The mysterious role of the NLRP9 pyrin domain in inflammasome assembly

The molecular foundation for the activation of proinflammatory caspases was unveiled by the discovery of a supramolecular protein platform known as the inflammasome [1]. Since this scientific breakthrough in innate immunity almost 20 years ago, different types of inflammasomes that respond to a variety of pathogens and cellular insults have been identified [2]. We now have a better understanding of the proteins involved in the formation of the inflammasome [3,4], their structural and dynamic features, and preferred interaction partners [4–7]. More recently, we have been able to identify important structural factors that are key in inflammasome assembly [8–11].

Pattern recognition receptors (inflammasome sensors) [12] become activated and oligomerize in the presence of specific pathogen-associated or danger-associated molecular patterns. The adaptor protein ASC [5] typically accompanies the activated sensor and, by means of its two protein-interacting death domains [pyrin domain (PYD) and CARD], connects the oligomerized sensor to procaspase-1 molecules. The recruitment of the procaspase into the inflammasome platform is assumed to facilitate its activation via the proximity-induced autoprocessing model [13]. The active caspase cleaves proinflammatory cytokines (interleukin-1β and interleukin-18) and the protein gasdermin D [14]. Upon cleavage, the latter leads to lytic cell death by pyroptosis and the concomitant release of the active cytokines [14]. The inflammatory response has been triggered.

Once assembled, the inflammasome adopts the shape of a filamentous ring with an outer diameter of approximately 0.5 μm [15]. Inflammasome assembly seems to be directional, leading to the formation of concentric rings formed by the sensor and the adaptor protein ASC, with a localized high concentration of the procaspase at the center of the ring [15]. However, we still do not understand how all the molecular pieces are connected together to form the inflammasome ring. Structural information obtained from different inflammasome sensors indicates that they can oligomerize forming macromolecular rings of ~13 to ~30 nm depending on the type of receptor [3]. These rings tend to stack-forming dimers or higher-order multimers with helicoidal macrostructures [10]. However, sensor self-association into linear oligomers has also been observed [16]. How these nanometer rings and helical or linear macrostructures lead to the micrometer-size complete inflammasome particle is not known. For ASC-dependent inflammasomes, it has been suggested that the sensor recruits ASC, prompting its self-association into filaments with a PYD core that leaves the CARDs free for further interaction with the procaspase [8]. However, structural studies of full-length ASC indicate that it can polymerize into two-sided filaments constituted of both PYD and CARD domains, which could confer the directionality of assembly and could facilitate the observed stacking of additional filaments, thus potentially explaining the thickness of the whole inflammasome particle [11,17,18].

Substantial amount of research is still needed to provide answers to these and other questions. For example, it is not understood why inflammasome assembly is necessary for procaspase-1 activation, whereas other caspases, such as caspase-4 and caspase-5, oligomerize and self-activate by direct binding to cytoplasmic lipopolysaccharide components of Gram-negative bacteria [19]. In addition, the role of many pattern recognition receptors remains unclear. Significant functional evidence has been reported for inflammasomes involving several members of the NOD-like receptors (NLR) family including NLRP1, NLRP2, NLRP3, NLRP6, NLRP7, NLRP12, and NLRC4, which respond to diverse stimuli [20]. However, it is becoming increasingly evident that a subgroup of the NLRP subfamily is involved in early embryonic development and could perform critical functions related to the reproductive system, as certain developmental disorders have been associated with mutations in NLR genes and high expression levels of several NLRs are observed in reproductive organs [21]. Moreover, NLRs from the subgroup including NLRP2, NLRP4, NLRP5, NLRP7, NLRP8, NLRP9, NLRP11, NLRP13, and NLRP14 form a phylogenetic cluster separated from other NLRP proteins [21].

A particularly mysterious NLR member is NLRP9, as it seemingly moonlights in the immune and developmental systems: NLRP9 is expressed in human oocytes and belongs to the development-related phylogenetic cluster [21], and the murine orthologue (NLRP9b) has been reported to form an inflammasome together with ASC and procaspase-1 against rotavirus infection in intestinal epithelial cells upon recognition of dsRNA by helicase DHX9 [22]. Interestingly, the NLRP9b inflammasome apparently includes DHX9 in its assembly, which adds a new aspect to the molecular mechanisms of inflammasome formation.

To improve our understanding on the modus operandi of the NLRP9 receptor, Park et al. [23] and Geyer et al. [24] have studied the biophysical behavior of the PYD of NLRP9 (NLRP9^PYD) in solution and its structural features by X-ray crystallography. Both studies show using size-exclusion chromatography, and dynamic and static light scattering that NLRP9^PYD is mainly monomeric in vitro under physiological conditions at millimolar concentration. Moreover, Park et al. indicate that although NLRP9^PYD shows some tendency to dimerize, it does not polymerize even after several hours of incubation. Importantly, Geyer et al. found that NLRP9^PYD and NLRP9b^PYD do not self-associate nor colocalize with ASC specks in cellular assays. In contrast, most PYDs tend to polymerize into filaments both in cell-free media and in cells. In fact, this work shows that under identical conditions, NLRP3^PYD is capable of polymerizing and colocalizing with ASC in their cell assays. Evidence presented by both groups suggests that particular structural attributes of NLRP9^PYD are involved in this anomalous behavior. Specifically, Park et al. indicate that the N-terminal loop comprising residues 5–11 in NLRP9^PYD forms stabilizing hydrophobic interactions with helix I, hence restricting its flexibility and facilitating its location into the surface comprised by helices I and IV, a unique feature not found in other three-dimensional (3D) structures of PYDs. By building a model that represents the potential polymerization of NLRP9^PYD into a helical tube using the known 3D polymeric structure of NLRP6^PYD as template [25], Park et al. suggest that the position of the N-terminal loop would obstruct the interacting surfaces necessary for the stabilization of the polymeric architecture. Based on these observations, this group proposes that the N-terminal loop in NLRP9^PYD could serve as an autoinhibitory mechanism for inflammasome assembly.

Geyer et al. point out that several amino acid charge inversions and substitutions of hydrophobic residues by charged ones in the NLRP9^PYD structure are located in the interacting surfaces that hold the polymerized helical tube of other PYDs. They conclude that these significant changes in the protein surface might preclude NLRP9^PYD self-association. In fact, Lys48 and Lys49 of different protomers would be positioned close to each other at the center of the helical tube should the domain polymerize, which leads Geyer at al. to argue that the potential repulsive forces could impede oligomerization. From the sequence and structural comparison of NLRP9^PYD with other PYDs, they conclude that monomeric NLRP9^PYD already adopts a conformation consistent with the polymerized protomer structure. This work also proposes that other factors such as post-translational modification, specifically phosphorylation, and oligomerization mediated by other NLR domains could play a role in the function of the NLRP9 inflammasome. Indeed, it has been suggested that NLRP6 and NLRP9b could be recruited to the same inflammasome [26], raising the question of whether NLRP6 is the oligomerization inducer and NLRP9b has a more passive role. Altogether, the studies by Park et al. and Geyer et al. point to different structural features that could be involved in the atypical behavior of NLRP9^PYD and suggest research directions that could help fill the gaps in our still scarce understanding of the inflammasome.

Abbreviations

NLRP

NOD-like receptor family, pyrin domain containing

PYD

pyrin domain