Multiubiquitin Chain Binding and Protein Degradation Are Mediated by Distinct Domains within the 26 S Proteasome Subunit Mcb1

,

The ubiquitin/26 S proteasome pathway is a major route for the selective degradation of eukaryotic proteins.Through the removal of key regulatory components, the pathway helps control many aspects of cell homeostasis, growth, and development (1)(2)(3)(4).Examples include cell cycle progression, maintenance of chromatin structure, DNA repair, enzymatic regulation, transcription, signal transduction, and programmed cell death.In addition, the ubiquitin pathway participates in cellular housekeeping and the stress response by removing abnormal and denatured proteins.
In the ubiquitin pathway, proteins are first enzymatically tagged for breakdown by the covalent attachment of one or more chains of ubiquitin monomers.This process is catalyzed by an enzymatic cascade, involving ubiquitin-activating enzymes (E1s), 1 ubiquitin-conjugating enzymes (E2s), and ubiquitin-protein ligases (E3s), that couples ATP hydrolysis to ubiquitin ligation (1)(2)(3).Attachment is via an isopeptide bond between the C-terminal glycine of ubiquitin and free lysines either in the target or in the preceding ubiquitin in the chain.Within the multiubiquitin chain, Lys 48 appears to be the preferred intermolecular linkage site (5,6) but genetic evidence has implicated several other lysines as well (e.g.Lys 29 and Lys 63 ) (7)(8)(9).Once assembled, the multiubiquitin chain functions as a recognition signal for degradation of the substrate by the 26 S proteasome, a multisubunit complex specific for multiubiquitinated proteins (10).The tagged proteins are broken down into short peptides, while the ubiquitin moieties are released intact for reuse.
Specificity within the ubiquitin pathway is achieved by at least three mechanisms.The most important determines which proteins should be ubiquitinated by the E1/E2/E3 cascade of reactions.Here, substrate specificity is primarily regulated by a large family of E2 and E3 isozymes working alone or in combination to recognize key degradation signals within the targets (1)(2)(3)11).The second regulatory mechanism involves control of the steady-state level of ubiquitin-protein conjugates prior to breakdown.Conjugate levels are affected not only by the rate of ubiquitination but also by the rate of deubiquitination; a reaction catalyzed by a diverse family of ubiquitinspecific proteases that cleaves the junction between ubiquitin and the protein moiety (2).Through such deubiquitination, proteins can be rescued from degradation (e.g.see Refs.[12][13][14]. The third mechanism to achieve specificity is the association and breakdown of ubiquitin-protein conjugates by the 26 S proteasome.The 26 S proteasome is composed of two subcomplexes, designated the 19 S regulatory complex and the 20 S proteasome (4,10,15,16).The 20 S proteasome contains the catalytic core of the protease; it exists as a hollow cylinder, created by the assembly of four stacked polypeptide rings, and confines the protease active sites within the lumen (17,18).The 19 S regulatory complex binds to one or both ends of the 20 S particle.It contains ϳ15 subunits and confers both ATP and ubiquitin dependence to the holoenzyme complex.The function(s) of most of these subunits are unknown; six belong to the AAA family of ATPases (19,20).Presumably, the 19 S complex recognizes appropriate targets through the multiubiquitin chain, unfolds the target moiety, and directs the unfolded polypeptide into the lumen of the 20 S complex for breakdown (4,10).During or after this process, the multiubiquitin chain is disassembled by ubiquitin-specific proteases.
Recognition of ubiquitinated substrates by the 26 S proteasome is proposed to be mediated by one or a few common multiubiquitin chain receptors located in the 19 S particle.We recently identified a candidate for one such receptor from diverse species, designated here as Mcb1 for Multiubiquitin chain-binding protein, based on its ability to bind Lys 48 -linked multiubiquitin chains and its presence in the 26 S proteasome (21,22).(Additional names include Mbp1 (21), ASF-1 and S5a (23,24), Sun1 (25), and p54 (26) for the Arabidopsis, human, yeast (Saccharomyces cerevesiae), and Drosophila homologs, respectively.)Most of these Mcb1 polypeptides range from 376 to 414 amino acids long, the exception being yeast Mcb1 (268 amino acids), which is missing a portion of the C-terminal end present in the other Mcb1 proteins (21,22).Consistent with a role in ubiquitin conjugate recognition, Mcb1 proteins preferentially associate with multiubiquitin chains versus the ubiquitin monomer in vitro, especially those chains containing three or more ubiquitins (21,22,27,28).
Genetic studies in yeast revealed that Mcb1 may not be the only ubiquitin receptor.This is based on the observations that yeast Mcb1 knockout strains display a mild phenotype.Unlike mutants in other essential ubiquitin components (2,4), ⌬mcb1 strains are viable, have near-wild-type growth rates, degrade the bulk of short-lived proteins (including an N-terminal end rule substrate) normally, and are not sensitive to UV radiation or heat stress (22,25).However, they are more sensitive to amino acid analogs and cold than wild-type yeast (22). 2 In addition, ⌬mcb1 mutants cannot degrade ubiquitin-proline-␤galactosidase (Ub-Pro-␤-gal), an artificial substrate that is degraded by a subpathway in the ubiquitin system (ubiquitin fusion degradation (UFD) pathway) (9).Taken together, the results suggest that Mcb1 proteins are involved in the recognition and degradation of only a subset of ubiquitin/26 S proteasome substrates.
To further define the role of Mcb1 in conjugate recognition and to identify the site(s) of multiubiquitin chain binding, we analyzed here a series of deletion and amino acid substitution mutants of yeast and Arabidopsis Mcb1.A stretch of highly conserved hydrophobic residues was found to be critical for recognizing Lys 48 -linked multiubiquitin chains in vitro.However, when the various Mcb1 derivatives were tested for their ability to complement yeast ⌬mcb1, the multiubiquitin chain recognition motif was found not to be required for conferring resistance to amino acid analogs or for restoring degradation of Ub-Pro-␤-gal.Instead, a conserved region near the N terminus was essential for these functions.These data show that a domain near the N terminus, and not the multiubiquitin chainbinding site, is most critical for Mcb1 function in vivo.

Construction of the Yeast and Arabidopsis Mcb1
Mutants-Constructions encoding the yeast and Arabidopsis Mcb1 derivatives (see Fig. 2) were made in the Escherichia coli expression vector pET28a (Novagen, Madison, WI) by PCR strategies.The derived genes were verified as correct by DNA sequence analysis.The 5Ј-amplification oligonucleotides were designed to add an NdeI site at the native start codon or, for N-terminal-deleted constructions, at the deletion point to create a new start codon.For the C-terminal-deleted constructions, the 3Ј-amplification oligonucleotides were designed to add an internal stop codon.Introduced restriction sites and stop codons in the various oligonucleotides listed below are underlined and italicized, respectively.Positions of new start and stop codons for the terminal-deleted constructions are indicated in Fig. 2. The PCR products were cloned either directly or through intermediate vectors into pET28a using the NdeI site at the 5Ј ends and a 3Ј site which was created by the design of the 3Ј-amplification oligonucleotide (EcoRI, underlined) or derived from intermediate cloning vectors (EcoRI, HindIII, SalI, or NdeI).
Arabidopsis MCB1 (AtMCB1) and its deletion mutants were generated by PCR from the AtMCB1 cDNA (21).For AtMCB1, CTGCTTATC-GACCATATGGTTCTCGAGGCG (primer 3) and the KS primer (Stratagene, La Jolla, CA) were used for PCR amplification.The PCR product, which also contained 160 bp of 3Ј-untranslated region, was cloned into pET28a using the 5Ј NdeI site and a 3Ј EcoRI site derived from the cDNA vector, pBluescript SKϩ (Stratagene).For Arabidopsis C⌬1, AtMCB1 in pET28a was digested with BamHI and religated to introduce a premature stop codon.For Arabidopsis C⌬2 and C⌬3, primer 3 and either AGCTAAAGCCAGATCTCAGTCCTCATCAGCC (primer 4) or CGAAGGGCAAGAGCAAGTTCTCAATCGATATTTGGGTCC, respectively, were used for PCR.Amplified DNA fragments were cloned 2 H. Fu and R. D. Vierstra, unpublished results.through intermediate vector pGEMT into pET28a using NdeI/SalI and NdeI/NdeI sites, respectively.For Arabidopsis N⌬1, N⌬2, N⌬3, and N⌬4, CAAAGGACATATGGTATTGACTACTCCTACCTCTG, ATTTC-GGGGAGGATGATCATATGGAAAAGCCTCAGAAA (primer 5), TTTG-GTGTGGACCCACATATGGATCCAGAACTTGCT, or GCGGCCGAT-GAGGCACATATGAAAGACAAAGATGG, respectively, was used as the 5Ј-oligonucleotide and the KS primer was used as the 3Ј-oligonucleotide.Corresponding PCR products were cloned separately into pET28a using the 5Ј NdeI site and a 3Ј HindIII site derived from the cDNA vector.For Arabidopsis I1, primer 5 and primer 4 were used for PCR amplification.The PCR product was cloned through intermediate vector pT7Blue-T (Novagen) into pET28a using the 5Ј NdeI site and the 3Ј EcoRI site derived from the intermediate pT7Blue-T vector.For Arabidopsis I2 and I3, primer 5 or GAGGGTGCAAGTGGCCATATGTCTGCG-GCAGCTGCT was used as the 5Ј-oligonucleotide and AACTGAATTCT-GTTAGGCGGAAGCTGTGTCCC was used as the 3Ј-oligonucleotide.The PCR products were cloned directly into pET28a using NdeI/EcoRI sites.
For all the site-directed mutants (i.e.yeast N5 and G and Arabidopsis D5/G, D5, N5/G, N5, D/G, D, and G) (see Fig. 2), the QuickChange TM mutagenesis strategy (Stratagene) was employed using either ScMCB1 or AtMCB1 present in pET28a.The oligonucleotides used for the mutagenesis were designed according to the manufacturer's guidelines to have the noncomplementary nucleotides bracketed by 10 -22 complementary nucleotides.
In Vitro Multiubiquitin Chain Binding Activity Assay-All yeast and Arabidopsis Mcb1 proteins were expressed in E. coli strain BL21 (DE3) as His 6 -tagged versions using pET28a.The His 6 tag, containing the amino acid sequence MGSSHHHHHHSSGLVPRGSH, was encoded by a nucleotide sequence 5Ј to the NdeI site in the pET28a vector.Induction of protein expression and preparation of bacterial lysates were performed according to manufacturer's protocols (Novagen).Total protein from cell lysates was fractionated by SDS-polyacrylamide gel electrophoresis (PAGE) and electroblotted onto nitrocellulose membranes (Millipore, Bedford, MA).The membranes were washed for 5 min in Tris-buffered saline (TBS; 20 mM Tris⅐HCl (pH 7.5, 25 °C), 0.5 M NaCl), blocked for 2 h in TBS containing 10 mg/ml of bovine serum albumin and washed for 5 min in TBS.The membranes were incubated for 1-2 h at room temperature in TBS containing 10 mg/ml bovine serum albumin and 3.5 ϫ 10 5 cpm/ml of Lys 48 -linked multiubiquitin chains labeled with 125 I.The membranes were then washed in TBS for 1-2 h and subjected to autoradiography.The Lys 48 -linked multiubiquitin chains were prepared with ZmUbc7 enzyme and were radiolabeled with carrier-free Na 125 I (Amersham Corp.) using IODO-BEAD (Pierce) as described previously (6).
Amino Acid Analog Sensitivity Assays-The various MCB1 derivatives were moved from pET28a vector into a high copy 2 yeast shuttle vector pRS424 (32) and expressed without the His 6 tag under the direction of the yeast ScMCB1 promoter.pRS424 was first modified to include the ScMCB1 promoter immediately followed by NdeI and EcoRI sites.The ScMCB1 promoter, containing 552 bp upstream of the start codon, was isolated from genomic DNA by PCR using oligonucleotides TGCATCATTGCGAATACCGAG and CTGTAGCTTCCAATACCATAT-GGCGGTTACTGC.With the exception of the gene encoding Arabidopsis C⌬2, the various MCB1 derivatives were moved from pET28a into the modified pRS424 using the 5Ј NdeI site and a 3Ј EcoRI site.Arabidopsis C⌬2 construction was moved using the 5Ј NdeI site and a 3Ј SalI (filled in) into the modified pRS424 digested with NdeI and EcoRI (filled in).
Wild-type yeast, ⌬mcb1 harboring an empty pRS424 vector, or ⌬mcb1 yeast strains expressing various MCB1 derivatives were grown in liquid selection medium to late logarithmic phase.Cultures were washed twice with selection medium (arginine and phenylalanine omitted) and resuspended in the same medium to A 600 ϭ 1.0.For qualitative visualization of amino acid analog sensitivity, resuspended cells were spotted in a 10-fold dilution series onto solidified selection medium supplemented with canavanine and p-fluorophenylalanine at 1.5 and 25 g/ml, respectively, and incubated for 6 days.For quantitative measurement, resuspended cells were plated and grown for 14 days on solidified selection medium with or without amino acid analogs.Survival rate was calculated by dividing the number of colonies formed on selection medium with analogs by the number formed on selection medium without analogs.
Purification of the Yeast 26 S Proteasome-The 26 S proteasome was partially purified from wild-type and ⌬mcb1 yeast strains by conventional chromatography as described previously (34).The peak of hydrolytic activity (as measured using the substrate Suc-LLVY) from the DEAE-Affi-Gel blue column (Bio-Rad) was loaded onto a MonoQ column (Pharmacia Biotech Inc.).Protein was eluted using a linear gradient from 200 mM to 450 mM NaCl.Fractions were collected and assayed for peptidase activity and the presence of Mcb1 and Sug1/Cim3 (34).The peak of peptidase activity eluted at around 350 mM NaCl.
26 S proteasome preparations were further resolved by nondenaturing PAGE using a modification of the protocol of Hoffman et al. (35).PAGE employed a single gel layer consisting of 0.18 M Tris borate (pH 8.3), 5 mM MgCl 2 , 1 mM ATP, 1 mM dithiothreitol, and 4% acrylamidebisacrylamide (at a ratio of 37.5:1), polymerized with 0.1% Temed and 0.1% ammonium persulfate.The running buffer was the same as above without acrylamide.Xylene cyanol was added to the samples, and the samples were electrophoresed at 100 -150 mV until the xylene cyanol migrated through the gel.The position of the 26 S proteasome was visualized by UV light following an overlay with Suc-LLVY-AMC using a modification of the protocol of Hoffman et al. (35).The PAGE gel was incubated for 10 min with 30 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, and 2 mM ATP supplemented with 0.1 mM Suc-LLVY-AMC.The fluorescent gels were transilluminated with a UV light and photographed with a Polaroid camera.Sug1/Cim3 and Mcb1 were detected by immunoblot analysis as described above.Anti-Sug1/ Cim3 serum was generously provided by Dr. Carl Mann (Center d'Etudes de Saclay, Gif-sur-Yvette Cedex, France).(26), humans (24), and moss3 revealed four highly conserved regions that may function in these capacities (designated domains I-IV, Fig. 1, A and B).Whereas the Drosophila, human, and moss sequences average 47% identity to Arabidopsis Mcb1 over their entire length, the four conserved domains average 72% identity.Yeast Mcb1 also contains the first three conserved domains but is lacking the C-terminal region that includes domain IV (22).Among all five proteins, domain III exhibits the greatest conservation (82% identity).It contains a short stretch of conserved hydrophobic residues (LAL/ MALRL/V) surrounded by charged amino acids and is preceded by an invariant sequence GVDP (Fig. 1C).Beal et al. (36) previously proposed, from binding studies of mutant multiubiquitin chains with the 26 S proteasome, that repeated hydrophobic patches formed within assembled multiubiquitin chains are important for binding to the complex.Based on their data, we have suggested that this conserved hydrophobic patch within Mcb1 could participate in this association (21).Moreover, because Arabidopsis Mcb1 also contains a second LALAL patch near the C terminus of the protein (residues 310 -314), it was possible that multiubiquitin chain binding was strengthened by the cooperative action of these motifs.

Mutational Analysis of Yeast and Arabidopsis
Based on these sequence comparisons, a series of N-and/or C-terminal deletion mutants were generated from both yeast and Arabidopsis Mcb1 with a special emphasis on the four conserved domains (Fig. 2).Domain III was further analyzed by a collection of site-directed mutants designed to alter the hydrophobicity of the LAL/MALRL/V motif.The mutant pro-teins were expressed in E. coli and tested for their ability to bind Lys 48 -linked multiubiquitin chains in vitro.Various Arabidopsis and yeast Mcb1 mutants were also expressed in yeast and examined for their ability to complement the phenotypic defects of ⌬mcb1 (22) and for their ability to integrate into the yeast 26 S proteasome.
The Conserved Hydrophobic Sequence in Mcb1 Is Critical for Binding Multiubiquitin Chains-Evaluating the chain binding activity of the various yeast and Arabidopsis Mcb1 derivatives was facilitated by the discovery that each derivative accumulated to high levels in a soluble form when expressed in E. coli (Fig. 3A).Similar to the parental molecules (21,22), the apparent molecular mass (as measured by SDS-PAGE) of the various deletions was substantially greater than their presumed mass.This discrepancy was especially strong for the N-terminal deletions, suggesting that the C-terminal portion has an unusual structure in the presence of SDS.The site-directed mutants that replaced all five of the LAL/MAL residues in domain III with either aspartic acid or asparagine (D5/G, N5/G, D5, and N5) (see Fig. 3) also migrated noticeably slower than the parental polypeptides.
Chain binding was assayed using total protein from induced cultures that had been subjected to SDS-PAGE and electrophoretically transferred onto nitrocellulose membranes (28).The transfer efficiency of the various Mcb1 derivatives differed substantially.This was especially true for the smaller deletion mutants (i.e.yeast N⌬3 and Arabidopsis N⌬3, N⌬4, I1, I2, and I3), which required much lower currents and shorter electrophoretic transfer times.To ensure that equal amounts of protein were transferred onto the membranes, the electrophoresis conditions were varied accordingly and the amount of each Mcb1 protein transferred was verified subsequently by protein staining of the membranes (data not shown).Duplicate membranes were then probed with Lys 48 -linked multiubiquitin chains prepared in vitro and radiolabeled with 125 I.It should be emphasized that this assay is a qualitative measure of chain binding and may not precisely reflect the relative affinities in solution.
Binding analysis of the N-and/or C-terminal deletions revealed that a region encompassing domains III of yeast and Arabidopsis Mcb1 (amino acids 204 -268 and 202-264, respectively) contains a site required for multiubiquitin chain binding.As shown in Fig. 3B, every derivative containing an intact domain III showed binding activity (e.g.Arabidopsis and yeast N⌬1 and N⌬2, and Arabidopsis C⌬1 and C⌬2); even polypeptides containing just domain III surrounded by a few additional residues were competent (i.e.yeast N⌬3 and Arabidopsis I1, I2, and I3).In contrast, every deletion that removed all or part of domain III failed to bind chains (e.g.yeast C⌬1 and Arabidopsis C⌬3, N⌬3, and N⌬4).Domains I and II were not essential as demonstrated by the strong binding activity of yeast and Arabidopsis N⌬1 and N⌬2.Loss of the first 60 amino acids of yeast Mcb1 actually improved the association of chains with the membrane-bound protein by as much as 5-fold (Fig. 3B).Of interest is the absence of chain binding activity of the Arabidopsis N-terminal deletion N⌬4, in which domain IV and the second LALAL patch are intact (residues 310 -314) (21), and the presence of chain binding activity for the Arabidopsis C⌬2, which is missing both domains but containing an intact domain III.These data are inconsistent with a major role for both domain IV and the second LALAL patch in chain recognition.
An essential role for domain III in multiubiquitin chain recognition was shown more conclusively through the analysis of various site-directed mutants (Figs. 2 and 3B).Substituting the hydrophobic LALAL (226 -230) sequence in AtMcb1 with five aspartic acids abolished chain binding, implicating the 10-fold.In addition to the five contiguous hydrophobic residues, an adjacent hydrophobic amino acid one residue C-terminal from the patch (Leu 234 or Val 232 in yeast and Arabidopsis Mcb1, respectively) (Fig. 1) was also critical.Conversion of this amino acid to glycine was sufficient to dramatically impair chain binding (mutant G, Fig. 3B).As expected, Arabidopsis mutations combining the G substitution with mutations D5, N5, and D (D5/G, N5/G, and D/G, Fig. 3B) also failed to bind chains.Although the hydrophobic patch in domain III was clearly essential for multiubiquitin chain binding, sequence flanking this patch also appeared to influence binding strength.Arabidopsis N⌬3, which contained the intact hydrophobic stretch but was missing 10 residues within the N-terminal portion of domain III (encompassing the conserved GVDP motif), showed no binding activity (Fig. 3B).Likewise, Arabidopsis I3, which was missing residues more N-terminal to the patch (residues 151-201), showed reduced binding activity as compared with I2 (Fig. 3B).
As reported previously, the Mcb1 family of proteins prefers binding Lys 48 -linked multiubiquitin chains over ubiquitin monomers, especially those chains containing three or more ubiquitins (21,22,24,28,37).To test whether domain III alone is sufficient for this selectivity, the profile of multiubiquitin chains bound to full-length yeast Mcb1 was compared with those of the mutants N⌬2 and N⌬3. 125I-Labeled chains were incubated with the nitrocellulose-bound proteins, and chains that associated were released subsequently by heating the corresponding regions of the membrane in an SDS-containing buffer.Radioactive chains eluted from equivalent surface areas of the nitrocellulose were subjected to SDS-PAGE and autoradiography.As shown in Fig. 4, the profile of chains bound to N⌬2 and N⌬3 was similar to that bound to wild-type yeast Mcb1; chains lengths Ն3 ubiquitins were preferentially enriched as compared with free ubiquitin.(It should be noted that the stronger signal obtained with chains released from N⌬2 reflected the greater amount of radiolabeled chains initially bound to the immobilized protein (see Fig. 3B).)Similar results were obtained when full-length Arabidopsis Mcb1 was compared with its mutants, C⌬2, N⌬2, and I3 (data not shown).
The N-terminal Region of Mcb1 Is Required for Amino Acid Analog Resistance-Although MCB1 is not essential in yeast, deletion of this gene results in several phenotypes (22).The most notable is an enhanced growth sensitivity to amino acid analogs, presumably due to the accumulation of abnormal proteins that are normally removed by the ubiquitin pathway (38).To identify region(s) of Mcb1 required for this function, we expressed a number of the yeast and Arabidopsis Mcb1 derivatives (see Fig. 2) in the yeast ⌬mcb1 strain and examined their ability to complement the growth defect on analog-containing medium.
The wild-type and mutant proteins were expressed from a high copy 2 plasmid (pRS424) under the control of the yeast MCB1 promoter.All but one of the yeast and all of the Arabidopsis Mcb1 derivatives could be expressed to levels easily detected by immunoblot analysis with anti-Arabidopsis Mcb1 sera (Fig. 5A).Protein levels, estimated from the immunoblots, ranged from approximately 0.1 (e.g.Arabidopsis C⌬1) to 10 times (e.g.yeast Mcb1 and G) the level of ScMcb1 found normally in wild-type yeast (WT).The notable exception was yeast N⌬2, which appeared to be expressed at extremely low levels as it could be detected only following extensive development of the immunoblots (data not shown).Although most polypeptides appeared to be stable in vivo, breakdown products of several were observed, especially those containing N-terminal deletions (e.g.yeast N⌬1 and Arabidopsis N⌬1 and N⌬2) and the Arabidopsis site-directed mutants D5/G, D5, and N5 (Fig. 5A).
The growth of yeast ⌬mcb1 strains expressing the various Mcb1 derivatives was examined on medium containing canavanine (CAN) and p-fluorophenylalanine (pFP), analogs of arginine and phenylalanine, respectively.As can be seen in Fig. 5B, the levels of CAN and pFP used were sufficient to reduce the growth of ⌬mcb1 by over 100-fold as compared with wildtype yeast.Analog resistance could be completely restored by reintroducing ScMcb1 and restored to at least 50% of the wildtype levels by ectopic expression of AtMcb1.When the mutant proteins were tested, we found that the hydrophobic patch in domain III, necessary for the multiubiquitin chain binding, was not required for amino acid analog resistance.C-terminal deletions removing most of domain III (yeast C⌬1 and Arabidopsis C⌬3) as well as site-directed mutants affecting residues within the hydrophobic patch (yeast N5 and G, and Arabidopsis D5/G, D5, N5/G, N5, D/G, D, and G) could rescue the growth defect even though none could bind chains by the solid-phase assay used in Fig. 3.The level of resistance provided by these mutants was comparable to that of wild-type yeast and those ⌬mcb1 strains expressing full-length yeast and Arabidopsis Mcb1.In contrast, the N-terminal region encompassing domain I was essential for analog resistance.Expression of the various N-terminal mutations missing domain I (yeast N⌬1 and N⌬2 and Arabidopsis N⌬1, N⌬2, N⌬3, and N⌬4) not only failed to restore resistance, it actually accentuated the analog sensitivity of ⌬mcb1 (Fig. 5B).This was especially true for the strains expressing yeast N⌬1 or N⌬2, which exhibited a severe growth defect on CAN/pFP-containing medium.
When we used a more quantitative assay involving plating efficiency as a measure of analog resistance, similar results were obtained (Fig. 6).At the levels of CAN/pFP used here, only ϳ10% of the yeast ⌬mcb1 cells harboring the 2 vector alone formed colonies.Plating efficiency was restored to near wildtype levels by expression of any of the site-directed mutations altering the hydrophobic patch in domain III or almost any of the C-terminal deletions.The only exception was Arabidopsis C⌬1, but its slightly lower efficiency was likely caused by the poor expression of this mutant protein in yeast (see Fig. 5A).In contrast, the strain expressing yeast N⌬1 was Ͼ100 times more sensitive to the analogs than ⌬mcb1, whereas the strains expressing yeast N⌬2 or Arabidopsis N⌬1, N⌬2, N⌬3, or N⌬4 were ϳ3-7 times more sensitive (Fig. 6).The hypersensitivity induced by yeast N⌬1 and N⌬2 was particularly surprising given the low levels of protein present (especially for N⌬2; see Fig. 5A) and suggested that the deletions were behaving in a dominant negative fashion.This negative effect was only seen when ⌬mcb1 was exposed to the amino acid analogs.When ⌬mcb1 strains expressing the N-terminal deletions were grown on complete medium without analogs, plating efficiency was indistinguishable from that of ⌬mcb1 and wild-type yeast (data not shown).
It was previously shown that purified Arabidopsis Mcb1 is a potent inhibitor of ubiquitin-dependent proteolysis in vitro, presumably by binding ubiquitin conjugates in a free form and blocking their subsequent interaction with Mcb1 and other receptors associated with the 26 S proteasome (39).Based on this inhibitory action, it was possible that the growth defect observed for the N-terminal deletions was caused by their ability to bind multiubiquitin chains through domain III even though another function(s) was now impaired (e.g. assembly into the 26 S proteasome, interaction with other proteasome subunits; see below).To test this possibility, we generated yeast and Arabidopsis double mutants combining the N⌬1 deletions with the site-directed L234G or V232G mutations (mutation G) that abolished multiubiquitin chain binding in vitro (see Fig. 3B).When yeast N⌬1/G was expressed in ⌬mcb1, the severe growth defect of N⌬1 was suppressed by over 10-fold (Fig. 6B).However, the N⌬1/G-expressing strain was still ϳ7 times more sensitive to CAN/pFP than ⌬mcb1.For Arabidopsis N⌬1/G, no decrease in analog sensitivity was seen as compared with that of Arabidopsis N⌬1 (Fig. 6B).These data suggest that the negative effect observed for N⌬1 does not require a functional multiubiquitin chain-binding site.
If the N-terminal deletions could behave in a dominant negative fashion, they could be useful tools to poison selected part(s) of the ubiquitin pathway in vivo (e.g.UFD pathway) (9).To test this possibility, we expressed yeast Mcb1 and N⌬1 from a high copy 2 plasmid in wild-type yeast and tested for growth inhibition on CAN/pFP plates.Whereas N⌬1 was expressed to nies formed by wild-type yeast and the various ⌬mcb1 strains on medium containing the amino acid analogs, canavanine, and p-fluorophenylalanine (1.5 and 25 g/ml, respectively).Cultures were grown in analog-free medium, resuspended in analog-containing medium to an A 600 ϭ 1.0, and then spotted in 10-fold serial dilutions (left to right) onto solidified medium supplemented with the analogs.Colony growth was observed after 6 days.levels similar to that of endogenous ScMcb1, ectopic expression of ScMcb1 resulted in a ϳ10-fold increase in ScMcb1 protein (Fig. 7A).Both the ScMcb1-and N⌬1-expressing wild-type strains showed the same growth resistance to the analogs as wild-type yeast expressing the plasmid alone (Fig. 7B).This lack of dominance suggests that the N-terminal mutants cannot interfere with the function of endogenous ScMcb1.Moreover, it showed that ScMcb1 does not hinder yeast cell survival when expressed above normal levels, thus diminishing its potential value as a pathway inhibitor in vivo.
N-terminal Domain Is Required for Degradation of Ub-Pro-␤-Gal-Prior phenotypic analysis of ⌬mcb1 (22) showed that ScMcb1 is essential for the breakdown of Ub-Pro-␤-gal, a synthetic ubiquitin fusion degraded by the ubiquitin pathway.Unlike ubiquitin fusion substrates bearing other amino acids besides proline at the junction, the ubiquitin moiety in Ub-Pro-␤-gal is not cleaved following synthesis (33).This ubiquitin then serves as an acceptor site for further ubiquitination by the UFD subpathway involving the E2/E3 pair encoded by UBC4/5 and UFD4 (9).To identify the domains within Mcb1 required for this breakdown, the stability of Ub-Pro-␤-gal was examined by pulse-labeling in ⌬mcb1 strains expressing various yeast Mcb1 derivatives.Whereas Ub-Pro-␤-gal was extremely stable in the ⌬mcb1 strain, it was rapidly degraded when the ScMCB1 gene was reintroduced (t 1/2 ϳ 12 min (Fig. 8)].This instability was similar to that obtained with wild-type yeast and was independent of whether a high copy 2 (pRS424) or a low copy CEN plasmid (pRS314) was used for expression (data not shown and Fig. 8).Each of the modifications tested that deleted or altered the hydrophobic patch in domain III (C⌬1, N5, and G) also restored rapid degradation of Ub-Pro-␤-gal to a rate nearly equivalent to that seen with ScMcb1 (Fig. 8).In contrast, Ub-Pro-␤-gal remained stable in strains expressing N⌬1.These results were consistent with those obtained with the amino acid analog sensitivity assays, suggesting that the Nterminal region containing domain I is required for both functions.B, colonies formed by the various yeast strains when plated on medium containing the amino acid analogs, canavanine and p-fluorophenylalanine (1.5 and 25 g/ml, respectively).Cultures were grown initially in analog-free medium, resuspended in analog-containing medium to an A 600 ϭ 1.0, and then spotted in 10-fold serial dilutions (left to right) onto solidified medium supplemented with the analogs.Colony growth was observed after 6 days.gels were determined by a peptidase overlay assay using the fluorogenic 20 S proteasome substrate Suc-LLVY-AMC (35).As shown in Fig. 9A, the yeast 26 S proteasome migrates as two species (26S a and 26S b ) on nondenaturing PAGE, corresponding to 20 S catalytic core particles capped by two or one regulatory particles (35). 4The mobility and peptidase activity of both species were similar in wild-type and ⌬mcb1 strains and in ⌬mcb1 strains expressing yeast C⌬1 or N⌬1, indicating that none of the derivatives substantially impaired assembly or the peptidase activity of the holoenzyme complex.

Assembly of Mutant Mcb1
Yeast Mcb1 and its derivatives were detected in the 26 S complex by immunoblot analysis of the PAGE gels with yeast Mcb1 antisera.For wild-type yeast, Mcb1 protein was found in both species of the 26 S proteasome (Fig. 9B, lane 1).As expected, this signal was not present in proteasome preparations from the ⌬mcb1 strain (Fig. 9B, lane 4).When a ⌬mcb1 strain expressing the C⌬1 deletion was examined similarly, the C⌬1 protein was detected in both 26 S proteasome species, indicating that this C-terminal truncation is incorporated into the complex (Fig. 9B, lane 2).In the case of N⌬1, a broad smear of immunoreactive material was observed reproducibly following nondenaturing PAGE, overlapping both species of the 26 S proteasome (Fig. 9B, lane 3).Sucrose gradient analysis suggested that this signal arose primarily from heterogeneous aggregates of unincorporated N⌬1 (data not shown).To separate these aggregates from the proteasome, the DEAE-Affi-Gel blue eluate was further fractionated by MonoQ FPLC (Fig. 9,  C-E).N⌬1 protein could be detected in these samples immunologically.It co-fractionated as a single peak with the 26 S proteasome whose position was determined by peptidase activity and by immunoblot analysis for Sug1/Cim3, a yeast ATPase subunit of the 19 S regulatory complex (34).To confirm that N⌬1 protein was actually associated with 26 S complex, the peak peptidase activity from the MonoQ FPLC (fraction 26) was further analyzed by nondenaturing PAGE.As can be seen in Fig. 9F, N⌬1 was clearly detected in both species of the more purified 26 S complexes.These complexes co-migrated with the complexes containing full-length Mcb1 similarly prepared from wild-type yeast.From these data, we concluded that both yeast N⌬1 and C⌬1 can be incorporated into the 26 S proteasome.DISCUSSION Degradation of proteins by the ubiquitin/26 S proteasome pathway presumably requires receptors for selectively directing multiubiquitinated proteins to the 26 S proteasome.A number of subunits within the 19 S regulatory complex have been resolved (10), but so far only Mcb1 has been tentatively identified as one such receptor based on its location in the 19 S regulatory complex of the 26 S proteasome and its affinity for multiubiquitin chains (21,22).Although genetic analyses in yeast clearly indicates that Mcb1 cannot be the sole ubiquitinconjugate receptor, these studies showed that Mcb1 has an important role in the degradation of specific ubiquitin pathway substrates (e.g.Ub-Pro-␤-gal) and in the resistance to amino acid analogs (22).To identify Mcb1 domains required for these activities, a structure/function analysis of the yeast and Arabidopsis Mcb1 proteins was initiated.This study was made possible, in part, by the ability of Arabidopsis Mcb1 to complement the functions of its yeast counterpart.Detailed mutational analysis using the Arabidopsis and yeast homologs yielded remarkably consistent results that defined a single essential domain for multiubiquitin chain binding in vitro.Unexpectedly, in vivo data revealed that this domain is separate from that required for amino acid analog resistance and Ub-Pro-␤-gal degradation.Neither domain is essential for the assembly of Mcb1 into the 26 S complex.
Analysis of both deletion and site-directed mutants localized a region essential for multiubiquitin chain binding to a hydrophobic patch (LAL/MALRL/V) within domain III, a highly conserved domain present in all Mcb1 proteins (Fig. 1) (22).While it is possible that other regions also participate, data presented here suggest that this motif provides the major interaction site.(i) Mutants missing sequences either C-or N-terminal to the patch (e.g.Arabidopsis C⌬1 and C⌬2 and yeast N⌬1 and N⌬2) show no significant reduction in chain binding.(ii) Deletion mutants containing just domain III flanked by short sequences (e.g.yeast N⌬3 and Arabidopsis I3 containing only 65 and 63 amino acids, respectively) were still capable of binding multiubiquitin chains, albeit at lower levels.(iii) Single amino acid substitutions in the hydrophobic patch, (changing residue Leu 234 (yeast) or Val 232 (Arabidopsis) to glycine) were sufficient to dramatically impair binding.(iv) The smallest yeast and Arabidopsis Mcb1 mutants containing only domain III had a binding preference for multiubiquitin chains comparable to their full-length counterparts.For Arabidopsis Mcb1 in particular, our data eliminate a significant role for the second LALAL patch near the C terminus of the protein (residues 310 -314).A deletion containing just this domain (N⌬4) failed to bind chains whereas a deletion missing this domain (C⌬2) had full binding activity by the solid-phase assay used here.
Our results are consistent with those recently published by Haracska and Udvardy (27) who found that domain III in Drosophila Mcb1 (p54) contains a major chain-binding site.In contrast to our work, they also detected a second chain-binding site within the conserved domain IV, but possibly with weaker affinity.Our mutational analysis do not support a major role for this second site.Domain IV is not present in yeast Mcb1.Although present in Arabidopsis Mcb1, the ability of a single amino acid substitution in domain III (e.g.Val 232 3 Gly) to completely block chain binding indicates that a second site, if it is present, cannot work autonomously.However, we cannot rule out the possibility that a second site exists but was not  2.The metabolic stability of the reporter protein was determined by pulse-chase analysis as described previously (22).Ub-Pro-␤-gal was immunoprecipitated from cell lysates and subjected to SDS-PAGE.The level of Ub-Pro-␤-gal was quantitated by Phosphor-Imager analysis and expressed as a percentage of that present at t ϭ 0. In the data shown here, ScMcb1 and mutant derivatives were expressed from a high copy 2 vector (pRS424) (31).For ScMcb1 and the sitedirected mutant N5, similar results were obtained (data not shown) with a low copy CEN vector (pRS314) (43).
detected by the different assay conditions used here (higher temperature incubations and higher salt washes).
How the binding site in domain III interacts with multiubiquitin chains is not understood.That Mcb1 can still bind multiubiquitin chains even after SDS denaturation and fixation onto nitrocellulose membrane indicates that the binding site probably has a stable secondary structure or that it can renature upon binding to the membrane.One possibility is that the hydrophobic patch forms a pocket or loops out of the Mcb1 protein, stabilized in part by the highly charged flanking regions.This site could interact directly with complementary hydrophobic patch(es) on multiubiquitin chains, some of which may involve Leu 8 and Ile 44 , previously shown to be required for the association of ubiquitin conjugates with the 26 S proteasome (36).Because multiubiquitin chains interact more strongly than free ubiquitin with Mcb1, we presume that the site of interaction with Mcb1 is present in the ubiquitin polymer and not found in free ubiquitin.It also suggests that the greater affinity of Mcb1 for longer multiubiquitin chains does not result from increased cooperative interactions among complementary repeated binding sites in both Mcb1 and multiubiquitin chains.Instead, the greater affinity could arise from the enhanced probability that one of the ubiquitin moieties will interact with domain III or from the possibility that longer multiubiquitin chains stabilize a structural motif that is absent in the ubiquitin monomer.In support of the latter, structural differences between the ubiquitin dimer, trimer, and tetramer have been noted (6,40).
Surprisingly, analysis of the various yeast and Arabidopsis Mcb1 mutants in yeast showed that the multiubiquitin chainbinding site in domain III is not essential for the phenotypic functions of Mcb1, i.e. resistance to amino acid analogs or for FIG. 9. Analysis of the 26 S proteasome in wild-type yeast and ⌬mcb1 strains expressing the yeast deletions N⌬1 or C⌬1.Equal amounts of yeast cells from wild-type, ⌬mcb1, and ⌬mcb1 strains expressing Mcb1 deletion C⌬1 or N⌬1, were lysed and fractionated by DEAE-Affi-Gel blue chromatography.Equal amounts of protein, as determined by Bradford assay, from fractions containing the 26 S proteasome were resolved by nondenaturing PAGE on a 4% acrylamide gel (panels A and B).DEAE-Affi-Gel blue fractions, containing the 26 S proteasome from the wild-type yeast and ⌬mcb1 strain expressing N⌬1, were further fractionated by MonoQ FPLC using a linear 200 -450 mM NaCl gradient (panels C-E for N⌬1 expressing strain).MonoQ FPLC fractions, containing equal amount of protein from wild-type yeast and ⌬mcb1 strain expressing N⌬1, were resolved by nondenaturing PAGE and subjected to immunoblot analyses (panel F).A, peptidase overlay assay of the 26 S proteasome preparations from wild-type, C⌬1, N⌬1, and ⌬mcb1 strains using the fluorogenic substrate Suc-LLVY-AMC (lanes 1-4, respectively).B, immunoblot of the gel shown in panel A using yeast Mcb1 antisera.Migration position of the two 26 S proteasome species (26S a and 26S b ) as well as the 20 S proteasome (20S) are indicated.C-E, fractions from the MonoQ FPLC, prepared from the ⌬mcb1 strain expressing N⌬1, were analyzed for the 26 S proteasome by peptidase assay using the substrate Suc-LLVY (C) and for the presence of Sug1/Cim3 (D) and N⌬1 protein (E) by immunoblot analysis.F, immunoblot of a native PAGE gel using yeast Mcb1 antisera.Lanes 1 and 2 correspond to the peak peptidase fractions from the MonoQ FPLC prepared from wild-type yeast and ⌬mcb1 strain expressing N⌬1, respectively.Migration positions of the two forms of the 26 S proteasome (26S a and 26S b ) are indicated.
the ability to degrade Ub-Pro-␤-gal.Instead, we found that a distinct domain within the first 60 N-terminal residues (domain I) was critical.Whereas mutations affecting domain III could fully complement the ⌬mcb1 deletion, mutants missing domain I failed to promote the degradation of Ub-Pro-␤-gal and actually made the ⌬mcb1 strain more sensitive to amino acid analogs.
Our ability to separate the phenotypic functions of Mcb1 from its ability to recognize multiubiquitin chains raises a question as to the role of Mcb1 in 26 S proteasome function.Because domain III mutants lack phenotypic consequences, it can be argued that the multiubiquitin chain binding activity seen in vitro is not relevant to the in vivo function of Mcb1.While definitive proof is not yet available, several observations point against this possibility.First, Mcb1 is the only subunit of purified rabbit and Drosophila 26 S proteasomes that displayed an affinity for multiubiquitin chains in vitro (27,28).Second, the region identified here as involved in ubiquitin binding (domain III) represents the most conserved region of the molecule, implicating it as a critical domain for function.Third, both the hydrophobic nature of domain III and its preference for multimeric over monomeric and dimeric ubiquitin conforms well with predictions made concerning the specificity of a ubiquitin receptor within the 26 S proteasome (5,28,36).Fourth, this structural motif is highly specific for multiubiquitin binding because even single amino acid substitutions within domain III were sufficient to substantially impair binding (mutants D and G).
A more likely possibility is that Mcb1 has multiple functions.One function involves multiubiquitin chain recognition but is dispensable by virtue of the presence of other multiubiquitin receptors.The other function, which is more apparent phenotypically, is unclear but presumably requires the N-terminal region.Interestingly, a recent study has also indicated that the N-terminal domain of Mcb1 is important.It showed that human Mcb1 (or S5a) can associate both in vitro and in vivo with Id1, a helix-loop-helix protein that interacts with and blocks the action of the DNA-binding protein MyoD (41).Mcb1 appears to interact with Id1 through its N-terminal half (which includes domain I) and appears to counteract the inhibitory role of Id1 in myogenesis.How this interaction may relate to protein turnover is unclear.
What is the in vivo function of the N-terminal region?One obvious possibility is that domain I is required for assembly of Mcb1 into the 26 S proteasome.The hypersensitivity of yeast Mcb1 mutant N⌬1 could be explained by its failure to integrate into the complex while still retaining its multiubiquitin chain binding activity.The partial loss of the hypersensitive phenotype of N⌬1 by affecting its chain-binding site (N⌬1/G) would support this notion.However, we can detect N⌬1 in the 26 S proteasome.While this association discounts a gross defect in 26 S proteasome assembly, more subtle pertubations are possible.Alternatively, domain I could be essential for appropriate interactions between Mcb1 and other 19 S subunits.Potential binding partners include Nin1 (25) and Soi1, 5  proteasome subunits have been shown to have an affinity for multiubiquitin chains using the solid-phase assay (21,27,28).It is possible that other receptors were not identified by this assay because they failed to survive SDS denaturation and fixation onto nitrocellulose membranes or because they exist as multisubunit complexes.Both Nin1 and Soi1 have some interaction with Mcb1, but neither have been shown yet to have any ubiquitin binding activity.It is also possible that other ubiquitin receptors are not tightly associated with the 26 S proteasome and dissociate upon purification of the 19 S complex.An attractive possibility is that ubiquitin receptors freely dissociate from the 19 S complex allowing them to shuttle multiubiquitinated proteins to the 26 S proteasome.The fact that a substantial portion of cellular Mcb1 can also be found in a free form, not associated with the 26 S proteasome, is consistent with this notion (21,22,26).One possible candidate of this type is p62, a human phosphotyrosine-independent SH2 domain ligand, that has an affinity for ubiquitin (42).Clearly, a combination of biochemical and genetic approaches will be required to identify these additional receptor elements.
Mcb1-To identify domains in Mcb1 necessary for multiubiquitin chain recognition, association with the 26 S proteasome, and in vivo functions, a parallel structure/function analysis of yeast and Arabidopsis Mcb1 was initiated by constructing various deletion and amino acid substitution mutants.Amino acid sequence comparisons of Mcb1 homologs from Arabidopsis (21), Drosophila

FIG. 1 .FIG. 3 .
FIG. 1.Amino acid sequence comparisons of the multiubiquitin chain-binding protein Mcb1 from various species.Mcb1 amino acid sequences were from the following sources: Arabidopsis (21), human (24), Drosophila (26), yeast (22), and moss (Physcomitrella patens) (H.Fu, P. Girod, and R. D. Vierstra, unpublished results).A, dot blot analysis comparing the amino acid sequence similarity of Arabidopsis and human Mcb1.The plot was generated from the UW-GCG programs COMPARE and DOTPLOT based on a window of 55 and a stringency of 55.From pairwise comparisons of all five Mcb1 proteins, four conserved domains were identified and designated I-IV.B, percent amino acid sequence identities of domains I-IV in Drosophila, human, moss, and yeast Mcb1s as compared with Arabidopsis Mcb1.C, amino acid sequence alignment of domain III.The comparison was generated with the computer program BoxShade 2.7; identical and similar residues are shaded with black and gray boxes, respectively.The consensus sequence denotes hydrophobic residues by a dash (-), charges residues by q, and prolines by P. Arrows indicate the hydrophobic residues which were identified as necessary for the binding of multiubiquitin chains in vitro (see Fig.3).In B and C, coordinates refer to the amino acid sequence positions of Arabidopsis Mcb1.

FIG. 4 .
FIG. 4. Profile of Lys 48 -linked multiubiquitin chains bound to yeast Mcb1 and deletions N⌬2 and N⌬3.Yeast Mcb1 (ScMcb1), N⌬2, and N⌬3 (see Fig. 2) were expressed in E. coli, subjected to SDS-PAGE, and transferred onto nitrocellulose membranes.The membranes were incubated with a mixture of monomeric ubiquitin and Lys 48 -linked multiubiquitin chains labeled with 125 I (Ub 1 -Ub n ).Regions of the membrane containing the Mcb1 proteins were excised, and the bound radioactivity was eluted by boiling in SDS-PAGE sample buffer.The eluted material was then subjected to SDS-PAGE and autoradiography.Radioactivity eluted from an equal amount of nitrocellulose membrane was subjected to SDS-PAGE for each mutant.Background denotes the profile of chains eluted from an equivalent amount of nitrocellulose membrane not containing Mcb1 protein.The left lane shows the profile of the 125 I-labeled multiubiquitin chains (Ub chains) used in the assay.

FIG. 5 .
FIG. 5. Ability of various Mcb1 derivatives to complement the growth sensitivity of the yeast ⌬mcb1 mutant to amino acid analogs.A, expression of various yeast (ScMcb1) and Arabidopsis (AtMcb1) Mcb1 derivatives in yeast.Crude extracts (15 g) from wild-type (WT) yeast, the ⌬mcb1 strain, or the ⌬mcb1 strain expressing various Mcb1 derivatives (see Fig.2) were subjected to SDS-PAGE and immunoblotted with an Arabidopsis Mcb1 antiserum.The migration positions of the various Mcb1 proteins are indicated by asterisks.B, colonies formed by wild-type yeast and the various ⌬mcb1 strains on medium containing the amino acid analogs, canavanine, and p-fluorophenylalanine (1.5 and 25 g/ml, respectively).Cultures were grown in analog-free medium, resuspended in analog-containing medium to an A 600 ϭ 1.0, and then spotted in 10-fold serial dilutions (left to right) onto solidified medium supplemented with the analogs.Colony growth was observed after 6 days.
FIG. 6.Amino acid analog sensitivity of yeast ⌬mcb1 expressing various Mcb1 derivatives as measured by cell survival.Wildtype yeast, the ⌬mcb1 strain, or the ⌬mcb1 strain carrying various yeast (ScMcb1) and Arabidopsis (AtMcb1) Mcb1 derivatives (see Fig. 2) were grown in analog-free liquid medium, washed, and plated onto solidified medium with or without the amino acid analogs, canavanine and p-fluorophenylalanine (1.5 and 25 g/ml, respectively).After a 14-day incubation, cell survival rate was calculated by dividing the number of colonies formed on medium containing analogs by the number of colonies formed on medium without the analogs.A, percent cell survival rates of yeast ⌬mcb1 strains carrying the various Mcb1 derivatives.B, expanded version of panel A focusing on the increased analog sensitivity of the ⌬mcb1 strains expressing various N-terminal deletion mutants of Mcb1.Mutants N⌬1/G combined the N⌬1 deletion with the yeast Leu 232 or Arabidopsis Val 234 to Gly substitution (see Fig. 2).Error bars indicate the S.D. FIG. 7. N⌬1 does not increase amino acid analog sensitivity of wild-type yeast.Yeast Mcb1 (ScMcb1) or the N-terminal deletion N⌬1 were expressed in wild-type (WT) yeast from the high copy 2 vector.A, expression of the ScMcb1 and N⌬1 proteins in yeast.Crude extracts (15 g) from the ⌬mcb1 strain, wild-type yeast, and wild-type yeast ectopically expressing ScMcb1 or N⌬1 proteins were subjected to SDS-PAGE and immunoblotted with an Arabidopsis Mcb1 antiserum.The migration positions of the ScMcb1 and N⌬1 proteins are indicated by arrows.B, colonies formed by the various yeast strains when plated on medium containing the amino acid analogs, canavanine and p-fluorophenylalanine (1.5 and 25 g/ml, respectively).Cultures were grown initially in analog-free medium, resuspended in analog-containing medium to an A 600 ϭ 1.0, and then spotted in 10-fold serial dilutions (left to right) onto solidified medium supplemented with the analogs.Colony growth was observed after 6 days.

FIG. 8 .
FIG. 8. Degradation of Ub-Pro-␤-gal in yeast ⌬mcb1 strains containing various ScMcb1 derivatives.Various ScMcb1 derivatives were introduced into a ⌬mcb1 yeast strain expressing the Ub-Pro-␤-gal reporter protein.Descriptions of the various Mcb1 mutants are shown in Fig.2.The metabolic stability of the reporter protein was determined by pulse-chase analysis as described previously(22).Ub-Pro-␤-gal was immunoprecipitated from cell lysates and subjected to SDS-PAGE.The level of Ub-Pro-␤-gal was quantitated by Phosphor-Imager analysis and expressed as a percentage of that present at t ϭ 0. In the data shown here, ScMcb1 and mutant derivatives were expressed from a high copy 2 vector (pRS424)(31).For ScMcb1 and the sitedirected mutant N5, similar results were obtained (data not shown) with a low copy CEN vector (pRS314)(43).
In vitro binding studies indicate that the Mcb1 family of proteins can bind multiubiquitin chains assembled through various Lys isopeptide linkages (21, 27, 28, 37).Whether ubiquitin chains linked through lysines other than Lys 48 also interact with this hydrophobic patch in Mcb1 is not known.Binding studies of the various Mcb1 mutants with alternatively linked chains and the structural solution of Mcb1/multiubiquitin complexes will be useful in this regard.The identity of a single dominant binding site in Mcb1 is inconsistent with a model that binding involves multiple patches within Mcb1 interacting with complementary sites formed by the repeated ubiquitin monomers within the chains.
identified by genetic screens and yeast two-hybrid analysis, respectively, as 19 S subunits that interact with Mcb1.Expression of Mcb1 mutants missing the N-terminal domain could interfere with proper assembly of these and potentially other 19 S subunits into the 26 S complex.The location of key residues in domain I as well as the identification of proteins that interact with this region will be critical to clarify the function(s) of Mcb1.If Mcb1 is not the only ubiquitin receptor in the 26 S proteasome, what are other candidates?Besides Mcb1, no other 26 S