Fig. 1. Three domains of NOD2. The CARD domain interacts with RIP2. The NBD domain is responsible for the oligomerization of NOD. The LRR domain recognizes MDP.

2.2 The MDP dependent activation of NOD2

The activation of NOD2 signaling pathways can be divided into two categories: MDP dependent and MDP independent. The MDP dependent pathways have been extensively investigated. Cells phagocytose bacteria to form phagosome, which fuses with lysosome and further matures into phagolysosomes, where the bacteria are digested into PGN, the major component of bacteria cell wall[15]. The basic unit of PGN, MDP, should be phosphorylated by N-acetylglucosamine kinase (NAGK) before it is recognized by NOD2[16]. In macrophages and dendritic cells, the SLC15 family endosomal peptide transporter protein 1 PHT1 (SLC15A4) enables transmembrane transport of MDP, recruiting NOD2. And endosomes serve as a signaling platform for triggering the innate immune response[3,17–19]. Disrupting endosomal acidic chemicals inhibits NOD2-induced NF-κB activation, which may be due to the dependence of ligand transport on proton gradients as the energy source[20]. When bacteria are not invading the cytoplasmic region of the host, it can also cross the plasma membrane and localize to the cytoplasm through the plasma membrane transporter protein Human peptide transporter 1 hPepT1 (SLC15A2), which is expressed in intestinal epithelial cells[21,22]. Heat shock protein 70 (HSP70) is a positive regulator of NOD2 signaling, which prolongs the half-life of NOD2 and enhances NF-κB activity [23].

Although studies initially described NOD2 as a cytoplasmic receptor, it has been shown that membrane targeting of NOD2 is required for NF-κB activation[24]. This was illustrated by the CD-related NOD2 loss-of-function mutation 3020insC, which, despite being fully capable of binding MDP, fails to bind to the membrane and therefore cannot detect bacterial invasion [25,26]. Besides the anchorage of NOD2 to membranes by cytoskeletal components (vimentin) and binding proteins (FRMPD2, FERM and PDZ structural domain 2) or to endosomes by endosomal proteins (SLC15A)[27,28,18], studies have shown that palmitoylation of NOD1/2 oligomerization is needed not only for steady-state membrane binding but also for MDP-induced signaling pathways. During this process, the presence of Zinc Finger DHHC-Type Palmitoyltransferase 5 (ZDHHC5) and its enzymatic activity are essential for proper recruitment of NOD1/2 to the site of bacterial entry and the formation of phagosomes[29]. But how ZDHHC5 is attracted to bacterial entry sites is still unknown[29].

2.3 The MDP independent activation of NOD2

In contrast to MDP dependent activation, the mechanism of MDP-independent activation of NOD2 is ambiguous. It has been shown that viruses can activate NOD2[30,31], such as respiratory syncytial virus (RSV) and Middle East respiratory syndrome coronavirus (MERS-CoV). Once viral ssRNA is recognized, NOD2 can enhance IFN-β expression through activation of IRF3 by mitochondrial antiviral signaling protein (MAVS)[30]. Keestra-Gounder et al. also proposed a link between endoplasmic reticulum (ER) stress and the increase of IL-6 secretion via NOD1/NOD2 signaling, with MDP non-dependence[32]. Accumulation of unfolded or misfolded proteins, as well as viral or bacterial infections, can induce an unfolded protein response (UPR) in the endoplasmic reticulum, which is a protective response that limits cellular damage caused by ER stress[33]. UPR involves three transmembrane receptors, the protein kinase RNA-like endoplasmic reticulum kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE1α). In mice, intraperitoneal injection ofBifidobacterium abortus triggered the type IV secretion system, which secreted the effector protein system effector protein VceC, and induced ER stress. IRE1α recruited Tumor necrosis factor receptor-associated factor 2 (TRAF2) to the ER membrane and induced NOD1/NOD2 activation to promote IL-6 production via the NF-κB pathway[32] (Figure 2) .

The cytoskeletal protein vimentin performs cellular functions by interacting with the LRR structural domain in NOD2, which is recruited into the cell membrane, affecting NF-κB activation, autophagy, and bacterial processing[34]. The interaction of the cytoskeletal regulator RAC1 and ARHGEF7 (also known as PAK3BP) negatively regulates NOD2 activation. And the downregulation of the expression of both proteins enhances MDP-driven NOD2 activation[35,36]. In line with this, in bone marrow monocytes and intestinal epithelial cells , the disruption of the actin cytoskeleton by cytochrome D also increases NOD1- and NOD2-mediated NF-κB activation even in the absence of peptidoglycan, suggesting that cytoskeletal alterations can trigger the activation of NOD2[36].

2.4 Signaling pathways following NOD2 activation

Activation of NOD2 mainly initiates downstream gene expression through two signaling pathways, the NF-κB and MAPK signaling pathways(Figure 2) . In response to ligand binding, RIP2 recruited by NOD2 is extensively and selectively ubiquitinated, in which methionine 1-linked (M1) ubiquitin binds the IκB kinase complex (IKK) complex (IKKβ, IKKα, and NEMO). And the E3 ubiquitin ligases X-linked inhibitor of apoptosis protein (XIAP), cellular inhibitor of apoptosis protein 1 (CIAP1) , CIAP2, TNFR-associated factor 2 (TRAF2), TRAF5, and the linear ubiquitin assembly complex (LUBAC) are also involved in this process[37–40]. Indeed, the RING domain in XIAP recruits LUBAC, the trimeric complex contributing to the linear ubiquitination modification, to RIP2, which is essential for the NOD2-dependent response[40]. In line with this observation, mutations (Leu207Pro, Val198Met) in the BIR2 domain of XIAP lead to Ub ligase inactivation and impaired NOD2-dependent immune signaling[41]. Besides, several deubiquitinating enzymes (DUBs) negatively regulate the activation of NOD2, including CYLD, A20 and OTULIN[42,43]. Interferon regulatory factor 4 (IRF4) works as a negative regulator by inhibiting RIP2 ubiquitination in human DCs and thereby reducing NOD2-dependent NF-κB activation[44]. Next the activated IKK complex acts as an IκB kinase to phosphorylate IκB protein, followed by auto-ubiquitination and proteasomal degradation, releasing the NF-κB complex, which is translocated into the nucleus to promote cytokine expression[45].

In parallel, NOD2 signaling via lysine 63-linked (K63) ubiquitin initiates the recruitment of the TAK1-TAB2-TAB3 complex[46], enabling the activation of extracellular signal-regulated kinase 1 (ERK1), ERK2, JUN N-terminal kinase (JNK), and p38, which regulate the transcription factor AP-1 through MAPK phosphorylation[47]. The transcription factors NF-κB and AP-1 alone or in combination induce cytokine expression in the nucleus[48–50].