A Human Ubiquitin-conjugating Enzyme Homologous to Yeast UBC8*

Ubiquitin-conjugating enzymes catalyze the covalent attachment of ubiquitin to cellular substrates. Here we describe the isolation of a novel ubiquitin-conjugating enzyme from human placenta and the cloning of the cor- responding cDNA. DNA sequencing revealed that this gene, UbcH2, encodes a protein with significant se- quence similarity to yeast UBCS. In contrast to a previous report (&in, S., Nakajima, B., Nomura, M., and Arfin, S. M. (1991) J. Biol. Chem. 266, 15549-155541, we discov-ered that UBC8 is interrupted by a single intron bearing an unusual branch point sequence. The revised amino acid sequence of yeast UBCS exhibits 54% amino acid sequence identity to human UbcH2. Moreover, full- length UbcH2 and UBCS enzymes expressed from their cDNAs show similar enzymatic activities in vitro by catalyzing the ubiquitination of histones, suggesting that the two enzymes may fulfill similar functions in vivo. Interestingly, comparison of the enzymatic activities of a truncated UBCS

Ubiquitin-conjugating enzymes catalyze the covalent attachment of ubiquitin to cellular substrates. Here we describe the isolation of a novel ubiquitin-conjugating enzyme from human placenta and the cloning of the corresponding cDNA. DNA sequencing revealed that this gene, UbcH2, encodes a protein with significant sequence similarity to yeast UBCS. In contrast to a previous report (&in, S., Nakajima, B., Nomura, M., and Arfin, and of the full-length enzyme (this report) suggests, that the first 12 amino-terminal residues of UBCS are required for ubiquitination of histones in vitro but not for thiolester formation with ubiquitin. This suggests that the N H 2 terminus of UBCS may be necessary either for substrate recognition or for the transfer of ubiquitin onto substrates. The UbcH2 gene is located on chromosome 7 and shows a complex expression pattern with at least five different mRNAs.
A major pathway for protein degradation in eukaryotes is ubiquitin-dependent. Substrate specific ubiquitin-conjugating enzymes (E2 enzymes) and accessory factors recognize specific * 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 U.S.C. Section 1734 solely to indicate this fact.

to the GenBankTMIEMBLData Bank with accession numberls) 229328-
The nucleotide sequence(s1 reported in this paper has been submitted 229331. signals on proteolytic substrates and attach ubiquitin, a small and highly conserved protein, to defined lysine residues of substrate proteins. Ubiquitin-protein conjugates are then degraded by the 26s protease. Ubiquitin conjugation is highly selective and is required for a surprising variety of cellular functions. Genetic studies in yeast showed that ubiquitin-conjugating enzymes are required for DNA repair, sporulation, repression of retrotransposition, cell cycle progression, heat shock resistance, cadmium tolerance, and peroxisome biogen- Several in vivo substrates of the ubiquitin system have been identified, including histones (4-6), actin (71, cell surface receptors (8)(9)(10)(11), the MATa2 transcriptional repressor ( E ? ) , the tumor suppressor protein p53 (13), the Mos kinase (14), and cyclins (15). Substrate selectivity by the ubiquitin-conjugating system is thought to be mediated by the recognition of degradation signals on proteolytic substrates by ubiquitin-conjugating enzymes or accessory substrate recognition proteins known as E3s. Previous studies indicated that conjugation of ubiquitin to the model substrate histone in vitro does not depend on a n E3 protein. Studies with the yeast enzymes UBC2/RAD6 (16) and UBC3/CDC34 showed that the highly acidic carboxyl (CO0H)-terminus of the two enzymes is required for this in vitro activity (17,18).
In this report we describe the isolation and characterization of a human ubiquitin-conjugating enzyme. cDNA cloning revealed that this enzyme, UbcH2, is structurally homologous to the yeast UBCS enzyme (19). Both UbcH2 and UBC8 have COOH-terminal extensions enriched in acidic residues, and we show that both enzymes are capable of conjugating ubiquitin to histones in uitro. Interestingly, we found that amino-terminal sequences of UBC8 appear to be required for ubiquitin-histone conjugation but not for ubiquitin thiolester formation.

EXPERIMENTAL PROCEDURES
Cloning of the UBCB Gene and cDNA-The UBC8 gene was isolated from a yeast genomic DNA library in EMBL3A using a radiolabeled UBCB probe (plasmid kindly provided by S. Arflnl. The DNA sequence of a 562-base pair EcoRI-SphI fragment carrying the 5' portion of the UBCB gene was determined after subcloning into M13mp18/mp19 vectors. UBCB cDNA was synthesized with M-MuLV reverse transcriptase (New England Biolabs, Beverley, MA) under conditions recommended by the supplier after annealing the antisense primer WS53 (5'-GGCAGCTTCGTTATTCAAGGG-3') to total yeast RNA(50 pg). A UBCB cDNA clone was amplified by PCR' with primers WS52 (5"GGAATAT-TGGAAGAAAGGAGCG-3') and WS53, rendered blunt end, and cloned in the SmaI site of M13mp18 for sequencing.
Protein Purification and Peptide Sequencing-Components of the ubiquitin-conjugating system were purified from human term placenta by covalent affinity chromatography on an ubiquitin-Sepharose column, and peptide sequencing was done as described (201.
The abbreviation used is: PCR, polymerase chain reaction.

8798
The UBC8 Ubiquitin-conjugating Enzyme Family cDNA Cloning and Sequencing-Plasmids were isolated as described (21). Cloning, screening, and PCR methods were according to standard protocols (22).
We obtained the following peptide sequence by protein sequence analysis of the purified human UbcH2: FYGPEGTPYEG. The underprobe mixture TTT(C) TAT(C) GGN CCN GAA(G) GG for screening of a lined sequence was used to design the degenerated oligonucleotide human cDNA library, prepared from HeLa cells (20). A partial cDNA clone with striking similarity to the wheat germ E223k (23) was isolated.
Further screening using this cDNA as a radiolabeled probe succeeded in the isolation of clone 23W3.
To clone the 5' end of the cDNA, we used two oligonucleotides and a human cDNA prepared from HeLa cells and cloned into plasmid YEp35lGalAdh (24) to amplify a full-length cDNA by PCR. Amplification was done with oligonucleotide 1 (CACCATTGAGAGGATCTATGG) reverse complementary to the 5' end of 23W3, oligonucleotide 2 (AAACTTCTTTGCGTCCATCC) corresponding to the sequence of YEp351GalAdh, and the SalI-linearized cDNA library. The amplified 500-base pair fragment was labeled with DIG-dUTP (Boehringer Mannheim) and used to screen the cDNA library prepared in YEp351GalAdh. The DIG-dUTP-labeled probe was detected by chemiluminiscence using a n anti-DIG-Fab fragment conjugated to alkaline phosphatase and the chemiluminiscent substrate AMPPD (Boehringer Mannheim) (25). The screening procedure succeeded with the isolation of clones 23W1 and 23W2.
cDNA inserts were subcloned into the Bluescript plasmid (Stratagene, San Diego, CA), and sequences were determined by the dideoxy chain termination method (26).
The coding sequence of the yeast UBCB gene was isolated by PCR A (ATGAGTAGCTCTAAAAGAAGGATC) and B (GTCGTACGTGTGG-using genomic yeast DNA. Amplification was done with oligonucleotides TATAG). Oligonucleotide A fused the first five nucleotides (ATGAG) located upstream from the intron with the remaining coding sequence located downstream from the intron to obtain the complete undisrupted coding sequence of yeast UBC8. The PCR product was cloned inframe into the unique StuI site of the expression plasmid pMALTM-p (New England Biolabs, Beverley, M A ) to obtain UBC8IpMAL. Expression of U6cH2 a n d UBCB in E. coli-cDNA 23W1 was cloned inframe into the unique StuI site of the expression plasmid pMALTM-p. The obtained recombinant plasmid UbcH2ipMAL and yeast UBCW pMAL were transformed into Escherichia coli strain SURETM (Stratagene). Expression and purification of the recombinant fusion proteins was performed essentially as described by the manufacturer. The maltose-binding protein was cut off by factor Xa (Boehringer Mannheiml. Using a weight/weight ratio of 0.001, fusion protein cleavage was completed after 2 h at room temperature. Assay for UbcH2 and UBC8 Actiuity-100 ng of recombinant human UbcH2 or recombinant yeast UBC8, 100 ng of affinity purified human placental El, 1 pg of biotinylated ubiquitin, and in case of histone ubiquitination 2 pg of histone H2A (Boehringer Mannheim) were incubated in a total volume of 25 pl of 100 mM Tris, pH 7.5, 0.2 m M dithiothreitol, 5 mM ATP, and 5 mM MgCI, for 30 min a t 37 "C. Reaction products were separated by 12.5% polyacrylamide SDS-gel electrophoresis and transferred to a nitrocellulose membrane as described (27). The membrane was incubated in TBS (20 mM Tris, pH 7.5, 150 m M NaCI) supplemented with 1.5% (wivl commercial skim milk powder for 30 min a t room temperature, washed with TBS for 2 min, and further incubated for at least 2 h with alkaline phosphatase conjugated to streptavidin (Boehringer Mannheim) diluted 1:5000 in TBS. Unbound streptavidin was removed during incubation with TBS/milk powder for 15 min and TBS for 5 min. Detection of bound streptavidin alkaline phosphatase conjugate was performed with the substrates nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate. Thiolester linkage was confirmed by the observed instability of the product after boiling the sample prior to electrophoresis for 5 min in 4% 2-mercaptoethanol.
Northern Blot Analysis-Isolation of RNA from HeLa cells and separation in a formaldehyde, 1.2% agarose gel were performed as described (22). The cDNAinsert ofclone 23W1 was cloned into Bluescript plasmid, and DIG-dUTP-labeled antisense RNA was prepared as probe by in vitro transcription as described (28). Northern blot hybridization and chemiluminiscence detection of the hybridized probe was performed as previously described (25).
Biotin Labeling of Ubiquitin--1.5 mg of activated biotin (biotin-X-NHS, Boehringer Mannheim) was dissolved in 20 pl of dimethyl formamide and mixed with 2.6 mg of ubiquitin (Sigma) dissolved in 1 ml of phosphate-buffered saline, pH 7.5 (137 mM NaCI, 2 mM KCI, 4.6 mM Na,HP04, 1.5 mM NaH,HPO,). After 4 h of labeling at room tempera- were amplified in 30 cycles (1 mid95 "C, 1 mid60 "C, and 30 s/72 "C). 15 pl of each reaction mix were separated by agarose gel electrophoresis, and the DNA fragments were transferred onto Hybond N' membrane (Amersham, Puckinghamshire, United Kingdom). The membrane was probed with a "'P-labeled UbcH2 cDNA by standard procedures.

Cloning of the Human UbcH2
cDNA-Using ubiquitin-Sepharose affinity chromatography (201, we have purified from human placenta an ubiquitin-conjugating enzyme of an estimated size of 23 kDa. We obtained a partial sequence information of a tryptic fragment of the protein and designed a degenerate oligonucleotide for library screening. Using this oligonucleotide as a probe, we isolated from an oligo(dT1primed human cDNA library three cDNAs with identical nucleotide sequences but which differed in the length of their 3'untranslated regions. As the region of sequence identity also included the untranslated sequences, we conclude that the three cDNAs derived from the same gene. The gene was designated UbcH2 (UbcHl, encodes human E217K; (20); UbcHl is identical to HHRGB (29)). Fig. 1 shows the DNA sequence and predicted amino acid sequence derived from two of the longest clones. Northern analysis confirmed a complex transcript pattern in HeLa cells with transcripts of sizes ranging from -0.8 The to more than 5 kilobases (Fig. 2).
UbcH2 Is Structurally Homologous to East UBC8-The open reading frame of UbcH2 predicts a protein of 183 amino acids with a molecular mass of 20.6 kDa. The difference between the calculated size of UbcH2 and the size estimated by its migration in SDS gels may be caused by an abnormal running behavior in gels due to the high proline content of the protein.
Comparison of the deduced amino acid sequence of UbcH2 with sequences in current data bases revealed extensive sequence similarities to all published ubiquitin-conjugating enzymes, but the strongest similarity was with the yeast UBC8 (19) and wheat E223K (23) ubiquitin-conjugating enzymes. In addition to the core domain characteristic for all ubiquitin-conjugating enzymes, these three related proteins have COOH-terminal extensions enriched for acidic residues (Fig. 4). In contrast to the overall similarity of the three enzymes, the published predicted amino acid sequence of yeast UBC8 (19) indicated that this protein might be smaller, lacking 12 NHz-terminal residues found in both the human and the wheat enzymes. However, inspection of the published nucleotide sequence of UBC8 revealed an out-of-frame ATG codon only two bases upstream of the previously assigned initiator codon. Since translation commonly starts at the first AUG of a mRNA (30), we suspected that the initiator codon may be erroneously assigned. Isolation and sequencing of a larger genomic yeast DNA fragment confirmed the previously published DNA sequence. However, the reason for the alleged discrepancy became apparent when we isolated the UBC8 cDNA. A comparison of the DNA sequences of the genomic and the cDNA clones revealed the presence of an intron in the UBC8 coding region (Fig. 3). Whereas both 5' and 3' splice sites follow the consensus (31), the putative branch point sequence of the UBC8 intron (CACTAAC) deviates from the canonical TACTAAC sequence (31). As the corrected version of the predicted UBC8 sequence shows a good match with the NHz termini of both UbcH2 and wheat E223K (Fig. 41, we conclude that the ATG codon shown in Fig. 3 is the authentic initiator codon. After introducing some small gaps in a sequence alignment, UbcH2 shows more than 50% sequence identity to wheat E 2 2 3~ and the revised sequence of yeast UBC8. UbcH2 and UBC8 Ubiquitinate Histones in Vitro-Two ubiquitin-conjugating enzymes of yeast, UBC2/RAD6 and UBC3/ CDC34, are known to be capable of conjugating ubiquitin to histones in vitro (16,18). This reaction does not require addi-  c g t t t a a t g a a a a A G C T C T A A A A G A A G G   tional substrate recognition proteins (E3s). Both ubiquitin-conjugating enzymes have COOH-terminal extensions enriched in acidic residues that are required for histone conjugation in vitro. As UbcH2 and yeast UBC8 have similar COOH-terminal sequences, we investigated whether these two enzymes may also ubiquitinate histones. We expressed both proteins in E. coli using cDNA clones and assayed for enzyme activity.

S S K R R A T C G A G A C A G A T G T T m
Ubiquitin-conjugating activity was detected with a novel nonradioactive assay using biotinylated ubiquitin. Thiolester formation between ubiquitin and the ubiquitin-conjugating enzymes was tested in the presence of human ubiquitin-activating enzyme (20) and ATP. Using recombinant UbcH2 in this assay, we noticed the appearance of two closely spaced bands in the range of 30 kDa that were both sensitive to P-mercaptoethanol as predicted for a thiolester complex of UbcH2 and ubiquitin (Fig. 5). The reason for the double band is unclear (also known for other ubiquitin-conjugating enzymes) but may be due to an abnormal running behavior of these proteins in gels.
Addition of bovine histone H2A in these assays showed that UbcH2 can ubiquitinate histones in vitro. This reaction does not require additional factors (such as E3) but depends on the activating enzyme E l and ATP (Fig. 6A ). Low concentration of histones in the test tube led to the formation of monoubiquitinated histones whereas a t high concentration of both histone and UbcH2 low levels of multiubiquitinated histones were detected. We got very similar reactivities toward histones when we used recombinant yeast UBC8 expressed from its cDNA (Fig. 6B ). This result was somewhat surprising as in previously published assays UBC8 was shown to form a thiolester with ubiquitin but was unable to ubiquitinate histones (19). Since the enzyme used in these assays is a truncated version of UBC8 lacking the 12 NH2-terminal residues we conclude that these residues are required for ubiquitin-histone conjugation in vitro.
Chromosomal Localization of the Human UbcH2 G'ene-Twenty-seven different genomic DNA samples from hamsterhuman hybrid cells were used as templates for amplifying DNA fragments by PCR with oligonucleotides specific for the human UbcH2 cDNA. The reaction resulted in two samples producing a DNA fragment with the same size as produced from human genomic DNA. Hybridization with the human cDNA confirmed the identity of the amplified DNA fragments. The only human chromosome contained in these hybrid cell DNA samples is chromosome 7 (Table I). These data indicate that the human UbcH2 gene is localized on chromosome 7.

8800
The UBC8 Ubiquitin-conjugating Enzyme Family The arrow and the arrowhead indicate migration of Eland UbcH2 thiolester adducts, respectively.

DISCUSSION
In this paper we report the cDNA cloning and characterization of a human ubiquitin-conjugating enzyme. This enzyme, UbcH2, is structurally homologous to wheat E223k (23) and yeast uBC8 with more than 50% amino acid identity. Moreover, these three related proteins have COOH-terminal extensions enriched in acidic residues. Comparable acidic COOH-terminal tails are found in the UBC2 1 -6 (16) and UBC3JCDC34 (18) gene products where they were shown to be involved in the H2A and H2B in an E3-independent reaction, was tested in uitro by a novel nonradioactive assay. In contrast to published results (19), which were based on studies using a truncated UBC8, we demonstrated that the full-length yeast UBC8 also catalyzes the transfer of ubiquitin moieties to histones in uitro. Obviously, the first 12 NHz-terminal amino acids, missing in the recombinant UBC8 used by &in et al., are important for ubiquitination of histones in uitro. The NHz-terminal amino acids are highly homologous between the human and yeast enzymes and may be involved in substrate recognition or enzyme function. The importance of the NHz terminus of ubiquitin-conjugating enzymes was recently demonstrated for UBCZ/RADG. In this case the NHz terminus is important for RAD6 function in sporulation and DNA repair (32).
The yeast UBC8 gene contains an intron which separates the first five translated nucleotides from the remaining coding sequence by 123 spliced nucleotides. Within this noncoding sequence we localized a 5' and a 3' splice site and an unusual branchpoint sequence. There are three other yeast UBC genes containing an intron, namely UBC4, UBC5 (331, and UBCS (3).
The relative frequent occurrence of introns in this yeast gene family is suggestive of a possible regulation at the splicing level. Interestingly, the introns in UBC8 and UBCS contain non-consensus splice or branchpoint sites. The non-consensus splice signals have been implicated in the regulation of yeast genes (34, 35).
The human UbcH2 gene shows a complex expression pattern. We detected a t least five UbcH2 specific mFiNAs of different size by Northern blot analysis. The length of these transcripts are between 800 and more than 5000 nucleotides. cDNAs corresponding to the three shorter transcripts were isolated, and comparison of their primary structures revealed that they are generated from a single primary transcript by the usage of alternative polyadenylylation sites. The corresponding gene is located on chromosome 7, as shown by using hamsterhuman cell hybrids. The biological function for this complex transcription pattern might reflect post-transcriptional regulation by differing RNA stability.
Ubiquitinated histones (H2A and H2B) are the most abun-dant ubiquitin conjugates known in higher eukaryotes. Despite extensive studies in the past 15 years the function of this chromatin modification remains elusive. The capacity of the UBC8I UbcH2 enzymes to ubiquitinate histones in uitro raises the intriguing possibility that these enzymes may be involved in this process in uiuo. Support for an important cellular function of these enzymes comes from the finding that they are remarkably conserved during evolution. Surprisingly, however, the yeast ubc8 deletion mutant does not exhibit a detectable deleterious phenotype. This may indicate that functionally overlapping enzymes may exist in yeast or that this yeast enzyme has a more specialized function. Studies addressing these possibilities are currently underway.