tivity of the double mutant K167R/K193R. Taken together these results demonstrate that duck TRIM25 is activating duck RIG-I through a mechanism that does not require anchored K63-linked polyubiquitin chains. Discussion Duck RIG-I CARD K167R/K193R Double Mutant is not Ubiquitinated To ascertain the importance of sites for attachment of anchored polyubiquitin chains in the MedChemExpress 480-44-4 activation of duck CARD domains in DF1 chicken cells, we created a set of mutants and examined their ubiquitination. We first created a mutant, Q170K, to restore the equivalent of the human RIG-I lysine 172. This mutant did not show an increased ability to activate the RIG-I pathway and ubiquitinated bands look identical to those of the wild type construct. We then mutated each of the lysines, K167, and K193 independently, and together in a double mutant. Both of the single mutants showed identical patterns of bands demonstrating ubiquitination, but the K167R/K193R double mutant lost all larger bands indicating it was not ubiquitinated, nor were K63 polyubiquitin chains attached. These results demonstrate that both K167 and K193 are attachment sites for bound ubiquitin and polyubiquitin chains in duck RIG-I CARD domains. Duck RIG-I CARD K167R/K193R Mutant can be Activated by TRIM25 To determine whether the duck CARD domains with the mutated lysines could interact with TRIM25, we cotransfected the mutant with duck TRIM25 in DF-1 cells. We confirmed the interaction of duck CARD domains bearing K167 and K193 mutations with duck TRIM25 using GST-pulldown, suggesting these lysines are not essential for duck TRIM25 interaction. No larger bands were detected in the GST-pulldown samples in the presence of duck TRIM25 indicating that duck TRIM25 does not attach ubiquitin or polyubiquitin chains at any lysine other than K167 and K193. Exploring the importance of K167 and K193 in the activation of RIG-I, we used the dual luciferase assay and examined chIFN-b promoter induction in cells transfected with the different mutants. None of the mutants lost its ability to activate the RIG-I pathway, as indicated by 400-fold to 800-fold induction of the chicken IFN-b promoter. These results suggest that ubiquitination of K167 and K193 is not essential for activation of RIG-I CARD domains. We show the ability of duck TRIM25 to increase the TRIM25 Activation of Duck RIG-I 7 TRIM25 Activation of Duck RIG-I linked unanchored polyubiquitin chains, while UbcH5a catalyzes the formation of covalently attached polyubiquitin chains. Both E2 ligases are highly conserved in chickens, and the human and duck TRIM25 are interchangeable for activation of duck RIG-I. An alternative hypothesis is that human CARD domains cannot interact with chicken MAVS, and this interaction is needed prior to attached monoubiquitination. While this idea is contrary to the existing models for RIG-I activation, the divergence of 8 TRIM25 Activation of Duck RIG-I chicken and human MAVS also explains PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19649022 the previously observed inability 
of chicken MDA5 to function in human cells or human MDA5 to function in chicken DF-1 cells. It is worth noting that ducks, and other birds, have an alanine residue at D122, and the D122A mutant of human RIG-I does not activate the IFN pathway even though it interacts with polyubiquitin chains, presumably because it cannot engage MAVS. The conservation of a RIG-I splice variant present in tissue samples of influenza-infected ducks provides indirect evidence of the importance of the TRIM25 RIG-I phy