Supplementary Results

Previous work on TRIM71 had shown that TRIM71 ubiquitinates AGO2 thereby targeting it for proteasomal degradation (1). The mammalian TRIM71 paralog TRIM32 also interacts with AGO proteins and has ubiquitin ligase activity but was reported not to mediate AGO ubiquitination (2). In fact, TRIM32 was described as a positive regulator of the pathway enhancing miRNA mediated repression (2), while TRIM71 was proposed to antagonize miRNA mediated repression (1). Our findings that identify TRIM71 as a repressor of mRNAs co-operating with miRNAs to repress certain mRNA targets resemble the situation described for TRIM32. With the aim to get some understanding what determines the differences between TRIM71 and TRIM32, we first tested the activity of TRIM71 and TRIM32 to ubiquitinate AGO2. We adapted the in vivo ubiquitination protocol established by Rybak et al. (1), which examines ubiquitination of overexpressed tagged AGO2 upon co-expression of TRIM-NHL proteins in HEK293 cells. FLAG-tagged AGO2 was immunopurified from whole cell lysates and analyzed by western blotting using an antibody against either AGO2 or ubiquitin. AGO2 ubiquitination was induced upon expression of either TRIM71 or TRIM32, as indicated by a high molecular weight smear visible with both the AGO2 and the ubiquitin antibodies (Supplementary Fig. S8a, lane 2 and 3). To show that TRIM32 mediated ubiquitination of AGO2 is dependent on TRIM32’s ubiquitin ligase activity, we performed the assay with a TRIM32 mutant that lacks catalytic activity due to the mutation of two cysteine and one histidine residues in the RING domain (TRIM32 mut)(3). As expected, mutation of the catalytic domain abolished the ability of TRIM32 to ubiquitinate AGO2 (Supplementary Fig. S8b). A similar experiment demonstrating that the TRIM71 ability to ubiquitinate AGO2 is dependent on its ligase activity has already been described (1). Collectively, these data indicate that TRIM71 and TRIM32 are both able to ubiquitinate AGO2.