J Biol Chem, Vol. 275, Issue 11, 7462-7465, March 17, 2000

ACCELERATED PUBLICATION
A Protein Conjugation System in Yeast with Homology to Biosynthetic Enzyme Reaction of Prokaryotes^*

Kazuaki Furukawa^§, Noboru Mizushima^¶, Takeshi Noda, and Yoshinori Ohsumi

From the  Department of Cell Biology, National Institute for Basic Biology, Okazaki 444-8585, ^¶ PRESTO, Japan Science and Technology Corporation, Okazaki 444-8585, and ^§ Suntory Institute for Medicinal Research and Development, Gunma 370-0503, Japan

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Protein conjugation, such as ubiquitination, is the process by which the C-terminal glycine of a small modifier protein is covalently attached to target protein(s) through sequential reactions with an activating enzyme and conjugating enzymes. Here we report on a novel protein conjugation system in yeast. A newly identified ubiquitin related modifier, Urm1 is a 99-amino acid protein terminated with glycine-glycine. Urm1 is conjugated to target proteins, which requires the C-terminal glycine of Urm1. At the first step of this reaction, Urm1 forms a thioester with a novel E1-like protein, Uba4. urm1 and uba4 cells showed a temperature-sensitive growth phenotype. Urm1 and Uba4 show similarity to prokaryotic proteins essential for molybdopterin and thiamin biosynthesis, although the Urm1 system is not involved in these pathways. This is the fifth conjugation system in yeast, following ubiquitin, Smt3, Rub1, and Apg12, but it is unique in respect to relation to prokaryotic enzyme systems. This fact may provide an important clue regarding evolution of protein conjugation systems in eukaryotic cells.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In recent years, it has been unveiled that protein-protein conjugation plays an indispensable role in several cellular processes in eukaryotes. The best example is the ubiquitin system (1-4). Ubiquitin is activated by the ubiquitin-activating enzyme (E1)¹ with consumption of ATP and then forms a thioester with E1 between the C-terminal glycine of ubiquitin and the active site cysteine of E1. Ubiquitin is subsequently transferred to one of the ubiquitin carrier proteins (E2). It is finally conjugated to substrate proteins via an isopeptide bond between the C-terminal carboxy group of glycine in ubiquitin and an -amino group of lysine in substrates. E3 enzymes or complexes often catalyze this final step. Ubiquitination is the tag for selective degradation by the 26 S proteasome and for endocytosis of many cell surface proteins. In the yeast, Saccharomyces cerevisiae, three other conjugation systems have been reported. Smt3, which was originally isolated as a high-copy suppressor of mif2 mutations, is covalently attached to several targets (5). Some of the targets were recently identified to be septins (6, 7). This conjugation reaction requires Uba2 and Ubc9 as E1- and E2-like enzymes, respectively (5, 8). Rub1 is also a modifier protein that is conjugated to Cdc53 in a Uba3- and Ubc12-dependent manner (9, 10). In the course of studies on Apg proteins essential for autophagy, we recently discovered the fourth protein conjugation system in yeast. The Apg12 modifier protein is covalently attached to Apg5 through a series of covalent intermediates with Apg7 and Apg10 (11-13). Amino acid sequences of Smt3 and Rub1 are similar to that of ubiquitin, whereas Apg12 is not homologous to ubiquitin. This finding led us to speculate that more protein conjugation systems exist than ever expected. Here we describe the fifth protein conjugation system in yeast, which has affinity to the molybdopterin and thiamin biosynthetic pathways in various species, including prokaryotes.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Yeast Strains and Media-- An S. cerevisiae strain used for gene disruption was KA31 (MAT ura3 leu2 his3 trp1) (14). Characterization of phenotype and immunochemical analysis were performed with urm1 (MAT ura3 leu2 his3 trp1 urm1::HIS3), uba4 (MAT ura3 leu2 his3 trp1 uba4::LEU2), and urm1uba4 (MAT ura3 leu2 his3 trp1 urm1::HIS3 uba4::LEU2) strains. For two-hybrid screening or analysis, PJ69-4A strain (MATa ura3 leu2 his3 trp1 gal4 gal80 LYS2::GAL1-HIS3 GAL2-ADE2 met2::GAL7-lacZ) was used (15). Cells were grown either in YPD medium (1% yeast extract, 2% peptone, and 2% glucose) or in SD medium containing nutritional supplements.

Plasmid Construction-- The 1.6-kb AccI-AccI genomic DNA fragment containing URM1 was amplified by PCR and was cloned into the SmaI site of pRS426. A BglII site was created immediately after the initiation codon (ATG), where 3× hemagglutinin (HA) epitopes were introduced to generate pHA-URM1. Similarly, the UBA4 gene was PCR-amplified, subcloned into pRS424, and tagged with 3× Myc epitopes at its N terminus. For gene disruption, pUC18urm1, containing the URM1 gene fragment in which the SacI-SacI fragment was replaced with HIS3, and pUC18uba4, containing the UBA4 gene in which the whole UBA4 gene was replaced with LEU2, were used.

Two-hybrid Screen-- A bait plasmid (pGBD-URM1) encoding Urm1 fused in-frame to the C-terminal end of the Gal4 DNA binding domain was constructed (15). Two-hybrid screen was performed using the system described by James et al. (15). The strain PJ69-4A was sequentially transformed with pGBD-URM1 and a mixture of yeast genomic two-hybrid libraries fused to the Gal4 activation domain (Y2HLA-C1, -C2, and -C3) (15). Transformants were selected for growth on Ade Trp Leu plates. An insert in pGAD plasmid of each positive clone was isolated by colony PCR and identified by DNA sequencing and Southern blot hybridization. Positive interaction was verified by co-transformation of pGBD-URM1 and each of the recovered prey plasmids.

Immunochemical Procedures-- Whole cell extracts were prepared by suspending cells in 0.2 M NaOH, 0.5% 2-mercaptoethanol and precipitated with acetone. The extracts were separated by SDS-polyacrylamide gel electrophoresis followed by immunoblotting using anti-HA monoclonal antibody (16B12; BAbCO). For immunoprecipitation, cell lysates were prepared by homogenizing with glass beads and were precipitated with 16B12 or anti-Myc monoclonal antibody (9E10; BAbCO) as described previously (13, 16).

Site-directed Mutagenesis-- Mutation and deletion constructs were generated by PCR-based site-directed mutagenesis and confirmed by automated DNA sequencing.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Identification of a Novel Modifier, Urm1-- In the yeast S. cerevisiae, whose whole genome sequence has been uncovered, a simple BLAST search could not identify any more ubiquitin-related modifier proteins. We then took notice of putative prototypes of ubiquitination-like protein activation systems. In Escherichia coli, the C-terminal glycine residues of small proteins, MoaD (a small subunit of molybdopterin (MPT) synthase) and ThiS, are activated in an ATP-dependent manner by E1-like enzymes, MoeB (MPT synthase sulfurylase) and ThiF, respectively (17, 18). These reactions are essential for MPT and thiamin biosynthesis, and the MoaD system is also conserved in eukaryotes except for S. cerevisiae (19-21). We performed a PSI BLAST search using the sequences of MoaD and ThiS and identified an uncharacterized open reading frame (ORF), YIL008w, in yeast. It encodes a 99-amino acid protein with a predicted relative molecular mass of 11.0 kDa. As it would function in a pathway other than those of the MPT and thiamin biosynthesis (discussed below), we named it Urm1 (ubiquitin related modifier 1). Urm1 shows 23 and 20% identity to MoaD and ThiS, respectively. Significant homology is observed in the C-terminal region of these proteins (Fig. 1A). Urm1, ThiS, and MoaD possess the C-terminal glycine-glycine motif, which is a common feature of ubiquitin and ubiquitin-related modifiers. Unlike other ubiquitin-related modifiers, Urm1 does not have any C-terminal extension after the glycine-glycine residues. Although Urm1 does not show apparent overall homology to ubiquitin and other modifiers, it shows limited homology to Smt3 at its C-terminal region. Possible counterparts of Urm1 exist in higher eukaryotes, including human. A human cDNA (AI816106) encodes a protein that is 42% identical to amino acids 10-99 of Urm1 (Fig. 1B). Thus, Urm1 fulfills a basic function in eukaryotic cells.

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Sequence analysis of Urm1. Amino acid comparison of Urm1 with E. coli modifier-like proteins (A) and a potential human counterpart (B). The sequence of Urm1 is available from GenBankTM under accession number P40554 (ORF YIL008w). A human cDNA (AI816106) encodes a protein that is 64% similar and 42% identical to amino acids 10-99 of Urm1.

A Novel Protein Activation Enzyme, Uba4-- Because Urm1 has the characteristic C-terminal domain, we postulated that Urm1 would be activated by E1-like enzymes. Using Urm1 as a bait, we performed a two-hybrid screen and obtained several positive clones containing fragments of ORF YHR111w. Strong two-hybrid interaction was actually observed between Urm1 and the full sequence of YHR111w protein. YHR111w encodes a 440-amino acid protein with a predicted molecular mass of 49.4 kDa. We named it Uba4, because the region containing residues 46-196 shows high similarity to the corresponding regions in Uba1 (ubiquitin-activating enzyme, E1) (22) (Fig. 2A) and other E1-like enzymes (5, 9, 12, 23) (not shown) including a conserved ATP binding motif (GXGXXG). Intriguingly, the overall sequence of Uba4 is closely similar to MPT synthase sulfurylase of various species including E. coli MoeB and E. coli ThiF (Fig. 2B).

View larger version (84K): [in this window] [in a new window]   Fig. 2.   A novel E1-like enzyme, Uba4. A, comparison of amino acid sequence between Uba4 and Uba1, S. cerevisiae E1 enzymes. The box I (I b) region of Uba1 (amino acids 416-574) is 47% similar and 30% identical to amino acids 46-196 of Uba4 with the conserved ATP binding motif (GXGXXG). An active-site cysteine of Uba1 is located in the box III (III) region. The similarity boxes correspond to those described previously (12). The sequence of Uba4 is available from GenBankTM under accession number AAB68852 (ORF YHR111w). B, sequence alignment of Uba4, molybdopterin synthase sulfurylase (E. coli MoeB and human Mocs3), and E. coli ThiF. Open and closed circles indicate an ATP binding motif and a metal binding motif (CXXC), respectively, which are also conserved in Uba1 and the E1-like enzymes (Uba2, Uba3, and Apg7) (not shown). The asterisk shows a putative active-site cysteine.

If Uba4 is a Urm1-activating enzyme, Urm1 and Uba4 would form a thioester on the analogy of the ubiquitin-related systems. The lysates of urm1uba4 strains expressing a 3× HA-tagged Urm1 (^HAUrm1) and/or 3× Myc-tagged Uba4 (^MycUba4) were immunoprecipitated with anti-Myc antibody, and the resulting precipitates were analyzed by Western blotting. A conjugate linked by a thioester bond can be detected as a reducing reagent-sensitive band on a Western blot. In the case of the cells expressing both ^HAUrm1 and ^MycUba4, a 76-kDa band was detected with anti-Myc antibody, in addition to a 58-kDa band corresponding to free ^MycUba4 under non-reduced condition (Fig. 3, lanes 7 and 8). The 76-kDa band was also precipitated with anti-HA antibody (not shown). The intensity of this band was decreased with increasing concentrations of a reducing reagent, DTT (Fig. 3, lanes 8-12). In contrast, an 18-kDa band corresponding to free ^HAUrm1 appeared in the precipitate after the DTT treatment (Fig. 3, lanes 8-12). These results suggest that ^HAUrm1 interacts with ^MycUba4 through a thioester bond, and Uba4 functions as a Urm1-activating enzyme.

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Thioester formation between Urm1 and Uba4. Total lysates from the urm1uba4 cells expressing Urm1 bearing the N-terminal 3× HA tag (HAUrm1) and/or Uba4 bearing the N-terminal 3× Myc epitope tag (MycUba4) were immunoprecipitated with anti-Myc monoclonal antibody (9E10). Precipitates were eluted by SDS loading buffer lacking reducing reagents (resultant eluates are controls). Aliquots of each eluted sample were reacted with the indicated concentrations of DTT through boiling for 5 min followed by SDS-polyacrylamide gel electrophoresis and Western blotting with anti-HA monoclonal antibody (16B12) or 9E10. Cont., control.

Conjugation of Urm1 to Target Proteins-- We next determined whether Urm1 is actually conjugated to some target proteins. The extract of the urm1 cells expressing ^HAUrm1 by a 2µ-based plasmid was treated with 2-mercaptoethanol and subjected to Western blotting. A major band at 37 kDa was detected in addition to the 18-kDa band that corresponds to ^HAUrm1 (Fig. 4A, WT). The 37-kDa band was not observed in the cells expressing ^HAUrm1G in which the C-terminal glycine was deleted (Fig. 4A, G). This indicates that Urm1 is covalently attached to at least one target protein, most probably through an isopeptide bond between the C-terminal glycine of Urm1 and a lysine residue in the target. As shown in Fig. 4A, there were many other faint bands in the WT, most of which were weakly observed in the G cells. However, a band at 33 kDa (arrow head) disappeared in the G cells, suggesting that the 33-kDa molecule is another conjugate or a degradation product of the 37-kDa conjugate.

View larger version (62K): [in this window] [in a new window]   Fig. 4.   Formation of Urm1 target protein conjugate(s). A, Urm1 is covalently attached to target protein(s) via the C-terminal glycine of Urm1. urm1 cells were transformed with a vector (2µ plasmid) alone, HAURM1 (WT), or HAURM1G (deletion of the C-terminal glycine; G). The extract from each transformant was treated with 2-mercaptoethanol and subjected to Western blot analysis with anti-HA monoclonal antibody (16B12). B, Cys225 of Uba4 is essential for the conjugate formation. urm1uba4 cells were co-transformed with UBA4 (WT), vector alone (uba4), UBA4C225S (C225S), or UBA4C225A (C225A) together with HAURM1. Immunoblot analysis was performed as above. The positions of free HAUrm1 or HAUrm1G (*) and a 37-kDa conjugate (**) are indicated. Arrow heads indicate another possible conjugate (33 kDa). Treatment of the samples with 100 mM DTT gave essentially the same results (not shown).

These conjugate(s) were not found in the uba4 cells (Fig. 4B, uba4), confirming that Urm1 activation by Uba4 is a prerequisite for the conjugate formation. As shown in Fig. 2B, a metal binding motif (CXXC) of Uba4 is conserved in MPT synthase sulfurylase and ThiF and also in other E1-like enzymes (Uba2, Uba3, and Apg7) (not shown). In the case of MoeB, it was shown that MoeB contains stoichiometric zinc (17). As the active-site cysteines of these E1-like enzymes in eukaryotes are 10-20 amino acid residues downstream from the metal binding motif (5, 9, 12), a Cys²²⁵ of Uba4 is the most possible active-site cysteine. Indeed, the mutants in which the Cys²²⁵ was replaced by serine or alanine were unable to catalyze the conjugate formation (Fig. 4B, C225S and C225A).

Temperature-sensitive Growth of Mutants Defective in the Urm1 System-- URM1 was not essential because the urm1 mutant was able to grow on YPD or non-fermentable carbon source media (glycerol, ethanol, and acetic acid). However, the growth was slightly retarded at 23 or 30 °C and severely impaired at 37 °C (Fig. 5A). This growth defect was suppressed by addition of an osmotic stabilizer (1 M sorbitol) in the medium (Fig. 5B). As shown in Table I, the phenotype of the uba4 strain was the same as the urm1 strain, and the urm1uba4 strain also showed the same temperature sensitivity without additive effect. Furthermore, Urm1G, Uba4C225S, and Uba4C225A mutants, which are defective in the conjugate formation (Fig. 4), showed the same phenotype. These results imply that Urm1 and Uba4 act in the same pathway, and conjugate formation is essential to fulfill a role of the system.

View larger version (83K): [in this window] [in a new window]   Fig. 5.   Temperature-sensitive growth of urm1. The wild-type (KA31) and urm1 strains were streaked on a YPD plate (A) or a YPD containing 1 M sorbitol plate (B), and incubated at 37 °C for 4 days.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Here we have shown the novel Urm1 conjugation system in yeast. This is the fifth protein conjugation system in yeast, successive to the ubiquitin, Smt3 (5, 8, 24), Rub1 (9, 10), and Apg12 (11-13, 25) systems. Although the ubiquitin system is well known, the other systems are rather new. Smt3 and Rub1 have been considered to be modifier proteins (5, 26), and some of their substrates were recently identified (6, 7, 9, 10). As the fourth yeast system, we previously found the Apg12 system in our apg mutant collection (11-13). These discoveries increasingly revealed the importance of protein conjugation systems. However, this field is still in the early stage of development.

Among the five conjugation systems in yeast, only the Urm1 system has high affinity to the prokaryotic enzyme systems. So far, the ubiquitination-like protein conjugation systems have been discovered and focused almost exclusively in eukaryotic cells. However, at least the first step, i.e. ATP-dependent protein activation by E1-like proteins, apparently has ancient prototypes such as the MoaD and ThiS activation systems in E. coli (17). Primary function of the MoaD and ThiS systems has been assigned to sulfur transfer. After activation by MoeB, MoaD receives sulfur to its C-terminal glycine and donates it to MPT precursor Z. Pitterle and Rajagopalan (27) suggested that MoaD might be covalently attached to MoaE (a large subunit of MPT synthase). However, the relationship between generation of the MoaD-MoaE complex and the ATP-dependent enzyme reaction remains unknown. The Urm1 conjugation system that has similarity to these systems may provide a missing link between ATP-dependent cofactor sulfuration and ATP-dependent protein conjugation. Furthermore, our results could suggest that MoaD may first form an adenylate and then a thioester with MoeB prior to sulfuration as previously suggested (17), and eukaryotes have probably utilized such a kind of enzyme reaction systems to create a world of the ubiquitination-type protein conjugation systems.

The exact function of the Urm1 system is still to be determined, and future identification of substrate(s) will disclose its physiological roles. We assume that the Urm1 system must function in a distinctive pathway from the MPT and thiamin biosynthetic systems by the following reasons: it was suggested that the yeast S. cerevisiae lost MPT and its precursor, although the MPT biosynthetic system is well conserved in eukaryotes (20, 21). In higher eukaryotes, MoaD homologues are distinct from Urm1 homologues (20) (Fig. 1B and data not shown). As for thiamin biosynthesis, a yeast counterpart of ThiF is assigned to Thi4 (28, 29) that is different from Uba4. Furthermore, we observed that the temperature sensitivity of the urm1 and uba4 cells was not supressed by thiamin addition (data not shown). Finally, although humans do not synthesize thiamin, Urm1 is highly conserved in human.

Our successful discovery of the new conjugation system implies that protein conjugation is more prevailing in the eukaryotic cells than ever predicted from simple sequence similarity. Actually, we have recently reported that Apg10 acts as a protein-conjugating enzyme (E2) for the Apg12 modifier, although it shows no homology to ubiquitin E2 enzymes (13). Now it is reasonable to speculate that protein conjugation provides more generalized apparatus to modify the fate and function of target proteins.

	ACKNOWLEDGEMENTS

We thank Philip James (University of Wisconsin) for providing the yeast strain, vectors, and genomic libraries for the two-hybrid screening and analysis.

	FOOTNOTES

^* This work was supported in part by grants-in-aids for scientific research from the Ministry of Education, Science and Culture of Japan and by the Joint Research Program of the Graduate University for Advanced Studies, Hayama, Japan.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Cell Biology, National Inst. for Basic Biology, 38 Nishigonaka, Myodaiji, Okazaki 444-8585, Japan. Tel.: 81-564-55-7515; Fax: 81-564-55-7516; E-mail: yohsumi@nibb.ac.jp.

	ABBREVIATIONS

The abbreviations used are: E1, ubiquitin-activating enzyme; E2, ubiquitin carrier protein; E3, ubiquitin-protein isopeptide ligase; PCR, polymerase chain reaction; HA, hemagglutinin; MPT, molybdopterin; ORF, open reading frame; DTT, dithiothreitol; ^HAUrm1, HA-tagged Urm1; ^MycUba4, Myc-tagged Uba4; WT, wild-type.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES