Bacterial cGAS-like enzymes produce 2',3'-cGAMP to activate an ion channel that restricts phage replication.

The mammalian innate immune system uses cyclic GMP-AMP synthase (cGAS) to synthesize the cyclic dinucleotide 2',3'-cGAMP during antiviral and antitumor immune responses. 2',3'-cGAMP is a nucleotide second messenger that initiates inflammatory signaling by binding to and activating the stimulator of interferon genes (STING) receptor. Bacteria also encode cGAS/DncV-like nucleotidyltransferases (CD-NTases) that produce nucleotide second messengers to initiate antiviral (antiphage) signaling. Bacterial CD-NTases produce a wide range of cyclic oligonucleotides but have not been documented to produce 2',3'-cGAMP. Here we discovered bacterial CD-NTases that produce 2',3'-cGAMP to restrict phage replication. Bacterial 2',3'-cGAMP binds to CD-NTase associated protein 14 (Cap14), a transmembrane protein of unknown function. Using electrophysiology, we show that Cap14 is a chloride-selective ion channel that is activated by 2',3'-cGAMP binding. Cap14 adopts a modular architecture, with an N-terminal transmembrane domain and a C-terminal nucleotide-binding SAVED domain. Domain-swapping experiments demonstrated the Cap14 transmembrane region could be substituted with a nuclease, thereby generating a biosensor that is selective for 2',3'-cGAMP. This study reveals that 2',3'-cGAMP signaling extends beyond metazoa to bacteria. Further, our findings suggest that transmembrane proteins of unknown function in bacterial immune pathways may broadly function as nucleotide-gated ion channels.

A type VII secretion system in Group B Streptococcus mediates cytotoxicity and virulence.

Type VII secretion systems (T7SS) have been identified in Actinobacteria and Firmicutes and have been shown to secrete effector proteins with functions in virulence, host toxicity, and/or interbacterial killing in a few genera. Bioinformatic analysis indicates that isolates of Group B Streptococcus (GBS) encode at least four distinct subtypes of T7SS machinery, three of which encode adjacent putative T7SS effectors with WXG and LXG motifs. However, the function of T7SS in GBS pathogenesis is unknown. Here we assessed the role of the most abundant GBS T7SS subtype during GBS pathogenesis. In a murine model of hematogenous meningitis, mice infected with GBS lacking a functional T7SS or lacking the secreted WXG100 effector EsxA exhibited less mortality, lower bacterial burdens in tissues, and decreased inflammation in the brain compared to mice infected with the parental GBS strain. We further showed that this T7SS induces cytotoxicity in brain endothelium and that EsxA contributes to these cytotoxicity phenotypes in a WXG motif-dependent manner. Finally, we determined that EsxA is a pore-forming protein, thus demonstrating the first role for a non-mycobacterial EsxA homolog in pore formation. This work reveals the importance of a T7SS in host-GBS interactions and has implications for T7SS effector function in other Gram-positive bacteria.

Toxin secretion and trafficking by Mycobacterium tuberculosis.

The tuberculosis necrotizing toxin (TNT) is the major cytotoxicity factor of Mycobacterium tuberculosis (Mtb) in macrophages. TNT is the C-terminal domain of the outer membrane protein CpnT and gains access to the cytosol to kill macrophages infected with Mtb. However, molecular mechanisms of TNT secretion and trafficking are largely unknown. A comprehensive analysis of the five type VII secretion systems of Mtb revealed that the ESX-4 system is required for export of CpnT and surface accessibility of TNT. Furthermore, the ESX-2 and ESX-4 systems are required for permeabilization of the phagosomal membrane in addition to the ESX-1 system. Thus, these three ESX systems need to act in concert to enable trafficking of TNT into the cytosol of Mtb-infected macrophages. These discoveries establish new molecular roles for the two previously uncharacterized type VII secretion systems ESX-2 and ESX-4 and reveal an intricate link between toxin secretion and phagosomal permeabilization by Mtb.

Pore-forming Esx proteins mediate toxin secretion by Mycobacterium tuberculosis.

Mycobacterium tuberculosis secretes the tuberculosis necrotizing toxin (TNT) to kill host cells. Here, we show that the WXG100 proteins EsxE and EsxF are essential for TNT secretion. EsxE and EsxF form a water-soluble heterodimer (EsxEF) that assembles into oligomers and long filaments, binds to membranes, and forms stable membrane-spanning channels. Electron microscopy of EsxEF reveals mainly pentameric structures with a central pore. Mutations of both WXG motifs and of a GXW motif do not affect dimerization, but abolish pore formation, membrane deformation and TNT secretion. The WXG/GXW mutants are locked in conformations with altered thermostability and solvent exposure, indicating that the WXG/GXW motifs are molecular switches controlling membrane interaction and pore formation. EsxF is accessible on the bacterial cell surface, suggesting that EsxEF form an outer membrane channel for toxin export. Thus, our study reveals a protein secretion mechanism in bacteria that relies on pore formation by small WXG proteins.

The tuberculosis necrotizing toxin is an NAD+ and NADP+ glycohydrolase with distinct enzymatic properties.

Upon host infection, Mycobacterium tuberculosis secretes the tuberculosis necrotizing toxin (TNT) into the cytosol of infected macrophages, leading to host cell death by necroptosis. TNT hydrolyzes NAD+ in the absence of any exogenous cofactor, thus classifying it as a ß-NAD+ glycohydrolase. However, TNT lacks sequence similarity with other NAD+ hydrolyzing enzymes and lacks the essential motifs involved in NAD+ binding and hydrolysis by these enzymes. In this study, we used NMR to examine the enzymatic activity of TNT and found that TNT hydrolyzes NADP+ as fast as NAD+ but does not cleave the corresponding reduced dinucleotides. This activity of TNT was not inhibited by ADP-ribose or nicotinamide, indicating low affinity of TNT for these reaction products. A selection assay for nontoxic TNT variants in Escherichia coli identified four of six residues in the predicted NAD+-binding pocket and four glycine residues that form a cradle directly below the NAD+-binding site, a conserved feature in the TNT protein family. Site-directed mutagenesis of residues near the predicted NAD+-binding site revealed that Phe727, Arg757, and Arg780 are essential for NAD+ hydrolysis by TNT. These results identify the NAD+-binding site of TNT. Our findings also show that TNT is an NAD+ glycohydrolase with properties distinct from those of other bacterial glycohydrolases. Because many of these residues are conserved within the TNT family, our findings provide insights into understanding the function of the >300 TNT homologs.

NAD+ Depletion Triggers Macrophage Necroptosis, a Cell Death Pathway Exploited by Mycobacterium tuberculosis.

Mycobacterium tuberculosis (Mtb) kills infected macrophages by inhibiting apoptosis and promoting necrosis. The tuberculosis necrotizing toxin (TNT) is a secreted nicotinamide adenine dinucleotide (NAD+) glycohydrolase that induces necrosis in infected macrophages. Here, we show that NAD+ depletion by TNT activates RIPK3 and MLKL, key mediators of necroptosis. Notably, Mtb bypasses the canonical necroptosis pathway since neither TNF-α nor RIPK1 are required for macrophage death. Macrophage necroptosis is associated with depolarized mitochondria and impaired ATP synthesis, known hallmarks of Mtb-induced cell death. These results identify TNT as the main trigger of necroptosis in Mtb-infected macrophages. Surprisingly, NAD+ depletion itself was sufficient to trigger necroptosis in a RIPK3- and MLKL-dependent manner by inhibiting the NAD+ salvage pathway in THP-1 cells or by TNT expression in Jurkat T cells. These findings suggest avenues for host-directed therapies to treat tuberculosis and other infectious and age-related diseases in which NAD+ deficiency is a pathological factor.