Muf1, a novel Elongin BC-interacting leucine-rich repeat protein that can assemble with Cul5 and Rbx1 to reconstitute a ubiquitin ligase.

The heterodimeric Elongin BC complex has been shown to interact in vitro and in mammalian cells with a conserved BC-box motif found in a growing number of proteins including RNA polymerase II elongation factor Elongin A, SOCS-box proteins, and the von Hippel-Lindau (VHL) tumor suppressor protein. Recently, the VHL-Elongin BC complex was found to interact with a module composed of Cullin family member Cul2 and RING-H2 finger protein Rbx1 to reconstitute a novel E3 ubiquitin ligase that activates ubiquitylation by the E2 ubiquitin-conjugating enzymes Ubc5 and Cdc34. In the context of the VHL ubiquitin ligase, Elongin BC functions as an adaptor that links the VHL protein to the Cul2/Rbx1 module, raising the possibility that the Elongin BC complex could function as an integral component of a larger family of E3 ubiquitin ligases by linking alternative BC-box proteins to Cullin/Rbx1 modules. In this report, we describe identification and purification from rat liver of a novel leucine-rich repeat-containing BC-box protein, MUF1, which we demonstrate is capable of assembling with a Cullin/Rbx1 module containing the Cullin family member Cul5 to reconstitute ubiquitin ligase activity. In addition, we show that the additional BC-box proteins Elongin A, SOCS1, and WSB1 are also capable of assembling with the Cul5/Rbx1 module to reconstitute potential ubiquitin ligases. Taken together, our findings identify MUF1 as a new member of the BC-box family of proteins, and they predict the existence of a larger family of Elongin BC-based E3 ubiquitin ligases.

The mammalian Elongin BC complex is a heterodimer composed of the 112-amino acid Elongin C protein and the 118amino acid, ubiquitin-like Elongin B protein. The Elongin BC complex was initially identified as a positive regulator of RNA polymerase II elongation factor Elongin A, which is one of several transcription factors capable of stimulating the rate of elongation by RNA polymerase II in vitro (1). The Elongin BC complex was subsequently found to be a component of the multiprotein von Hippel-Lindau (VHL) 1 tumor suppressor complex (2,3). Interaction of Elongin BC with Elongin A and VHL depends on binding of Elongin C to an ϳ10 amino acid degenerate sequence motif, referred to as the BC-box, which has the consensus sequence ((A,P,S,T)LXXXCXXX(A,I,L,V)) and which is the only sequence shared between Elongin A and VHL (2)(3)(4). Analysis of the crystal structure of the VHL-Elongin BC complex revealed that binding of Elongin BC to the BC-box is governed by interaction of the highly conserved leucine at position 2 in the N terminus of the BC-box motif with a hydrophobic pocket created by residues in the C-terminal half of Elongin C (5). Elongin B binds to a short N-terminal Elongin C region and does not appear to interact directly with the BC-box. In addition to Elongin A and VHL, Elongin BC binds to a large number of additional proteins including members of the SOCSbox protein family (6,7), each of which includes an Elongin BC-binding site linked to protein-protein interaction motifs such as SH2 domains, ankyrin repeats, WD repeats, SPRY domains, or Ras-like domains (8).
A role for the Elongin BC complex in ubiquitylation was brought to light by the discovery that the VHL tumor suppressor complex is an E3 ubiquitin ligase (9, 10) that targets the ␣ subunits of the hypoxia-inducible transcription factors HIF1 and HIF2 for ubiquitylation (11)(12)(13)(14). In the context of the VHL complex, the VHL protein functions as a substrate recognition subunit (12), which binds directly to HIF1␣ or HIF2␣ that is hydroxylated at a critical proline within the oxygen-dependent degradation domain (15,16). Elongin BC functions as an adaptor that links the VHL protein to a Cul2/Rbx1 module that can function as a potent activator of HIF␣ ubiquitylation by the E2 ubiquitin conjugating enzyme Ubc5 (12). Notably, the VHL complex bears a striking resemblance to SCF-ubiquitin ligase complexes, which contain one of a number of F-box protein substrate recognition subunits linked to a Cul1(Cdc53)/Rbx1 module by the Elongin C-like Skp1 protein (17)(18)(19)(20)(21)(22). Taken together, these observations raise the possibility that the Elongin BC complex could function as an integral component of a larger family of E3 ubiquitin ligases by linking additional BCbox containing substrate recognition subunits to Cullin/Rbx1 modules.
To address this possibility, we sought to determine whether additional BC-box proteins might also function as components of ubiquitin ligases. In a previous study, we identified and partially purified multiple, chromatographically distinct Elongin BC-containing species from rat liver (6). Here we report purification of one of these species to near homogeneity and identification of a novel leucine-rich repeat-containing BC-box protein, MUF1, which we demonstrate is capable of assembling with a Cullin/Rbx1 module containing Cullin family member Cul5 to reconstitute ubiquitin ligase activity. In addition, we show that the additional BC-box proteins Elongin A, SOCS1, and WSB1 are also capable of assembling with the same Cul5/ Rbx1 module to reconstitute potential ubiquitin ligases. Taken together, our findings identify MUF1 as a new member of the BC-box family of proteins, and they predict the existence of a larger family of Elongin BC-based ubiquitin ligases.
Purification of a MUF1-Elongin BC-containing Complex from Rat Liver-A postnuclear supernatant was prepared from the livers of ϳ350 male Harlan-Sprague Dawley rats and fractionated by precipitation with 38% (NH 4 ) 2 SO 4 as previously described (23). Protein fractions containing Elongin B were identified by Western blotting using an anti-Elongin B polyclonal antibody (6). The 0 to 38% (NH 4 ) 2 SO 4 fraction was resuspended in Buffer A (40 mM Hepes-NaOH (pH 7.9), 1 mM DTT, 0.5 mM EDTA, and 10% (v/v) glycerol) containing 0.5 mM phenylmethylsulfonyl fluoride, 10 g/ml antipain, and 10 g/ml leupeptin and brought to a conductivity equivalent to that of Buffer A containing 0.1 M KCl by a 3-h dialysis against buffer A containing 0.5 mM phenylmethylsulfonyl fluoride and dilution with the same buffer. Following centrifugation for 30 min at 28,000 ϫ g, the resulting supernatant was mixed with 0.8 liter of phosphocellulose (P11, Whatman) pre-equilibrated in Buffer A containing 0.1 M KCl and 0.5 mM phenylmethylsulfonyl fluoride. After 45 min, the slurry was filtered at 500 ml/h in a 10.5-cm diameter column and then washed at the same flow rate with Buffer A containing 0.1 M KCl. The phosphocellulose column was eluted stepwise at 800 ml/h with Buffer A containing 0.3 M KCl, and 160-ml fractions were collected. Fractions containing Elongin B were concentrated by precipitiation with 0.3 g/ml (NH 4 ) 2 SO 4 , resuspended in ϳ10 ml of Buffer A containing 10 g/ml antipain and 10 g/ml leupeptin, and dialyzed against Buffer A containing 0.3 M KCl to a conductivity equivalent to that of Buffer A containing 0.4 M KCl. Following centrifugation for 15 min at 12,000 ϫ g, the resulting supernatant was applied at 20 ml/h to a 500-ml 2.6-cm diameter Ultrogel AcA 34 gel filtration column (IBF Biotechnics) pre-equilibrated in Buffer A containing 0.4 M KCl. The AcA 34 column was eluted at 20 ml/h, and 10-ml fractions were collected. Fractions containing Elongin B, which eluted as a discrete species with an apparent native molecular mass between 150 and 250 kDa, were pooled and dialyzed against Buffer C (40 mM Tris-HCl (pH 7.9), 1 mM DTT, 0.5 mM EDTA, and 10% (v/v) glycerol) to a conductivity equivalent to that of Buffer C containing 80 mM KCl. Following centrifugation for 20 min at 60,000 ϫ g, the resulting supernatant was applied at 5 ml/min to a 21.5 ϫ 150-mm Spherogel TSK DEAE 5-PW HPLC column (Toso Haas) pre-equilibrated in Buffer C containing 80 mM KCl. The TSK DEAE 5-PW column was eluted at the same flow rate with a 250-ml linear gradient from 80 to 500 mM KCl in Buffer C, and 5-ml fractions were collected. Fractions containing Elongin B, which eluted between 100 and 140 mM KCl, were pooled and dialyzed against Buffer A to a conductivity equivalent to that of Buffer A containing 40 mM KCl. Following centrifugation for 20 min at 60,000 ϫ g, the resulting supernatant was applied at 1 ml/min to a 7.5 ϫ 75-mm TSK SP 5-PW HPLC column (Toso Haas) pre-equilibrated in Buffer A containing 40 mM KCl. The TSK SP 5-PW column was eluted at the same flow rate with a 30-ml linear gradient from 40 to 500 mM KCl in Buffer A, and 1-ml fractions were collected. Fractions containing Elongin B, which eluted between 220 and 260 mM KCl, were pooled and diluted with an equal volume of Buffer C containing 2.0 M (NH 4 ) 2 SO 4 . Following centrifugation for 20 min at 60,000 ϫ g, the resulting supernatant was applied at 0.5 ml/min to a 7.5 ϫ 75-mm TSK phenyl 5-PW HPLC column (Toso Haas) preequilibrated in Buffer C containing 1.0 M (NH 4 ) 2 SO 4 . The TSK phenyl-5-PW column was eluted at the same flow rate with a 20-ml linear gradient from 1.0 to 0 M (NH 4 ) 2 SO 4 in Buffer C, and 0.5-ml fractions were collected. Fractions containing Elongin B, which eluted between 60 and 30 mM (NH 4 ) 2 SO 4 , were pooled and dialyzed against Buffer C containing 30 mM KCl to a conductivity equivalent to that of Buffer C containing 40 mM KCl. Following centrifugation for 20 min at 60,000 ϫ g, the resulting supernatant was applied at 0.2 ml/min to a 4.6 ϫ 35-mm TSK DEAE-NPR HPLC column (Toso Haas) pre-equilibrated in Buffer C containing 40 mM KCl. The TSK DEAE-NPR column was eluted at the same flow rate with a 6-ml linear gradient from 40 to 350 mM KCl in Buffer C, and 0.2-ml fractions were collected. Fractions containing Elongin B eluted between 70 and 110 mM KCl.
Cloning of MUF1 cDNA-The MUF1-Elongin BC-containing complex was fractionated by 13% SDS-polyacrylamide gel electrophoresis. Proteins were visualized by Coomassie Blue staining, excised, and subjected to in-gel reduction, S-carboxyamidomethylation, and tryptic digestion. Peptide sequences were determined in a single run by microcapillary reversed-phase chromatography coupled to the electrospray ionization source of a quadrupole ion trap mass spectrometer (Finnigan LCQ, San Jose, CA). Identification of human and mouse expressed sequence tags that encoded MUF1 peptide sequences was facilitated by the algorithm SEQUEST and by programs developed in the Harvard Microchemistry and Proteomics Analysis Facility (24). IMAGE Consor- tium cDNA clones (accession numbers AA023471 and AA111647) encoding the mouse MUF1 ORF were obtained from Research Genetics (Huntsville, AL).
Expression of Recombinant Proteins in Sf21 Insect Cells-cDNA encoding wild type mouse MUF1 and mouse MUF1 double point mutant MUF1[L24P;C28F] containing N-terminal 6-histidine and Flag tags, mouse WSB1 containing N-terminal 6-histidine and Flag tags, and human Cul1 and Cul2 containing N-terminal HA tags were subcloned into pBacPAK 8. cDNA encoding rat Elongin A containing an N-terminal Flag tag and human Cul3, mouse Cul4, and human Cul5 containing N-terminal HA tags were subcloned into pBacPAK 9. Recombinant baculoviruses were generated with the BacPAK baculovirus expression system (CLONTECH). Baculoviruses encoding human VHL, human Elongin B, human Elongin C, and mouse Rbx1 containing an N-terminal Myc tag, and mouse SOCS1 containing N-terminal 6-histidine, T7, and Xpress tags were previously described (6,12).
Immunoprecipitations and Western Blotting-Sf21 cell lysates were incubated for 2 h at 4°C with protein A-Sepharose and the antibodies indicated in the figures. Protein A-Sepharose was washed 3 times in buffer containing 40 mM Hepes-NaOH (pH 7.9), 150 mM NaCl, 1 mM DTT, 0.5 mM EDTA, and 0.5% (v/v) Triton X-100. Immunoprecipitated proteins were analyzed by SDS-polyacrylamide gel electrophoresis and transferred to Hybond P membranes (Amersham Pharmacia Biotech) and visualized by Western blotting with Supersignal West Pico chemiluminescent reagent (Pierce).

RESULTS AND DISCUSSION
Purification and Cloning of Novel Elongin BC-box Protein MUF1-A multiprotein Elongin BC-containing complex was purified from a post-nuclear supernatant prepared from ϳ3 kg of rat liver by 0 -38% (NH 4 ) 2 SO 4 fractionation, followed by chromatography on consecutive phosphocellulose, Ultrogel AcA 34, TSK DEAE 5-PW, TSK SP 5-PW, TSK phenyl-PW, and TSK DEAE-NPR columns (Fig. 1A). Purification of the Elongin BCcontaining complex was monitored by Western blotting of aliquots of column fractions using anti-Elongin B antibodies. Analysis of the final TSK DEAE-NPR column fractions by   FIG. 2. Sequence of MUF1. The MUF1 ORF was derived from mouse expressed sequence tags (accession numbers AA023471 and AA111647). MUF1 peptide sequences determined by ion trap mass spectrometry are underlined. An asterisk (*) is placed above amino acids that fall within the LXXLXL consensus sequence for leucine-rich repeats (38). The BC-box is boxed. SDS-polyacrylamide gel electrophoresis revealed that the ϳ19-Da Elongin B protein copurified with five additional polypeptides of ϳ90, ϳ45, ϳ30, ϳ20, and ϳ15 kDa (Fig. 1B). The identities of the ϳ19and ϳ15-Da proteins as Elongin B and Elongin C were confirmed by ion trap mass spectrometry (data not shown). The ϳ45-, ϳ30-, and ϳ20-kDa polypeptides were identified by ion trap mass spectrometry as the EAP45, EAP30, and EAP20 proteins, which had previously been independently purified in association with RNA polymerase II elongation factor ELL and shown to assemble into an EAP20/30/45 subcomplex in vitro and in cells (26). Sequence analysis of the ϳ90-Da polypeptide by ion trap mass spectrometry revealed that it was a previously uncharacterized mammalian protein. Two overlapping mouse expressed sequence tags (accession numbers AA023471 and AA111647), which included all of the peptide sequences from the ϳ90-kDa polypeptide, were combined to generate a 790-amino acid ORF (Fig. 2). 5Ј-Rapid amplification of cDNA ends carried out with several cDNA libraries failed to identify any longer ORFs. Because a search of the GenBank TM nonredundant data base identified an expressed sequence tag designated MUF1 (accession no. CAA60013), which codes for a C-terminal portion of the predicted ORF from amino acids 226 to 776, we refer to the ϳ90-kDa polypeptide as MUF1. Further analysis revealed that MUF1 is a leucine-rich repeat protein (Fig. 2) that is highly conserved in mammalian species, but failed to identify potential MUF1 orthologs in lower eukaryotes.

MUF1 Binds Directly to the Elongin BC Complex through an N-terminal BC-box Motif Resembling the SOCS-box-A search of the Conserved Domain
Database with reverse position specific BLAST (27) revealed that MUF1 contains an N-terminal sequence motif resembling the SOCS-box (Fig. 3A), which was previously shown to contain a functional BC-box and to bind the Elongin BC complex (6, 7). The SOCS-box is an ϳ50-amino acid sequence motif composed of an N-terminal Elongin BC-box and a short C-terminal proline-rich region of unknown function (6 -8). The SOCS-box was originally identified as a conserved sequence motif in the C terminus of the SH2 domain-containing suppressors of cytokine signaling (SOCS) proteins, which act as negative regulators of cytokine-induced Jak/STAT signaling (28 -30). SOCS proteins appear to inhibit phosphorylation and activation of STATs by binding to and inhibiting Jak or receptor tyrosine kinases. In addition, recent studies have shown that overexpression of SOCS1 in mammalian cells can increase the rate of ubiquitin-dependent proteolysis of Vav, TEL-JAK2, and JAK2, consistent with the possibility that SOCS1 is a component of a ubiquitin ligase (31,32). Subsequently, a larger collection of SOCS-box proteins was identified and found to include nearly 30 previously uncharacterized members of the SH2, WD-40 repeat, ankyrin repeat, SPRY, and Ras-like protein families (8). The Elongin BC complex has been shown to interact in vitro and in cells with representative members of each of these families (6,7).
To determine which of the proteins in the MUF1-Elongin BC complex is capable of interacting with the Elongin BC complex, Sf21 insect cells were co-infected with baculoviruses encoding Elongins B and C and Flag-MUF1, T7-EAP20, HA-EAP30, or Myc-EAP45. As expected from the presence of a SOCS box-like motif in the N terminus of MUF1, the Elongin BC complex could be specifically co-immunoprecipitated with MUF1 from Sf21 cell lysates (Fig. 3B). The Elongin BC complex did not co-immunoprecipitate with EAP20, EAP30, or EAP45 (data not shown).
To determine whether MUF1 binds the Elongin BC complex through its N-terminal SOCS box-like motif, a MUF1 double Anti-Flag immunoprecipitations were performed as described under "Experimental Procedures." Total cell lysates and anti-Flag immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis and analyzed by Western blotting as described under "Experimental Procedures." Anti-Flag immunoprecipitates were assayed for their ability to activate ubiquitylation by Ubc5 as described under "Experimental Procedures." HC, heavy chain. point mutant MUF1[L24P;C28F] containing BC-box mutations shown previously to disrupt binding of Elongin BC to SOCS1 (6) was coexpressed in Sf21 cells with Elongins B and C. As shown in Fig. 3B, binding of MUF1 to Elongin BC was strongly dependent on the presence of an intact BC-box within the MUF1 N-terminal SOCS box-like motif; whereas wild type MUF1 was co-immunoprecipitated from Sf21 cell lysates with Elongins B and C, the MUF1 mutant was not.
The MUF1-Elongin BC Complex Is Capable of Assembling with a Cul5/Rbx1 Module to Reconstitute a Ubiquitin Ligase-In light of evidence that the Elongin BC-containing VHL tumor suppressor complex is a ubiquitin ligase (9,10,13,14) and, furthermore, that Elongin BC functions in the context of the VHL complex as an adaptor that links the VHL protein to a Cul2/Rbx1 module that can activate formation of polyubiquitin chains by the E2 ubiquitin conjugating enzymes Ubc5 and Cdc34 (12), we asked whether Elongin BC could also function as an adaptor to link MUF1 to a Cullin/Rbx1 module to reconstitute ubiquitin ligase activity. In mammalian cells the Cullin protein family includes at least 5 members referred to as Cul1, Cul2, Cul3, Cul4, and Cul5 (33). Each of these Cullin proteins has been shown to assemble with Rbx1 to reconstitute a module that is capable of activating ubiquitylation by E2 ubiquitin conjugating enzymes (20,21). To determine whether Elongin BC is capable of recruiting a Cullin/Rbx1 module to the MUF1-Elongin BC complex, anti-Flag and anti-HA immunoprecipitations were carried out on lysates of Sf21 cells co-infected with baculoviruses encoding Flag-MUF1, Elongins B and C, Rbx1, and individual Cullin proteins containing N-terminal HA epitope tags. In contrast to the VHL-Elongin BC complex, the MUF1-Elongin BC complex did not interact stably with the Cul2/Rbx1 module. In addition, the MUF1-Elongin BC complex did not interact stably with Cullin/Rbx1 modules containing Cul1, Cul3, or Cul4. The MUF1-Elongin BC complex did, however, assemble stably with a Cul5/Rbx1 module to reconstitute a multiprotein complex containing MUF1, Elongins B and C, Cul5, and Rbx1 (Fig. 4).
To determine whether the MUF1-Elongin BC-Cul5-Rbx1 complex possesses ubiquitin ligase activity, the complex was immunoaffinity purified and assayed for its ability to activate formation of polyubiquitin chains by the E2 ubiquitin conjugating enzyme Ubc5 in the presence of ATP, the E1 ubiquitin activating enzyme Uba1, and GST-ubiquitin K48R . As shown in Fig. 5A, the MUF1 complex stimulated formation of a ladder of GST-ubiquitin K48R conjugates by Ubc5. In GST-ubiquitin K48R , ubiquitin lysine 48 is mutated to arginine. As a consequence, GST-ubiquitin K48R is unable to form multi-ubiquitin chains linked through lysine 48. At the present time, we do not know whether the observed ladder of GST-ubiquitin K48R conjugates is due to formation of alternate forms of polyubiquitin linked through lysines 6, 11, 29, or 63 of ubiquitin (34) or to attachment of single GST-ubiquitin K48R molecules to multiple sites within the GST portion of the fusion protein, to subunits of the MUF1 complex, or to the E1 or E2 enzymes.
In a previous study, we observed that VHL BC-box mutants that do not bind Elongin BC or assemble with the Cul2/Rbx1 module do not activate ubiquitylation by Ubc5 (12). Likewise, the MUF1 BC-box mutant MUF1[L24P;C28F], which does not bind Elongin BC, did not efficiently assemble with the Cul5/ Rbx1 module and did not activate ubiquitylation by Ubc5 (Fig.  5B).
The Elongin BC Complex Can Link the Additional BC-box Proteins Elongin A, SOCS1, and WSB1 to the Cul5/Rbx1 Module-Our finding that Elongin BC is capable of recruiting Cullins Cul2 and Cul5 into multiprotein VHL and MUF1 complexes possessing ubiquitin ligase activity raised the possibility that Elongin BC could also function as an adaptor to link additional BC-box proteins to an E2-activating Cullin/Rbx1 module. To investigate this possibility, anti-HA immunoprecipitations were carried out on lysates of Sf21 insect cells co-infected with baculoviruses encoding the BC-box proteins VHL, Elongin A, SOCS1, or WSB1, Elongins B and C, Rbx1, and either HA-Cul2 or HA-Cul5. As shown in Fig. 6, the VHL-Elongin BC complex was capable of assembling with both the Cul2/Rbx1 and Cul5/Rbx1 modules; in contrast, the Elongin A-Elongin BC, SOCS1-Elongin BC, and WSB1-Elongin BC complexes all failed to assemble efficiently with the Cul2/Rbx1 module, but were able to assemble with the Cul5/Rbx1 module. Like the MUF1 complex, all of these complexes are capable of activating formation of GST-ubiquitin conjugates by the E2 ubiquitin-conjugating enzyme Ubc5 (data not shown).
In summary, in this report we have identified the novel Elongin BC-box protein MUF1 and demonstrated that it is capable of assembling with the Ubc5-activating Cul5/Rbx1 module to reconstitute a multiprotein complex with ubiquitin ligase activity. In addition, we have shown that the additional BC-box proteins VHL, Elongin A, SOCS1, and WSB1 are also capable of assembling with the Cul5/Rbx1 module. Cul5 is the founding member of the Cullin protein family and was originally identified as a cytoplasmic arginine-vasopressin receptor, referred to as the vasopressin-activated calcium-mobilizing-1 protein (35). Subsequently Cul5 (vasopressin-activated calcium-mobilizing-1) was shown to have a role in signal transduction and to be capable of mobilizing calcium, stimulating D-FIG. 6. The Elongin BC complex links the additional BC-box proteins Elongin A, SOCS1, and WSB1 to the Cul5/Rbx1 module. Anti-HA immunoprecipitations were performed as described under "Experimental Procedures." Total cell lysates and anti-HA immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis and analyzed by Western blotting as described under "Experimental Procedures." EloA, Elongin A. myo-inositol 1,4,5-triphosphate production, and inhibiting cAMP production in mammalian cells (36). Although it is presently not known how Cul5 carries out these functions, our finding that Cul5 is capable of interacting with BC-box proteins raises the possibility that one or more of these proteins may also have roles in Cul5-dependent signal transduction. In addition, our finding that RNA polymerase II elongation factor Elongin A is capable of assembling with the Cul5/Rbx1 module (Fig. 6, upper right panel) raises the intriguing possibility that at least one of the functions of Cul5 may be to participate together with Elongin A in ubiquitin-dependent destruction of components of the RNA polymerase II transcriptional machinery. Finally, although we presently do not understand the functional relationship between MUF1 and the EAP20, EAP30, and EAP45 components of the MUF1-Elongin BC complex, it is noteworthy that the potential S. cerevisiae orthologs of these proteins are encoded by class E vacuolar protein sorting (vps) genes, which have recently been linked to cellular ubiquitylation pathways by Hochstrasser and co-workers (37) through their genetic interactions with the S. cerevisiae Doa4 deubiquitylating enzyme.