The binding of OTULIN restrains LUBAC activity to prevent TNF-driven immunopathology

17 Met1-linked ubiquitin chains (Met1 -Ub), synthesised by the linear ubiquitin chain 18 assembly complex (LUBAC) and disassembled by the deubiquitinase OTULIN, 19 critically regulate inflammatory signalling. Although OTULIN’s activity is essential to 20 prevent TNF -driven autoinflammatory pathology and embryonic lethality, the 21 regulatory significance of its direct interaction with LUBAC remains unclear. Here, we 22 reveal that mice harbouring a point mutation (OTULINY56A) in the OTULIN PUB -23 interacting motif , which disrupts OTULIN -LUBAC interaction , are viable without 24 spontaneous immunopathology. However, OtulinY56A/Y56A mice exhibited 25 hypersensitivity to TNF-induced toxicity, which was not prevented by inhibiting RIPK1 26 kinase-mediated cell death. Mechanistically, disruption of the OTULIN -LUBAC 27 interaction led to Met1-linked autoubiquitination, which enhanced LUBAC’s activity 28 and increased Met1 -Ub accumulation at the TNF receptor signalling complex . This 29 stabilised the signalling complex even after dissociation from TNF , increased NF-κB 30 signalling and, contrary to loss of OTULIN or its activity, protected cells from TNF -31 induced apoptosis. During systemic Listeria monocytogenes infection, the increased 32 response to TNF in OtulinY56A/Y56A mice exaggerated pathology without affecting 33 bacterial burden. Collectively, we identify the physical association of OTULIN to 34 LUBAC as a critical brake that restricts LUBAC’s function and Met1 -Ub-dependent 35 inflammatory signalling , thereby preserving tissue integrity and promoting disease 36 tolerance during acute immune activation. 37 38 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 3

39 Inflammation is an essential protective mechanism during infection but can cause tissue 40 damage and fatality when dysregulated 1, 2 . Tumour necrosis factor (TNF) promotes 41 inflammation through activation of mitogen-activated protein kinase ( MAPK) and 42 nuclear factor-kappa B (NF-κB) signalling, and indirectly by inducing cell death2, 3. The 43 latter exacerbates inflammation by releasing damage-associated molecular patterns 44 (DAMPs) from dying cells 2, 3 . Therefore, appropriate regulation of TNF -driven 45 cytokine production and cell death is critical for preventing sterile inflammation and 46 immune pathology in response to infection. 47 The l inear ubiquitin chain assembly complex (LUBAC), composed of HOIP, 48 HOIL-1 and SHARPIN, is recruited to the TNF receptor 1 (TNFR1) signalling complex 49 (TNF-RSC), where it conjugates Met1 -linked ubiquitin (Ub) chains (Met1-Ub) onto 50 various ubiquitinated substrates, including TNFR14, 5 . The Met1 -Ub function s as a 51 scaffold t o facilitate NF -κB signalling and suppress es TNF-induced and RIPK1 -52 mediated cell death through the recruitment or retention of Met1 -Ub binding proteins 53 such as the NEMO-IKK complex, A20 and ABIN1/25-8. 54 Met1-Ub assembly by LUBAC is counterbalanced by the Met1 -Ub-specific 55 deubiquitinase (DUB) OTULIN (OTU DUB with linear linkage specificity) 9. 56 Accordingly, OTULIN dysfunction causes OTULIN-related autoinflammatory 57 syndrome (ORAS), a TNF-driven disease in humans5, 10-12. In mice, OTULIN activity 58 is required during embryogenesis by protecting against aberrant cell death and 59 regulating angiogenesis, and OTULIN deficiency or ablation of OTULIN activity in 60 adult mice leads to systemic autoinflammation 10, 13 -15. Mechanistically, OTULIN 61 prevents the accumulation of Met1 -Ub on LUBAC subunits and maintain s normal 62 LUBAC levels in a cell type-specific manner10-13. 63 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 4 LUBAC is also regulated by the DUB CYLD, which preferentially cleaves Lys63- 64 and Met1-Ub16, 17. CYLD regulates ubiquitination at receptor signalling complexes and 65 promotes TNF-induced cell death, but its role in regulating Met1-Ub is ambiguous18, 19. 66 OTULIN and CYLD (via its adaptor SPATA2 ) both interact with the HOIP 67 peptide:N-glycanase/UBA- or UBX-containing proteins (PUB) domain via a conserved 68 PUB-interacting motif (PIM) in OTULIN and SPATA2 20, 21. TNF stimulation rapidly 69 recruits LUBAC in complex with SPATA2 -CYLD to the TNF-RSC where CYLD 70 promotes the retention of LUBAC at the TNF-RSC20. In contrast, OTULIN appears not 71 to be recruited with LUBAC to receptor signalling complexes 4, 20, 22 . Thus, the 72 physiological role of the OTULIN-LUBAC interaction and how it influences TNF 73 signalling outcomes is not known.12 74 In this study, we reveal that th e OTULIN -LUBAC interaction, in contrast to 75 OTULIN activity, is dispensable for embryonic development and immune homeostasis 76 in unchallenged mice. Rather, the interaction restricts LUBAC and limits TNF-induced 77 signalling and cytokine production , which protects against pathology during immune 78 activation. 79 80

81 The OTULIN-LUBAC interaction prevents LUBAC auto -ubiquitination but is 82 dispensable for embryonic development 83 Tyrosine 56 (Y56) within the evolutionarily conserved PIM of OTULIN mediates the 84 binding to the HOIP PUB domain and is essential for OTULIN’s interaction with 85 LUBAC21, 23 . To investigate the physiological relevance of th is interaction, we 86 generated knock-in mice carrying a Y56A point mutation in OTULIN ( Figure S1A). 87 Unexpectedly, homozygous OtulinY56A/Y56A mice displayed normal viability (Figure 88 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 5 S1B), distinct from the embryonic lethality caused by OTULIN deficiency or catalytic 89 inactivation10, 13-15. Genotype distribution of o ffspring followed Mendelian ratios and 90 adult OtulinY56A/Y56A mice gained body weight (BW) comparabl y to wildtype (WT) 91 littermates (Figures 1A and 1B). Spleen-to-BW ratios were also unchanged in contrast 92 to mice with myeloid-specific ablation of OTULIN10 (Figure S1C). In accordance, flow 93 cytometric analysis showed comparable numbers of myeloid and lymphoid cell 94 populations, as well as similar T cell sub set distributions between genotypes (Figures 95 1C, S1D, and S1E ), indicating that OtulinY56A/Y56A mice did not develop spontaneous 96 systemic i nflammation as described in adult mice following ablation of Otulin or 97 OTULIN activity10, 13. 98 Disruption of the OTULIN -LUBAC interaction in mice was confirmed by 99 immunoprecipitation (IP) of LUBAC from OtulinY56A/Y56A splenocytes, mouse dermal 100 fibroblasts (MDFs) and bone marrow-derived macrophages (BMDMs) (Figures 1D, 1E, 101 and S2A), and by biotin proximity labelling using TurboID coupled to OTULIN that 102 was stably expressed in OTULIN -deficient NIH 3T3 cells 24 (Figure S2B). TurboID-103 OTULINWT extensively biotinylated HOIP, HOIL -1 and SHARPIN but no 104 biotinylation of LUBAC subunits was detected in TurboID-OTULINY56A cells (Figure 105 S2B). Notably, the disruption of OTULIN -binding did not affect the interaction of 106 LUBAC with CYLD or with p97, which also contains a PIM that can bind the HOIP 107 PUB domain21, 23 (Figures 1D and 1E ). Thus, the interaction between OTULIN and 108 LUBAC is dispensable for mouse embryogenesis and immune homeostasis. 109 Primary OtulinY56A/Y56A cells showed accumulation of Ub-modified HOIL-1 and 110 increased LUBAC-associated Met1-Ub relative to WT cells (Figures 1D, 1E, and S2A). 111 However, this did not lead to a reduction in the level of LUBAC subunits as has been 112 reported in various cell types deficient for OTULIN or OTULIN activity (Figures 1D 113 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 6 and 1E)10-13. Intrigued by this, we investigated the distinct role of the OTULIN PIM 114 versus loss of OTULIN or OTULIN activity in Otulin-knockout NIH 3T3 cells 115 (OtulinKO) reconstituted with OTULIN variants (Figure 1F). As expected, OTULINWT 116 efficiently suppressed the elevated Met1 -Ub levels and auto-ubiquitination of HOIP 117 and HOIL-1 in OTULIN-deficient cells, and led to increased HOIP and HOIL-1 levels10, 118 13 (Figure 1F). OTULIN Y56A also substantially reduced Met1 -Ub levels and 119 ubiquitination of HOIP albeit less efficiently than OTULINWT, and increased HOIP and 120 HOIL-1 levels (Fig ure 1F). However, the Met1-Ub on HOIL -1 w as only modestly 121 reduced by OTULINY56A relative to OTULIN-deficient cells, in line with the increased 122 ubiquitination of HOIL-1 in OtulinY56A/Y56A MDFs and BMDMs (Figures 1F, 1G, and 123 S2A). In contrast to OTULINY56A, catalytically inactive OTULINC129A enhanced Met1-124 Ub levels and LUBAC auto-ubiquitination and led to a further reduction in HOIP and 125 HOIL-1 relative to the levels in OTULIN -deficient cells, in a manner that was 126 dependent on the interaction with HOIP (Fig ure 1F). This suggest ed a) that the 127 OTULIN-HOIP interaction is critical for removing the Met1-Ub conjugated to the auto-128 monoubiquitinated HOIL-1 (monoUb-HOIL-1) 25, and b) that inactive OTULIN bound 129 to LUBAC interferes with the removal of Met1 -Ub by other cellular DUBs 21, 26. We 130 obtained similar results in mouse embryonic fibroblasts ( Otulindel/del MEFs) 131 reconstituted with the same OTULIN variants although the effects were less 132 pronounced, possibly due to low expression of the reintroduced OTULIN variants 133 (Figure S2C). Treatment of SHARPIN IP samples with recombinant OTULIN 134 confirmed that Ub chains on HOIP and HOIL-1 were Met1-linked (Figure S2C). 135 CYLD also can disassemble Met1-Ub but disruption of the interaction between 136 LUBAC and CYLD through genetic ablation of Spata2 did not increase LUBAC auto-137 ubiquitination or LUBAC-associated Met1-Ub in MDFs (Figure 1G). Also, purification 138 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 7 of CYLD-associated LUBAC from OTULIN -mutated MEFs showed clear Met1-Ub-139 modification of HOIL-1 (Figure S2D). Thus, the OTULIN-HOIP interaction has a non-140 redundant function in removing Met1-Ub from HOIL-1. 141 Since ablation of OTULIN or its catalytic activity sensitises to TNF-induced cell 142 death11, 13, 27, 28 , we next examined if disruption of the OTULIN-LUBAC interaction 143 affected the sensitivity to TNF. OTULINY56A protected OTULIN-deficient NIH 3T3 144 cells and MEFs better or to a similar extent as OTULIN WT, whereas OTULINC129A, as 145 reported13, accelerated TNF-induced cell death (Figures 1H, S2E, S3A, and S3B). The 146 sensitisation to TNF by OTULIN C129A was reversed by the Y56A mutation 147 (OTULINYA/CA), supporting that disruption of the OTULIN -LUBAC interaction 148 protects from TNF-induced cell death (Figures 1H, S2E, S3A, and S3B). Accordingly, 149 MDFs and BMDMs from OtulinY56A/Y56A mice were less sensitive than WT 150 counterparts to apoptosis when treated with TNF plus the TAK1 inhibitor 5Z -7-151 Oxozeaenol (TAKi) whereas TNF alone did not induce cell death in either genotype 152 (Figures 1I, S3C, and S3D). Necroptotic cell death by TNF in combination with caspase 153 inhibition (zVAD) was induced similarly in both genotypes , which, in line with 154 previous studies11, 27, 28, indicates that OTULIN primarily regulates apoptotic cell death 155 (Figures 1I, S3C, and S3D) . Inhibition of RIPK1 activity by Necrostatin 2 (Nec1s) 156 largely prevented the cell death of both WT and OtulinY56A/Y56A MDFs (Figures 1I and 157 S3C). 158 159 The OTULIN-LUBAC interaction protects from TNF pathology 160 OTULIN critically protects from TNF-driven inflammatory pathology10, 11, 27, 28 . We 161 therefore sought to determine the role of the OTULIN-LUBAC interaction in TNF 162 responses in adult mice. Strikingly, intraperitoneal (i.p.) injection of mouse TNF 163 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 8 (mTNF) caused rapid hypothermia within 2-4 hours in OtulinY56A/Y56A mice (Figure 2A), 164 suggestive of systemic inflammatory response syndrome (SIRS) 29. Hypothermia of 165 OtulinY56A/Y56A mice progressively worsened and they became moribund within hours, 166 whereas Otulin+/+ and OtulinY56A/+ mice remained asymptomatic with minimal 167 reduction in body temperature (Figure 2A). 168 Haematological analysis showed a comparable reduction of peripheral white blood 169 cell (WBC) and lymphocyte counts between genotypes, indicating equivalent exposure 170 to mTNF (Figure S4A). OtulinY56A/Y56A mice had a modest but significant increase in 171 red blood cell counts, likely as a result of TNF-driven vascular permeability and fluid 172 loss 30, 31 (Figure S4B). In line with this, cytokine profiling revealed elevated levels of 173 pro-inflammatory cytokines , including TNF, IL -6 and CXCL1 , in the serum of 174 OtulinY56A/Y56A mice compared with WT littermates prior to the onset of hypothermia 175 (2.5 h; Figures 2B and 2C). To determine if this exacerbated response was mediated by 176 TNFR1 alone or also by TNFR2, mice were administered human TNF (huTNF), which 177 only activates mouse TNFR132. Akin to mTNF, huTNF resulted in heightened cytokine 178 levels in OtulinY56A/Y56A mice relative to WT littermates, which was accompanied by a 179 rapid onset of hypothermia (Figures 2D, S4C, and S4D) . Thus, the TNF -180 hypersensitivity of OtulinY56A/Y56A mice was driven by TNFR1 signalling. 181 Hepatocyte cell death and liver damage is an early event after i.v. injection of TNF 182 in WT mice 29, 33 . However, only few apoptotic hepatocytes were detected in 183 OtulinY56A/Y56A mice at 2.5 hours after mTNF injection , while widespread c leaved-184 Caspase-3 positivity was evident by 5 hours (Figure 2E). At this time point livers of 185 OtulinY56A/Y56A mice were visibly darkened, indicative of severe vascular congestion 186 (Figure S4E). Cell death of hepatocytes in OtulinY56A/Y56A mice correlated with 187 significant elevation of serum alanine transaminase (ALT) and aspartate transaminase 188 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 9 (AST) levels at 5 hours after mTNF (Figure 2F), indicating that liver pathology 189 occurred after the systemic increase in cytokines and development of hypothermia. WT 190 mice did not show signs of liver damage at either time point (Figures 2E, 2F, and S4E). 191 Paradoxically, OtulinY56A/Y56A mice displayed increased crypt-specific apoptosis of 192 intestinal epithelial cells (IECs) relative to WT littermates already at 2.5 hours after 193 mTNF as evidenced by pyknotic and hyperchromatic nuclei and positive staining for 194 cleaved-Caspase-334 (Figures 2G and S5A ). huTNF similarly induced early crypt-195 specific apoptosis of IECs in OtulinY56A/Y56A mice with little/no liver cell death (Figures 196 S5B and S5C). The localised cell death was reminiscent of mice with ablation of Otulin 197 in IECs28, albeit less pronounced, which may indicate a distinct role for the OTULIN -198 LUBAC interaction in protecting crypt IECs in the small intestine against cytotoxic 199 effects of TNF. 200 RIPK1 activity-mediated cell death is responsible for the development of SIRS in 201 WT mice in response to i.v. injection of TNF 29, 33, which prompted us to address if the 202 hypersensitivity of OtulinY56A/Y56A mice to TNF was driven by RIPK1 activity. In 203 accordance with previous reports , Nec1s completely prevented hypothermia and 204 lethality of WT mice in response to i.v. injected TNF29, 35 (Figure S5D). However, 205 Nec1s did not prevent lethality of OtulinY56A/Y56A mice following i.p. injection of TNF 206 although it delay ed the progression of hypothermia, showing that RIPK 1 activity-207 mediated cell death accelerated the pathology but was not the primary cause of the TNF 208 hypersensitivity (Figure 2H). Together, these data reveal a key role for the OTULIN-209 LUBAC interaction in protecting against TNF-induced cytokine storm, cell death and 210 tissue damage. 211 212 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 10 The OTULIN-LUBAC interaction restricts TNF-induced Met1-Ub and cytokine 213 production 214 To understand mechanistically how the OTULIN -LUBAC interaction regulates TNF 215 responses, the TNF-RSC was purified using biotin-labelled mTNF (biotin -mTNF). 216 OtulinY56A/Y56A MDFs and BMDMs exhibited substantially more Met1-Ub at the TNF-217 RSC and more extensive TNFR1 ubiquitination, evident from slow-migrating polyUb-218 TNFR1 smears, than their WT counterparts (Figures 3A and 3B ). This was 219 accompanied by increased abundance of signalling components that rely on Met1 -Ub, 220 namely NEMO, IKKβ, TBK1, ABIN -1/2 and A20 (Fig ures 3A, 3B, S6A, and S6B) . 221 The recruitment of LUBAC components and CYLD to the TNF -RSC was also 222 increased in OtulinY56A/Y56A cells relative to WT cells (Figures 3A, 3B, S6A, and S6B). 223 This was despite the elevated auto-ubiquitination of LUBAC in OtulinY56A/Y56A cells, 224 which was further increased in response to TNF, indicating that auto-ubiquitination did 225 not compromise LUBAC ’s function in TNF signalling. RIPK1 ubiquitination was 226 comparable between genotypes (Figures 3A, 3B, S6A, and S6B) , consistent with the 227 Ub chains on RIPK1 consisting predominantly of linkages other than Met1-Ub17, 36, 37. 228 In line with the partial protection of OtulinY56A/Y56A cells from TNF-induced cell 229 death (Fig ures 1H, 1I, and S3D ) and the protective role of the TBK1 and IKK 230 checkpoints38, 39, phosphorylation of both kinases was increased at the TNF -RSC and 231 in cell lysates of OtulinY56A/Y56A cells compared with WT cells (Figures 3A, 3B, S6A, 232 and S6B). OtulinY56A/Y56A BMDMs also showed a modest but consistent increase in 233 TNF-induced phosphorylation of IKK-substrates IκBα and the NF -κB subunit RelA 234 relative to WT BMDMs whereas phosphorylation of MAP kinases p38, MK2, ERK1/2 235 and JNK was comparable between genotypes (Figures 3C and S6C ). This translated 236 into an increase in IL-6 and TNF production by IFN-γ or M-CSF primed OtulinY56A/Y56A 237 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 11 BMDMs relative to WT BMDMs in response to TNF whereas unprimed BMDMs 238 produced negligible amounts of cytokines (Figures 3D and S6D) . The cell surface 239 TNFR1 abundance was comparable between genotypes and was increased substantially 240 by IFN-γ and M-CSF, which likely contributed to the increased cytokine production in 241 both genotypes after priming (Figure S6E). 242 Stimulation of IFN -g-primed BMDMs with the NOD2 ligand L18 -MDP showed 243 enhanced TNF and IL -6 production in OtulinY56A/Y56A cells relative WT, in line with 244 the essential role of Met1-Ub in NOD2 signalling40-42 (Figure S6F). In contrast, TLR2- 245 and TLR4-mediated cytokine production was comparable between genotypes ( Figure 246 S6F). These findings show that the OTULIN -LUBAC interaction restrains the 247 deposition of Met1-Ub at the TNF-RSC (and likely the NOD2 signalling complex) to 248 regulate signalling outcomes by restricting the accumulation of Met1 -Ub-dependent 249 signalling factors at the receptor complex. 250 251 Disruption of OTULIN-binding stabilises LUBAC-association with the TNF-RSC 252 Given the role of Met1-Ub in stabilising the TNF-RSC43, we sought to determine if the 253 enhanced Met1-Ub accumulation in OtulinY56A/Y56A cells would affect the disassembly 254 of the TNF -RSC. TNFR1 is i nternalised within minutes upon TNF sensing, which 255 ultimately leads to disassembly of the complex and is required for the formation of 256 complex II at later timepoints 44-46. In line with previous studies of the TNF -RSC19, 43, 257 we noted a pronounced reduction in signalling complex components within the TNF -258 bound TNF-RSC at 15 min compared to 5 min after TNF stimulation (Fig ures 3A and 259 3B). To track the fate of the signalling complex, we performed sequential purification 260 of the TNF-bound TNF-RSC followed by enrichment of Met1-Ub in the flow-through 261 fraction. This revealed that, in addition to the increased Met1-Ub deposition at the TNF-262 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 12 RSC in OtulinY56A/Y56A cells relative to WT cells, OtulinY56A/Y56A cells also accumulated 263 substantially more Met1 -Ub not contained within the TNF -bound TNF-RSC (Figure 264 4A). This prompted us to a nalyse the sequential pulldown fractions by mass 265 spectrometry. 266 In accordance with the biochemical data (Figures 3A and 3B ), the pulldown of 267 biotin-mTNF enriched most known TNF-RSC components and showed a mild increase 268 of LUBAC subunits and Met1-Ub-associated proteins in samples from OtulinY56A/Y56A 269 cells in comparison with WT cells (Figure S7A). Ubiquitin linkage analysis of the 270 Met1-Ub pulldown from the flow-through fraction revealed a substantial and selective 271 increase of Met1 -Ub in OtulinY56A/Y56A cells relative to WT cells , which was further 272 increased by TNF treatment (Figure 4B). This was accompanied by enrichment of most 273 TNF-RSC components, including TNFR1, TRADD, RIPK1, LUBAC, and several 274 proteins recruited via Met1 -Ub, namely NEMO-IKK, A20, ABIN-1/2, TBK1, TANK 275 in OtulinY56A/Y56A cells after TNF treatment (Figures 4C, 4D, S7B, and S7C). TNFR1, 276 TRADD and RIPK1 were also enriched in the Met1-Ub pulldown from WT cells but to 277 a lesser degree than in OtulinY56A/Y56A cells (Figures 4C, 4D, and S7C). This increase 278 of TNF-RSC components in OtulinY56A/Y56A cells was not due to residual TNF-bound 279 TNF-RSC since TNF was efficiently captured in the biotin-mTNF pulldown, with no 280 significant difference in TNF abundance in the Met1-Ub pulldown samples (Figure 4D). 281 IP of HOIP following depletion of the TNF-bound TNF-RSC showed that LUBAC 282 associated with several TNF-RSC components as well as Met1-Ub that, in response to 283 TNF, were enriched in OtulinY56A/Y56A cells relative to WT cells (Figures 4E and 4F). 284 Moreover, after TNF treatment HOIP co-purified ubiquitinated forms of TNFR1 and 285 RIPK1 along with NEMO and phosphorylated forms of IKKβ and CYLD, indicating 286 that the complex emanated from the TNF -activated TNF-RSC (Figures 4E, 4F, S7D, 287 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 13 and S7E). This showed that the TNF-RSC remains assembled after dissociation from 288 TNF and suggest s that disruption of the OTULIN-LUBAC interaction modulates 289 signalling outcomes by stabilising both the TNF-bound and TNF-dissociated TNF-RSC. 290 291 Disruption of OTULIN-binding increases LUBAC activity 292 Our data posed a conundrum since b iochemical purification indicates that OTULIN is 293 not present at the TNF -RSC20, 22 (Figures 3A and 3B ). In line with this, we did not 294 detect TNF-induced biotinylation of TNFR1 or RIPK1 in cells expressing TurboID -295 OTULIN (Figure S8A). This implies that the enhanced deposition of Met1 -Ub at the 296 TNF-RSC in OtulinY56A/Y56A cells resulted from increased LUBAC function, yet auto-297 ubiquitination of LUBAC was proposed to be inhibitory13, 25. 298 To directly address how LUBAC auto -ubiquitination in OtulinY56A/Y56A cells 299 influences its enzymatic activity, we performed in vitro ubiquitination assays. 300 Endogenous auto-ubiquitinated LUBAC was purified from OtulinY56A/Y56A MDFs by IP 301 of HOIP and was then incubated with USP21 to remove the ubiquitin chains or with 302 buffer as a control (Figure 5A, lane 3 and 4). USP21 treatment led to a clear increase in 303 monoUb-modified HOIL-1, consistent with HOIL-1 depositing the first Ub moiety via 304 an oxyester bond or, alternatively, that the linkage is inaccessible to cleavage by 305 USP2125, 47 . To rule out that residual USP21 activity in the deubiquitinated samples 306 would interfere Met1-Ub accumulation, USP21 -treated LUBAC was incubated with 307 tetra-Met1-Ub, which showed negligible cleavage after 20 min of incubation (Fig ure 308 5A, lane 9 and 10). Strikingly, when the ubiquitination reaction was started, the auto-309 ubiquitinated LUBAC assembled Met1 -Ub more efficiently than did USP21-treated 310 LUBAC (Figure 5A). HOIP and HOIL -1 were both extensively ubiquitinated during 311 the reaction but, interestingly, most of HOIL-1 was modified by 10 or less Ub moieties, 312 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 14 suggesting that HOIP preferentially extended monoUb or short Ub chains on HOIL-1. 313 This was not due to a general preference for LUBAC to generate short Met1-Ub chains 314 since very slow-migrating Met1-Ub was readily detected (Figure 5A, compare HOIL-315 1 and Met1 -Ub blots). Consistently, inhibition of HOIP activity by Hoipin -848 in 316 OTULIN deficient NIH -3T3 cells reconstituted with OTULIN variants led to a rapid 317 decrease in short Met1 -Ub chains on HOIL-1 and in slow -migrating Met1-Ub not 318 conjugated to HOIL-1 (Figures 5B and S8B). Notably, the Hoipin-8 treatment led to a 319 clear increase in monoUb-HOIL-1 in cells with Met1 -Ub-modified HOIL -1 and 320 suppressed TNF-induced Met1 -Ub accumulation irrespective of the OTULIN status 321 (Figures 5B and S8C). This showed that HOIP activity is required continuously to 322 maintain the Met1-Ub modification of HOIL-1 and that auto-ubiquitination of LUBAC 323 in OtulinY56A/Y56A cells promotes, rather than inhibits, its ability to assemble Met1-Ub. 324 Together, our data suggest that OTULIN, through its interaction with LUBAC, restricts 325 Met1-Ub-mediated TNF signalling by preventing Met1 -Ub extension on 326 monoubiquitinated HOIL -1 (and other LUBAC subunits) and thereby limits the 327 propensity of LUBAC to conjugate Met1-Ub on non-LUBAC substrates (Figure 5C). 328 329 OTULIN-LUBAC interaction protects from systemic pathology in response to 330 Listeria Monocytogenes infection 331 TNF and IFN g are essential for the early -phase control of Listeria Monocytogenes 332 (Listeria) infection by macrophages 49-51. This prompted us to investigate the 333 pathophysiological role of the OTULIN-LUBAC interaction during Listeria infection. 334 WT mice challenged with a sublethal dose of Listeria experienced a transient drop in 335 BW at day 2 post-infection, which normalised by day 3, whereas OtulinY56A/Y56A mice 336 progressively lost BW (Figure 6A). At this point, h aematological analysis indicated 337 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 15 impaired host fitness in OtulinY56A/Y56A mice, characterised by a pronounced reduction 338 in WBC, lymphocyte and platelet counts relative to WT mice (Figure 6B). 339 Correspondingly, Listeria-infected OtulinY56A/Y56A mice exhibited elevated levels of 340 several proinflammatory cytokines and chemokines relative to WT mice , namely IL-341 1β, IL-6, IL-17A/F and CXCL1 (Figure 6C). Despite this, bacterial loads in spleen and 342 liver were comparable between WT and OtulinY56A/Y56A mice, suggesting that the ability 343 to control Listeria was not compromised by the loss of LUBAC -OTULIN interaction 344 (Figure 6D). Supporting this, Listeria-infection of BMDMs in vitro showed similar 345 intracellular bacterial growth in WT and OtulinY56A/Y56A BMDMs, which was 346 suppressed equally well by IFN-γ-priming in both genotypes (Figures 6E and S9A)52, 347 53. Additionally, i nfection-induced cell death of BMDMs was similar between 348 genotypes (Figure S9B). 349 To test if the reduced fitness of OtulinY56A/Y56A mice in response to Listeria was 350 caused by TNF-driven pathology , we administered mice with TNF -neutralising 351 antibodies. This significantly improved BW maintenance in OtulinY56A/Y56A mice at day 352 2 post-infection compared to isotype controls (Figure 6F). However, TNF neutralisation 353 induced abrupt BW loss in both genotypes at day 3 (Fig ure 6F), which was 354 accompanied by a dramatic accumulation of Listeria in spleen and liver, consistent with 355 the critical role of TNF in bacterial control (Figure 6G)49. Collectively, this indicates 356 that the OTULIN -LUBAC interaction protects from TNF -driven immunopathology 357 during infection without compromising pathogen control. 358 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 16

359 The cellular machinery that assembles and disassembles Met1-Ub has emerged as 360 crucially important for regulation of immune responses and inflammation. This is 361 particularly evident for OTULIN whose DUB activity is required to prevent TNF-362 driven autoinflammatory diseases and lethality of mice during embryogenesis10, 13-15. 363 The fact that OtulinY56A/Y56A mice were viable without signs of spontaneous 364 inflammation indicates that the PIM -PUB interaction between OTULIN and HOIP 365 serves a discrete regulatory function that is mechanistically distinct from OTULIN’s 366 catalytic role in controlling Met1 -Ub homeostasis. Supporting this, patients with a 367 biallelic OTULINR57C mutation that attenuates the OTULIN-LUBAC interaction were 368 recently identified with recurrent pyoderma gangrenosum (PG) without the systemic 369 autoinflammatory symptoms characteristic for ORAS patients with mutations that 370 interfere with the DUB activity54, 55. Also, ablation of the OTULIN-LUBAC interaction 371 led to a much less pronounced auto-ubiquitination of LUBAC subunits and 372 accumulation of Met1-Ub than observed in cells without OTULIN activity and , 373 importantly, did not result in depletion of LUBAC levels as observed in the absence of 374 OTULIN activity13. 375 RIPK1 kinase activity mediates TNF -induced SIRS in WT mice and the 376 spontaneous TNF-driven inflammatory pathologies described in mice with skin - and 377 liver-specific ablation of Otulin27-29, 56 . In contrast, t he exquisite sensitivity of 378 OtulinY56A/Y56A mice to TNF was associated with a rapid increase in serum cytokines 379 levels but was not driven by RIPK1 activity-mediated cell death albeit RIPK1 inhibition 380 delayed the hypothermia. This likely reflects that disruption of the OTULIN -LUBAC 381 interaction promotes NF-kB-driven expression of inflammatory mediators whereas the 382 absence of OTULIN or OTULIN activity sensitises to cell death11, 13, 27, 28. Intriguingly, 383 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 17 OtulinY56A/Y56A mice displayed early crypt -specific cell death of IECs after TNF 384 challenge which shared features with the intestinal cell death in mice with IEC-specific 385 ablation of Otulin or Tnfaip3 (encodes A20) 28, 34 . Whether this reflects a particular 386 sensitivity of OtulinY56A/Y56A crypt-associated IECs to TNF -induced cell death 387 independently of RIPK1 activity, or an enhanced inflammatory response that disrupted 388 intestinal barrier integrity, will be interesting to delineate in future studies. 389 OtulinY56A/Y56A mice displayed impaired fitness during Listeria infection despite a 390 similar bacterial burden in liver and spleen as their WT counterpart, which points to a 391 critical role for OTULIN-binding in restricting LUBAC-mediated responses to promote 392 disease tolerance during infection by limiting immune -mediated pathology 57. This 393 appears to be distinct from the role of CYLD during listeriosis as ablation of Cyld is 394 shown to reduce the bacterial burden 58. Whether this role of CYLD is mediated by its 395 interaction with LUBAC via SPATA2 remains to be determined. 396 The generation of Met1-Ub by LUBAC in cells with catalytically active OTULIN 397 protects from TNF-induced cell death2. This together with the observation that ablation 398 of OTULIN or OTULIN activity leads to extensive Met1-Ub accumulation on LUBAC 399 subunits has led to the notion that auto -ubiquitination inhibits LUBAC function. Our 400 data show that this is not necessarily the case since the auto -ubiquitinated LUBAC in 401 cells with OTULINY56A generated more Met1-Ub at the TNF-RSC and protected from 402 cell death similarly or better than LUBAC in cells with WT OTULIN. Moreover, in 403 OTULINY56A expressing cells, the association of LUBAC with TNF-RSC components 404 was increased after TNF treatment, which was particularly evident in our analysis of 405 the TNF -bound TNF -RSC followed by enrichment of the LUBAC -associated TNF -406 RSC. This implicates the OTULIN PIM -mediated interaction with the HOIP PUB 407 domain as a regulatory axis that counterbalances LUBAC activity. 408 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 18 Work by Iwai and colleagues demonstrated that monoubiquitination of LUBAC 409 subunits by HOIL-1 generates a preferred substrate for HOIP to extend Met1-Ub in cis, 410 which limits LUBAC -mediated Met1 -Ub conjugation on other substrates 25. This 411 together with our findings suggests that the OTULIN-LUBAC complex is continuously 412 assembling and disassembling auto -Met1-Ub, which suppresses the propensity for 413 LUBAC to assemble Met1 -Ub on other substrates. When the OTULIN -LUBAC 414 interaction was disrupted predominantly short HOIP-generated Met1-Ub accumulated 415 on HOIL-1, suggestively because OTULINY56A removes the Ub chains less efficiently. 416 An appealing model to explain the increased Met1 -Ub activity of auto -ubiquitinated 417 LUBAC is that the Met1-Ub on HOIL-1 become poor substrates for HOIP as the chains 418 are extended, which redirects HOIP to conjugate Met1 -Ub in trans on substrates such 419 as Ub -modified TNFR1 , akin to the scenario without HOIL -1-mediated 420 monoubiquitnation25 (Figure 5C). This mechanism may also explain the observation 421 that short Ub chains transiently accumulated on LUBAC subunits in response to TNF 422 as a consequence of recruitment of LUBAC to the TNF-RSC without OTULIN (Figures 423 3A, 3B, 4E, and 4F)22. 424 In summary, we here uncover an immunoregulatory role for the LUBAC-OTULIN 425 interaction in restricting LUBAC and TNF signalling, which protects tissue integrity 426 during immune activation and promotes host fitness during infection. Further studies of 427 how the interaction is regulated may provide new therapeutic opportunities for 428 inflammatory disorders. 429 430 431 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 19 Supplemental Information 432 SI document. Materials and Methods, Supplementary Table 1, Supplementary Figures 433 1-9 and Figure Legends, SI references. 434 435 Acknowledgments 436 We thank Kit Lee (M.G-H. group) for breeding of mice during the initial phase of the 437 project, members from the M.G-H. and R.B.D. group for helpful advice and suggestions, 438 and Søren Riis Paludan (Aarhus University) for scientific discussion. We acknowledge 439 the valuable contribution to this study made by the Core Facility for Transgenic 440 Animals, the Animal Housing and Breeding Facility at the Department of Experimental 441 Medicine, the Flow Cytometry and Single Cell Core Facility, the Histolab and 442 Veterinary Diagnostic Laboratory at the University of Copenhagen. Mass spectrometry 443 analyses were performed by the Proteomics Research Infrastructure (PRI) at the 444 University of Copenhagen. 445 446 Funding 447 This work was supported by the LEO foundation (University of Copenhagen; Grant No. 448 LF18500) and the Ludwig Institute for Cancer Research Ltd (University of Oxford) . 449 Work in the M.G-H. lab was supported by a Wellcome Trust Fellowship 450 (215612/Z/19/Z) and the Novo Nordisk Foundation (NNF200C0059392). Mass 451 spectrometry analyses at PRI were supported by the Novo Nordisk Foundation 452 (NNF19SA0059305). 453 454 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 20 Author Contributions 455 Conceptualization: MGH, WL 456 Investigation: WL, BKF 457 Methodology: WL, BKF, JR, MK, MBS, BM, MFJ 458 Formal analysis: WL, BKF, MBS 459 Visualization: WL, BKF, MBS, MGH 460 Funding acquisition: MGH 461 Resources: MGH, RBD 462 Supervision: MGH 463 Writing—original draft: WL, MGH 464 465 Declaration of Interests 466 The authors declare no competing interests. 467 468

& Correspondence 469 Correspondence relating to the article and material requests should be addressed to 470 Mads Gyrd-Hansen, e-mail: [email protected] 471 472 Data availability 473 The mass -spec proteomics data have been deposited to the ProteomeXchange 474 Consortium via the PRIDE59 partner repository with the dataset identifier PXD070197. 475 476 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 21

477 1. Angus Derek, C. & van der Poll, T. Severe Sepsis and Septic Shock. New 478 England Journal of Medicine 369, 840-851. 479 2. van Loo, G. & Bertrand, M.J.M. Death by TNF: a road to inflammation. Nature 480 Reviews Immunology 23, 289-303 (2023). 481 3. Huyghe, J., Priem, D. & Bertrand, M.J.M. Cell death checkpoints in the TNF 482 pathway. Trends in Immunology 44, 628-643 (2023). 483 4. Fiil, B.K. & Gyrd -Hansen, M. The Met1 -linked ubiquitin machinery in 484 inflammation and infection. Cell Death & Differentiation 28, 557-569 (2021). 485 5. Hrdinka, M. & Gyrd -Hansen, M. The Met1 -Linked Ubiquitin Machinery: 486 Emerging Themes of (De)regulation. Molecular Cell 68, 265-280 (2017). 487 6. Ikeda, F. et al. SHARPIN forms a linear ubiquitin ligase complex regulating 488 NF-κB activity and apoptosis. Nature 471, 637-641 (2011). 489 7. Tokunaga, F. et al. SHARPIN is a component of the NF -κB-activating linear 490 ubiquitin chain assembly complex. Nature 471, 633-636 (2011). 491 8. Gerlach, B. et al. Linear ubiquitination prevents inflammation and regulates 492 immune signalling. Nature 471, 591-596 (2011). 493 9. Keusekotten, K. et al. OTULIN antagonizes LUBAC signaling by specifically 494 hydrolyzing Met1-linked polyubiquitin. Cell 153, 1312-1326 (2013). 495 10. Damgaard, R.B. et al. The deubiquitinase OTULIN is an essential negative 496 regulator of inflammation and autoimmunity. Cell 166, 1215 -1230. e1220 497 (2016). 498 11. Damgaard, R.B. et al. OTULIN deficiency in ORAS causes cell type‐specific 499 LUBAC degradation, dysregulated TNF signalling and cell death. EMBO 500 molecular medicine 11, e9324 (2019). 501 12. Zhou, Q. et al. Biallelic hypomorphic mutations in a linear deubiquitinase 502 define otulipenia, an early-onset autoinflammatory disease. Proceedings of the 503 National Academy of Sciences 113, 10127-10132 (2016). 504 13. Heger, K. et al. OTULIN limits cell death and inflammation by deubiquitinating 505 LUBAC. Nature 559, 120-124 (2018). 506 14. Fu, Y. et al. OTULIN allies with LUBAC to govern angiogenesis by editing 507 ALK1 linear polyubiquitin. Molecular Cell 81, 3187-3204.e3187 (2021). 508 15. Rivkin, E. et al. The linear ubiquitin -specific deubiquitinase gumby regulates 509 angiogenesis. Nature 498, 318-324 (2013). 510 16. Komander, D. et al. Molecular discrimination of structurally equivalent Lys 63‐511 linked and linear polyubiquitin chains. EMBO reports 10, 466-473 (2009). 512 17. Elliott, P.R. et al. Regulation of CYLD activity and specificity by 513 phosphorylation and ubiquitin -binding CAP -Gly domains. Cell Reports 37 514 (2021). 515 18. Massoumi, R., Chmielarska, K., Hennecke, K., Pfeifer, A. & Fässler, R. Cyld 516 inhibits tumor cell proliferation by blocking Bcl-3-dependent NF-κB signaling. 517 Cell 125, 665-677 (2006). 518 19. Wei, R. et al. SPATA2 regulates the activation of RIPK1 by modulating linear 519 ubiquitination. Genes & development 31, 1162-1176 (2017). 520 20. Elliott, P.R. et al. SPATA2 links CYLD to LUBAC, activates CYLD, and 521 controls LUBAC signaling. Molecular cell 63, 990-1005 (2016). 522 21. Elliott, P.R. et al. Molecular basis and regulation of OTULIN -LUBAC 523 interaction. Molecular cell 54, 335-348 (2014). 524 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 22 22. Draber, P. et al. LUBAC-recruited CYLD and A20 regulate gene activation and 525 cell death by exerting opposing effects on linear ubiquitin in signaling 526 complexes. Cell reports 13, 2258-2272 (2015). 527 23. Schaeffer, V. et al. Binding of OTULIN to the PUB Domain of HOIP Controls 528 NF-κB Signaling. Molecular Cell 54, 349-361 (2014). 529 24. Cho, K.F. et al. Proximity labeling in mammalian cells with TurboID and split-530 TurboID. Nature Protocols 15, 3971-3999 (2020). 531 25. Fuseya, Y. et al. The HOIL-1L ligase modulates immune signalling and cell 532 death via monoubiquitination of LUBAC. Nature Cell Biology 22, 663 -673 533 (2020). 534 26. Davidson, S. et al. Dominant negative OTULIN -related autoinflammatory 535 syndrome. Journal of Experimental Medicine 221 (2024). 536 27. Hoste, E. et al. OTULIN maintains skin homeostasis by controlling keratinocyte 537 death and stem cell identity. Nature communications 12, 5913 (2021). 538 28. Verboom, L. et al. OTULIN protects the intestinal epithelium from apoptosis 539 during inflammation and infection. Cell Death & Disease 14, 534 (2023). 540 29. Duprez, L. et al. RIP Kinase -Dependent Necrosis Drives Lethal Systemic 541 Inflammatory Response Syndrome. Immunity 35, 908-918 (2011). 542 30. Rivers, E. et al. Early Goal-Directed Therapy in the Treatment of Severe Sepsis 543 and Septic Shock. New England Journal of Medicine 345, 1368-1377. 544 31. Fajgenbaum, D.C. & June, C.H. Cytokine storm. New England Journal of 545 Medicine 383, 2255-2273 (2020). 546 32. Lewis, M. et al. Cloning and expression of cDNAs for two distinct murine tumor 547 necrosis factor receptors demonstrate one receptor is species specific. 548 Proceedings of the National Academy of Sciences 88, 2830-2834 (1991). 549 33. Zelic, M. et al. RIP kinase 1 –dependent endothelial necroptosis underlies 550 systemic inflammatory response syndrome. The Journal of Clinical 551 Investigation 128, 2064-2075 (2018). 552 34. Vereecke, L. et al. Enterocyte-specific A20 deficiency sensitizes to tumor 553 necrosis factor –induced toxicity and experimental colitis. Journal of 554 Experimental Medicine 207, 1513-1523 (2010). 555 35. Takahashi, N. et al. Necrostatin-1 analogues: critical issues on the specificity, 556 activity and in vivo use in experimental disease models. Cell Death & Disease 557 3, e437-e437 (2012). 558 36. Hospenthal, M.K., Mevissen, T.E. & Komander, D. Deubiquitinase -based 559 analysis of ubiquitin chain architecture using Ubiquitin Chain Restriction 560 (UbiCRest). Nature protocols 10, 349-361 (2015). 561 37. Annibaldi, A. et al. Ubiquitin-mediated regulation of RIPK1 kinase activity 562 independent of IKK and MK2. Molecular cell 69, 566-580. e565 (2018). 563 38. Lafont, E. et al. TBK1 and IKKε prevent TNF -induced cell death by RIPK1 564 phosphorylation. Nature Cell Biology 20, 1389-1399 (2018). 565 39. Dondelinger, Y. et al. NF-κB-independent role of IKKα/IKKβ in preventing 566 RIPK1 kinase -dependent apoptotic and necroptotic cell death during TNF 567 signaling. Molecular cell 60, 63-76 (2015). 568 40. Damgaard, R.B. et al. The ubiquitin ligase XIAP recruits LUBAC for NOD2 569 signaling in inflammation and innate immunity. Molecular cell 46, 746 -758 570 (2012). 571 41. Hrdinka, M. et al. CYLD limits Lys63 -and Met1-linked ubiquitin at receptor 572 complexes to regulate innate immune signaling. Cell reports 14, 2846 -2858 573 (2016). 574 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 23 42. Fiil, B.K. et al. OTULIN restricts Met1-linked ubiquitination to control innate 575 immune signaling. Molecular cell 50, 818-830 (2013). 576 43. Haas, T.L. et al. Recruitment of the linear ubiquitin chain assembly complex 577 stabilizes the TNF -R1 signaling complex and is required for TNF -mediated 578 gene induction. Molecular cell 36, 831-844 (2009). 579 44. Schütze, S., Tchikov, V. & Schneider -Brachert, W. Regulation of TNFR1 and 580 CD95 signalling by receptor compartmentalization. Nature Reviews Molecular 581 Cell Biology 9, 655-662 (2008). 582 45. Schneider-Brachert, W. et al. Compartmentalization of TNF receptor 1 583 signaling: internalized TNF receptosomes as death signaling vesicles. Immunity 584 21, 415-428 (2004). 585 46. Schneider-Brachert, W. et al. Inhibition of TNF receptor 1 internalization by 586 adenovirus 14.7 K as a novel immune escape mechanism. The Journal of 587 clinical investigation 116, 2901-2913 (2006). 588 47. Kelsall, I.R., Zhang, J., Knebel, A., Arthur, J.S.C. & Cohen, P. The E3 ligase 589 HOIL-1 catalyses ester bond formation between ubiquitin and components of 590 the Myddosome in mammalian cells. Proceedings of the National Academy of 591 Sciences 116, 13293-13298 (2019). 592 48. Katsuya, K. et al. Small-molecule inhibitors of linear ubiquitin chain assembly 593 complex (LUBAC), HOIPINs, suppress NF -κB signaling. Biochemical and 594 Biophysical Research Communications 509, 700-706 (2019). 595 49. Rothe, J. et al. Mice lacking the tumour necrosis factor receptor 1 are resistant 596 to IMF -mediated toxicity but highly susceptible to infection by Listeria 597 monocytogenes. Nature 364, 798-802 (1993). 598 50. Unanue, E.R. Inter -relationship among macrophages, natural killer cells and 599 neutrophils in early stages of Listeria resistance. Current opinion in immunology 600 9, 35-43 (1997). 601 51. Buchmeier, N.A. & Schreiber, R.D. Requirement of endogenous interferon -602 gamma production for resolution of Listeria monocytogenes infection. 603 Proceedings of the National Academy of Sciences 82, 7404-7408 (1985). 604 52. Huang, S. et al. Immune Response in Mice That Lack the Interferon-γ Receptor. 605 Science 259, 1742-1745 (1993). 606 53. Jessop, F. et al. Interferon Gamma Reprograms Host Mitochondrial Metabolism 607 through Inhibition of Complex II To Control Intracellular Bacterial Replication. 608 Infection and Immunity 88, 10.1128/iai.00744-00719 (2020). 609 54. Bhattad, S. et al. Profile of 208 patients with inborn errors of immunity at a 610 tertiary care center in South India. Clinical and Experimental Medicine 23, 611 5399-5412 (2023). 612 55. Gil, H., Mariskanish, C., Bhattad, S. & Markle, J. Pediatric Pyoderma 613 Gangrenosum in Patients with a Novel Biallelic Mutation in OTULIN. Journal 614 of Human Immunity 1, CIS2025abstract. 2124 (2025). 615 56. Verboom, L. et al. OTULIN Prevents Liver Inflammation and Hepatocellular 616 Carcinoma by Inhibiting FADD - and RIPK1 Kinase -Mediated Hepatocyte 617 Apoptosis. Cell Reports 30, 2237-2247.e2236 (2020). 618 57. Medzhitov, R., Schneider, D.S. & Soares, M.P. Disease Tolerance as a Defense 619 Strategy. Science 335, 936-941 (2012). 620 58. Nishanth, G. et al. CYLD enhances severe listeriosis by impairing IL-6/STAT3-621 dependent fibrin production. PLoS pathogens 9, e1003455 (2013). 622 59. Perez-Riverol, Y. et al. The PRIDE database at 20 years: 2025 update. Nucleic 623 acids research 53, D543-D553 (2025). 624 625 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 24 FIGURES and LEGENDS 626 627 Figure 1. OTULIN-LUBAC interaction is dispensable for embryonic development, 628 immune homeostasis and protection from cell death by TNF despite LUBAC auto-629 ubiquitination. 630 (A) Observed numbers and ratios of offspring from intercrosses of OtulinY56A/+ mice. 631 (B) Body weight of adult male (7-8 weeks old) and female (10-12 weeks old) mice. 632 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 25 (C) Total cell number of the indicated immune cell populations in spleen analysed by 633 flow cytometry. 634 (D-G) Immunoblot analysis of immunoprecipitated LUBAC complexes via SHARPIN 635 from splenocytes (D), immortali sed MDFs ( E, G) or OtulinKO NIH 3T3 cells 636 reconstituted with OTULIN variants (F). 637 (H) Cell death of OtulinKO NIH 3T3 cells reconstituted with OTULIN variants 638 stimulated with 10 ng/mL mTNF determined by YOYO3 positivity and normalised to 639 cell number. 640 (I) Cell death of primary MDFs pre-treated for 30 minutes with inhibitors (100 nM 641 TAKi, 10 µM Nec1s, 10 µM zVAD or DMSO as vehicle control), followed by 642 stimulation with 10 ng/mL mTNF for 24 hours. Cell death determined by Sytox Green 643 positivity normalised to cell confluence. 644 Data are presented as mean ± SEM , and open circles show data from individual mice 645 or biological replicates. ‘n’ indicates the number of mice or biological replicates in each 646 group. Data in (F, G) are representative of three independent experiments with similar 647 results. Statistical analysis; one-way ANOVA (B), multiple unpaired t-test (C) and two-648 way ANOVA with Turkey’s multiple comparisons (H, I). 649 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 26 650 Figure 2 . Disruption of the OTULIN -LUBAC interaction sensiti ses to TNF -651 induced SIRS. 652 (A) Body temperature and survival of mice injected i.p. with 50 µg/kg mTNF. 653 (B, C, E-G) Mice injected i.p. with PBS or 50 µg/kg mTNF were euthanized after 2.5 654 h or 5 h and analysed for serum cytokine levels by multiplex cytokine profiling (B) or 655 ELISA (C), H&E or anti -cleaved caspase-3 staining of liver (E) and ileum (G), and 656 serum levels of AST and ALT (F). 657 (D) Serum levels of mTNF measured by ELISA of mice injected i.p. with PBS or 500 658 µg/kg huTNF euthanized after 2 h. 659 (H) Body temperature and survival of OtulinY56A/Y56A mice injected i.v. with 5% DMSO 660 or 6 mg/kg Nec1s, followed by i.p. injection of 50 µg/kg mTNF. 661 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 27 Data are presented as mean ± SEM , and open circles show data from individual mice . 662 ‘n’ indicates the number of mice or biological replicates in each group . Data in (B) 663 represents mean of 3 biological replicates for PBS groups and 5 biological replicates 664 for mTNF groups. Images in (E, G) are representative of three biological replicates with 665 similar results. Arrowheads indicate damaged cells in the crypts. Scale bar: 400 µm (E) 666 and 100 µm (F). Statistical analysis; Log-rank (Mantel-Cox) test (A, H) and two-way 667 ANOVA with multiple comparisons (C, D, F). 668 669 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 28 670 671 Figure 3. OTULIN-LUBAC interaction restricts TNF-induced Met1-Ub and NF-672 κB signalling. 673 (A, B) Immunoblot analysis of TNF-RSC purified after stimulation with 50 ng/mL 674 biotin-mTNF for the indicated time points from immortalised MDFs (A) and BMDMs 675 (B). 0 min samples are from cells treated on ice for 2 minutes. 676 (C) Immunoblot analysis of whole cell lysates from BMDMs stimulated with 10 677 ng/mL mTNF for the indicated time points. 678 (D) Cytokine concentrations in supernatants of BMDMs primed overnight with 20 679 ng/mL mIFNγ and treated with 10 ng/mL mTNF or 100 ng/mL huTNF for 24 hours. 680 ‘n’ indicates the number of biological replicates in each group. 681 Representative results from three biological replicates (A-C) are shown. Data in (D) 682 are mean ± SEM. Statistical analysis; multiple unpaired t-test (D). 683 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 29 684 685 Figure 4. Disruption of OTULIN-LUBAC interaction stabilises the association of 686 LUBAC with TNF-RSC components. 687 (A-D) Immortalised MDFs stimulated with 50 ng/mL biotin -mTNF for the indicated 688 time points. 0 min samples are from cells treated on ice for 2 minutes. Biotin -mTNF 689 pulldowns and Met1 -Ub pulldowns from flow-through fractions analysed by 690 immunoblotting (A) or by m ass spectrometry -based proteomic analysis of ubiquitin 691 linkages (B), protein enrichment (C), and MaxLFQ intensities (D). 692 (E-F) Immunoblot analysis of immunoprecipitated LUBAC complexes via HOIP from 693 flow-through fraction from samples in (Figures 3A and 3B). 694 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 30 Data are presented as mean ± SEM from three biological replicates ( B-D). 695 Representative data from three ( A, F) or two (E) biological replicates are shown. 696 Statistical analysis; two-way ANOVA with Turkey's multiple comparisons (B, D). 697 698 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 31 699 700 Figure 5. OTULIN counteracts LUBAC auto-ubiquitination to restrict its activity. 701 (A) In vitro ubiquitination assays with LUBAC complexes immunoprecipitated via 702 HOIP from immortali sed MDFs derived from OtulinY56A/Y56A mice. Samples were 703 incubated with buffer or treated with 1 µM USP21 to remove auto-ubiquitination. Tetra-704 Met1-Ub was added to samples to determine residual DUB activity. Formation of Met1-705 Ub was analysed by immunoblotting. 706 (B) Immunoblot analysis of Met1-Ub associated with LUBAC immunoprecipitated via 707 SHARPIN from OtulinKO NIH 3T3 cells reconstituted with OTULIN variants treated 708 with 3 µM Hoipin-8 for the indicated time points. 709 (C) Model of how the LUBAC-OTULIN interaction restricts LUBAC activity. Via the 710 PIM-mediated interaction with the HOIP PUB domain, OTULIN counters the Met1-711 Ub extension o n auto-monoubiquitinated HOIL -1 (and possibly other LUBAC 712 subunits) by HOIP in cis to maintain HOIL-1 in a mostly auto-monoubiquitinated form, 713 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 32 which is a preferred substrate for HOIP. This restricts the propensity of LUBAC to 714 conjugate Met1-Ub on non -LUBAC substrates. Disruption of the OTULIN -LUBAC 715 interaction, e.g. by a Y56A mutation in the OTULIN PIM, enables accumulation of 716 Met1-Ub chains on HOIL-1, limiting the ability of HOIP to further extend the chains. 717 This promotes the generation of Met1 -Ub by LUBAC on non -LUBAC substrates in 718 trans, leading to enhanced Met1-Ub formation at the TNF receptor signalling complex, 719 increased gene induction, and suppression of apoptosis. 720 Data are representative of three independent experiments with similar results. 721 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint 33 722 Figure 6. OTULIN-LUBAC interaction protect s against L. monocytogenes 723 infection-induced pathology driven by TNF. 724 (A) Body weight of mice infected i.p. with 600,000 CFU of Listeria monitored daily 725 for 3 days. 726 (B-D) Peripheral blood cell counts (B), Serum cytokine levels (C), and bacterial burden 727 (CFU) in spleen and liver (D) on day 3 of mice in (A). 728 (E) Intracellular bacterial burden (CFU) of BMDMs infected with Listeria (MOI=1). 729 BMDMs were primed overnight, or not, with 20 ng/mL mIFN-γ as indicated. 730 (F, G) Body weight (F) and bacterial burden (CFU) in spleen and liver on day 3 (G) of 731 mice infected with Listeria as in (A) with or without i.p. injection of 100 µg anti-mTNF 732 antibodies or IgG1 isotype control 12 hours before and 36 hours after infection. 733 Data are presented as mean ± SEM of combined results from two separate experiments, 734 and open circles show data from individual mice. ‘n’ indicates the number of mice in 735 each group. Data in (D, G) are shown as median with 95% confidence intervals. Data 736 in (C) is the mean of 4 biological replicates for Control group and 5 biological replicates 737 for Listeria group. Statistical analysis ; two-way ANOVA with Turkey's multiple 738 comparisons (A, E, F) and unpaired two-tailed Student’s t-test (B, D, G). 739 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 28, 2026. ; https://doi.org/10.64898/2026.02.27.708452doi: bioRxiv preprint