1 (54629, Addgene; a gift from M. Davidson) 52 and p65 ORF from Rela–cFlag–pcDNA3 (20012, Addgene; a gift from S. Smale) 53 , and the amplicons were cloned into BamHI-HF (R3136, NEB)-digested pLenti-Tet 51 using Gibson assembly. Obtained pLenti-Tet-mPlum-p65 plasmid was transduced into iBMDM MyD88–AGF + Tet target cells with lentiviral particles and the cells selected with 15 μg ml −1 blasticidin S and 200 μg ml −1 zeocin ( R25001 , Thermo Fisher Scientific) to generate MyD88–AGF + Tet-mPlum-p65 iBMDMs. Successful transduction was validated by immunoblotting after doxycycline stimulation overnight and the cells were sorted for uniform mPlum-p65 expression.

Generation of cells expressing IRAK4–mScarlet

To generate an IRAK4–mScarlet expression plasmid, Irak4 ORF was amplified from WT iBMDM cDNA and mScarlet ORF from pmScarlet-I_peroxisome_C1 (85065, Addgene; a gift from D. Gadella) 54 using Q5 DNA polymerase and inserted into BamHI-HF- and XhoI-digested (R0146, NEB) pLenti-P2A- PuroR 55 using Gibson assembly. Obtained pLenti–IRAK4–mScarlet–P2A–PuroR plasmid was transduced into MyD88–AGF iBMDMs using lentiviral particles and cells were selected using 7.5 μg ml −1 puromycin. Successful transduction was validated by immunoblotting and the cells were sorted for uniform IRAK4–mScarlet expression.

Flow cytometry and cell sorting

For analysis of TLR surface localization, 2 × 10 6 iBMDMs or THP-1 cells were seeded per well of a six-well plate, differentiated and treated as described above. Cells were collected by scraping in ice-cold PBS, washed twice with PBS, passed through a 35 μm nylon mesh strainer (352235, Corning) and resuspended in PBS containing either human TruStain FcX (Fc Receptor Blocking Solution; 422302, BioLegend) or mouse TruStain FcX PLUS (anti-mouse CD16/32; 156604, BioLegend), LIVE/DEAD Fixable Violet Dead Cell Stain ( L34964 , Thermo Fisher Scientific) and 1:100 diluted primary antibodies against TLRs or isotype control ( Supplementary Table 5 ). Cells were stained by incubation at 4 °C for 30 min in the dark, washed twice with ice-cold PBS and analysed on a LSRFortessa flow cytometer using FACSDiva v.9.7 software (BD Biosciences). The proportion of surface-localized TLRs was determined by calculating the mean fluorescence intensity using FlowJo v.10.8.2 and comparing to untreated cells after gating for single cells based on forward- and side-scatter characteristics and live cells. For sorting cells on the basis of fluorescent protein expression, cells were collected, washed in PBS, resuspended in PBS + 1% FCS to a concentration of 4 × 10 6 cells per ml and passed through a 35 μm nylon mesh strainer. Subsequently, cells were sorted into 500 μl complete medium on a BD FACSMelody Cell Sorter using BD FACSChorus v.3.0 software (BD Biosciences) by gating on singlets and fluorescent protein positive cells. Immediately after sorting, cells were seeded into six-well plates with complete medium. SDS–PAGE, native PAGE and immunoblotting For immunoblotting, 1 × 10 6 cells were seeded per well of a 24-well plate, differentiated and stimulated as described above. Cells were washed with ice-cold PBS and lysed for 15 min on ice in 250 μl RIPA buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with protease inhibitors (Protease Inhibitor Cocktail set III, EDTA free; 539134, Millipore) and 1 mM activated sodium orthovanadate (567540, Sigma Aldrich), 5 mM sodium fluoride (201154, Sigma Aldrich) and 3 mM β-glycerophosphate (G9422, Sigma-Aldrich) phosphatase inhibitors. Lysates were cleared by centrifugation at full speed for 10 min at 4 °C. The BCA assay (Pierce BCA protein assay kit; 23225, Thermo Fisher Scientific) was performed to determine protein concentrations. 15–20 μg of total protein per sample was mixed with 5× loading dye (125 mM Tris-HCl pH 6.8, 10% SDS, 50% glycerol, 0.06% bromophenol blue) supplemented with 5% β-mercaptoethanol, denatured at 95 °C for 10 min and separated by SDS–PAGE. For immunoblots against phosphorylated proteins, cells were washed with PBS and then directly lysed in 250 μl 1× loading dye supplemented with 5% β-mercaptoethanol. Prepared cell lysates were homogenized by several passages through a 1 ml sub-Q syringe attached to a 29 gauge needle, denatured at 95 °C for 10 min and separated by SDS–PAGE using home-made running buffer (25 mM Tris-HCl pH 8.3, 192 mM glycine, 0.1% (w/v) SDS) and 10%, 12% or 15% polyacrylamide gels for around 90 min at 140 V. For native PAGE, cells were washed with PBS and lysed in 200 μl lysis buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.05% Triton X-100) supplemented with protease and phosphatase inhibitors for 15 min on ice. Lysates were cleared by centrifugation at 4 °C for 10 min at full speed. Cleared supernatants were collected and protein concentration determined with BCA assays. Samples were prepared by mixing lysate containing 20 μg protein with 4× native-PAGE sample buffer (BN2003, Thermo Fisher Scientific) and loaded onto 4–16% Bis-Tris native PAGE gradient gels (BN2112BX10, Thermo Fisher Scientific). Native PAGE was performed using prechilled running buffers (BN2001 and BN2002, Thermo Fisher Scientific) and the gel was run at 120 V for around 4 h at 4 °C. After SDS–PAGE, proteins were transferred onto methanol-activated PVDF membranes (IPVH00010, Millipore) using wet-transfer at 400 mA for 1 h in Towbin buffer (25 mM Tris-HCl pH 8.3, 192 mM glycine, 20% (v/v) methanol). Transfer after native PAGE was performed at 40 V overnight at 4 °C. Depending on the primary antibody used for immunoblotting, the membranes were blocked with 5% BSA or 5% non-fat dried milk in TBS-T (0.05% Tween-20) for >30 min at room temperature. Incubation with primary antibodies ( Supplementary Table 6 ) was performed at a dilution of 1:1,000 for all of the antibodies, except for anti-actin, which was diluted 1:5,000, and at 4 °C overnight. Blots were developed by washing with TBS-T, probing with 1:10,000 diluted HRP-conjugated secondary antibodies ( Supplementary Table 7 ), washing again and imaging on the ChemiDoc MP imaging system (Bio-Rad) with Image Lab v.6.1 (Bio-Rad) software using SuperSignal West Pico PLUS Chemiluminescent Substrate (34580, Thermo Fisher Scientific). For reprobing membranes with different primary antibodies, membranes were stripped using Restore PLUS western blot stripping buffer (46430. Thermo Fisher Scientific). Densitometry analysis was performed using quantification of protein band intensities with ImageJ/Fiji (v.2.14.0) 56 and normalization to the loading control.

Immunoprecipitations Immunoprecipitation of endogenous

MyD88 was performed as previously described 57 . In brief, 5 × 10 6 cells were seeded in six-well plates, differentiated and treated as described above. The cells were washed in ice-cold PBS and scraped from the plates. Whole-cell lysates were prepared by adding 500 μl lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5% glycerol, 1 mM sodium deoxycholate and 1% NP-40) containing protease and phosphatase inhibitors and incubation for 15 min on ice. Lysates were cleared by centrifugation at 4 °C for 10 min at full speed. The cleared lysates were collected and 100 μl of the supernatants was saved as sample inputs. The remaining supernatants were supplemented with 0.5 μg of anti-MyD88 antibodies ( Supplementary Table 6 ) and 15 μl (bed volume) of protein G Sepharose beads (for endogenous MyD88; P3296, Millipore) or with 15 μl (bed volume) of anti-Flag(M2) agarose beads (for MyD88–AGF; A2220, Millipore). Immunoprecipitation was performed by incubation at 4 °C for 6 h on a tube rotator. The beads containing the protein complexes were washed five times with lysis buffer. Proteins were eluted from Flag(M2)-beads using 200 ng ml −1 3×Flag peptide (F4799, Sigma-Aldrich) in lysis buffer by incubation on an orbital shaker (1,400 rpm) for >1 h at room temperature. Elution after immunoprecipitation of endogenous proteins was performed by acidification with 50 μl of 0.2 M glycine pH 2.0 followed by an immediate wash with 50 μl of lysis buffer. Input, unbound and elution fractions were kept during immunoprecipitations and analysed using immunoblotting. For quantifying the proportion of activated signalling effector proteins within myddosomes, fractions were maintained at equal volumes to obtain semi-quantitative results.

Differential detergent extraction

To quantify the proportion of MyD88 contained within myddosomes, 1 × 10 6 iBMDMs or THP-1 cells were seeded in 24-well plates, differentiated and treated as described above. Cells were collected by scraping in ice-cold PBS and lysed in 150 μl buffer IGEPAL (25 mM HEPES pH 7.4, 100 mM NaCl, 1 mM MgCl 2 , 1% (v/v) IGEPAL CA-630) containing protease and phosphatase inhibitors and then incubating on ice for 30 min. Cell lysates were homogenized by passing through a 29 gauge needle. BCA assays were performed to determine protein amounts and concentrations equalized. A total of 30 μl of lysate was saved as whole-cell lysate (fraction W) and the remaining lysate was centrifuged at 4 °C and 21,000 g for 45 min. In total, 30 μl of the cleared supernatant was saved as the IGEPAL-soluble fraction (fraction S). The protein pellet was washed with 250 μl ice-cold wash buffer (25 mM HEPES pH 7.4, 100 mM NaCl, 1 mM MgCl 2 ) and dissolved in 90 μl buffer SDS (25 mM HEPES pH 7.4, 100 mM NaCl, 1 mM MgCl 2 , 1% (v/v) SDS) containing protease and phosphatase inhibitors by shaking at 1,200 rpm at room temperature for 1 h. Finally, the protein suspension was centrifuged at 21,000 g for 45 min at room temperature and the cleared supernatant saved as IGEPAL-insoluble fraction (fraction I). Obtained fractions (W, S and I) were analysed by SDS–PAGE followed by immunoblotting and proportion of MyD88 in each fraction determined by densitometry.

RT–qPCR analysis of TLR response gene expression

RNA was extracted from 0.5 × 10 6 cells using Trizol (15596026, Thermo Fisher Scientific) and purified by isopropanol precipitation, supplementing with 5 μg ml −1 GlycoBlue (AM9516, Thermo Fisher Scientific). The purified RNA was reverse transcribed using the high-capacity cDNA synthesis kit (4368813, Thermo Fisher Scientific). qPCR was performed using the iTaq Universal SYBR Green Supermix (1725121, Bio-Rad), 20 ng cDNA in a 10 μl reaction and primers at 1 μM final concentration (sequences are provided in Supplementary Table 8 ). Data were recorded on the CFX384 Real-Time system (Bio-Rad) and analysed using CFX Manager v.3.1 (Bio-Rad). Primer specificity was ensured by designing primers to span exon–exon junctions whenever possible. For each primer pair, a melt curve was recorded, and amplicon sizes were analysed by agarose gel electrophoresis ( Supplementary Fig. 4 ). Recorded C t values were normalized to the recorded C t of human HPRT1/ mouse Hprt1 housekeeping genes, and data were plotted as Δ C t (relative expression) or as the fold change compared with the untreated controls. Cytokine ELISAs, LDH, ROS and Griess assays ELISA was performed to measure TNF and IL-6 secretion. Cell culture supernatants were cleared of cell debris by centrifugation at 400 g for 5 min. Cleared supernatants were diluted 1:20–1:50 with ELISA diluent to be within the assay range. Concentrations of TNF and IL-6 were measured according to the manufacturer’s protocols using uncoated human IL-6 (88–7066-88), uncoated human TNF (88–7346-88), uncoated mouse IL-6 (88–7064-88) or uncoated mouse TNF (88–7324-88, all from Thermo Fisher Scientific) ELISA kits. Cell culture supernatants were assayed for LDH release using CyQUANT LDH Cytotoxicity Assays ( C20301 , Thermo Fisher Scientific) according to the manufacturer’s instructions. A microplate reader (Tecan Spark, Tecan) controlled by Tecan SPARKCONTROL v.3.1SP1 software was used to measure LDH release by recording absorbance at 490 nm and 680 nm for correction. Detergent-treated cells and untreated cells were used as positive and negative controls. ROS production was measured by loading cells seeded in black-wall 96-well plates with 10 μM H 2 DCFDA substrate (D399, Thermo Fisher Scientific) for 2 h. Cells were treated as described above and the fluorescence of oxidized substrate was recorded at 488 nm at the indicated times after treatment using bottom optics on a microplate reader. Fluorescence values of cells not loaded with the substrate were used for background correction and values from untreated cells to calculate fold changes. Production of RNS was monitored with colorimetric Griess assays. In brief, 50 μl cleared cell culture supernatants were mixed with 50 μl reagent A (1% (w/v) sulfanilamide in 2.5% (w/v) H 3 PO 4 ) and 50 μl reagent B (0.1% (w/v) N -(1-naphthyl)ethylenediamine dihydrochloride in 2.5% (w/v) H 3 PO 4 ) and incubated at room temperature for 15 min. Absorbance was measured at 550 nm with a microplate reader and the concentration of RNS was calculated using a NaNO 2 reference curve (200–0.78 μM). Myddosome mapping using spatiotemporally resolved proteomics In total, 20 × 10 6 MyD88–AGF iBMDM or THP-1 cells were seeded onto 10 cm dishes in triplicates (18 plates in total per cell line and PAMP treatment condition). Seeded cells were treated with 100 ng ml −1 LPS or P3C for 12, 3, 1 or 0.5 h in reverse order. Six 10 cm dishes were left untreated as controls. Biotin phenol (biotinyl tyramide; SML2135 Sigma-Aldrich) was added to a final concentration of 2.5 mM for THP-1 cells or 5 mM for iBMDMs 2 h before the end of the time course. The cells were washed 3 times with PBS ++ (PBS pH 7.4, 50 mM MgCl 2 , 100 mM CaCl 2 ) and the biotinylation reaction was induced by adding 8 ml of 0.5 mM H 2 O 2 in PBS ++ for 30 s and quenched by adding 8 ml 2× stop/wash solution and placing the cells on ice. Three dishes of untreated cells were kept aside without H 2 O 2 treatment as controls. Subsequently, the cells were washed twice with 1× stop/wash solution (PBS pH 7.4, 50 mM MgCl 2 , 100 mM CaCl 2 , 5 mM Trolox, 10 mM sodium ascorbate, 10 mM sodium azide), collected by scraping and lysed for 30 min on ice in 1 ml RIPA lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM EDTA, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, 1% Triton X-100, 5 mM Trolox, 10 mM sodium ascorbate, 10 mM sodium azide) supplemented with protease and phosphatase inhibitors. Obtained lysates were cleared by centrifugation at full speed for 15 min at 4 °C. The cleared lysates were collected and 20 μl used to perform SDS–PAGE and immunoblot analysis with streptavidin-HRP (N100, Thermo Fisher Scientific) to validate successful biotinylation. BCA assays were performed and 10 mg total protein from each sample was transferred to a new tube and volumes were equalized with RIPA lysis buffer. Proteins were reduced and alkylated by adding 5 mM TCEP (77720, Thermo Fisher Scientific) and 10 mM iodoacetamide (I1149, Sigma Aldrich) and incubating at 37 °C for 2 h on a tube rotator in the dark. The reaction was quenched with 15 mM DTT (final concentration) and the samples added to 100 μl Pierce Streptavidin magnetic beads (slurry; 88817, Thermo Fisher Scientific) per sample that had been equilibrated by three washes with RIPA lysis buffer. Immunoprecipitation was performed at 4 °C for 4 h on a tube rotator. The beads containing the biotinylated proteins were washed twice with RIPA lysis buffer, once with 1 M KCl, once with 0.1 M Na 2 CO 3 , once with 2 M urea in 10 mM Tris-HCl pH 8.0 and twice with RIPA lysis buffer without detergents and any excess liquid removed. Beads were resuspended in 50 μl 200 mM EPPS pH 8.5. Lys-C was added at an estimated 1:100 protease-to-peptide ratio overnight at room temperature with gentle shaking. Trypsin was used for further digestion for 6 h at 37 °C at the same ratio as Lys-C. After digestion, 12 μl acetonitrile was added into each sample to 30% final volume. 50 μg TMTpro reagent (126, 127N, 127C, 128N, 128 C, 129N, 129C, 130N, 130C, 131N, 131C, 132N, 132C, 133N, 133C, 134N, 134C and 135) in 10 μl acetonitrile was added to each sample. After 1 h of labelling, 2 μl of each sample was combined, desalted and analysed using MS. Total intensities were determined in each channel to calculate normalization factors. After quenching using 0.3% hydroxylamine, 11 samples were combined in 1:1 ratio of peptides based on normalization factors. Samples were combined, dried and fractionated with basic pH reversed-phase (BPRP) high-performance liquid chromatography (HPLC) as described before. The fractions were collected onto a 96-well plate and combined for 24 fractions in total. Twelve fractions were desalted and analysed using liquid chromatography–tandem MS (LC–MS/MS). Proteome data were collected on an Orbitrap Fusion Lumos mass spectrometer coupled to a Proxeon NanoLC-1200 UHPLC. The peptides were separated using a 100 μm capillary column packed with around 35 cm of Accucore 150 resin (2.6 μm, 150 Å; Thermo Fisher Scientific). The mobile phase was 5% acetonitrile, 0.125% formic acid (A) and 95% acetonitrile, 0.125% formic acid (B). In brief, the Thermo FAIMS Pro device was operated with the default parameters (inner and outer electrode were set at 300 °C, yielding a full width at half maximum of between 10 and 15 V and dispersion voltage was set at −5,000 V). Each fraction was eluted using a 90 min method over a gradient from 6 to 24% B. Peptides were ionized with a spray voltage of 2,700 kV. The instrument method included Orbitrap MS1 scans (resolution of 0.6 × 10 5 ; mass range 350–1,350 m / z ; automatic gain control target 4 × 10 5 ) and ion-trap MS2 scans (HCD collision energy of 36%; AGC target 1.0 × 10 5 ; rapid-scan mode). This data acquisition includes high-field asymmetric-waveform ion-mobility spectrometry (FAIMS). The dispersion voltage for FAIMS was set at 5,000 V, the compensation voltages (CVs) were set at −40 V, −60 V and −80 V, and TopSpeed parameter was set at 1 s per CV. Detailed parameters for MS2 are embedded in the RAW files. MS data were processed using a Comet-based pipeline. Spectra were converted to mzXML using a modified version of ReAdW.exe. Database search included all entries from a human UniProt database (accessed 24 November 2021) or mouse UniProt database (accessed 26 January 2022). This database was concatenated with one composed of all protein sequences in the reversed order. Searches were performed using a 50 ppm precursor ion tolerance for total protein-level analysis. The product ion tolerance was set to 50 ppm for MS2 analysis. TMT tags on lysine residues, peptide N termini (+304.207 Da), and carbamidomethylation of cysteine residues (+57.021 Da) were set as static modifications, while oxidation of methionine residues (+15.995 Da) was set as a variable modification. Peptide-spectrum matches (PSMs) were adjusted to false-discovery rate (FDR) < 0.01. PSM filtering was performed using a linear discriminant analysis, as described previously, while considering the following parameters: XCorr, ΔCn, missed cleavages, peptide length, charge state and precursor mass accuracy. For TMT-based reporter ion quantitation, we extracted the summed signal-to-noise ratio for each TMT channel and found the closest matching centroid to the expected mass of the TMT reporter ion. For protein-level comparisons, PSMs were identified, quantified and collapsed to a peptide FDR < 0.01 and then collapsed further to a final protein-level FDR < 0.01, which resulted in a final peptide level FDR < 0.001. Moreover, protein assembly was guided by principles of parsimony to produce the smallest set of proteins necessary to account for all observed peptides. PSMs with poor quality, MS2 spectra with a TMT reporter summed signal-to-noise ratio of less than 100 or with no MS2 spectra were excluded from quantification. The MS proteomics data have been deposited to the ProteomeXchange Consortium through the PRIDE 58 partner repository with the dataset identifier PXD047416. For the final data analysis, endogenously biotinylated proteins were removed from the MS data. Next, the average intensity of all samples was calculated and used to remove proteins that were 3 larger than the single GFP molecules. The average myddosome position over time was extracted by determining the centroid of each myddosome label within single cells, and merging the cells over time, thereby generating the overall cell outline/canvas and a topographical map depicting the movement of myddosomes. Other object detection was performed using the BactFluorOmni neural network, as implemented in the Omnipose (v.1.0.6) 67 framework for detection of Listeria . Nuclei were detected using the ImageJ plugin for StarDist (v.0.7.0) 68 . Label images of nuclei and cells were used for tracking of objects using the ImageJ plugin Trackmate (v.7.12.1) 69 , 70 and the simple LAP tracker (cells did not divide after PAMP stimulation). Tracking was performed using the following empirically determined parameters: max. linking distance: 8 μm; max. gap closing max distance: 10 μm; and max. frame gap: 2. As myddosomes were highly dynamic and were merging/splitting over time, we used the myddosome label images and tracked the complexes using the LAP tracker. Tracking was performed using the following empirically determined parameters: max. linking distance: 2 μm; max. gap closing max distance: 4 μm; max. frame gap: 3; max distance (split): 0.5 μm; and max distance (merge): 0.5 μm. Calculated tracks were filtered for a minimal length of 3 frames for the myddosomes and 50 frames for the cells and nuclei to remove artefacts. The obtained tracking schemes were further used to determine the average myddosome lifetime as well as merge and split rates of myddosomes per cell. To produce myddosome fate maps (plotting size versus IRAK4:MyD88 ratio), myddosomes were tracked as described above but filtered for a minimum track length of 175 frames. Tracked labels were then classified on the basis of whether or not they exceed 2 μm 2 for more than 10 consecutive frames at any point during the live imaging (major myddosomes), and 50 randomly selected myddosomes or major myddosomes were analysed for their IRAK4:MyD88 ratio and sizes over time. Detected nuclei and cell labels were used to analyse NF-κB activation in PAMP-treated macrophages. To do so, we tracked the labels over time, and used labelling arithmetic to match the cell labels with the corresponding nuclei labels. The labels were used as masks to calculate the total mPlum–p65 fluorescence in the respective areas and the ratio between nuclear and cytosolic fluorescence on a per cell basis. Cells with a nuclear:cytosolic ratio of >1.0 were considered to have active NF-κB signalling. To classify myddosomes on the basis of recruitment of core and effector proteins, we first detected the cells and myddosomes and then used the myddosome labels as seeds to crop areas of interest surrounding the myddosomes from the immunofluorescence images. Cropped images were normalized and passed through an in-house-trained neural network classifier (MyddoNet) to sort myddosomes into three categories: positive, negative or artefacts. MyddoNet is derived from a previously trained deep convolutional neural network for classification of protein recruitment to subcellular structures 71 and was retrained using a dataset of 2,000 manually annotated myddosome images. Cross-validation showed that MyddoNet classified myddosomes with an overall accuracy of 94.9%. As biological immunofluorescence imaging data are insensitive to orientation, myddosome images were passed through the classifier in five different orientations (rotation and flip) and the mode of classification was used as overall result to increase accuracy. After discarding the images classified as artefacts, the proportion of stain-positive myddosomes was calculated. Pixel-based colocalization analysis was performed by calculating the Pearson’s correlation coefficient with the Coloc2 plugin (v.3.0.6) as implemented in ImageJ. The analysis was restricted to the cell cytosol or the myddosomes using previously determined cell and myddosome labels. Colocalization was calculated using z -stack images with 15 planes and 0.5 μm spacing and on a per-cell basis. Cells were stained with deep/far-red dyes and channels were imaged sequentially to minimize bleed-through to the GFP channel (MyD88–AGF). Moreover, calibration images were acquired using 0.5 μm TetraSpeck Microspheres (T7281, Thermo Fisher Scientific) and used to determine chromatic-shift between the channels. Shift was corrected with sub-pixel resolution using the TransformJ: Translate plugin in ImageJ prior to analysis. To quantify myddosome association with LPS-coated beads, a circular region of interest was drawn around the detected bead labels extending outward by 0.5 μm in all directions. Myddosomes were classified as bead associated if >50% of their label overlapped with this region of interest. Recombinant protein production and TLR4 polyclonal antibody generation The DNA sequence encoding the cytosolic tail of mouse TLR4 (TLR4 cyto ; amino acids 668–835) was amplified by PCR, cloned into the protein expression vector pGEX-6P-1 and transformed into E. coli XL1-Blue. For protein production, 30 ml of an overnight culture of bacteria expressing GST–TLR4 cyto was transferred to 750 ml LB medium containing 100 μg ml −1 ampicillin and grown to an OD 600 of 0.6–0.8, when protein expression was induced by adding isopropyl β-d1-thiogalactopyranoside to a final concentration of 0.4 mM. The cultures were further incubated at 18 °C for 16–18 h with shaking at 250 rpm. Bacteria were collected by centrifugation at 6,000 g and lysed by sonication in the presence of protease inhibitors. The soluble fractions were collected by centrifugating twice at 12,000 g and 4 °C. GST–TLR4 cyto was purified with Glutathione Sepharose 4 Fast Flow resin (17513201, GE Healthcare) and captured proteins were eluted with 25 mM reduced glutathione in elution buffer (20 mM Tris/HCl pH 8.0, 100 mM NaCl). After GST-tag removal using PreScission Protease (27084301, GE Healthcare), the recombinant TLR cyto protein was used as an antigen to raise polyclonal antibodies in rabbits by Pocono Rabbit Farm and Laboratory. Polyclonal antibodies specific to TLR4 cyto were affinity-purified from immunized rabbit sera against the antigen coupled to Affigel (1536099, Bio-Rad) according to the manufacturer’s instructions. In brief, 1 mg of TLR4 cyto antigen was coupled to activated Affigel at 4 °C for >6 h. Unbound antigen was washed off with cold PBS-T. The sera (10–15 ml) from immunized animals were then incubated with antigen-coupled Affigel at 4 °C on a tube rotor overnight. After extensive washing with PBS-T, the captured polyclonal antibody against TLR4 cyto was eluted with elution buffer (glycine/HCl pH 2.5, 20 mM HCl). The elution fraction was immediately dialysed against PBS to neutralize the acidic pH and remove excess glycine, and the concentration of purified antibody was determined by Braford assay.

Primary antibody validation

All primary antibodies used in this study were validated for specificity in immunoblot and immunofluorescence analysis ( Supplementary Fig. 2 ). The newly generated mouse-specific TLR4 antibody was validated by immunoblotting with lysates from WT or Δ Tlr4 iBMDMs and with lysates of human HEK293T cells that were transfected with a Flag-mTLR4 expression plasmid or pcDNA3 empty vector control ( Supplementary Fig. 2a ). The rabbit monoclonal anti-pIRAK4 (Thr345/Ser346) antibody was a gift from S. De 28 . Antibodies for ATG5, ATG12, IKKα, IKKβ, IKKγ, IRAK1, IRAK2, IRAK4, human MyD88, mouse MyD88, p38α, p62, p65, TBK1 and TRAF6 were validated by immunoblotting using knockout cells generated using CRISPR–Cas9 (Figs. 4a and 5a and Extended Data Figs. 1d and 10a ). Antibodies for total protein or a phosphorylated species of ATG5, ATG12, cGAS, IKKα/β, IKKγ, IRAK2, IRAK4, IRF5, LAMP1, LC3B, p38α, p62, p65, TBK1 and TRAF6 were validated by performing immunofluorescence staining and imaging with the respective knockout cells ( Supplementary Fig. 2b ). The antibody for G3BP1 was validated for use in immunofluorescence assays by staining MyD88–AGF iBMDMs treated with 3 μM thapsigargin (SML1845, Sigma-Aldrich) for 3 h ( Supplementary Fig. 2c ). Cell staining with Proteostat for protein aggregates was validated using MyD88–AGF iBMDMs treated with 10 μM MG132 proteasome inhibitor (part of ENZ-51035, Enzo) for 12 h ( Supplementary Fig. 2d ). All other primary antibodies were obtained from other research groups and were validated in previous studies, including GM130 72 ; HDAC6, NLRP3, γ-tubulin and TGN38 73 ; and ATG16L1, K63-ubiquitin, M1-ubiquitin and total ubiquitin 74 .

Data handling and statistics

Data were plotted using Prism v.10.1.0 (GraphPad) and presented as the means of multiple experiments, with error bars showing the s.d. unless stated otherwise. The significance of results was determined using nonparametric one-way ANOVA, unpaired t -tests or two-way ANOVA, as indicated in the figure legends. Benjamini, Krieger and Yekutieli false-discovery-rate-based correction ( Q = 5%) for multiple comparisons as implemented in Prism was used when making three or more comparisons. Uncropped immunoblot images are provided in Supplementary Data 1 and raw data are provided as source data.

Material availability

All unique and newly generated materials (cell lines, plasmids, bacteria strains) are available from the corresponding author on request.

Supplementary Material Supplementary Video 1 Supplementary Video 2 Supplementary Video 3 Supplementary Video 4 Supplementary Video 5 Supplementary Video 6 Supplementary Video 7 Supplementary data 1_uncropped gels Supplementary data 2_MassSpec_data Source data Supplementary Figures and Tables Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41586-024-07614-7 .