Immunoproteasomes Are Important for Proteostasis in Immune Responses

The regulated degradation of ubiquitin (ub)-tagged proteins by the ubiquitin proteasome system (UPS) represents amajor pathway not only for the maintenance of proteostasis, but also for supplying peptides for MHC class I antigen presentation. This immune function of the UPS can be improved by the immunoproteasome (IP), an isoenzyme of the 26S proteasome with altered and enhanced peptide hydrolyzing activity (Figures 1A and 1E) (Deol et al., 2007; Seifert et al., 2010; Huber et al., 2012). In their Matters Arising (Nathan et al., 2013 [this issue of Cell]), Nathan et al. questioned several results of our original study regarding the role of IP in preserving proteostasis upon interferon (IFN)-induced oxidative stress (Seifert et al., 2010). We showed that IFN-induced radical production augments the formation of poly-ub-conjugated, oxidantdamaged nascent proteins (defective ribosomal products [DRiPs]). The formation of DRiPs is transiently higher than their degradation rate, resulting in a transient accumulation of poly-ub-conjugates (Figure 1B). In comparison to standard proteasomes (SP), IPs eliminate DRiPs more efficiently and thereby concomitantly facilitate the MHC class I peptide supply (Figures 1A and 1E). Nathan et al. were not able to detect an IFN-g-induced transient increase in polyub-conjugates or the accumulation of ub-rich aggregates in the absence of IP. They also did not detect any effect of IP deficiency on disease manifestation in experimental autoimmune encephalomyelitis (EAE) or accelerated degradation of poly-ub-conjugates by IP. The reasons for the discrepancies are not discussed. In the majority of their experiments, Nathan et al. applied different protocols, thereby limiting a direct comparison with our results. They further did not address oxidative stress in their experiments, and in addition, a number of technical issues arise. To visualize IFN-induced accumulation of poly-ub-conjugates, which was also observed by other groups, it is essential to counteract deubiquitylation and degradation. Moreover, the kinetics and the increase in poly-ub-conjugate formation are strictly dependent on cell viability, the latter being affected by IFN-g-signaling-induced apoptosis (see Figure 1C and Supplemental Information). The notion that conflicting data presented by Nathan et al. could simply reflect distinct experimental setups is reinforced by the fact that one of the coauthors had replicated our findings in our laboratory when using our stimulation and buffer conditions (Spinnenhirn, 2009). The transient accumulation of ubconjugates is, in part, due to the increased expression of UBE2L6 (Seifert et al., 2010), which is known to conjugate both ub and the ub-like modifier ISG15 (Buchwald et al., 2010). Nathan et al. argue that IFN-g-induced UBE2L6 is preferentially involved in ISG15 modification; however, they fail to detect IFN-ginduced ISG15 conjugates. Our IFN response experiments (Figure 1D) clearly demonstrate that ISG15 protein conjugates do accumulate in response to IFN but that ISGylation occurs much later (24 hr) than ubiquitylation (8 hr). Nathan et al. also claim that b5i/ LMP7 / mice are not more susceptible to EAE than wild-type mice. However, they used the abdominal area for immunization, bearing the risk of severe side effects, and, most strikingly, investigated mice from independent colonies. Genetic differences in mouse colonies and microbiotic environments do affect the disease outcome in EAE. Likewise, heterozygous breeding needs to be performed to obtain related wild-type littermates as optimal controls in EAE studies (see Extended Experimental Procedures in Supplemental Information).