Identification of a family of closely related human ubiquitin conjugating enzymes.

Two very closely related human E2 ubiquitin conjugating enzymes, UbfH5B and UbcH5C, have been identified. These enzymes are products of distinct genes and are 88-89% identical in amino acid sequence to the recently described human E2, UbcH5 (now designated UbcH5A), UbcH5A-C are homologous to a family of five ubiquitin conjugating enzymes from Arabidopsis thaliana, AtUBC8-12. They are also closely related to Saccharomyces cerevisiae ScUBC4 and ScUBC5, which are involved in the stress response, and play a central role in the targeting of short-lived regulatory proteins for degradation. mRNAs encoding UbcH5A-C were co-expressed in all cell lines and tissues evaluated, with UbcH5C transcripts generally expressed at the highest levels. Analysis of Southern blots suggests that there are likely to be other related members of this family. Both UbcH5B and UbcH5C form thiol ester adducts with ubiquitin, and have activities similar to UbcH5A and AtUBC8 in the conjugation of ubiquitin to target proteins in the presence of the human ubiquitin protein ligase E6-AP. These results establish the existence of a highly conserved, and widely expressed, family of human ubiquitin conjugating enzymes.

Two very closely related human E2 ubiquitin conjugating enzymes, UbcH5B and UbcH5C, have been identified. These enzymes are products of distinct genes and are 88 -89% identical in amino acid sequence to the recently described human E2, UbcH5 (now designated UbcH5A). UbcH5A-C are homologous to a family of five ubiquitin conjugating enzymes from Arabidopsis thaliana, AtUBC8 -12. They are also closely related to Saccharomyces cerevisiae ScUBC4 and ScUBC5, which are involved in the stress response, and play a central role in the targeting of short-lived regulatory proteins for degradation. mRNAs encoding UbcH5A-C were co-expressed in all cell lines and tissues evaluated, with UbcH5C transcripts generally expressed at the highest levels. Analysis of Southern blots suggests that there are likely to be other related members of this family. Both UbcH5B and UbcH5C form thiol ester adducts with ubiquitin, and have activities similar to UbcH5A and AtUBC8 in the conjugation of ubiquitin to target proteins in the presence of the human ubiquitin protein ligase E6-AP. These results establish the existence of a highly conserved, and widely expressed, family of human ubiquitin conjugating enzymes.
The modification of proteins with ubiquitin constitutes an important cellular mechanism for targeting proteins for degradation by the 26 S proteasome (reviewed in Ref. 1). Proteins known to be degraded in this fashion include abnormal polypeptides and a number of short-lived regulatory proteins including plant phytochrome A (2), c-Myc (3), c-Jun (4), cyclins (5), p53 (6), and components of the NF-B complex (7). In Saccharomyces cerevisiae, this modification has been extensively studied and found to lead to the degradation of abnormal and test proteins, with the susceptibility of some proteins to ubiquitin-mediated degradation dependent on the nature of their amino termini (reviewed in Ref. 8). Several transmembrane receptors are also ubiquitinated specifically in response to receptor engagement (9 -12), but in these cases, the relationship between ubiquitination and degradation is less clear. Ubiquitinated proteins are found at increased levels in neuropathological states including Alzheimer's disease (13).
At least three classes of enzymes are involved in the conjugation of ubiquitin to proteins. Ubiquitin activating enzyme (E1), is responsible for the ATP-dependent charging of ubiquitin through the formation of a high energy thiol ester bond between the carboxyl terminus of ubiquitin and a cysteine within E1. The thiol ester-linked ubiquitin is transferred from E1 to a cysteine residue in an E2, or ubiquitin conjugating enzyme. E2 enzymes either by themselves or in combination with E3 enzymes (ubiquitin protein ligases) then transfer ubiquitin monomers or multiubiquitin chains to target proteins, where stable isopeptide linkages are formed (reviewed in Refs. 1 and 14 -17).
E1 enzymes have been characterized in several species including yeast (18), wheat (19), and man (20). Thus far, two gene products encoding E3 enzymes have been cloned, one is from S. cerevisiae (21), and the other is a human gene product termed E6-AP, named because it associates with the human papilloma virus E6 oncoprotein (22). This E3 has been shown to catalyze the E6-dependent ubiquitination of p53 (6). E3 activities have also been characterized in rabbit reticulocytes (23) and in wheat (24).
A multitude of E2s exist. In S. cerevisiae there are at least 10 E2 genes (16), whereas in Arabidopsis thaliana over 30 are likely present (17). Functions for these enzymes include roles in DNA repair, cell cycle progression, organelle biogenesis, secretion, and stress response (reviewed in Ref. 16). There is at least one example where two ubiquitin conjugating enzymes function in concert to transfer ubiquitin to a specific target protein (25). Two closely related S. cerevisiae E2s, ScUBC4 and ScUBC5, play important roles in the turnover of normal and abnormal proteins (16,26). The levels of ScUBC4/5 are increased in response to stress, and their loss has severe effects on cellular functions; the concomitant loss of a third E2, ScUBC1, is lethal. A single homolog of ScUBC4/5 has been characterized in Caenorhabditis elegans (27) and in Drosophila melanogaster (28). In A. thaliana, a set of five closely related gene products with over 88% similarity to ScUBC4/5 and over 94% similarity to each other have been identified, with genomic evidence for at least one additional family member (29).
More recently a human gene product that is 78 -79% identical to ScUBC4/5 in amino acid sequence has been identified. This enzyme, UbcH5 (now designated UbcH5A), stimulates the conjugation of ubiquitin to the tumor suppressor p53 in the presence of E6-AP and E6 (30). At least one member of the A. thaliana family of ScUBC4/5-related ubiquitin conjugating enzymes, AtUBC8, also can serve in this role, while an unrelated A. thaliana E2 does not (30). In this study we report the characterization of two additional human members of this class of closely related and highly conserved ubiquitin conjugating enzymes.

MATERIALS AND METHODS
Generation of UBC Clones-First strand cDNA was synthesized from 0.5 g of RNA purified from human PBLs 1 using a cDNA Cycle kit (Invitrogen, San Diego, CA) with random primers and Moloney leukemia virus reverse transcriptase (LTI, Gaithersburg, MD). PCR amplification of 10% of the reaction mixture was carried out for 35 cycles using previously described degenerate primers based on conserved sequences in other E2s (27). These oligonucleotides contained built-in 5Ј BamHI and 3Ј SalI restriction sites. PCR was carried out for 35 cycles with the following parameters: 94°C for 30 s, 55°C for 90 s, and 72°C for 90 s, with a final 10-min extension at 72°C. The product was purified using QIAquick Spin PCR Purification Kit (Qiagen, Chatsworth CA), re-amplified as above, and ligated into pGEM7zϩ (Promega, Madison, WI). A human PBL cDNA library made from cells stimulated for 4 h with phytohemagglutinin and phorbol myristate acetate and cloned into Zap (Stratagene, La Jolla, CA) was a gift from Drs. K. Kelly and U. Siebenlist, National Institutes of Health. Phage DNA was prepared from 240,000 plaques by plate lysis (31). Plasmid DNA from a human natural killer cell line (YT-1) library was provided by Dr. J.-X. Lin and W. Leonard, NIH. PCR amplification from template prepared from cDNA libraries was carried out as above except that when T3 or T7 oligonucleotides were used in combination with specific primers for amplification, the T3/T7 primers were added after 5 cycles. This was done to increase the relative frequency of cDNAs encoding the proteins of interest. All specific oligonucleotides used for amplification contained built-in restriction sites to allow for cloning into pGEM7zϩ.
Characterization of UBC Expression and Genomic Analysis-Cell lines were grown at 37°C in 5% CO 2 . HeLa (32) was grown in Dulbecco's modified Eagle's-based media (Biofluids, Rockville, MD), while Jurkat (33) was grown in media containing RPMI 1640 (Biofluids) (34). Outdated human PBL were obtained from the NIH Blood Bank. Tissue RNA samples were provided by F. Collins, M. Erdos, and R. Pozzatti (NIH). For analysis of E2 expression, first strand cDNA was synthe-sized with reverse transcriptase using oligo(dT) as primer; 0.1 g of poly(A) ϩ RNA or 3 g of total RNA was used in these reactions. PCR amplification was carried out on 1 l of the cDNA product and on multiple dilutions thereof, using 50 ng of each of two primers in a final volume of 50 l. PCR conditions were as follows: 94°C for 1 min; 55°C for 2 min; 72°C for 3 min for 25 cycles. After resolution on 1.5% agarose gels and transfer to nitrocellulose, hybridization was carried out with a 32 P-end-labeled antisense oligonucleotide corresponding to bases 169 to 228 of UbcH5A-C. This probe was 81.6% identical to each of the three cDNAs, bases that are mismatched against all three cDNAs are underlined: tggatgttaaattattgttgtcaaggcaagcttcggtggtttcaatggttactctgtggg. Hybridization was carried out using standard conditions (31) at 40°C. Blots were washed at 0.2 ϫ SSC at 30°C prior to autoradiography and quantitation by PhosphorImager (Molecular Dynamics, Sunnyvale, CA), using ImageQuant software. Oligonucleotides used for amplification were: for UbcH5A (30) cgccatccctgacccatggc and tgatacagtcagagctggt; for UbcH5B caccgcatcacaccatggctctg and ggttatccaataatttgtttaattacatcgc; and for UbcH5C tgaggagccagacgacaagca and caggttattctgactttaaggtagc.
Genomic DNA from human placenta (Oncor, Gaithersburg, MD) and from Balb/c mice was exhaustively digested prior to separation on 0.8% agarose gels. Blots were transferred to nylon-backed nitrocellulose (Schleicher and Schuell) and hybridized as above with PCR-generated probes (Bios, New Haven CT) prior to a final wash at 0.2 ϫ SSC at 63°C.
Expression of Recombinant Human Ubiquitin Conjugating Enzymes-UbcH5A and UbcH5B in pGEM7zϩ were digested with NcoI and BamHI and ligated into pET-15b (Novagen, Madison, WI). UbcH5C was amplified by PCR with a 3Ј-oligonucleotide corresponding to bases 457 to 479 and a 5Ј-oligonucleotide which created an NcoI site at the 5Ј end without modifying the coding sequence. The resultant product was digested with NcoI and BamHI and ligated into pET-15b. Purified plasmids were transfected into BL21 Escherichia coli (Novagen). Induction of recombinant protein with isopropylthio-␤-galactosidase (Life Technologies, Gaithersburg, MD) was carried out per the manufacturer's instructions. Crude E. coli extracts were prepared from frozen cell pellets by adding 1 ml of sonication buffer (50 mM Tris (pH 8.0), 1 mM EDTA, 5 mM dithioerythritol, 2 mM phenylmethylsulfonyl fluoride) to pelleted bacterial cells from a 50-ml culture. After probe sonication for three 15-s bursts, separated by 30 s at 4°C, the material was clarified at 15,000 ϫ g and aliquots stored at Ϫ70°C. Protein content of the crude extracts was estimated by 12% SDS-PAGE followed by staining with Coomassie Blue.
Radiolabeling of Ubiquitin-Bovine ubiquitin (Sigma) was radiolabeled with carrier-free 125 I (Amersham) by utilization of Iodobeads (Pierce) to an initial specific radioactivity of 3.33 ϫ 10 5 cpm/ng. Unincorporated 125 I was removed via a Hi-Trap desalting column (Pharmacia Biotech Inc., Piscataway, NJ).
Assay for Ubiquitin-E2 Thiol Ester Linkage-Thiol ester assays were carried out using modifications of a previously described approach (18). For studies using anti-ubiquitin immunoblotting, equal amounts (estimated by Coomassie staining) of each E2 were combined with 1 l of E1 crude extract, and 200 ng of ubiquitin (Sigma) in 20 mM Tris (pH 7.5), 50 mM NaCl, 5 mM ATP, and 5 mM MgCl 2 in a final volume of 25 l. E1 activity was purified from a crude cell lysate derived from E. coli expressing wheat UBA1 (19). After 5 min at 22°C, samples were heated to 95°C in SDS-PAGE sample buffer in the presence or absence of ␤-mercaptoethanol prior to resolution on 12% SDS-PAGE. Gels were transferred to Immobilon-P (Millipore, Bedford, MA), incubated with affinity-purified polyclonal anti-ubiquitin followed by incubation with 125 I-protein A (ICN, Costa Mesa, CA) and autoradiography (35). For thiol ester assays using radiolabeled ubiquitin, reaction mixtures containing the various E2s (0.05 to 10 l), 1.0 g of purified recombinant wheat E1, 0.9 g of 125 I-ubiquitin (2 l), and 1 unit of inorganic pyrophosphate in 20 l of 50 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 1 mM ATP, 0.1 mM dithioerythritol were incubated for 20 s (or as indicated) at 0°C. Reactions were terminated by incubation in 25 mM Tris-HCl (pH 6.8), 5% glycerol, 4% lithium dodecyl sulfate, 4 M urea for 15 min at 30°C. Samples were subjected to SDS-PAGE and the gels were stained by Coomassie Brilliant Blue, dried between cellophane, and visualized by autoradiography. Quantification of E2-ubiquitin thiol ester adducts was accomplished by counting gel slices in a Packard Multi-Prias 1 ␥-counter, following localization by autoradiography.
Assay for Ubiquitin Conjugate Formation Using E6-AP-Expression and preparation of E6-AP was generally as described (36). Recombinant baculovirus expressing a 95-kDa truncated E6-AP was used to infect Trichoplusa Ni insect cells plated at a multiplicity of infection ϭ 10. A "mock" infection to which no baculovirus was added was carried out as a control. Infected cells were grown in Tc-100 (Life Technologies) ϩ 10% fetal bovine serum. At 36 h postinfection, expressed protein was harvested by sonication in 1 ml of 100 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1% Nonidet P-40, 1 mM dithioerythritol, 10% glycerol per 8 ϫ 10 6 cells.
Reaction mixtures containing the various E2s (0.05 to 10 l of bacterial lysate), 0.1 g of purified recombinant wheat E1, 0.5 units of phosphocreatine kinase, and 0.9 g of 125 I-ubiquitin were incubated with 0.1 l of either E6-AP extract, extract from mock infected cells, or FIG. 3. Expression of mRNA for UbcH5A-C in human cell lines. A, schematic representation of oligonucleotides used for selective amplification of UbcH5A-C. Oligo pairs are designated A, B, or C, based on their predicted ability to amplify UbcH5A-C. The position of the 60-base probe used for hybridization is indicated by the hatched area (see "Materials and Methods" for specific sequences). B, specificity of amplification. Plasmid DNA encoding UbcH5A-C was linearized and 0.2 pg subject to PCR amplification for 25 cycles with each of the three oligo pairs. Amplified fragments were resolved on 1.5% agarose gels followed by transfer to nitrocellulose membrane and hybridization with an end-labeled 60-base probe that was equally mismatched against the three members of the family (see "Materials and Methods" for details regarding hybridization and wash conditions). C, total RNA from colon DNA was reverse transcribed and various dilutions of first strand cDNA subject to 25 cycles of PCR amplification with each set of oligonucleotide pairs as described under "Materials and Methods." This was followed by hybridization with the oligonucleotide probe described above. D, quantification of the relative expression of UbcH5A-C. PCR products generated as described for C were quantified by Phos-phorImager. After subtraction of background the values obtained for UbcH5B and UbcH5C for each tissue or cell type were normalized to UbcH5A (represented by the thick line). The data shown are representative of at least two independent PCR reactions carried out at a dilution of 1:16 of first strand cDNA. In all cases multiple dilutions of first strand cDNA were amplified to assure that the results used for this analysis were within a linear range. p, UbcH5B; Ⅵ, UbcH5C. H 2 O in 20 l of 50 mM Tris-HCl (pH 7.5), 0.2 mM ATP, 0.5 mM MgCl 2 , 0.1 mM dithioerythritol, 1 mM creatine phosphate for the indicated times at 30°C. Reactions were terminated by boiling samples for 10 min in SDS-PAGE sample buffer containing 4% ␤-mercaptoethanol. Samples were subjected to SDS-PAGE and gels were stained by Coomassie Brilliant Blue, dried between cellophane, and visualized by autoradiography.

RESULTS
Prior to the initiation of this study, no mammalian homologs of the yeast ScUBC4/5 and the A. thaliana AtUBC8 -12 E2 enzymes had been identified. To determine if there are related members of this E2 class in man, oligonucleotides based on conserved regions in ScUBC5 and the related D. melanogaster DmUbcD1 were synthesized. Similar oligonucleotides had previously been used for the identification of a C. elegans E2, CeUbc-2, by PCR (27). The use of these oligonucleotides in the PCR with first strand cDNA from PBL as template resulted in a 325-base pair fragment. Cloning and sequencing of this fragment revealed a product with an open reading frame that was most homologous to the region encoding amino acids 52 to 139 of DmUbcD1 and CeUbc-2. Using DNA prepared from a human PBL library as template, PCR was carried out in which sense and antisense oligonucleotides synthesized based on the sequence of the 325-base pair fragment were paired with oligonucleotides corresponding to the T3 or T7 regions flanking the Zap polylinker. Overlapping fragments extending 5Ј and 3Ј were identified, sequenced, and found to contain parts of the original product. Oligonucleotides based on regions predicted to be the 5Ј-and 3Ј-untranslated regions were then synthesized and used to clone the entire open reading frame by PCR (Fig. 1). A search of the GenBank data base found that the deduced amino acid sequence of this gene product was 89% identical to a recently described human E2, UbcH5 (30) (now designated UbcH5A), and was thus designated UbcH5B.
Oligonucleotides generated based on the original 325-base pair PCR fragment also yielded overlapping fragments that were distinct from UbcH5B. This second sequence was truncated at its 5Ј end prior to the predicted site of translation initiation, thus it did not encode a complete E2 enzyme. However, it appeared to overlap a partial sequence found in Gen-Bank that had been designated EST06924. EST06924 had previously been identified by random sequencing of a human infant brain cDNA library (accession number: T09032) (37), and had been noted to have homology to the 5Ј ends of cDNAs encoding ubiquitin conjugating enzymes. A 5Ј-oligonucleotide based on EST06924 was synthesized and used in combination with an antisense oligonucleotide based on the 3Ј region of our sequence in the PCR, with DNA from the PBL library as template. This resulted in the generation of a full-length clone encoding an E2 related to UbcH5A and UbcH5B that we denote UbcH5C. The correct sequence of this clone was independently confirmed by PCR amplification and cloning from a cDNA library made from YT-1, a human natural killer cell line.
The cDNAs encoding UbcH5B and UbcH5C are 87% identical on a nucleotide level within the predicted coding region. In contrast, their 3Ј-untranslated regions are only 23% conserved. Both UbcH5B and UbcH5C are 78% identical to UbcH5A within their coding regions. Alignment of the deduced amino acid sequences of UbcH5B and UbcH5C shows only four amino acid differences, the only non-conservative change being amino acid 11 where there is an asparagine in UbcH5B and a serine in UbcH5C. When compared to the amino acid sequence of UbcH5A, UbcH5B is 89%, and UbcH5C is 88% identical (Fig. 2). These three human E2s are 77-80% identical to ScUBC4/5 (26) and to members of the AtUBC8 -12 family of E2s (29). Although the homology between UbcH5A and the two new family members is substantial, the amino acid sequences of UbcH5B and UbcH5C are even more closely related to the D. melanogaster DmUbcD1 (28) Conjugate assay was carried out in the presence or absence of E1 from wheat (19). Samples were heated to 95°C in the presence or absence of ␤-mercaptoethanol prior to resolution on SDS-PAGE, and immunoblotting as described (35). quence were generated and used to isolate cDNAs encoding UbcH5A from both the PBL and YT libraries. The PBL isolate was sequenced and found to be identical to the published sequence (30).
To determine the tissue distribution and relative expression of transcripts encoding UbcH5A-C, first strand cDNA was synthesized from RNA followed by PCR amplification with pairs of oligonucleotides specific to individual members of the family (Fig. 3A). The resultant products were hybridized with a 32 Plabeled 60-base oligonucleotide probe that was equally mismatched against UbcH5A-C (Fig. 3A), and therefore should hybridize equally well with PCR products encoding UbcH5A-C. To establish the specificity of the oligonucleotide pairs and the ability of the probe to hybridize with the PCR products, amplification of linearized plasmid DNA encoding each of the three E2s was carried out. As shown (Fig. 3B), the oligonucleotide pairs behaved as expected, with products detected only in samples having the appropriate combination of oligonucleotide and template. To assess the expression of UbcH5A-C in samples reverse-transcribed from human RNA, conditions were optimized to achieve a degree of amplification that was within a linear range. A typical result using 25 cycles of PCR and three different dilutions of template is shown (Fig. 3C). Results obtained using 1:16 dilutions of first strand cDNA are presented in Fig. 3D. For each source of RNA, the values obtained for UbcH5B and UbcH5C were normalized to UbcH5A. The results of this analysis demonstrate that messages encoding UbcH5A-C are co-expressed in a number of different tissue and cell lines (Fig. 3D). While UbcH5C was consistently found at higher levels than UbcH5A, the ratio of these two family members varied from 9:1 (PBL) down to 2:1 (prostate and HeLa). In most cases UbcH5B was found at levels between UbcH5A and UbcH5C. However, there was significant variation in the expression of UbcH5B relative to UbcH5A and UbcH5C. For example, in PBL, the level of UbcH5B was almost equal to UbcH5C, while in other samples, such as Jurkat and prostate, UbcH5B more closely approximates UbcH5A.
To determine if there are other closely related members of the UbcH5A-C gene family, Southern blots were carried out on genomic DNA from human placenta and from Balb/c mice. DNA was digested with either HindIII or EcoRI and the resultant blots hybridized first with a radiolabeled probe correspond-FIG. 6. Ubiquitin conjugation of UbcH5A-C. To carry out conjugation assays, volumes of lysate from E. coli expressing the various E2s being assayed were determined that resulted in equivalent thiol ester adduct formation in the presence of E1. Reaction volumes were equalized using lysate from E. coli not expressing an E2. Panel A demonstrates equivalent thiol ester adduct formation between 125 I-ubiquitin and the indicated E2s. Adducts of 125 I-ubiquitin with E2s and with E1 are indicated. Panel B illustrates the conjugation of cellular proteins with 125 I-ubiquitin in an E6-AP and E2-dependent fashion using equivalent thiol ester adduct activity for the four different E2s. After incubation as indicated for 2 h at 30°C, samples were heated to 95°C in sample buffer containing ␤-mercaptoethanol, and resolved by SDS-PAGE. E. coli transformed with empty pET-15b vector is shown as a control. Assays were carried out with lysates from Trich Ni cells that express E6-AP. Lysates from cells not expressing E6-AP are shown as controls (Mock). As an additional control, samples to which no E. coli lysate was added are shown (No E2 lanes). The species seen in the AtUBC8 lanes at ϳ25 and 32 kDa were found reproducibly, and are of unknown significance. In Panel C, conjugate formation was carried out for the indicated times. The area of the gel corresponding to molecular masses greater than 50 kDa was excised and quantified by ␥-counting. E, UbcH5B; q, UbcH5C; Ⅺ, UbcH5A; f, AtUBC8.
ing to nucleotides 12 to 213 within the coding region of UbcH5C, and then re-hybridized with a probe corresponding to nucleotides 203 to 442 of UbcH5C (Fig. 4). Since the two probes have only an 11-base overlap, fragments that hybridized with both probes should give the minimal number of genes in this family. Up to five common bands were found in man and nine in mouse. The presence of fewer bands in the human HindIII lane than in the EcoRI lane may indicate that members of this family are closely linked in the human genome, and contained within a single fragment. Alternatively, preserved HindIII sites may exist between two genes in this family, resulting in co-migrating fragments. While some of the larger hybridizing bands present in murine DNA may represent distinct genes, given their relative intensity, they may also represent partially digested DNA. Even if these higher molecular weight murine species are discounted, it would appear that mammalian genomes encode more than three members of this E2 family.
To evaluate the ability of these enzymes to function in ubiquitin conjugation, UbcH5A-C were expressed in E. coli using the pET expression system (38). The expression of these gene products in E. coli is shown in Fig. 5A. While all three cDNAs resulted in species of the predicted size (ϳ16 kDa), lower levels of UbcH5A were consistently found. A more slowly migrating major species (ϳ28 kDa) was also seen in lysates of E. coli expressing UbcH5A. A similar species of unknown significance was found in the original description of this enzyme (30). An essential characteristic of E2s is their ability to form thiol ester adducts with ubiquitin in an E1-dependent manner. All three of these proteins were able to form thiol ester adducts with ubiquitin in the presence of ATP and E1 (Fig. 5B). As expected for a thiol ester linkage, the adducts were unstable in the presence of the reducing agent ␤-mercaptoethanol.
The data presented in Fig. 5 demonstrates that these enzymes are able to form thiol ester linkages with ubiquitin. To determine whether they are able to function with an E3 enzyme in the ubiquitination of target proteins, UbcH5A-C were tested for their ability to conjugate ubiquitin to cellular proteins in an E6-AP dependent manner. They were compared to AtUBC8, an E2 previously demonstrated to function with E6-AP in a cell-free system (6,30). Using a quantitative thiol ester assay, the volume of E. coli extract that had an equal activity for each E2 was determined (Fig. 6A). When equivalent amounts of thiol ester forming activity was added to conjuga-tion assays, all four E2s demonstrated similar abilities to catalyze the transfer of ubiquitin to higher molecular weight proteins in an E6-AP-dependent fashion (Fig. 6B). Over the time period in which conjugate formation was linear (Fig. 6C), there was no substantial difference in the ability of these enzymes to catalyze the ubiquitination of cellular substrates. DISCUSSION An increasing number of cellular proteins are being recognized as substrates for conjugation to ubiquitin (reviewed in Ref. 1). In some cases, the nature of the amino terminus determines the potential susceptibility for ubiquitin-mediated degradation, a concept known as the N-end rule (8). For the mitotic cyclins and c-Jun, specific internal sequences have been identified that confer susceptibility to ubiquitination (4,5). However, in most cases, the molecular mechanisms responsible for targeting proteins for ubiquitination are unknown. Specificity with regard to the recognition of targets would seem to lie with the E2 and E3 enzymes. In yeast and in plants, a multitude of E2s exist (16,17). In humans, several distinct E2s have been characterized. Two very closely human related E2s, HHR6A and HHR6B, have been identified that are homologous to the product of the yeast DNA repair gene RAD6 (39 -41). In addition to its function in DNA repair (42), the RAD6 gene product is also believed to be involved in the recognition of substrates via the N-end rule (43,44). Another human E2, identified by complementation studies, is the human homolog of the S. cerevisiae CDC34, which is important in the G 1 to S transition in yeast (45,46). UbcH2, a human homologue of AtUBC4 -6 (47) and ScUBC8 (48), may be involved in the ubiquitination of histones in an E3-independent manner (49). These E2s are characterized by acidic COOH termini that help in the recognition of basic histones. Studies of auto-antibodies expressed in pemphigus foliaceus has resulted in the identification of a human cDNA denoted EPF-5, which, in an alternative reading frame (EPF5-ORF2), encodes an E2 enzyme (50). The significance of this reading frame remains to be determined.
It is clear from the present study that there are at least three related human E2s that are homologous to yeast ScUBC4/5 and A. thaliana AtUBC8 -12. All three of these enzymes are expressed in the same cells, and all function in an equivalent manner in a cell-free system with the human E3 enzyme E6-AP. These findings naturally raise the issue of whether these FIG. 7. Evolutionary relationship of ubiquitin conjugating enzymes. This illustrates the likely evolutionary relationship between human E2s, designated by an asterisk (*), and E2s related to UbcH5A-C from other species (only two of the five A. thaliana members of this family are shown). Rn refers to Rattus norvegicus. This analysis was generated using GeneWorks (Intelligenetics, Mountainview, CA), using a progressive alignment method. enzymes perform redundant, partially overlapping, or distinct in vivo functions. Pertinent to a discussion of this issue is one of two related studies published while this manuscript was being prepared. In this study, Wing and Jain (51) describe two rat E2s, RnE2/2E and RnE2/4A, cloned from testis, that are identical in amino acid sequence to UbcH5B and UbcH5C, respectively. There is also greater than 95% nucleic acid conservation between each pair of human and rat cDNAs, this includes both the coding regions and the 3Ј-untranslated regions. A dendrogram, based on deduced amino acid sequence, illustrating the relationship between UbcH5A-C, related E2s from other species, and other human E2s is depicted in Fig. 7. It shows the close relationship that these ScUBC4/5-related E2s have to each other, and also suggests that genes encoding UbcH5B/C and UbcH5A likely arose by duplication quite early in evolution. The total conservation of amino acid sequence between UbcH5B/C and their rat equivalents, the early divergence of UbcH5B/C from the closely related UbcH5A, together with the finding that the relative levels of transcripts encoding UbcH5A-C varies among tissues and cell types makes it reasonable to surmise that there are likely to be in vivo differences in functions among these enzymes.
Based on genomic analysis, and the precedent in A. thaliana, there are probably other related members of this family of E2 enzymes. A second, recently published relevant study reports the cloning of a human cDNA from the a human cervical carcinoma cell line, HeLa. This cDNA encodes a protein identical to UbcH5B except for the substitution of a glutamine for a lysine at position 128 due to a single base substitution (52). This lysine is conserved in all of the ScUBC4/5-related E2s (Fig. 2) and was found in four distinct cDNAs encoding UbcH5B, isolated from three sources, including two normal individuals. There are only two additional base changes in the coding regions between the HeLa-derived cDNA and UbcH5B, both of which are silent. Thus, while it is conceivable that this (128K3 Q) form of UbcH5B represents the product of a distinct gene, it more likely represents an allele of the UbcH5B gene.
The family of structurally related E2s to which UbcH5A-C belong is believed to play key roles in the catabolism of cellular proteins. Both UbcH5A and the related AtUBC8 have been shown to catalyze the E6 and E6-AP-dependent ubiquitination of p53 (30). In in vitro studies in plant, homologues of UbcH5A-C are the most active class of E2 enzymes in crude cell extracts (24). The importance of this E2 class is underscored by the critical role played by the ScUBC4/5, and their up-regulation in response to stress (16). Establishing whether there are other members of this family expressed in man, and the range of substrates and E3 enzymes with which members of this family of E2s interact, are issues that await further study.