EspL is a bacterial cysteine protease effector that cleaves RHIM proteins to block necroptosis and inflammation

Abstract

Cell death signalling pathways contribute to tissue homeostasis and provide innate protection from infection. Adaptor proteins such as receptor-interacting serine/threonine-protein kinase 1 (RIPK1), receptor-interacting serine/threonine-protein kinase 3 (RIPK3), TIR-domain-containing adapter-inducing interferon-β (TRIF) and Z-DNA-binding protein 1 (ZBP1)/DNA-dependent activator of IFN-regulatory factors (DAI) that contain receptor-interacting protein (RIP) homotypic interaction motifs (RHIM) play a key role in cell death and inflammatory signalling^1-3. RHIM-dependent interactions help drive a caspase-independent form of cell death termed necroptosis^4,5. Here, we report that the bacterial pathogen enteropathogenic Escherichia coli (EPEC) uses the type III secretion system (T3SS) effector EspL to degrade the RHIM-containing proteins RIPK1, RIPK3, TRIF and ZBP1/DAI during infection. This requires a previously unrecognized tripartite cysteine protease motif in EspL (Cys47, His131, Asp153) that cleaves within the RHIM of these proteins. Bacterial infection and/or ectopic expression of EspL leads to rapid inactivation of RIPK1, RIPK3, TRIF and ZBP1/DAI and inhibition of tumour necrosis factor (TNF), lipopolysaccharide or polyinosinic:polycytidylic acid (poly(I:C))-induced necroptosis and inflammatory signalling. Furthermore, EPEC infection inhibits TNF-induced phosphorylation and plasma membrane localization of mixed lineage kinase domain-like pseudokinase (MLKL). In vivo, EspL cysteine protease activity contributes to persistent colonization of mice by the EPEC-like mouse pathogen Citrobacter rodentium. The activity of EspL defines a family of T3SS cysteine protease effectors found in a range of bacteria and reveals a mechanism by which gastrointestinal pathogens directly target RHIM-dependent inflammatory and necroptotic signalling pathways.

Figure 1. EspL is a T3SS cysteine protease that degrades RIPK1 and RIPK3.

a, Schematic representation of EPEC E2348/69 genomic integrative element 6 (IE6) harbouring espL, nleB1, and nleE and immunoblot showing RIPK1 degradation in HT-29 cells infected with derivatives of EPEC E2348/69 as shown. Representative immunoblot from at least 3 independent experiments. Actin; loading control. b, Immunoblot showing RIPK1 degradation in HeLa cells uninfected or infected with EPEC E2348/69; untreated, or treated with either MG132 or z-VAD-FMK (z-VAD). Representative immunoblot from at least three independent experiments. Actin; loading control. c, Schematic representation of cysteine protease motif and secondary structure predicted by Phyre in YopT from Y. pestis KIM and EspL from EPEC E2348/69. d, Immunoblot showing levels of RIPK1, RIPK2 and RIPK3 in HT-29 cells infected with derivatives of EPEC E2348/69. Representative immunoblot from at least three independent experiments. Actin; loading control.

Figure 2. Distribution of EspL in Gram negative pathogens and substrate specificity.

a, Phylogeny of EspL homologues from a range of Gram negative pathogens which was midpoint rooted. Different genera are highlighted by background colour and the tips are coloured by species or pathotype. b, Coomassie Brilliant Blue stain of SDS PAGE gel showing in vitro cleavage of purified RHIM-containing regions of RIPK1, RIPK3, TRIF and ZBP1 (expressed as recombinant His6-ubiquitin-tagged proteins) by purified recombinant EspL. Representative gel from at least two independent experiments. Purified recombinant EspL_C47S was used as a negative control. White arrows indicate cleavage products. Black arrows indicate EspL and the band corresponding to free His₆-ubiquitin (His-Ub) Note, observed cleavage product derived from ZBP1/DAI was consistent with EspL cleavage in first RHIM of ZBP1/DAI. c, EspL cleavage site indicated by an arrow in the RHIM containing regions of RIPK1, RIPK3, TRIF and ZBP1/DAI. Alignment was performed using Clustal Omega and ESPript3. Cleavage sites in RIPK3 and TRIF were determined experimentally by mass spectrometry and N-terminal sequencing.

Figure 3. EspL inhibits TNF-induced necroptosis.

a, Cell death visualised by propidium iodide (PI) staining in HT-29 cells infected with derivatives of EPEC and treated with TNF, compound A (Cp.A) and z-VAD-FMK (z-VAD). Hoechst; stain for nucleic acid. Scale bar, 20 μm. Representative images shown from at least three independent experiments. b, Quantification of PI staining from microscopic analysis in HT-29 cells infected with derivatives of EPEC E2348/69 and treated with TNF, Cp.A and z-VAD. Results are mean ± s.e.m. percentage of cells positive for PI staining from three independent experiments counting ~200 cells in triplicate. *P<0.0001 compared to EPEC E2348/69 infected cells, one-way ANOVA with Holm-Sidak multiple comparison. c, Blue native PAGE analysis of MLKL membrane translocation in HT-29 cells infected with derivatives of EPEC E2348/69 treated with TNF, Cp.A and z-VAD. Representative immunoblot from at least three independent experiments. GAPDH; cytosolic fraction loading control, VDAC; membrane fraction loading control.

Figure 4. EspL activity inhibits TLR3/4-mediated signalling and contributes to in vivo persistence.

a, Ifnβ expression in doxycycline-inducible iBMDMs stably expressing either Flag-EspL or Flag-EspL_C47S and treated with either LPS or Poly I:C for 3 h as indicated. Results are mean ± s.e.m of at least three independent experiments performed in triplicate. Ifnβ expression relative to uninduced, unstimulated cells. *P<0.0005 compared to Flag-EspL induced with doxycycline and treated with LPS or poly I:C, unpaired, two-tailed t-test. b, MTT reduction in doxycycline-inducible iBMDMs stably expressing either Flag-EspL or Flag-EspL_C47S and treated with either LPS/z-VAD or Poly I:C/z-VAD for 20 h. Results are mean ± s.e.m. of absorbance at 540 nm from three independent experiments performed in triplicate. *P< 0.05 compared to Flag-EspL_C47S induced with doxycycline and treated with LPS or poly I:C, unpaired, two-tailed t-test. c, Immunoblot showing degradation of endogenous RIPK1 by EspL from C. rodentium (Flag-CREspL) but not EspL_C42S (Flag-CREspLC42S) expressed ectopically in HEK293T cells. Representative immunoblot from at least three independent experiments. d, Bacterial load in the faeces of mice 16 days after infection with derivatives of C. rodentium, including wild type C. rodentium ICC169 (CR), an espL deletion mutant (ΔespL) and ΔespL complemented with espL (EspL) or espL_C42s (EspL_C42S) by Tn 7 transposition. Each data point represents log₁₀ c.f.u. per g faeces per individual animal (c.f.u., colony forming units). Mean ± s.e.m. are indicated. Data was combined from three independent experiments. P values from Mann–Whitney U-test.