VprBP binds full-length RAG1 and is required for B-cell development and V(D)J recombination fidelity
2011; Springer Nature; Volume: 31; Issue: 4 Linguagem: Inglês
10.1038/emboj.2011.455
ISSN1460-2075
AutoresMichele D. Kassmeier, Koushik Mondal, Victoria Palmer, Prafulla Raval, Sushil Kumar, Greg A. Perry, Dirk K. Anderson, Paweł Ciborowski, Sarah Jackson, Yue Xiong, Patrick C. Swanson,
Tópico(s)Genetics and Neurodevelopmental Disorders
ResumoArticle13 December 2011free access Source Data VprBP binds full-length RAG1 and is required for B-cell development and V(D)J recombination fidelity Michele D Kassmeier Michele D Kassmeier Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Koushik Mondal Koushik Mondal Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Victoria L Palmer Victoria L Palmer Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Prafulla Raval Prafulla Raval Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Department of Chemistry, Creighton University, Omaha, NE 68178, USA Search for more papers by this author Sushil Kumar Sushil Kumar Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, NE 68198, USA Search for more papers by this author Greg A Perry Greg A Perry Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Dirk K Anderson Dirk K Anderson Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Accuri Cytometers 173 Parkland Plaza, Ann Arbor, MI 48103, USA Search for more papers by this author Pawel Ciborowski Pawel Ciborowski Department of Pharmacology and Experimental Neuroscience, University of Nebraska Medical Center, Omaha, NE, USA Search for more papers by this author Sarah Jackson Sarah Jackson Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center, Chapel Hill, NC, USA Search for more papers by this author Yue Xiong Yue Xiong Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center, Chapel Hill, NC, USA Search for more papers by this author Patrick C Swanson Corresponding Author Patrick C Swanson Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Michele D Kassmeier Michele D Kassmeier Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Koushik Mondal Koushik Mondal Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Victoria L Palmer Victoria L Palmer Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Prafulla Raval Prafulla Raval Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Department of Chemistry, Creighton University, Omaha, NE 68178, USA Search for more papers by this author Sushil Kumar Sushil Kumar Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, NE 68198, USA Search for more papers by this author Greg A Perry Greg A Perry Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Dirk K Anderson Dirk K Anderson Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USAPresent address: Accuri Cytometers 173 Parkland Plaza, Ann Arbor, MI 48103, USA Search for more papers by this author Pawel Ciborowski Pawel Ciborowski Department of Pharmacology and Experimental Neuroscience, University of Nebraska Medical Center, Omaha, NE, USA Search for more papers by this author Sarah Jackson Sarah Jackson Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center, Chapel Hill, NC, USA Search for more papers by this author Yue Xiong Yue Xiong Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center, Chapel Hill, NC, USA Search for more papers by this author Patrick C Swanson Corresponding Author Patrick C Swanson Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA Search for more papers by this author Author Information Michele D Kassmeier1,‡, Koushik Mondal1,‡, Victoria L Palmer1,‡, Prafulla Raval1,‡, Sushil Kumar1, Greg A Perry1, Dirk K Anderson1, Pawel Ciborowski2, Sarah Jackson3, Yue Xiong3 and Patrick C Swanson 1 1Department of Medical Microbiology and Immunology, Creighton University Medical Center, Omaha, NE, USA 2Department of Pharmacology and Experimental Neuroscience, University of Nebraska Medical Center, Omaha, NE, USA 3Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center, Chapel Hill, NC, USA ‡These authors contributed equally to this work *Corresponding author. Department of Medical Microbiology and Immunology, Creighton University Medical Center, 2500 California Plaza, Omaha, NE 68178, USA. Tel.: +1 402 280 2716; Fax: +1 402 280 1875; E-mail: [email protected] The EMBO Journal (2012)31:945-958https://doi.org/10.1038/emboj.2011.455 Present address: Department of Chemistry, Creighton University, Omaha, NE 68178, USA PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The N-terminus of full-length RAG1, though dispensable for RAG1/2 cleavage activity, is required for efficient V(D)J recombination. This region supports RING E3 ubiquitin ligase activity in vitro, but whether full-length RAG1 functions as a single subunit or a multi-subunit E3 ligase in vivo is unclear. We show the multi-subunit cullin RING E3 ligase complex VprBP/DDB1/Cul4A/Roc1 associates with full-length RAG1 through VprBP. This complex is assembled into RAG protein–DNA complexes, and supports in-vitro ubiquitylation activity that is insensitive to RAG1 RING domain mutations. Conditional B lineage-specific VprBP disruption arrests B-cell development at the pro-B-to-pre-B cell transition, but this block is bypassed by expressing rearranged immunoglobulin transgenes. Mice with a conditional VprBP disruption show modest reduction of D–JH rearrangement, whereas VH–DJH and Vκ–Jκ rearrangements are severely impaired. D–JH coding joints from VprBP-insufficent mice show longer junctional nucleotide insertions and a higher mutation frequency in D and J segments than normal. These data suggest full-length RAG1 recruits a cullin RING E3 ligase complex to ubiquitylate an unknown protein(s) to limit error-prone repair during V(D)J recombination. Introduction B and T lymphocytes have the unique capacity for antigen-specific recognition, which is mediated by immunoglobulins (Igs) and T-cell receptors (TCRs), respectively. The exons encoding the antigen-binding domains of Igs and TCRs must be assembled from variable (V), joining (J) and sometimes diversity (D) gene segments to gain functionality, which is achieved through a site-specific DNA rearrangement process called V(D)J recombination. V(D)J recombination is understood to occur in two distinct phases: the ‘cleavage phase’ and the ‘joining phase’ (Fugmann et al, 2000; Gellert, 2002). In the cleavage phase, two different gene segments are brought into close proximity by two lymphoid cell-specific proteins, called RAG1 and RAG2, which bind a conserved sequence element flanking each gene segment, called the recombination signal sequence (RSS). The RAG proteins then introduce a DNA double-strand break (DSB) coordinately at the 5′ end of each RSS, liberating two blunt signal ends, and two coding ends terminating in covalently sealed DNA hairpin structures. After cleavage, the DSBs are subsequently transitioned into the joining phase, during which the DNA ends are reorganized, processed and joined by components of the non-homologous end joining (NHEJ) repair pathway (Lieber, 2010). Typically as a result, the two gene segments become joined to one another to form a ‘coding joint’ that is often imprecise, and the two RSSs become precisely joined to create a ‘signal joint’. RAG protein structure–function analysis reveals that RAG1 can support V(D)J recombination of extrachromosomal and integrated substrates even when ∼1/3rd of the protein is removed from the N-terminus (residues 1–383 of 1040 amino acids) (Sadofsky et al, 1993; Silver et al, 1993; Kirch et al, 1996). Although dispensable for the basic catalytic activity of the recombinase, the N-terminus of RAG1 is evolutionarily conserved (Sadofsky et al, 1993), and is necessary to support rearrangement of endogenous loci with high efficiency and fidelity (Dudley et al, 2003; Talukder et al, 2004). The RAG1 N-terminus was recognized to contain a C3HC4 RING finger domain (Yurchenko et al, 2003), a motif commonly found in members of a large family of E3 ubiquitin (Ub) ligases that catalyse Ub transfer in the Ub modification system (Deshaies and Joazeiro, 2009). In this system, Ub is used to modify intracellular proteins to regulate their stability or function as a mechanism to control cellular responses. This process is achieved through three steps: (i) ATP-dependent loading of Ub to an Ub-activating enzyme (E1); (ii) transfer of Ub from the E1 enzyme to a Ub-conjugating enzyme (E2) and (iii) transfer of Ub from the E2–Ub complex to a lysine residue of a substrate protein catalysed by an E3 Ub ligase. Given the presence of a RING finger domain in the RAG1 N-terminus, this region was speculated to regulate V(D)J recombination by functioning as an E3 Ub ligase to target itself or other proteins. In support of this possibility, two studies showed that an isolated RAG1 N-terminal fragment containing the RING domain supports auto-ubiquitylation (Jones and Gellert, 2003), and ubiquitylation of a test substrate in vitro (Yurchenko et al, 2003). Subsequent studies revealed that RAG1 can promote ubiquitylation of two different N-terminal RAG1-interacting proteins, KPNA1 (Simkus et al, 2009) and histone H3 (Grazini et al, 2010), and perhaps more specifically the acetylated form of histone H3.3 (Jones et al, 2011). The functional relevance of RAG1-mediated ubiquitylation of these substrates remains obscure. It is notable in this regard that both of these RAG1-interacting proteins were initially identified in assays using RAG1 only in the absence of RAG2 (Cortes et al, 1994; Grazini et al, 2010). Whether RAG1 interactions with RAG2 or other factors alter the spectrum of targets ubiquitylated by RAG1 is unclear. Moreover, since many RING-type E3 Ub ligases consist of multi-subunit assemblies containing various adaptor and substrate receptor proteins (Deshaies and Joazeiro, 2009), whether RAG1 functions physiologically as an E3 Ub ligase independent of accessory proteins, or whether such factors are required for its ubiquitylation activity towards physiological substrates remains to be clarified. Here, we show that components of a cullin E3 Ub ligase complex that includes the scaffold protein Cul4A and its associated RING finger domain protein Roc1 (also called Rbx1), the adaptor protein damaged DNA binding protein 1 (DDB1), and the substrate receptor protein Vpr binding protein (VprBP; also called DCAF1) are all co-purified with full-length RAG1, and that VprBP is required for B-cell development and high-fidelity V(D)J recombination. Results A complex containing VprBP, DDB1, Cul4A and Roc1 interacts with full-length RAG1 We previously showed that co-expressed N-terminal maltose binding protein (MBP)-tagged full-length RAG1 and core RAG2 purified under mild conditions (FLMR1/cMR2) support formation of a RAG–RSS complex containing two NHEJ factors, Ku70 and Ku80, detectable by electrophoretic mobility shift assay (EMSA) (this complex is called RAG/Ku–RSS) (Raval et al, 2008). In these experiments, Ku70/Ku80 failed to supershift a lower order RAG–RSS complex assembled with FLMR1/cMR2, suggesting an unidentified bridging molecule or cofactor mediates the RAG–Ku interaction (Raval et al, 2008). In support of this hypothesis, FLMR1/cMR2 preparations routinely showed two additional protein bands on SDS–PAGE gels, called p120 and p170, which were absent in RAG preparations containing core RAG1 and either core or full-length RAG2 (Figure 1A). Using mass spectrometry, p170 was identified as VprBP (also called DCAF1) (Figure 1B). VprBP is a member of the WD40-repeat family of proteins (Lee and Zhou, 2007), and reportedly functions as a substrate receptor in at least two different E3 Ub ligase complexes in non-lymphoid cells: the RING-type DDB1/Cul4A/Roc1 complex (Huang and Chen, 2008) and the HECT-type DDB1/DYRK2/EDD complex (Maddika and Chen, 2009). Because DDB1 is a 127-kDa protein that co-purifies with VprBP in nearly equivalent amounts (Wen et al, 2007), we speculated that p120 was DDB1. Immunoblotting confirmed the presence of VprBP, DDB1 and Cul4A and Roc1 (weakly) in preparations of FLMR1/cMR2, but not cMR1/cMR2 and cMR1/FLMR2 (Figure 1C). DYRK2 and EDD were not detected in purified FLMR1/cMR2 preparations (Supplementary Figure S1A), suggesting they are not stably integrated into an E3 Ub ligase complex containing full-length RAG1. Figure 1.VprBP and DDB1 associate with full-length RAG1 and core RAG2 and assemble a higher order RAG–RSS complex. (A) SYPRO orange-stained SDS–PAGE gels of core and full-length RAG1 and RAG2 preparations purified from HEK293 cells revealed two co-purifying proteins, p120 and p170, in the FLMR1/cMR2 preparation (arrows). Consistent with previous results (Raval et al, 2008), the expression and recovery of full-length RAG1 are slightly lower than for core RAG1. Protein standards to determine apparent molecular weight (M) were run in lane 1. Forms of maltose binding protein (MBP)-tagged RAG1 and RAG2 are designated below the gel. (B) Peptide sequences identified by mass spectroscopy of trypsin-digested p170. The position of the peptides in the VprBP sequence, the number of times they are represented in the sample, and their presence in multiple, independent samples are indicated in brackets, parentheses, and by an asterisk, respectively. (C) Immunoblots of RAG protein preparations in (A) using antibodies against MBP, VprBP, DDB1, Cul4A and Roc1, as indicated at right. A sample of whole cell lysate prepared from HEK293 cells (293 WCL) was run along with the RAG protein preparations as a positive control. By SDS–PAGE, all four full-length RAG1-associated proteins migrated to positions consistent with their predicted molecular weight (VprBP, 169 kDa; DDB1, 127 kDa; Cul4A, 80 and 82 kDa; and Roc1, 12 kDa). (D) cMR1/cMR2 or FLMR1/cMR2 was incubated with a radiolabelled 12-RSS in the absence or presence of pre-immune or anti-VprBP anti-serum (Zhang et al, 2001), and RAG–RSS complex formation visualized by EMSA. (E) cMR1/cMR2 or FLMR1/cMR2 binding to a 12-RSS was analysed by EMSA in the absence or presence of purified VprBP/DDB1 and VprBP/DDB1/Cul4A/Roc1 (expressed in insect cells; VD and VDCR, respectively) and with or without subsequent addition of anti-VprBP antibody as indicated. Figure source data can be found in Supplementary data. Download figure Download PowerPoint VprBP/DDB1 associates with FLMR1/cMR2 and Ku in a higher order protein–DNA complex Having established that VprBP, DDB1, Cul4A and Roc1 (hereafter termed the VDCR complex) co-purify with FLMR1/cMR2, we asked whether any of these components are present in the higher order RAG/Ku–RSS complex that is detected in binding reactions containing FLMR1/cMR2 but not cMR1/cMR2 (Figure 1D, compare lanes 1 and 4). We focused on VprBP because of its putative role as substrate receptor for E3 Ub ligases. We found that anti-VprBP anti-serum, but not pre-immune serum, specifically depleted the RAG/Ku–RSS complex, but not the lower order RAG–RSS complex in the same lane (Figure 1D, compare lanes 5 and 6). In addition, VprBP/DDB1 purified from insect cells with or without Cul4A and Roc1 did not supershift RAG–RSS complexes assembled with cMR1/cMR2, but stimulated formation of the higher order RAG/Ku–RSS complex in samples containing FLMR1/cMR2 (Figure 1E, compare lane 4 with lanes 5 and 6, and lane 7 with lanes 8 and 9). This result suggests VprBP/DDB1 is stably integrated into the RAG/Ku–RSS complex. Consistent with this notion, VprBP-specific antibody supershifted the RAG/Ku–RSS complex in all samples containing FLMR1/cMR2 regardless of whether Cul4A/Roc1 was present (Figure 1E, compare lanes 7–9 with 10–12). These data suggest Cul4A and Roc1 are dispensable for stable association of VprBP/DDB1 with the RAG/Ku–RSS complex. RAG1 interactions with VprBP are primarily mediated by the WD40 motif of VprBP and the far N-terminus of full-length RAG1 Because VprBP functions as a substrate receptor for other E3 Ub ligases, we suspected that VprBP directly interacts with FLMR1/cMR2. To test this possibility, we incubated purified FLMR1/cMR2 with purified full-length human VprBP (FL), an N-terminal truncation mutant lacking the C-terminal WD40 repeats and acidic tail (D1), or truncation mutants containing only the WD40 repeats (D5), or the acidic tail (D8) (Figure 2A), and then immunoprecipitated the RAG proteins and probed for VprBP by immunoblotting. We found that both full-length VprBP and the D5 fragment readily co-immunoprecipitated (co-IP) with the RAG proteins, but the D1 and D8 fragments did so poorly (D1) or not at all (D8) (Figure 2B). Interestingly, in samples containing full-length VprBP, a contaminating N-terminal VprBP fragment is recovered after IP. This fragment is present in the input and is smaller than the D1 fragment, yet D1 itself associates only weakly with FLMR1/cMR2 (Figure 2B). This outcome is most easily explained if full-length VprBP not only interacts with FLMR1/cMR2 through its WD40 motif, but also associates directly or indirectly with the truncated VprBP fragment. This interaction is most likely indirect, because the LisH motif, which reportedly mediates VprBP oligomerization (Ahn et al, 2011), should not be present in the truncated fragment based on its apparent molecular mass (Figure 2B). Alternatively, full-length VprBP may associate with the truncated VprBP fragment through DDB1, because DDB1 co-purifies with both full-length VprBP and the D1 fragment (Supplementary Figure S1A and B). Notably, these experiments rule out the possibility that the RAG proteins directly interact with DDB1, because the D1 but not D5 preparation was found to contain DDB1 (Supplementary Figure S2A and B), yet D5 readily bound to FLMR1/cMR2, but D1 did so poorly (Figure 2B). Figure 2.Full-length RAG1 primarily interacts with the WD40 motif of VprBP. (A) Diagram of full-length and truncated forms of VprBP used in this study. A summary of VprBP binding activity from (B) is shown at right. The location of a proteolytic site in full-length VprBP is indicated by an asterisk. (B) Full-length or truncated VprBP fragments in (A) were incubated with purified FLMR1/cMR2 in vitro and VprBP binding was assessed by immunoprecipitation (IP) using anti-MBP antibodies followed by immunoblotting (IB) with the indicated antibodies; 5% of the binding reaction was probed directly as a control (input). Note that a proteolytic VprBP fragment present in the input full-length VprBP preparation (indicated by an asterisk) is also recovered after IP (see text for details). (C, D) Preparations of MBP-tagged RAG1 truncation mutants co-purified with cMR2 described previously (Raval et al, 2008) were normalized for RAG1 and analysed by IB to detect RAG1 and VprBP. An individually expressed and purified form of MBP–RAG1 containing only the N-terminal region (MR11–383) was similarly analysed. Download figure Download PowerPoint To further narrow the region within the RAG1 N-terminus required for VprBP association, we probed a previously prepared panel of N-terminal MBP–RAG1 truncation mutants (Raval et al, 2008) (co-purified with cMR2 and normalized for RAG1 levels) for the presence of endogenous VprBP (Figure 2C). We found that VprBP levels are substantially reduced when the first 150 residues of RAG1 are removed; truncation of an additional 60 residues largely abolishes VprBP association (Figure 2C). Thus, VprBP primarily interacts with the first 150 residues of RAG1. We also probed an individually expressed and purified form of MBP–RAG1 containing only the N-terminal region (residues 1–383; called MR11–383) and found that this region was sufficient to mediate the association with VprBP (Figure 2C). Immunoblotting also confirmed the presence of DDB1, Cul4A and Roc1 in this preparation (unpublished observations). FLMR1/cMR2 supports RAG1 RING-independent E3 Ub ligase activity in vitro An isolated RAG1 N-terminal fragment containing the RING domain has been reported to undergo auto-ubiquitylation and mediate ubiquitylation of KPNA1 in vitro (Simkus et al, 2009). Histone H3 was subsequently identified as a putative target of RAG1-mediated ubiquitylation (Grazini et al, 2010), with the acetylated form of the histone variant H3.3 possibly a more specific target (Jones et al, 2011). In addition, both RAG2 and Ku70 are reported targets for ubiquitylation by other Ub ligases (Mizuta et al, 2002; Jiang et al, 2005; Gama et al, 2006). To determine whether one or more of these proteins are ubiquitylated by FLMR1/cMR2, we established an in-vitro ubiquitylation reaction using conditions similar to those reported by others (Jones and Gellert, 2003; Yurchenko et al, 2003; Simkus et al, 2009). Because Ub ligase activity is influenced by the choice of E2 Ub-conjugating enzyme used in the reaction, we first tested a panel of E2 enzymes, chosen based on these previous studies, for their ability to promote FLMR1/cMR2 ubiquitylation activity in vitro. We found that purified FLMR1/cMR2, but not cMR1/cMR2, supported robust E2-dependent de-novo ubiquitylation in vitro in the presence of the E2 enzymes UbcH5a and H5b, but not the other E2 enzymes tested (Figure 3A, top). When these blots were reprobed using MBP-specific antibodies (to detect the RAG proteins), no evidence of de-novo ubiquitylation was detected, suggesting that neither RAG1 nor RAG2 is efficient target of RAG1-mediated (auto)ubiquitylation under these conditions (Figure 3A, bottom). Figure 3.Purified FLMR1/cMR2 supports RING-independent ubiquitylation of an unknown target protein, but does not undergo efficient auto-ubiquitylation in vitro. (A) In-vitro ubiquitylation reactions containing wild-type (WT) FLMR1/cMR2, myc–Ub, E1 and various E2 carrier proteins were assembled as indicated, incubated as described in Materials and methods, and probed by IB using the antibodies at left. (B) In-vitro ubiquitylation reactions containing WT, mC325Y, or mC328Y FLMR1/cMR2 preparations were assembled and analysed as in (A) using UbcH5a as the E2 carrier. (C) In-vitro ubiquitylation reactions were assembled as in (A) using UbcH5b as the E2 carrier protein, except that some reactions were further supplemented with 293 cell-purified FLAG-tagged VprBP (lanes 5, 7 and 9) and/or Cul4A (lanes 6, 7 and 9). Reactions were analysed as in (A). (D, E) In-vitro ubiquitylation reactions were assembled and analysed as in (A), except that FLMR1/cMR2 was replaced by individually expressed and purified FLMR1 (D) or MR11–383 (E). (F) In-vitro ubiquitylation reactions containing wild-type FLMR1 or FLMR1/cMR2 preparations were assembled and analysed as in (A) using UbcH5a as the E2 carrier protein. Figure source data can be found in Supplementary data. Download figure Download PowerPoint To determine whether ubiquitylation was mediated directly by RAG1, we tested whether an RAG1 RING domain mutation (hC328Y) associated with human Omenn's syndrome (Villa et al, 2001) impaired Ub conjugation in our in-vitro system. Previous studies have shown that, compared with wild-type (WT) RAG1, a murine equivalent of the human mutant, mC325Y RAG1, exhibits defects in protein folding and reduced activity in assays of auto-ubiquitylation and V(D)J recombination (Simkus et al, 2007). For these studies, we prepared not only the mC325Y RAG1 mutant, but also generated an mC328Y RAG1 mutant, because this cysteine residue also participates in zinc ion coordination (Rodgers et al, 1996). Consistent with previous results, we found that both RAG1 RING domain mutants showed evidence of protein misfolding as indicated by an increase in endogenous proteolysis and an associated reduction in protein recovery after purification (Supplementary Figure S3A), as well as severely impaired V(D)J recombination activity in assays of signal and coding joint formation (Supplementary Figure S3C and D). Interestingly, both RAG1 mutants supported in-vitro ubiquitylation at levels comparable to WT FLMR1/cMR2, suggesting that full-length RAG1 does not directly mediate ubiquitylation in this system (Figure 3B), and raises the possibility that the VDCR complex instead mediates ubiquitylation. In support of this hypothesis, both purified VprBP/DDB1 and Cul4A/DDB1 supported in-vitro ubiquitylation in the presence of UbcH5b, but adding FLMR1/cMR2 slightly stimulates this activity in vitro (Figure 3C). To determine whether ubiquitylation activity in this experimental system is RAG2 dependent, or dependent on the core region of RAG1, we tested whether purified FLMR1 alone or MR11–383 mediated ubiquitylation in vitro. We found that, like FLMR1/cMR2, both FLMR1 alone and MR11–383 supported a similar pattern of E2-dependent ubiquitylation in the presence of UbcH5a and UbcH5b (Figure 3D and E), suggesting that neither RAG2 nor the core portion of RAG1 is required for ubiquitylation activity in this system. Interestingly, when FLMR1 and FLMR1/cMR2 preparations were directly compared in this assay (Figure 3F), FLMR1 showed higher levels of ubiquitylation, suggesting RAG2 may partially inhibit full-length RAG1 ubiquitylation activity in vitro. The failure of RAG1 RING domain mutations to impair in-vitro ubiquitylation activity suggests that such mutations disrupt V(D)J recombination by an alternative mechanism. Consistent with this possibility, we found that the RAG1 RING domain mutants exhibit a defect at the cleavage step of V(D)J recombination, as both the mC325Y and mC328Y FLMR1/cMR2 preparations supported lower levels of signal end break (SEB) formation in cells compared with WT FLMR1/cMR2 as detected by ligation-mediated PCR (LM-PCR; Supplementary Figure S3E and F). We attribute the poor cleavage activity of the mutants to a DNA binding defect, because neither the mC235Y or mC328Y FLMR1/cMR2 preparation supported detectable RAG–RSS complex formation by EMSA (Supplementary Figure S3B). Since KPNA1 and histone H3 have been reported to associate with and be ubiquitylated by RAG1 (Simkus et al, 2009; Grazini et al, 2010), we tested whether either protein was present in purified FLMR1/cMR2, and indeed detected both proteins by immunoblotting (Supplementary Figure S1B). However, in-vitro ubiquitylation experiments failed to provide convincing evidence that KPNA1 or histone H3 (variant H3.1, 3.2 or 3.3) undergoes full-length RAG1-dependent ubiquitylation, either in reactions supplemented with recombinant substrates (Supplementary Figure S4A–C), or in reactions containing only the endogenous co-purified proteins (unpublished observations). We also performed in-vitro ubiquitylation reactions using Ku70 and Ku80 as test substrates based on our previous studies showing that both proteins associate with full-length RAG1 (Raval et al, 2008), but these experiments yielded equivocal results as well (Supplementary Figure S4D). Since the evidence suggests that KPNA1, histone H3 and Ku70/80 are not robust ubiquitylation targets in this experimental system, we performed a time-course study to determine whether mono-ubiquitylation of a smaller protein is detectable early in the reaction. However, even at the earliest time points, only high molecular weight ubiquitylated products were detected, which accumulate over time and progress towards higher order polyubiquitylated products that fail to enter the polyacrylamide gel (Supplementary Figure S4E). Taken together, these experiments favour a model in which FLMR1 recruits the VDCR complex to target one or more proteins for ubiquitylation that are distinct from those previously shown to be ubiquitylated by RAG1. However, we cannot formally exclude the possibility that in this experimental system, a small protein in limited abundance is targeted for ubiquitylation, but cannot be readily detected until it becomes polyubiquitylated. Such a possibility would make it difficult to identify the target. Conditional disruption of VprBP expression in the B lineage blocks B-cell development at the pro-B-to-pre-B cell transition by triggering cell-cycle arrest and apoptosis If the VDCR complex plays an essential role in V(D)J recombination, disrupting expression of one or more of these components
Referência(s)