Similarly the MyDDosome, formed by MyD88, IRAK-4 and IRAK-2 shows all three types of interactions to occur suggesting that they constitute a common mode of interaction within the DD superfamily. At present DDs constitute the only subfamily that has been shown to be able to promote the formation of large multi-protein assemblies through homo-oligomerization. In contrast, the CARD and pyrin domain-containing NLR proteins are believed to form higher order complexes via oligomerization of their NACHT domains, while the CARDs or PYDs are assumed to interact in a 1:1 fashion with their downstream effectors. Unlike other NLRs, NOD2 a protein that regulates nuclear factor kB activation and mutations in which have been linked to a predisposition to Crohn��s disease contains 2 CARDs in tandem. Ligand sensing by the LRRs of NOD2 and subsequent NACHT domain-mediated self-oligomerization is thought to induce the recruitment and polyubiquitination of the downstream effector kinase RIP2 that in turn activates the NF-kB signaling pathway and subsequent transcription of proinflammatory genes. Complex formation between NOD2 and RIP2 relies on the specific recognition of their respective CARDs, an interaction which is not understood on a molecular level. Here we provide the first biophysical characterization of the tandem CARDs of NOD2. Our study uncovered that the two CARDs interact with one another in an intramolecular fashion. We present a biophysical analysis of the isolated and tandem domains, which is aimed at understanding the molecular basis of this intramolecular interaction and provide data from a mutational analysis that suggest that different sites are used for the intra and intermolecular CARD-CARD interactions of NOD2. r may require each other for their stability. We now show that in addition to recognizing RIP2 the tandem CARDs interact in an intramolecular fashion with an affinity that is similar to that identified for other CARD-CARD complexes. Importantly, we have identified a mutation in NOD2 CARDa, R86A that fully disrupts the interaction with RIP2 but has no effect on binding to CARDb. These results strongly indicate that the surfaces used by NOD2 for intra- and intermolecular interactions differ and hence are not mutually exclusive. Furthermore, our thermal DL-Carnitine hydrochloride unfolding studies showed that the interaction between the two CARDs significantly increases their stability, suggesting that the two domains do not act independently of one another. Based on these observations it is tempting to speculate that NOD2 may use an extended surface created by the interaction between the two CARDs to interact with RIP2 in a manner that is different from typical CARD-mediated interactions, which could explain the discrepancy between our data and those reported for the NOD1-RIP2 interaction. Mutual stabilization of tandem protein interaction modules, in which only one module may be able to interact with ligands, has been observed in other proteins. For example two of the six PDZ domains of Timosaponin-BII glutamate receptor-interacting protein pack against each other to form a stable, supramodular structure that supports binding of one of the two domains to its target. Similarly, adjacent WW domains in Suppressor of deltex ) interact with one another and this arrangement is stabilized upon ligand binding to one of the tandem domains, while the other is lacking a functional ligand binding site. Furthermore, adjacent protein interaction modules may interact to form a higher order structure that has ligand binding properties that are distinct from those of the individual domains as observed for the tandem SH3 domains of the NADPH oxidase subunit p47phox which form a superSH3 domain that contains only a single ligand binding site.