Intracellular lipopolysaccharide from Gram-negative bacteria including Escherichia coli, Salmonella typhimurium, Shigella flexneri, and Burkholderia thailandensis activates mouse caspase-11, causing pyroptotic cell death, interleukin-1β processing, and lethal septic shock. How caspase-11 executes these downstream signalling events is largely unknown. Here we show that gasdermin D is essential for caspase-11-dependent pyroptosis and interleukin-1β maturation. A forward genetic screen with ethyl-N-nitrosourea-mutagenized mice links Gsdmd to the intracellular lipopolysaccharide response. Macrophages from Gsdmd(-/-) mice generated by gene targeting also exhibit defective pyroptosis and interleukin-1β secretion induced by cytoplasmic lipopolysaccharide or Gram-negative bacteria. In addition, Gsdmd(-/-) mice are protected from a lethal dose of lipopolysaccharide. Mechanistically, caspase-11 cleaves gasdermin D, and the resulting amino-terminal fragment promotes both pyroptosis and NLRP3-dependent activation of caspase-1 in a cell-intrinsic manner. Our data identify gasdermin D as a critical target of caspase-11 and a key mediator of the host response against Gram-negative bacteria.

Caspase-1 activation by inflammasome scaffolds comprised of intracellular nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) and the adaptor ASC is believed to be essential for production of the pro-inflammatory cytokines interleukin (IL)-1β and IL-18 during the innate immune response. Here we show, with C57BL/6 Casp11 gene-targeted mice, that caspase-11 (also known as caspase-4) is critical for caspase-1 activation and IL-1β production in macrophages infected with Escherichia coli, Citrobacter rodentium or Vibrio cholerae. Strain 129 mice, like Casp11(-/-) mice, exhibited defects in IL-1β production and harboured a mutation in the Casp11 locus that attenuated caspase-11 expression. This finding is important because published targeting of the Casp1 gene was done using strain 129 embryonic stem cells. Casp1 and Casp11 are too close in the genome to be segregated by recombination; consequently, the published Casp1(-/-) mice lack both caspase-11 and caspase-1. Interestingly, Casp11(-/-) macrophages secreted IL-1β normally in response to ATP and monosodium urate, indicating that caspase-11 is engaged by a non-canonical inflammasome. Casp1(-/-)Casp11(129mt/129mt) macrophages expressing caspase-11 from a C57BL/6 bacterial artificial chromosome transgene failed to secrete IL-1β regardless of stimulus, confirming an essential role for caspase-1 in IL-1β production. Caspase-11 rather than caspase-1, however, was required for non-canonical inflammasome-triggered macrophage cell death, indicating that caspase-11 orchestrates both caspase-1-dependent and -independent outputs. Caspase-1 activation by non-canonical stimuli required NLRP3 and ASC, but caspase-11 processing and cell death did not, implying that there is a distinct activator of caspase-11. Lastly, loss of caspase-11 rather than caspase-1 protected mice from a lethal dose of lipopolysaccharide. These data highlight a unique pro-inflammatory role for caspase-11 in the innate immune response to clinically significant bacterial infections.

Recombineering allows DNA cloned in Escherichia coli to be modified via lambda (λ) Red-mediated homologous recombination, obviating the need for restriction enzymes and DNA ligases to modify DNA. Here, we describe the construction of three new recombineering strains (SW102, SW105 and SW106) that allow bacterial artificial chromosomes (BACs) to be modified using galK positive/negative selection. This two-step selection procedure allows DNA to be modified without introducing an unwanted selectable marker at the modification site. All three strains contain an otherwise complete galactose operon, except for a precise deletion of the galK gene, and a defective temperature-sensitive λ prophage that makes recombineering possible. SW105 and SW106 cells in addition carry l-arabinose-inducible Cre or Flp genes, respectively. The galK function can be selected both for and against. This feature greatly reduces the background seen in other negative-selection schemes, and galK selection is considerably more efficient than other related selection methods published. We also show how galK selection can be used to rapidly introduce point mutations, deletions and loxP sites into BAC DNA and thus facilitate functional studies of SNP and/or disease-causing point mutations, the identification of long-range regulatory elements and the construction of conditional targeting vectors.

The small intestine epithelium renews every 2 to 5 days, making it one of the most regenerative mammalian tissues. Genetic inducible fate mapping studies have identified two principal epithelial stem cell pools in this tissue. One pool consists of columnar Lgr5-expressing cells that cycle rapidly and are present predominantly at the crypt base1. The other pool consists of Bmi1-expressing cells that largely reside above the crypt base2. However, the relative functions of these two pools and their interrelationship are not understood. Here, we specifically ablated Lgr5-expressing cells using a diphtheria toxin receptor (DTR) gene knocked into the Lgr5 locus. We found that complete loss of the Lgr5-expressing cells did not perturb homeostasis of the epithelium, indicating that other cell types can compensate for elimination of this population. After ablation of Lgr5-expressing cells, progeny production by Bmi1-expressing cells increased, suggesting that Bmi1-expressing stem cells compensate for the loss of Lgr5-expressing cells. Indeed, lineage tracing showed that Bmi1-expressing cells gave rise to Lgr5-expressing cells, pointing to a hierarchy of stem cells in the intestinal epithelium. Our results demonstrate that Lgr5-expressing cells are dispensable for normal intestinal homeostasis. In the absence of these cells, the Bmi1-expressing cells can serve as an alternative stem cell pool, providing the first experimental evidence for the interrelationship between these populations. The Bmi1-expressing stem cells may represent both a reserve stem cell pool in case of injury to the small intestine epithelium and a source for replenishment of the Lgr5-expressing cells under non-pathological conditions.

Lipopolysaccharide from Gram-negative bacteria is sensed in the host cell cytoplasm by a non-canonical inflammasome pathway that ultimately results in caspase-11 activation and cell death. In mouse macrophages, activation of this pathway requires the production of type-I interferons, indicating that interferon-induced genes have a critical role in initiating this pathway. Here we report that a cluster of small interferon-inducible GTPases, the so-called guanylate-binding proteins, is required for the full activity of the non-canonical caspase-11 inflammasome during infections with vacuolar Gram-negative bacteria. We show that guanylate-binding proteins are recruited to intracellular bacterial pathogens and are necessary to induce the lysis of the pathogen-containing vacuole. Lysis of the vacuole releases bacteria into the cytosol, thus allowing the detection of their lipopolysaccharide by a yet unknown lipopolysaccharide sensor. Moreover, recognition of the lysed vacuole by the danger sensor galectin-8 initiates the uptake of bacteria into autophagosomes, which results in a reduction of caspase-11 activation. These results indicate that host-mediated lysis of pathogen-containing vacuoles is an essential immune function and is necessary for efficient recognition of pathogens by inflammasome complexes in the cytosol.