The NLRP1 inflammasome: new mechanistic insights and unresolved mysteries

Abstract

Nucleotide-binding domain, leucine-rich repeat (NLR) proteins constitute a diverse class of innate immune sensors that detect pathogens or stress-associated stimuli in plants and animals. Some NLRs are activated upon direct binding to pathogen-derived ligands. In contrast, we focus here on a vertebrate NLR called NLRP1 that responds to the enzymatic activities of pathogen effectors. We discuss a newly proposed 'functional degradation' mechanism that explains activation and assembly of NLRP1 into an oligomeric complex called an inflammasome. We also discuss how NLRP1 is activated by non-pathogen-associated triggers such as the anti-cancer drug Val-boroPro, or by human disease-associated mutations. Finally, we discuss how research on NLRP1 has led to additional biological insights, including the unexpected discovery of a new CARD8 inflammasome.

Figure 1.. Domain architecture of representative NLRs.

NLRs that form an inflammasome contain Nucleotide Binding Domain (NBD, purple) and Leucine-Rich Repeat (LRR, green) domains. Upon pathogen detection, NLRs assemble and recruit CASP1 via N-terminally located death domains: either directly through a CARD (grey) or indirectly through a PYD (black) via the adaptor ASC. Distinct from all other NLRs, NLRP1 has a C-terminal CARD and an N-terminal PYD, which has been lost in some species, including mouse. NLRP1 also contains a Function-to-find domain (FIIND), comprised of ZU5 and UPA subdomains, which undergoes constitutive self-cleavage (i.e., auto-processing). mNLRP1B, mouse NLRP1B; hNLRP1, human NLRP1; BIR, baculovirus inhibitor of apoptosis protein repeat.

Figure 2.. NLRP1 recognizes and responds to diverse pathogen-encoded effectors via ‘functional degradation.’

In this model, the FIIND domain, consisting of ZU5 and UPA subdomains, undergoes constitutive auto-processing, resulting in two non-covalently associated peptides (1). N-terminal proteolysis by Lethal Factor (LF) protease exposes a destabilizing N- degron (2a), which is recognized by a UBR2-containing complex resulting in NLRP1B ubiquitylation (2b). Subsequent proteasomal degradation of the N-terminal domains (NBD-LRR- ZU5) (3, 4) liberates the C-terminal UPA-CARD (5), which allows for its assembly and recruitment of CASP1 (6). The Shigella E3 ubiquitin ligase IpaH7.8 can directly ubiquitylate NLRP1B (2c), which also promotes NLRP1B degradation and inflammasome activation. Ub, ubiquitin; Nt, N-terminus.

Figure 3.. Pathogen recognition via an integrated decoy.

In the ‘integrated decoy’ strategy of pathogen detection, innate sensors harbor a domain that is similar to the ‘intended’ target of pathogen effectors. Host sensor activation occurs when the integrated decoy is triggered by the same enzymatic activity used by the pathogen effector to antagonize other host proteins. In plants (left), immune signaling requires numerous WRKY-containing transcription factors (TFs). The pathogen effectors AvrRps4 and PopP2 inhibit WRKY-containing TFs through post- translational modification. This same enzymatic activity is sensed by the host sensor RRS1 through its WRKY-integrated decoy, resulting in its activation and a protective immune response. We hypothesize that the NLRP1 PYD may also function as an integrated decoy for pathogen detection (right). Note that any NLRP1 N-terminal domain could theoretically serve as an integrated decoy. TIR, Toll–interleukin-1 receptor domain; Ub, ubiquitin.

Figure 4.. FIIND (ZU5-UPA) containing proteins.

The Function-to-find domain (FIIND) is comprised of two subdomains, ZU5 and UPA. In NLRP1 and CARD8, the FIIND is followed by a C-terminal CARD, a death-fold domain. This domain architecture is found in several other human proteins. The NLRP1 and CARD8 FIIND is structurally similar to that of UNC5B (pdb 3G5B) and other Netrin receptors, MACC1 and SH3BP4. In contrast, ANK2 (pdb 4D8O), DTHD1 and PIDD all have a second ZU5 preceding the FIIND, which may itself be sufficient for a second auto-processing event. FIIND auto-processing is required for the function of NLRP1, CARD8 and PIDD, but is apparently not essential for the function of UNC5CL. Likewise, some FIIND-containing proteins, like ANK2, do not appear to undergo auto-processing (closed diamond). FIIND auto-processing has not been experimentally tested in many cases (indicated by a “?”). In addition, FIIND-containing proteins are decorated with a variety of accessory domains, suggesting that ‘functional degradation’ may be utilized across diverse biological processes. Ig, immunoglobulin domain; TSP1, Thrombospondin type 1 repeats; AnkR, ankyrin repeat; NP55, Neuroendocrine-specific golgi protein P55; SH3, Src Homology 3 domain.