The Elongin B Ubiquitin Homology Domain

Mammalian Elongin B is a 118-amino acid protein composed of an 84-amino acid amino-terminal ubiquitin-like domain and a 34-amino acid carboxyl-terminal tail. Elongin B is found in cells as a subunit of the heterodimeric Elongin BC complex, which was originally identified as a positive regulator of RNA polymerase II elongation factor Elongin A and subsequently as a component of the multiprotein von Hippel-Lindau tumor suppressor and suppressor of cytokine signaling complexes. As part of our effort to understand how the Elongin BC complex regulates the activity of Elongin A, we are characterizing Elongin B functional domains. In this report, we show that the Elongin B ubiquitin-like domain is necessary and sufficient for interaction with Elongin C and for positive regulation of Elongin A transcriptional activity. In addition, by site-directed mutagenesis of the Elongin B ubiquitin-like domain, we identify a short Elongin B region that is important for its interaction with Elongin C. Finally, we observe that both the ubiquitin-like domain and carboxyl-terminal tail are conserved in Drosophila melanogaster and Caenorhabditis elegans Elongin B homologs that efficiently substitute for mammalian Elongin B in reconstitution of the transcriptionally active Elongin ABC complex, suggesting that the carboxyl-terminal tail performs an additional function not detected in our assays.

Elongin B is a subunit of the heterodimeric Elongin BC complex, which was originally identified as a positive regulator of RNA polymerase II elongation factor Elongin A (1,2). Subsequently the Elongin BC complex was identified as a component of both the multiprotein von Hippel-Lindau tumor suppressor complex (3,4), where it appears to function in tumor suppression (5) and regulation of expression of hypoxia-inducible genes (6 -8), and the suppressor of cytokine signaling-1 complex, where it may function at least in part to regulate the stability of the suppressor of cytokine signaling-1 protein (9).
Previous studies from our laboratory have shown that Elon-gin B and C perform distinct functions in regulation of Elongin A transcriptional activity (10). Elongin C is capable of binding directly to a site in the Elongin A elongation activation domain and inducing Elongin A transcriptional activity in vitro in the absence of Elongin B (2,10,11). Elongin B does not directly affect Elongin A activity in the absence of Elongin C; instead, Elongin B appears to bind directly to Elongin C and to promote its interaction with Elongin A. Molecular cloning of the mammalian Elongin B gene revealed that it encodes a 118-amino acid protein composed of an ϳ84 residue NH 2 -terminal ubiquitin-like domain fused to an ϳ34 residue COOH-terminal tail (2). The ubiquitin family comprises a diverse collection of proteins that fall into at least three classes. One class is composed of ubiquitin, which is conjugated by E2/E3 ubiquitin ligases to a variety of proteins involved in signal transduction and cell cycle control and targets them for degradation by the proteosome (12,13). A second class is composed of a small set of ribosomal proteins that contain NH 2terminal ubiquitin moieties, which are proteolytically removed following incorporation of these proteins into ribosomes (14,15). A third class includes Elongin B and a growing number of other ubiquitin-like proteins. Among the members of this class are NEDD8/Rub1p, SUMO, and Rad23, which have roles in cell cycle control, nucleocytoplasmic transport, and DNA repair, respectively (16 -19).
As part of our effort to understand the structure of the Elongin BC complex and its mechanism of action in regulation of Elongin A transcriptional activity, we are carrying out a systematic structure-function analysis of each of the Elongin subunits. In this report, we demonstrate that the Elongin B ubiquitin-like domain is necessary and sufficient for interaction of Elongin B with Elongin C and for regulation of Elongin A activity. In addition, by site-directed mutagenesis of the Elongin B ubiquitin-like domain, we identify a short Elongin B region that is important for its interaction with Elongin C and that is evolutionarily conserved in Elongin B homologs from D. melanogaster and C. elegans.

EXPERIMENTAL PROCEDURES
Materials-Unlabeled ultrapure ribonucleoside 5Ј-triphosphates were purchased from Pharmacia Biotech Inc. [␣-32 P]CTP (Ͼ650 Ci/ mmol) was obtained from Amersham Corp. Ni 2ϩ -nitrilotriacetic acidagarose (Ni 2ϩ -agarose) 1 was from Invitrogen. Placental ribonuclease * This work was supported in part by National Institutes of Health Grant GM41628. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  inhibitor (RNasin) and acetylated bovine serum albumin were from Promega. Guanidine hydrochloride (sequanal grade) was purchased from Pierce. Phenylmethylsulfonyl fluoride and polyvinyl alcohol (average molecular weight 30,000 -70,000) were obtained from Sigma. Phenylmethylsulfonyl fluoride was dissolved in dimethyl sulfoxide to 1 M. Polyvinyl alcohol was dissolved in water to 20% (w/v) and centrifuged or filtered through a 0.2-m filter prior to use.
DNA Template for Transcription-pDN-AdML plasmid DNA (20) was isolated from Escherichia coli using the Triton-lysozyme method (21). Plasmid DNA was banded twice in CsCl-ethidium bromide density gradients, precipitated with ethanol, and dissolved in TE buffer (20 mM Tris-HCl (pH 7.6), and 1 mM EDTA). A restriction fragment prepared by digestion of pDN-AdML DNA with EcoRI and NdeI was used as template in transcription reactions. The fragment was purified from 1.0% low melting temperature agarose gels using GELase (Epicentre Technologies) according to the manufacturer's instructions. After phenolchloroform extraction and ethanol precipitation, purified DNA fragments were resuspended in TE buffer.
Assay of Run-off Transcription-ϳ200 ng of wild type or mutant Elongin B proteins were mixed with ϳ40 ng of wild type rat Elongin A and ϳ200 ng of rat Elongin C-(⌬41-50) and diluted ϳ2 fold to a final volume of 50 l with 40 mM Hepes-NaOH (pH 7.9), 100 mM KCl, 50 M ZnSO 4 , and 10% (v/v) glycerol. After 90 min on ice, the mixtures were dialyzed for 2 h at 4°C against 40 mM Tris-HCl (pH 7.9), 40 mM KCl, 0.1 mM EDTA, and 10% (v/v) glycerol. Elongin complexes were then assayed in 60-l reaction mixtures as follows. Preinitiation complexes were assembled by preincubation of ϳ10 ng of the EcoRI to NdeI fragment from pDN-AdML, ϳ10 ng of recombinant TFIIB, ϳ10 ng of recombinant TFIIF, ϳ7 ng of recombinant TFIIE, ϳ40 ng of rat TFIIH, ϳ20 ng of TBP, ϳ0.01 unit of RNA polymerase II, and 8 units of RNasin in 20 mM Hepes-NaOH (pH 7.9), 20 mM Tris-HCl (pH 7.9), 60 mM KCl, 2 mM DTT, 0.5 mg/ml bovine serum albumin, 2% (w/v) polyvinyl alcohol, and 3% (v/v) glycerol for 30 min at 28°C. 10 l of dialyzed Elongin complexes were then added to each reaction mixture. Transcription was initiated by addition of 7 mM MgCl 2 and 50 M ATP, 50 M GTP, 10 M CTP, 2 M UTP, and 10 Ci of [␣-32 P]CTP. After incubation of reaction mixtures for 12 min at 28°C, run-off transcripts were analyzed by electrophoresis through 6% polyacrylamide gels containing 7.0 M urea. Transcription was quantitated using a Molecular Dynamics PhosphorImager.
Construction of Elongin B Mutants-Elongin B mutants were constructed by oligonucleotide-directed mutagenesis (28) of M13mpET-Elongin B (2) with the Muta-Gene M13 in vitro mutagenesis kit (Bio-Rad) and confirmed by dideoxy DNA sequencing with the fmol DNA Sequencing System (Promega). Mutagenic oligonucleotides included 15 nucleotides from the parental rat Elongin B sequence on either side of the site of mutation.
Immunoprecipitation and Western Blotting-Anti-c-Myc epitope antibodies were from Roche Molecular Biochemicals and anti-HSV epitope antibodies were from Novagen. To immunoprecipitate Elongin B and associated proteins from 293T cell lysates, the lysates were incubated with anti-c-Myc antibody for 1 h at 4°C and then with protein A/G PLUS-agarose (Santa Cruz Biotechnology) for 1 h at 4°C. Protein A/G beads were washed two times in buffer containing 40 mM Hepes-NaOH (pH 7.9), 500 mM NaCl, 1 mM DTT, 0.5% (v/v) Triton X-100, 10% (v/v) glycerol and once in 40 mM Hepes-NaOH (pH 7.9), 150 mM NaCl, 1 mM DTT, 10% (v/v) glycerol. Immunoprecipitated proteins or proteins in total cell lysates were analyzed by electrophoresis through 13.5% SDSpolyacrylamide gels and transferred to polyvinylidine difluoride mem-branes (Millipore) and vizualized by Western blotting using the chemiluminescense reagent (NEN Life Science Products Inc.).
Expression and Purification of Elongin B Mutants and Elongin Subunits-NH 2 -terminal histidine-tagged, wild type rat Elongin A was overexpressed in E. coli using a pET16b expression vector (Novagen) and purified as described (29). Untagged wild type rat Elongin C and NH 2 -terminal histidine-tagged wild type rat Elongin C, rat Elongin C-(⌬41-50), wild type rat Elongin B, and rat Elongin B mutants were overexpressed in E. coli using the M13mpET bacteriophage expression system (27) and purified as described below. Briefly, 100-ml cultures of E. coli strain JM109(DE3) were grown at 37°C to an A 600 of 0.6 in Luria broth. Cells were infected with the appropriate M13mpET bacteriophage expression vectors at a multiplicity of infection of 10 -20. After an additional 2 h at 37°C, cells were induced with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside, and cultures were incubated an additional 3 h. Cells were harvested by centrifugation at 2000 ϫ g for 10 min at 4°C. The cell pellets were resuspended in 7 ml of 20 mM Tris-HCl (pH 8.0), 1 mg/ml lysozyme, and 10 mM imidazole (pH 8.0) and incubated for 30 min on ice. After one cycle of freeze-thaw, the suspensions were centrifuged at 100,000 ϫ g for 35 min. Inclusion bodies were solubilized by resuspension in 7 ml of ice-cold 6 M guanidine hydrochloride, 40 mM Tris-HCl (pH 8.0), 0.5 M KCl, 10 mM imidazole (pH 8.0), and 0.5 mM phenylmethylsulfonyl fluoride, and the resulting suspensions were clarified by centrifugation at 100,000 ϫ g for 35 min. All histidine-tagged proteins were further purified by Ni 2ϩ -agarose chromatography in guanidine hydrochloride as described (29).
Expression and Purification of D. melanogaster and C. elegans Elongin B Proteins-A TBLASTN search of the GenBank EST data base identified potential Elongin B homologs from D. melanogaster and C. elegans. EST yk121b3 encoding the potential C. elegans Elongin B homolog was obtained from Y. Kohara (National Institute of Genetics, Mishima, Japan). EST LD16189 encoding the potential Drosophila Elongin B homolog was obtained from Genome Systems, Inc. The entire open reading frames of the Drosophila and C. elegans proteins were overexpressed in E. coli with NH 2 -terminal histidine tags using the M13mpET bacteriophage expression system (27).
Assay of Elongin BC Complex Formation-ϳ2 g of histidine-tagged wild type rat Elongin B or rat Elongin B mutants were mixed with ϳ2 g of untagged wild type rat Elongin C and diluted 20-fold with 40 mM Hepes-NaOH (pH 7.9), 100 mM KCl, 50 M ZnSO 4 , and 10% (v/v) glycerol. After 30 min on ice, the mixtures were dialyzed for 2 h at 4°C against 40 mM Hepes-NaOH (pH 7.9), 0.1 mM EDTA, 100 mM KCl, and 10% (v/v) glycerol. Following dialysis, the mixtures were incubated for 1 h at 4°C with ϳ20 l of Ni 2ϩ -agarose pre-equilibrated in dialysis buffer containing 10 mM imidazole (pH 8.0) and then centrifuged for 20 s at 2000 rpm in a Micromax microcentrifuge (International Equipment Co.). Following centrifugation, the supernatants containing unbound proteins were collected, and the Ni 2ϩ -agarose was washed 3 times by resuspension in 500 l of 40 mM Hepes-NaOH (pH 7.9), 100 mM KCl, 0.1 mM EDTA, 10% (v/v) glycerol, and 40 mM imidazole (pH 8.0) and centrifugation for 20 s at 2000 rpm. Finally, bound material was eluted with 300 l of the same buffer containing 300 mM imidazole (pH 8.0). Aliquots of each fraction were analyzed by 13.5 or 18% SDSpolyacrylamide gel electrophoresis, and the proteins were visualized by silver staining.

RESULTS
Evolutionary Conservation of Elongin B-We previously described cloning of rat and human Elongin B cDNAs (2). Char-acterization of these cDNAs revealed that Elongin B is composed of an ϳ84 amino acid NH 2 -terminal ubiquitin-like domain fused to an ϳ34-amino acid COOH-terminal tail. The NH 2 -terminal ubiquitin-like domain can be modeled as ubiquitin (2), a compact globular structure containing a 3.5 turn ␣-helix (␣1), a short 3 10 helix, and a 5-stranded mixed ␤-sheet (30).
As an initial step in our analysis of the structure and function of Elongin B, we identified and characterized potential D. melanogaster and C. elegans Elongin B homologs. Comparison of the amino acid sequences of mammalian Elongin B and the potential Drosophila and C. elegans Elongin B homologs indicates that they are highly conserved (Fig. 1). According to the BESTFIT program of GCG, the rat and Drosophila proteins are 55% identical and 78% similar; the rat and C. elegans proteins are 39% identical and 62% similar; and the Drosophila and C. elegans proteins are 36% identical and 56% similar. A TBLASTN search of the S. cerevisiae data base revealed no S. cerevisiae open reading frames with significant sequence similarity to mammalian Elongin B, even though the S. cerevisiae genome includes an open reading frame that encodes a 100amino acid protein that exhibits significant sequence similarity to mammalian Elongin C (29,31).
To determine whether the potential Drosophila and C. elegans Elongin B homologs possess activities similar to those of mammalian Elongin B, we investigated their abilities to substitute for rat Elongin B in reconstitution of the Elongin complex. In previous studies, we showed that the transcriptionally active Elongin ABC complex could be reconstituted by recombining individual Elongin subunits purified from rat liver (1,32) or by refolding bacterially expressed Elongin subunits purified from guanidine hydrochloride-solubilized inclusion bod-ies (10). Formation of the Elongin ABC complex can be assayed by ion exchange HPLC using TSK SP-NPR (2,29). Elongin BC complexes and free Elongin B and C flow-through TSK SP-NPR at low ionic strength, whereas Elongin ABC complexes bind tightly to TSK SP-NPR and elute with ϳ0.3 M KCl.
To investigate the abilities of the potential Drosophila and C. elegans Elongin B homologs to assemble with mammalian Elongin A and C into chromatographically isolable ternary complexes, the Drosophila and C. elegans proteins were expressed in E. coli, purified from guanidine-solubilized inclusion bodies, refolded together with bacterially expressed rat Elongin A and C, and subjected to TSK SP-NPR HPLC. As shown in Fig. 2A, like rat Elongin B, both the potential Drosophila and C. elegans Elongin B homologs are capable of assembling with rat Elongin A and C to form ternary complexes that can be isolated by TSK SP-NPR HPLC; neither the Drosophila nor C. elegans proteins bound to TSK SP-NPR in the absence Elongin A (data not shown).
To determine whether the potential Drosophila and C. elegans Elongin B homologs function similarly to mammalian Elongin B in transcription, we assayed them for their abilities to promote activation of Elongin A transcriptional activity by Elongin C. We previously showed that Elongins B and C play different roles in activation of Elongin A activity (10). Elongin C functions as the inducing ligand and activates Elongin A by binding to a site in the Elongin A elongation activation domain. Although Elongin B is not essential for activation of Elongin A by Elongin C, it promotes interaction of Elongin C with Elongin A and, in so doing, increases both the yield and stability of the functional Elongin complex.
To devise an assay for Elongin transcriptional activity with the strongest possible dependence on Elongin B, we took ad- vantage of an Elongin C mutant (Elongin C-(⌬41-50)) (29) that does not bind stably to Elongin A in the absence of Elongin B (data not shown) and that does not detectably activate Elongin A transcriptional activity unless Elongin B is present (Fig. 2B).
In these experiments, Elongin complexes were assembled with rat Elongin A and Elongin C-(⌬41-50) in the absence of Elongin B or in the presence of either rat Elongin B or the potential Drosophila or C. elegans Elongin B homologs. The Elongin complexes were then assayed for their abilities to stimulate the rate of accumulation of run-off transcripts synthesized by RNA polymerase II from the AdML promoter in a purified basal transcription system reconstituted with the general initiation factors TBP, TFIIB, TFIIE, TFIIF, and TFIIH. As shown in Fig.  2B, like rat Elongin B, the potential Drosophila and C. elegans Elongin B homologs were capable of strongly promoting activation of Elongin A by Elongin C-(⌬41-50). Taken together, the results of both binding and transcription assays argue that the Drosophila and C. elegans proteins are homologs of mammalian Elongin B.
The Elongin B Ubiquitin-like Domain Is Sufficient for Elongin B Function in Vitro-To determine whether the NH 2 -terminal Elongin B ubiquitin-like domain, the COOH-terminal tail, or both are required for Elongin B function, a series of NH 2 -terminal, COOH-terminal, and internal deletion mutants of rat Elongin B were constructed (Fig. 3A), expressed in E. coli, purified from guanidine hydrochloride-solubilized inclusion bodies, and assayed for their abilities to assemble into Elongin BC and ABC complexes and to promote activation of Elongin A by Elongin C-(⌬41-50).
To assess the abilities of Elongin B deletion mutants to assemble with Elongin C into Elongin BC complexes, we assayed rat Elongin B mutants containing NH 2 -terminal histidine tags for their abilities to retain untagged wild type rat Elongin C on Ni 2ϩ -agarose. In these experiments, individual wild type Elongin B or Elongin B deletion mutants were refolded together with Elongin C and subjected to Ni 2ϩ -agarose chromatography. Unbound and bound protein fractions were collected, and equivalent amounts of each fraction were analyzed by SDS-polyacrylamide gel electrophoresis. As shown in Fig. 3B, untagged Elongin C does not bind to Ni 2ϩ -agarose, but is retained on the resin in the presence of histidine-tagged wild type Elongin B. Elongin C was quantitatively retained on Ni 2ϩagarose only by the COOH-terminal deletion mutant Elongin B-(1-84), which is the only deletion mutant with an intact ubiquitin-like domain. Elongin C was not retained on Ni 2ϩagarose by any Elongin B mutants with deletions of the NH 2terminal portion of the ubiquitin-like domain, and only very small amounts of Elongin C were retained on Ni 2ϩ -agarose by Elongin B mutants with deletions of the COOH-terminal portion of the ubiquitin-like domain. Taken together, these results argue that the Elongin B ubiquitin-like domain is necessary and sufficient for stable interaction with Elongin C in vitro.
To assess the abilities of the Elongin B deletion mutants to support formation of functional Elongin ABC complexes, we

FIG. 2. Characterization of D. melanogaster and C. elegans Elongin B homologs. Panel A, Drosophila and C. elegans
Elongin B homologs were assayed for their abilities to form Elongin ABC complexes by TSK SP-NPR HPLC as described under "Experimental Procedures." L, load; FT, flow-through. Panel B, the indicated Elongin B molecules were assayed for Elongin transcriptional activity as described under "Experimental Procedures." assayed Elongin B mutants for their abilities (i) to form Elongin ABC complexes isolable by TSK SP-NPR HPLC and (ii) to promote activation of Elongin A by Elongin C-(⌬41-50). In experiments investigating the abilities of Elongin B deletion mutants to support formation of the Elongin ABC complex, individual wild type Elongin B and Elongin B deletion mutants were refolded together with bacterially expressed wild type rat Elongins A and C and subjected to TSK SP-NPR HPLC. In experiments investigating the abilities of Elongin B deletion mutants to promote activation of Elongin A by Elongin C-(⌬41-50), bacterially expressed wild type rat Elongin A and Elongin C-(⌬41-50) were refolded together in the presence or absence of wild type rat Elongin B or Elongin B deletion mutants. Elongin complexes were then assayed for their abilities to stimulate the rate of accumulation of full-length run-off transcripts synthesized from the AdML promoter by RNA polymerase II and purified initiation factors TBP, TFIIB, TFIIE, TFIIF, and TFIIH.
As shown in Fig. 3, C and D, Elongin B-(1-84), which includes the entire ubiquitin-like domain and binds stably to Elongin C (Fig. 3B), was also capable of assembling with Elongin A and C into an isolable Elongin ABC complex and of strongly promoting activation of Elongin A transcriptional activity by Elongin C-(⌬41-50) mutant. In contrast, Elongin B deletion mutants lacking as few as 10 amino acids from the NH 2 terminus of the ubiquitin-like domain did not form Elongin ABC complexes and did not promote activation of Elongin A by Elongin C-(⌬41-50). Elongin B-(1-74), Elongin B-(1-64), and Elongin B-(⌬58 -66), which have deletions in the COOH terminus of the ubiquitin-like domain, did not promote detectable activation of Elongin A by Elongin C-(⌬41-50), but did form Elongin ABC complexes. Thus, the entire Elongin B ubiquitin-like domain is necessary and sufficient for promoting activation of Elongin A transcriptional activity by Elongin C-(⌬41-50). Elongin B residues 58 -84 of the ubiquitin-like domain are not essential for formation of Elongin ABC complexes, although deletion mutants lacking these residues are severely impaired in their abilities to form Elongin BC complexes. Elongin B-(⌬58 -66), Elongin B-(1-64), and Elongin B-(1-74) lack portions of predicted helix 2, the extended surface loop, and/or predicted ␤-sheet 5, regions of the protein that would be important for maintaining a ubiquitin-like tertiary structure (30,(33)(34)(35). Our results raise the possibility that contacts between Elongin B and the Elongin AC complex may contribute to proper folding of the Elongin B ubiquitin-like domain by compensating for loss of the predicted surface loop, portions of helix 2, or predicted ␤-sheet 5.
Mutagenesis of the Elongin B Ubiquitin-like Domain-Although ubiquitin is similar in sequence to Elongin B, ubiquitin neither binds detectably to Elongin C nor promotes activation of Elongin A transcriptional activity by Elongin C (2), suggesting that Elongin B sequences differing from those of ubiquitin are important for Elongin B function. As discussed above, the Elongin B ubiquitin-like domain can be modeled as ubiquitin. Notable features of the Elongin B model include its striking conservation of the ubiquitin hydrophobic core and conservation of the hydrophobic character of residues corresponding to ubiquitin surface residues Phe-4 and Leu-71. The Elongin B model also predicts Elongin B features not shared by ubiquitin; these include two additional hydrophobic surface residues Phe-15 and Phe-25, a prominent basic surface patch composed predominantly of residues Arg-8, Arg-9, His-10, and Lys-11, and an extended surface loop that falls between residues 62 and 70 and accommodates a 7-amino acid insertion in the ubiquitin-like domain.
To explore the relationship between Elongin B and ubiquitin and define in more detail Elongin B residues critical for its function, we investigated the activities of a collection of additional Elongin B mutants that were constructed either by replacing predicted Elongin B surface residues with those from the corresponding positions of ubiquitin, mutating the predicted Elongin B surface hydrophobic residues, or mutating Elongin B residues within the predicted basic patch. These mutants, which contained mutations spanning the entire ubiquitin-like domain, were tested for their abilities, (i) to assemble into isolable Elongin BC complexes in cells and (ii) to assemble into Elongin BC and ABC complexes in vitro and to promote activation of Elongin A by Elongin C-(⌬41-50) in vitro. To assay formation of Elongin BC complexes in cells, c-Myc-tagged Elongin B and Elongin B mutants were coexpressed with HSVtagged Elongin C in 293T cells and tested for their abilities to interact with one another by coimmunoprecipitation with an anti-c-Myc antibody. Formation of Elongin BC and ABC complexes and activation of Elongin transcription activity in vitro were assayed as described above. Results of these experiments can be summarized as follows.
(ii) Of the ubiquitin-like mutants that were capable of binding Elongin C in vivo, none assembled into Elongin BC complexes in vitro, suggesting that the mutations interfere with proper folding of Elongin B in vitro (Fig. 5A). Several of these mutants, however, were able to form Elongin ABC complexes and/or to promote activation of Elongin A by Elongin C-(⌬41- 50) (Fig. 5, A and B). The mutant Y45I/K46F/D47A/D48G/ Q49K, in which residues located at the COOH terminus of ␤-strand 3 and in the turn immediately following were mutated to the corresponding residues of ubiquitin, was able to form an Elongin ABC complex and to promote activation of Elongin A by Elongin C-(⌬41-50) as well as or better than wild type Elongin B. Another mutant, constructed by replacing Elongin B residues Lys-19 and Glu-20 with the predicted corresponding ubiquitin residues Glu-18 and Pro-19, which occupy a position in the turn between the second ␤-sheet and the first ␣-helix, failed to form isolable Elongin BC or ABC complexes in vitro, but did promote activation of Elongin A by Elongin C-(⌬41-50). A third mutant K11G, formed isolable Elongin ABC complexes in vitro, but did not promote activation of Elongin A by Elongin C-(⌬41-50).
(iii) The single Elongin B point mutations R8N, R9G, H10S, and K11N, which decrease the net positive charge in the predicted basic surface patch and which alter residues in a predicted turn between ␤-strands 1 and 2, had no detectable effect on the ability of Elongin B to form Elongin BC (Fig. 6A) and ABC complexes in vitro (Fig. 6B) or to promote activation of Elongin A transcriptional activity A by Elongin C-(⌬41-50) (Fig. 6C). Consistent with these observations, R9G could be immunoprecipitated from cell lysates with Elongin C (Fig. 4). An additional mutant, K11G, could also be immunoprecipitated from cell lysates with Elongin C (Fig. 4); however, when assayed for activity in vitro, this mutant was found to assemble into ABC complexes but not to promote activation of Elongin A by C-(⌬41-50) (Fig. 5, B and C). K11G was also unable to assemble into isolable Elongin BC complexes in vitro, suggesting that its failure to promote activation of Elongin A by C-(⌬41-50) might be due to improper folding of the protein purified from inclusion bodies.
(iv) Mutations of predicted Elongin B surface hydrophobic residues Phe-4, Phe-15, Phe-25, and Phe-62, which fall within predicted ␤-strand 1, ␤-strand 2, at the NH 2 terminus of predicted helix 1, and in predicted helix 2, respectively, had different effects on Elongin B function (Fig. 7). Elongin B point mutants F4N, F15N, and F25N could assemble into isolable Elongin BC and ABC complexes in vitro and promote activation of Elongin A by Elongin C-(⌬41-50), although Elongin B mutant F15N was less active than wild type Elongin B in promoting activation of Elongin A. Elongin B point mutant F62N, which contains a mutation at the NH 2 terminus of the predicted Elongin B surface loop not present in ubiquitin, assembled into Elongin BC and ABC complexes, but did not promote activation of Elongin A by Elongin C-(⌬41-50). Finally, F15T, which was constructed by replacing Elongin B residue Phe-15 with the predicted corresponding ubiquitin residue Thr-14, formed isolable Elongin BC complexes in cells but did not assemble into detectable Elongin BC or ABC complexes in vitro and did not promote activation of Elongin A by Elongin C-(⌬41-50), suggesting that the F15T mutation might interfere with proper folding of Elongin B in vitro.

DISCUSSION
In this report, we have investigated Elongin B sequences required for its interaction with Elongin C. Mammalian Elongin B is a 118-amino acid protein composed of an ϳ84 amino acid ubiquitin-like domain fused to an ϳ34 amino acid COOH-terminal tail (2). The Elongin B ubiquitin-like domain shares ϳ30% sequence identity with ubiquitin. Other ubiquitin-like proteins include NEDD8 and its yeast homologue Rub1p, SUMO, and Rad23, which have roles in cell cycle control, nucleocytoplasmic transport, and DNA repair, respectively (16 -19). These ubiquitin-like proteins exhibit a variable degree of similarity with ubiquitin, ranging from more than 50% sequence identity (NEDD8 and Rub1p) to less than 20% identity (SUMO). Despite the variation in their primary sequences, ubiquitin, NEDD8, Rub1p, and SUMO have very similar tertiary structures (33)(34)(35). Although a crystal structure of Elongin B has not yet been reported, Elongin B can be modeled as ubiquitin (2).
Analysis of the functions of Elongin B mutants has revealed that the Elongin B ubiquitin-like domain is sufficient for its interaction with Elongin C in vivo and in vitro and for reconstitution of the transcriptionally active Elongin complex in vitro. Furthermore, we observe that neither the predicted surface hydrophobic residues nor residues that make up the predicted basic surface patch are critical for reconstitution of functional Elongin complexes in vitro. We did, however, identify three mutations, R29A, G33D/I34K/L35E/K36G, and ⌬31-40, which fall within the ubiquitin-like domain and significantly reduce the interaction of Elongin B with Elongin C in vivo and in vitro. Notably, these mutations are confined to a short region in the COOH-terminal portion of predicted ␣-helix 1 and in the predicted turn between helix 1 and ␤-strand 3. Although we cannot rule out the possibility that these mutations affect the overall conformation of Elongin B, our results are consistent with the model that the COOH terminus of predicted ␣-helix 1 and the predicted turn between helix 1 and ␤-strand 3 may form a surface that interacts directly with Elongin C or may be important for maintaining Elongin B in a conformation that can interact with Elongin C. Two of the inactivating mutations, R29A and G33D/I34K/L35E/K36G, change Elongin B residues to the corresponding residues from ubiquitin. Sequence differences between Elongin B and ubiquitin in this region may be sufficient to account for the inability of ubiquitin to bind Elongin C, since eight additional ubiquitin-like mutations located throughout the ubiquitin-like domain had no significant effect on the interaction of Elongin B with Elongin C in cells.
Finally, by characterizing Drosophila and C. elegans Elongin B homologs that efficiently replace mammalian Elongin B in reconstitution of the transcriptionally active Elongin ABC complex, we show that both the Elongin B ubiquitin-like domain and COOH-terminal tail have been highly conserved during evolution. Although we have not yet identified a function for the Elongin B tail, it is noteworthy that it includes a PXXP motif that is a potential target for binding by SH3 domain proteins (36). Efforts to identify cellular proteins that interact with the Elongin B tail are underway.
FIG. 6. Analysis of Elongin B mutants with point mutations in the predicted basic patch. Panel A, wild type rat Elongin B and Elongin B mutants were assayed for their abilities to form Elongin BC complexes by Ni 2ϩ -agarose chromatography as described under "Experimental Procedures." Panel B, Elongin B deletion mutants were assayed for their abilities to form Elongin ABC complexes by TSK SP-NPR HPLC as described under "Experimental Procedures." L, load; FT, flow-through. Panel C, Elongin B mutants were assayed for transcriptional activity as described under "Experimental Procedures." FIG. 7. Analysis of Elongin B mutants with mutations of predicted surface hydrophobic residues. Panel A, wild type rat Elongin B and Elongin B mutants were assayed for their abilities to form Elongin BC complexes by Ni 2ϩ -agarose chromatography as described under "Experimental Procedures." Panel B, Elongin B deletion mutants were assayed for their abilities to form Elongin ABC complexes by TSK SP-NPR HPLC as described under "Experimental Procedures." L, load; FT, flow-through. Panel C, Elongin B mutants were assayed for transcriptional activity as described "Experimental Procedures."