Biophysical and functional study of CRL5Ozz, a muscle specific ubiquitin ligase complex

Abstract

Ozz, a member of the SOCS-box family of proteins, is the substrate-binding component of CRL5^Ozz, a muscle-specific Cullin-RING ubiquitin ligase complex composed of Elongin B/C, Cullin 5 and Rbx1. CRL5^Ozz targets for proteasomal degradation selected pools of substrates, including sarcolemma-associated β-catenin, sarcomeric MyHC_emb and Alix/PDCD6IP, which all interact with the actin cytoskeleton. Ubiquitination and degradation of these substrates are required for the remodeling of the contractile sarcomeric apparatus. However, how CRL5^Ozz assembles into an active E3 complex and interacts with its substrates remain unexplored. Here, we applied a baculovirus-based expression system to produce large quantities of two subcomplexes, Ozz-EloBC and Cul5-Rbx1. We show that these subcomplexes mixed in a 1:1 ratio reconstitutes a five-components CRL5^Ozz monomer and dimer, but that the reconstituted complex interacts with its substrates only as monomer. The in vitro assembled CRL5^Ozz complex maintains the capacity to polyubiquitinate each of its substrates, indicating that the protein production method used in these studies is well-suited to generate large amounts of a functional CRL5^Ozz. Our findings highlight a mode of assembly of the CRL5^Ozz that differs in presence or absence of its cognate substrates and grant further structural studies.

Figure 1

Expression and purification of CRL5^Ozz components. (a) Coomassie brilliant blue (CBB)–stained gels of baculovirus-produced Ozz (Plaque 1, Plaque 2), Elo B (Plaque 1, Plaque 2), Elo C (Plaque 1, Plaque 2), Cul 5 (Plaque 1, Plaque 2), and Rbx1 (Plaque 1, Plaque 2) from total lysates of individual plaques, compared to mock control (Control). (b) Western blot analysis of total lysates overexpressing the individual CRL5^Ozz components probed with of anti-Ozz, anti-Elo B, anti-Elo C, anti-Cul5 and anti-Rbx1 specific antibodies. (c,d) CBB-stained SDS‐polyacrylamide gels of Ni-NTA agarose affinity purified Ozz–EloBC and Cul5–Rbx1 subcomplexes.

Figure 2

Analysis of CRL5^Ozz assembly by gel filtration chromatography and coimmunoprecipitation. (a) Affinity purified Ozz–EloBC (~ 60 kDa), Cul5–Rbx1 (~ 100 kDa) and the mixture of the two subcomplexes were visualized on SDS‐polyacrylamide gels stained with SYPRO Ruby. (b) CRL5^Ozz chromatography profile showing the elution volume and the fractions corresponding to the three major peaks (Peak 1, Peak 2 and Peak 3). (c) Superose 6 10/300 size exclusion chromatography markers: thyroglobulin, 669 kDa; apoferritin, 443 kDa; β-amylase, 200 kDa; carbonic anhydrase, 29 kDa. (d) The 1:1 mixture was loaded onto a gel filtration chromatography column. Fractions from Peak 1, Peak 2 and Peak 3 were pooled, concentrated and their protein content visualized on SDS‐polyacrylamide gels stained with SYPRO Ruby. All 5 components of CRL5^Ozz eluted together mainly in Peak 2 (RV 14.17 = ~ 340 kDa), corresponding to 1D1–1E1 fractions. (e) Purified Ozz–EloBC and Cul5–Rbx1 subcomplexes were mixed in a 1:1 ratio and subjected to immunoprecipitation using either anti-EloC or anti-Rbx1 antibodies. Immunoprecipitated samples were analyzed on SDS–polyacrylamide gels stained with SYPRO Ruby. Both antibodies co-immunoprecipitated all CRL5^Ozz components, confirming the formation of the assembled complex by mixing the two subcomplexes in vitro.

Figure 3

Analysis of CRL5^Ozz by glycerol gradient ultracentrifugation microfractionation. (a) Sedimentation profiles of the individual Ozz–EloBC and Cul5–Rbx1 subcomplexes as well as the 1:1 mixture of the two (7.5 μg Ozz–EloBC and 7.5 μg Cul5–Rbx1) were obtained by densitometric measurement of band intensity of eluted proteins after glycerol gradient ultracentrifugation. Aliquots of the first 18 fractions of the total 25 fractions collected were analyzed on SDS–polyacrylamide gel stained with SYPRO Ruby. The Ozz–EloBC (~ 60 kDa, upper panel) and Cul5–Rbx1 (~ 100 kDa, middle panel) subcomplexes eluted in fractions corresponding to their calculated molecular weight. The assembled CRL5^Ozz elution pattern shifted to fractions corresponding to its monomeric molecular weight (~ 160 kDa, lower panel). (b) 15 μg of a CRL5^Ozz complex and 15 μg of albumin were mixed and run together on a glycerol gradient. The sedimentation profile of CRL5^Ozz was not altered by the presence of albumin. (c) Sedimentation profile of Apoferritin (~ 443 kDa), β-amylase (~ 200 kDa) and albumin (~ 60 kDa) protein markers.

Figure 4

Sedimentation velocity—AUC analysis of Ozz–EloBC, Cul5–Rbx1, and CRL5^Ozz complex. Panels (a–c) Top panel: Fringe displaced sedimentation velocity profiles (fringes vs radius), with superimposed solid lines showing the best fit to the model, and below residuals of the fits. Bottom panel: display of the continuous sedimentation coefficient distribution c(s) plots of Ozz–EloBC (red line), Cul5–Rbx1 (blue line), and a mixture of the former two complexes (black line); (d–f) Top panel: fringe displaced sedimentation velocity profiles (fringes vs radius), with superimposed solid lines showing the best fit to the model, and below residuals of the fits. Middle panel: Contour plots (heat maps) of the two-dimensional size-and-shape distributions, c(s,f/f₀) and Bottom panel: Molar mass distributions c(s,M), of Ozz–EloBC, Cul5–Rbx1 and a mixture of Ozz–EloBC and Cul5–Rbx1, respectively. The experiments were conducted in 137 mM NaCl, 0.27 mM KCl, 10 mM Na2HPO4 and 1.8 mM KH2PO4 pH 7.2 buffer at 20 °C and at a rotor speed of 50,000 rpm. The s-, f/f₀ and M-values of the protein complexes are listed in Tables 2 and 3.

Figure 5

Sedimentation velocity—analytical ultracentrifugation analysis, s_w isotherm of Ozz–EloBC. (a) The best-fit isotherm of the signal-weight-average s-values, s_w, obtained by integration of c(s) distributions of Ozz–EloBC over the s-range of 3 and 7 S for each loading concentration in a dilution series. The solid line is the fitted isotherm to a reversible monomer–dimer self-association model with the best-fit dissociation constants K_D12 value as well as confidence intervals and the root mean square deviation of the fits at all the concentrations listed (Table 4). (b) The species population plots of fraction protomer concentration vs log total concentration (Molar) with the amounts of monomer and dimer at specific concentrations determined by the K_D12 value of the self-association model are also shown.

Figure 6

Glycerol gradient ultracentrifugation microfractionation analysis of CRL5^Ozz and its substrates. Assembled CRL5^Ozz mixed with purified preparations of (a) GST- full length β-catenin, (b) GST-MyHC_emb fragment (1041–1941 a.a.) and (c) GST-full length Alix was fractionated from a post-ultracentrifuged glycerol gradient (8 h). Aliquots of each fraction were separated on SDS–polyacrylamide gels and their protein content visualized on gels stained with SYPRO Ruby. The densitometric measurement of band intensity of the proteins in each fraction showed a shift to a higher molecular weight when CRL5^Ozz was mixed with either of its substrates, compared to the molecular weights of CRL5^Ozz or its individual substrates: CRL5^Ozz–β-catenin (~ 272 kDa), CRL5^Ozz–MyHC_emb fragment (1041–1941 a.a.) (~ 280 kDa) or CRL5^Ozz–Alix (~ 282 kDa). (d) Sedimentation analysis of Alix mixed with albumin as internal control. The fractions were loaded on an SDS–polyacrylamide gel and stained SYPRO Ruby. Alix and albumin were visible in fractions 2–12 and the profile of either protein was not altered in presence of the other.

Figure 7

In vitro ubiquitination of recombinant β-catenin, MyHC_emb and Alix mediated by CRL5^Ozz. (a,d) GST-tagged-β-catenin, (b,e) MyHC_emb fragment (1041–1941 a.a.) and (c,f) Alix was incubated with CRL5^Ozz and either native ubiquitin or mutant Ub (K48R). In the presence of native ubiquitin and CRL5^Ozz, β-catenin or MyHC_emb or Alix were efficiently ubiquitinated. If the assay was performed in the presence of Ub K48R, the ubiquitination was reduced to background levels, demonstrating that CRL5^Ozz polyubiquitinated β-catenin, MyHC_emb and Alix. (a–c) CRL5^Ozz efficiently ubiquitinated β-catenin, MyHC_emb and Alix (lane 1). The specificity of the reaction was confirmed by omitting either the substrate (lane 2) or the CRL5^Ozz complex (lane 3) from the ubiquitination assay, which significantly reduced the Ub-substrate products. (d–f) In the presence of native ubiquitin, CRL5^Ozz efficiently ubiquitinated its substrates (lane 1). Instead, by using K48R Ub mutant in the ubiquitination reaction, the ubiquitination of the substrates was significantly reduced (lane 2).