The cell-cycle regulator p21(Cip1) is degraded by proteasomes independently of ubiquitination. We now show that degradation of p21 in vivo does not require the 19S proteasome lid, which contains the ubiquitin-binding subunit. Instead, the major proteasomal pathway for p21 degradation involves an alternative proteasome lid, the REGgamma complex. REGgamma binds to p21 in vivo, and deletion of p21's REGgamma-binding site greatly extends its half-life. Knockdown of REGgamma by RNA interference stabilizes p21, p21 has a significantly extended half-life in REGgamma(-/-) murine embryonic fibroblasts, and the p21 abundance is elevated in REGgamma(-/-) mice. The role of REGgamma in cell-cycle regulation may extend beyond p21 regulation, because p16(INK4A) and p19(Arf) also bind to REGgamma and are stabilized in REGgamma-deficient cells.

Protein complexes of the 28-kDa proteasome activator (PA28) family activate the proteasome and may alter proteasome cleavage specificity. Initial investigations have demonstrated a role for the IFN-γ-inducible PA28α/β complex in Ag processing. Although the noninducible and predominantly nuclear PA28γ complex has been implicated in affecting proteasome-dependent signaling pathways, such as control of the mitotic cell cycle, there is no previous evidence demonstrating a role for this structure in Ag processing. We therefore generated PA28γ-deficient mice and investigated their immune function. PA28γ−/− mice display a slight reduction in CD8+ T cell numbers and do not effectively clear a pulmonary fungal infection. However, T cell responses in two viral infection models appear normal in both magnitude and the hierarchy of antigenic epitopes recognized. We conclude that PA28γ−/− mice, like PA28α−/−/β−/− mice, are deficient in the processing of only specific Ags.

The proteasome catalytic β subunits LMP2, LMP7, and MECL-1 and two proteasome activator proteins, PA28 α and β, are induced following exposure to IFN-γ in vitro. Induction of these immunosubunits and the PA28 α/β hetero-oligomer alters proteasome catalytic functions and specificity and enhances production of certain MHC class I epitopes. We sought to determine whether and to what extent proteasome subunit composition is regulated in vivo and to elucidate the mechanisms of such regulation. We analyzed basal expression levels of these inducible genes in normal, IFN-γ-deficient, and Stat-1-deficient mice. Mice of all three genotypes display constitutive expression of the immunosubunits and PA28, demonstrating that basal expression in vivo is independent of endogenous IFN-γ production. However, basal expression levels are reduced in Stat-1−/− mice, demonstrating a role for Stat-1 independent of IFN-γ signaling. To demonstrate that IFN-γ can induce these genes in vivo, mice were infected with Histoplasma capsulatum. Elevated expression of these genes followed the same time course as IFN-γ expression in infected mice. IFN-γ-deficient mice did not display elevated protein expression following infection, suggesting that other inflammatory cytokines produced in infected mice are unable to influence proteasome expression. Cytokines other than IFN-γ also failed to influence proteasome gene expression in vitro in cell lines that had no basal expression of LMP2, LMP7, or MECL-1. Thus, both in vitro and in vivo data demonstrate that IFN-γ is essential for up-regulation, but not constitutive expression, of immunoproteasome subunits in mice.

REG-γ, a protein involved in protein degradation, binds to nuclear AID, and REG-γ–deficient B cells contain more AID and exhibit increased immunoglobulin class switching.

REGγ is a member of the 11S regulatory particle that activates the 20S proteasome. Studies in REGγ deficient mice indicated an additional role for this protein in cell cycle regulation and proliferation control. In this paper we demonstrate that REGγ protein is equally expressed throughout the cell cycle, but undergoes a distinctive subcellular localization at mitosis. Thus, while in interphase cells REGγ is nuclear, in telophase cells it localizes on chromosomes, suggesting a role in mitotic progression. Furthermore, we found that REGγ overexpression weakens the mitotic arrest induced by spindle damage, allowing premature exit from mitosis, whereas REGγ depletion has the opposite effect, thus reflecting a new REGγ function, unrelated to its role as proteasome activator. Additionally, we found that primary cells from REGγ -/-mice and human fibroblasts with depleted expression of REGγ or overexpressing a dominant negative mutant unable to activate the 20S proteasome, demonstrated a marked aneuploidy (chromosomal gains and losses), supernumerary centrosomes and multipolar spindles. These findings thus underscore a previously uncharacterized function of REGγ in centrosome and chromosomal stability maintenance.