SIRT7 is a CRL4-DCAF1 binding protein
SIRT7 was previously reported to interact with the CRL4-DCAF1 and modulate its ubiquitin ligase activity (11,17,18). Consistent with those findings, in co-immunoprecipitation experiments, we observed an interaction between DCAF1 and SIRT7, despite some non-specific interaction of SIRT7 and anti-FLAG antibody beads (Fig. 1A). Further, co-immunoprecipitation experiments against SIRT7 after co-expression of DDB1 or DDB1 and DCAF1 suggest that SIRT7 is a binding partner of the CRL4-DCAF1 (Fig. 1B).
We next wondered whether SIRT7 would still interact with the CRL4-DCAF1 in the presence of Vpr. Co-immunoprecipitation experiments showed that SIRT7 indeed forms a stable complex with DDB1, DCAF1 and Vpr (Fig. 1C). However, we did not see appreciable interaction between Vpr and SIRT7 (Fig. 1D). Together, these data suggest that both SIRT7 and Vpr form a stable complex with the CRL4-DCAF1.
SIRT7 is poly-ubiquitinated by the CRL4-DCAF1 in a Vpr-dependent manner for proteasome-dependent degradation.
Vpr can mediate proteasome-dependent degradation of several cellular targets by two distinctive mechanisms: Vpr, associated with DCAF1, directly interacts and recruits cellular targets for poly-ubiquitination by the CRL4, or Vpr, by binding DCAF1, activates the E3 ligase activity and mediates poly-ubiquitination of DCAF1-interacting proteins (See Fig. 6 and Discussion). Since we found that both SIRT7 and Vpr simultaneously interact with DDB1 and DCAF1 (Fig. 1C), we asked whether Vpr modulates the ubiquitin ligase activity of CRL4-DCAF1 for poly-ubiquitination of SIRT7. To that end, we performed in vitro ubiquitination assays with reconstituted CRL4-DCAF1 and CRL4-DCAF1-Vpr complexes (Fig. 2A). Here, we used the C-terminal domain of DCAF1 (residues 1045-1396), which is sufficient for interacting with DDB1, Vpr and SIRT7. Poly-ubiquitination of SIRT7 was apparent only in the presence of Vpr (Compare lane 1-3 versus lanes 4-6 in Fig. 2A). Transient expression of SIRT7 and Vpr resulted in down-regulation of SIRT7 in a Vpr-dose dependent manner (Fig. 2B). Treatment of those cells with proteasome inhibitor, MG132 moderated Vpr-dependent down-regulation of SIRT7 (Fig. 2C). Further, the level of endogenous SIRT7 protein decreased with increasing ectopic expression of Vpr (Fig. 2D). Taken together, these data suggest that Vpr enhances the activity of CRL4-DCAF1 to poly-ubiquitinate SIRT7 for proteasome-dependent degradation.
To further understand the role of Vpr in mediating SIRT7 down-regulation, we generated a Vpr mutant R62D/F69A. These two residues of Vpr reside at the DCAF1 binding interface (Fig. 3A). We previously showed that single R62D or F69A mutation affected Vpr interaction with DCAF1, respectively, and F69A mutation reduced Vpr-mediated Exo1 association with DDB1 and DCAF1 (15, 27). The Vpr mutant does not form a stable complex with DCAF1 (Fig. 3B) and cannot mediate down-regulation of SIRT7 (Fig. 3C). Further, treatment of cells ectopically expressing Vpr and SIRT7 with the CRL inhibitor, MLN4924 resulted in elevated SIRT7 level (Fig. 3D). Altogether, these data suggest that Vpr binding to DCAF1 is necessary to mediate SIRT7 degradation in a CRL4-dependent manner.
SIRT7 degradation is a conserved feature of HIV-1 Vpr and SIVcpz Vpr
The CRL4-DCAF1 is also usurped by an ortholog of HIV-1 Vpr, HIV-2/SIVmac (SIV isolated from macaque monkey) Vpx, which induces degradation of SAMHD1 and HUSH complex (19-22). Structural studies suggested that both Vpr and Vpx essentially bind DCAF1 in the same manner, using a common interface, and recruit their respective cellular targets using unique interfaces (15,23-25). Thus, we asked whether Vpx can also induce degradation SIRT7 by activating the CRL4-DCAF1. Surprisingly, SIVmac Vpx, even at a higher level than Vpr, did not enhance degradation of SIRT7 (Fig. 4A). In vitro ubiquitination confirmed that Vpx does not significantly enhance the CRL4-DCAF1 activity to mediate poly-ubiquitination of SIRT7 (Fig. 4B). Thus, the data suggest that CRL4-DCAF1 activation is a unique function inherent only to HIV-1 Vpr.
All the above studies were performed with Vpr isolated from the HIV-1 NL4-3 strain. To explore whether Vpr-dependent SIRT7 degradation is evolutionarily conserved among various HIV-1 strains, we co-expressed SIRT7 with Vpr isolated from HIV-1 YU2, LAI, Q23, M, N or O group (Fig. 4C). Each tested Vpr protein reduced the SIRT7 level in a dose-dependent manner. Further, Vpr isolated from SIVcpz PTT and PTS strains, evolutionary predecessors of HIV-1, also reduced the level of SIRT7 (Fig. 4D). However, Vpr isolated from HIV-2 A1, A2 or B strain, that are closely related SIV isolated from sooty mangabeys, did not efficiently reduce SIRT7 level (Fig. 4E). In line with our observation, recent quantitative proteomics with CEM-T4 T cells transduced with VSVg-pseudotyped HIV containing HIV-1/SIVcpz Vpr, but not other SIV or HIV-2 Vpr, showed reduction in SIRT7 cellular level, (26). In vitro ubiquitination assays of SIRT7 were performed with CRL4-DCAF1 in complex with various Vpr proteins and confirmed that SIRT7 is targeted for poly-ubiquitination (Fig. 4F). On the other hand, Vpx isolated from SIVmac and Vpr isolated from HIV-2 A1 did not siginificantly enhance the ubiquitin ligase activity of CRL4-DCAF1 (Fig. 4G). Thus, taken altogether, the data suggest that SIRT7 degradation is a conserved function of Vpr in HIV-1 and SIVcpz.
SIRT7 degradation and G2 arrest are independent functions of HIV-1 Vpr
One distinctive phenotype that has been associated with HIV-1 Vpr expression in cycling cells is G2 arrest. HEK293T cells transfected with HIV-1 Vpr, but not with SIVmac Vpx, display accumulation in G2 (Fig. 5A, upper panels). Mechanistically, the C-terminus of Vpr has been proposed to recruit unidentified cell cycle related factor(s) onto the CRL4-DCAF1 for proteasome-dependent degradation, activating damaged DNA response pathways. Cell cycle analyses of HEK293T cells transfected with the Vpr mutant that is deficient for DCAF1-binding (R62D/F69A) or Vpr C-terminal mutants (residue 1-79, and R80A) confirmed this model (Fig. 5A, lower panels). Since SIRT7 participates in the damaged DNA response pathway, we tested whether the C-terminus of Vpr interacts with SIRT7 for its degradation (Fig 5B). Both Vpr C-terminal mutants efficiently mediated SIRT7 degradation. Thus, Vpr-dependent SIRT7 degradation is independent of Vpr-induced G2 arrest.