Determinants of Nuclear and Cytoplasmic Ubiquitin-mediated Degradation of MyoD*

The ubiquitin-proteasome system is responsible for the regulation and turnover of many short-lived proteins both in the cytoplasm and in the nucleus. Degradation can occur via two distinct pathways, an N terminus-dependent pathway and a lysine-dependent pathway. The pathways are characterized by the site of initial ubiquitination of the protein, the N terminus or an internal lysine, respectively. MyoD, a basic helix-loop-helix transcription factor, is a substrate for the ubiquitin-proteasome pathway and is degraded in the nucleus. It is preferentially tagged for degradation on the N terminus and thus is degraded by the N terminus-dependent pathway. Addition of a 6× Myc tag to the N terminus of MyoD can force degradation through the lysine-dependent pathway by preventing ubiquitination at the N-terminal site. Modifications of the nuclear localization signal and nuclear export signal of MyoD restrict ubiquitination and degradation to the cytoplasm or the nucleus. Using these mutants, we determined which degradation pathway is dominant in the cytoplasm and the nucleus. Our results suggest that the lysine-dependent pathway is the more active pathway within the cytoplasm, whereas in the nucleus the two pathways are both active in protein degradation.

Degradation of many short-lived cellular proteins, such as transcription factors, tumor suppressors, and cell cycle regulators, occurs via the ubiquitin-proteasome pathway (1)(2)(3)(4). Through this pathway, proteins are targeted for degradation by the 26 S proteasome via the formation of a polyubiquitin chain. The process begins with activation of ubiquitin by the ubiquitin-activating enzyme (E1), 1 followed by transfer of ubiquitin to E2, a ubiquitin-conjugating enzyme. E2 shuttles the ubiquitin molecule to the substrate-specific ubiquitin ligase (E3), which then delivers the ubiquitin to the substrate to be degraded. Initially, it was thought that ubiquitination occurred only through a lysine-dependent ubiquitination pathway in which ubiquitin is covalently attached to the substrate protein via an amide linkage to the ⑀-amino group of an internal lysine (5). However, recent studies have shown that the N terminus of a protein substrate may also serve as the site of ubiquitination (6 -8), a pathway termed N terminus-dependent ubiquitination. Via either ubiquitination pathway, polyubiquitin chain formation continues by the conjugation of subsequent ubiquitin moieties to the attached ubiquitin, and the substrate-ubiquitin conjugate is then degraded by the 26 S proteasome in an ATP-dependent manner. Isopeptidases cleave the ubiquitin chain, and the single ubiquitin molecules are recycled (5). Currently, the relative contribution of each of these two pathways is unknown.
Among the short-lived proteins degraded by the ubiquitinproteasome system are several transcription factors, including MyoD. MyoD, a nuclear basic helix-loop-helix transcription factor necessary for skeletal muscle differentiation (9), is rapidly degraded by the ubiquitin-proteasome system both in vitro and in vivo (10). Ubiquitination on the N terminus of MyoD appears to occur in preference to ubiquitination on internal lysines (8), and MyoD appears to be ubiquitinated and degraded in the nucleus by the ubiquitin-proteasome system (11).
A nuclear ubiquitin-proteasome system may also be responsible for the degradation of several other transcription factors, including Smad2 and the basic helix-loop-helix/Per-ARNT-Sim homology domain dioxin receptor (12,13), as the activated forms of these proteins require translocation into the nucleus for degradation. In addition, Far1, a protein required for establishing cell polarity of mating yeast and for bringing about cell cycle arrest, is ubiquitinated and degraded in the nucleus (14). In contrast, cyclin D1 (15), p53 (16), p27 kip1 (17), IB␣ (18), and the aryl hydrocarbon receptor (19) all require nuclear export prior to degradation by the ubiquitin-proteasome system. Ligand-activated Smad3, although ubiquitinated in the nucleus, is also exported to the cytoplasm for degradation by the 26 S proteasome (20).
Transport of cellular proteins, including transcription factors such as MyoD, into and out of the nucleus is facilitated by nuclear localization and nuclear export sequences (NLS and NES), which facilitate transport of such proteins across the nuclear envelope (21)(22)(23)(24). NLS exist as single sequences (5-12 amino acids) of basic amino acids (lysine or arginine; Ref. 25) or as bipartite regions of basic residues separated by 10 -12 nonbasic residues (26). Less is known about NES, but they are generally comprised of regions of hydrophobic amino acids.
Thus, ubiquitin-mediated protein degradation occurs in both the cytoplasm and the nucleus, and both lysine-dependent and N terminus-dependent ubiquitination pathways exist for several proteins, including MyoD. The aim of the present study was to determine the relationship of these ubiquitin-mediated degradation pathways using NLS and NES mutants and inhibitors of nuclear uptake and export. Herein, we report that for MyoD, both the N terminus-dependent and the lysine-dependent pathways function within the nucleus, although the lysinedependent pathway appears to be more active in the cytoplasm. The importance of this observation lies in the elucidation of alterations in cellular protein degradation in physiological and pathophysiological states.

Plasmids and Construction of MyoD NLS and MyoD NES -Wild-type
MyoD and lysine-less MyoD, each of which is in the pCIneo vector, and N-terminal-blocked (6ϫ Myc-tagged) MyoD, which is in the pCS2ϩMT vector, have been described previously (8,11). Preparation of NLS and NES mutants of MyoD was accomplished using the QuikChange sitedirected mutagenesis kit (Stratagene) according to the manufacturer's instructions. DNA sequencing using Big Dye Version 2.0 (Applied Biosystems) was used to confirm all sequences.
Cell Culture-HeLa cells, which were selected because they do not express MyoD, were grown in Dulbecco's modified Eagle's medium, which was supplemented with 10% fetal calf serum, 2 mM L-glutamine, and antibiotics (100 units/ml penicillin G and 100 g/ml streptomycin) (Invitrogen) and was maintained in a humidified chamber at 37°C and 5% CO 2 . Transient transfections (efficiency ϳ40 -60%) were performed using the FuGENE 6 reagent (Roche Molecular Biochemicals), and cells were analyzed 16 -24 h later.
Immunofluorescent Localization of MyoD-Subcellular localization of MyoD and its mutants in HeLa cells was determined by direct immunofluorescence using the mouse monoclonal anti-MyoD antibody (1:100 dilution, NCLMyoD1; NovoCastra) followed by incubation with a TRITC-conjugated donkey anti-mouse IgG (heavy and light chain; Jackson ImmunoResearch) as described previously (11). MG132 (10 M; Peptides International) or LMB (10 nM; Sigma) were added to cells where indicated 2 h prior to fixation. MG132 was prepared as a 10 mM stock solution in Me 2 SO, and LMB was prepared as a 10 M solution in ethanol. Cells were observed using a Zeiss Axioskop microscope, and 10 -20 random fields of each culture condition were photographed (magnification ϫ40) using a Zeiss AxioCam digital camera. Cells were scored, and the percentage of cells expressing nuclear MyoD was determined by methods similar to those of Sachdev et al. (27) and Yagita et al. (28). Specifically, cells were scored according to the cellular localization of MyoD as predominately nuclear, predominately cytoplasmic, or distributed equally between the nucleus and the cytoplasm. For each determination over 100, cells were scored from 3 to 6 independent transfections.
Determination of Degradation of MyoD and MyoD Mutants in Vivo-As described previously (11), 16 -24 h after transfection, the HeLa cells were incubated with CHX (100 g/ml; Sigma) to inhibit further protein synthesis. MG132 (10 M) or LMB (10 nM) was added along with CHX as necessary. Following incubation for 0, 0.5, 1, 2, and 3 h, the cells were lysed for at least 30 min in phosphate-buffered saline containing 5% Igepal, 1 mM EDTA, 1 mM dithiothreitol, and 2 mM phenylmethylsulfonyl fluoride. Afterward the cells were sonicated and then centrifuged at 14,000 rpm for 10 min at 4°C in an Eppendorf microcentrifuge to remove cellular debris. The lysates were mixed with an equal amount of 2ϫ Laemmli sample buffer (Bio-Rad), and equal amounts of each sample were run on a 12% Tris-HCl gel (Bio-Rad) and were electroblotted onto nitrocellulose (Osmonics). The blots were probed with monoclonal anti-MyoD antibody (1:100 dilution; NovoCastra) followed by incubation with a secondary horseradish peroxidaseconjugated antibody and detection by chemiluminescence (Amersham Biosciences). The resulting bands were quantitated using the EADS system (Eastman Kodak Co.), and the data were graphed using the Excel graphing program (Microsoft). The degradation rate is expressed as half-life (t1 ⁄2 ), the time for degradation of 50% of the MyoD. Each of the constructs was evaluated by 3-8 independent determinations of t1 ⁄2 . The data are expressed as Ϯ S.D.

RESULTS
MyoD is localized to the cell nucleus in growing myoblasts (29) and in a variety of cells following transfection (11). To determine the subcellular localization of lysine-less MyoD (Lys 3 Arg mutations) and N-terminal-blocked (6ϫ Myctagged) MyoD, MyoD constructs were transfected into HeLa cells and were visualized 24 h later by immunofluorescence ( Fig. 1). Transfected cells were also treated either with MG132 or with LMB 2 h prior to fixation. MG132 inhibits the proteasome, whereas LMB inhibits CRM-1-dependent nuclear export FIG. 2. Half-life of wild-type MyoD with MG132. A, 24 h after transfection with wild-type MyoD, HeLa cells were treated with CHX or CHX plus MG132. Cells were lysed at 0, 0.5, 1, 2, or 3 h and were evaluated via SDS-PAGE gel and Western blot for MyoD. B, the pixels for each band were measured and normalized so that the number of pixels at t ϭ 0 was 100%. The log 10 of the % pixels was plotted versus time for each time point, and the t1 ⁄2 was calculated as the log of 50% (1.7 h). The t1 ⁄2 for MyoD was 0.8 h and 11.2 h when MG132 was added. Treatment of the cells with CHX plus LMB increased the t1 ⁄2 to 1.3 h (data not shown).

FIG. 1. Localization of wild-type
MyoD, lysine-less MyoD, and N-terminal-blocked MyoD. 24 h following transfection of HeLa cells, cells were treated with MG132 or LMB, and the localization of each construct was visualized by immunofluorescence. Wild-type MyoD, lysineless MyoD, and N-terminal-blocked MyoD were localized exclusively to the nucleus under all conditions. to retain MyoD within the nucleus (30). As seen in Fig. 1, wild-type MyoD, lysine-less MyoD, and N-terminal-blocked MyoD were all localized to the nucleus under basal conditions as well as in the presence of MG132 or LMB. Thus, at steady state, MyoD and its lysine-less and N-terminal-blocked mutants were found exclusively in the cell nucleus.
We next compared the rates of degradation of wild-type, lysine-less, and N-terminal-blocked MyoD in vivo. 16-24 h following transfection, HeLa cells were treated with CHX to inhibit further protein synthesis, and the amount of MyoD at 0, 0.5, 1, 2, and 3 h was determined via Western blot analysis. MG132 or LMB was added together with CHX in replicate samples to determine the degradation rate of these proteins under conditions in which the proteasome or nuclear export was inhibited. As seen in Fig. 2, wild-type MyoD was degraded with a t1 ⁄2 of 0.8 Ϯ 0.1 h, which was markedly increased to ϳ11 Ϯ 4 h in the presence of MG132. The half-life was slightly increased in the presence of LMB to 1.3 Ϯ 0.2 h (legend to Fig.  2). The half-lives of lysine-less MyoD and N-terminal-blocked MyoD were significantly increased over the half-lives of wild type, with t1 ⁄2 of ϳ2.7 Ϯ 0.6 h and 2.6 Ϯ 0.6 h, respectively, as seen in Table I. As with wild-type MyoD, the half-lives of lysine-less MyoD and N-terminal-blocked MyoD were markedly increased by incubation with MG132, although they were minimally affected by incubation with LMB (data not shown). Together these data suggest that wild-type MyoD, lysine-less MyoD, and N-terminal-blocked MyoD are each degraded via the ubiquitin-proteasome system within the nucleus.
Nuclear transcription factors, including MyoD, are synthesized in the cytoplasm and are targeted to the nucleus via NLS. To determine whether MyoD or its mutants are degraded within the cytoplasm either prior to nuclear uptake or during nucleocytoplasmic recycling, we sought to identify the NLS within MyoD to generate NLS-deficient mutants.
Based on results from Vandromme et al. (31), we mutated the basic residues within two regions of MyoD (amino acid regions 110 -112 and 130 -135) to alanine, as Vandromme et al. report that deletion of both of these regions is necessary to prevent nuclear localization of MyoD. However, as seen in Fig. 3, substitution mutagenesis yielded distinctly different results. Mutagenesis of K102A, R103A, K104A, R110A, R111A, and K112A (MyoD NLS1 ) was sufficient to markedly inhibit nuclear import, whereas mutagenesis of K133A and R134A (MyoD NLS2 ) did not FIG. 3. Localization and t1 ⁄2 of MyoD NLS mutants. Basic amino acids of putative NLS sequences in MyoD were mutated to alanines, and the effect on nuclear localization was determined by immunofluorescence. A total of six regions were mutated; MyoD NLS1 (amino acid region 102-112) and MyoD NLS3 (amino acid region 117-124) were found to be 83 and 79% in the cytoplasm, respectively. MyoD NLS1ϩNLS3 was found to be 87% cytoplasmic. MyoD NLS2 was 18% cytoplasmic. The rates of degradation of wild-type MyoD and its NLS mutants are described above, with the rate of MyoD NLS1ϩ3 not determined (ND). affect subcellular localization. Given these results, the entire amino acid sequence of MyoD was evaluated for additional basic residue-containing regions which may serve as NLS. Mutation of R117A, R119A, R120A, R121A, and K124A (MyoD NLS3 ) proved to be an additional NLS (Fig. 3). This same region was examined by Vandromme et al. (31) using deletion mutagenesis and was found not to serve as an NLS. Inhibition of nuclear import of MyoD NLS3 (79% cytoplasmic) was comparable with that seen with MyoD NLS1 (83% cytoplasmic). Furthermore, mutation of both NLS1 and NLS3 (MyoD NLS1ϩ3 ) did not inhibit import substantially more (i.e. ϳ87% cytoplasmic) than that observed with the single mutation. Therefore it is likely that NLS1 and NLS3 represent subdomains of a single large targeting sequence (amino acid region 102-124) within MyoD. Evaluation of three additional regions in MyoD (R143A, K146A, and R151A; R220A and R221A; R235A, R238A, and K241A) failed to reveal any additional NLS. The rates of degradation of several of these NLS mutants were examined. The t1 ⁄2 of wild-type MyoD and MyoD NLS2 , both of which localized to the nucleus (96 and 92% nuclear, respectively), were each found to be 0.8 Ϯ 0.1 h. However, the t1 ⁄2 of MyoD NLS1 , MyoD NLS3 , and MyoD NLS1ϩ2 , each of which localized predominately to the cytoplasm (83, 79, and 79% cytoplasmic, respectively), were found to be 1.3 Ϯ 0.3, 1.2 Ϯ 0.1, and 1.2 Ϯ 0.2 h, respectively. Thus, it appears that the rate of MyoD degradation correlates with its subcellular localization, in that nuclear localization is associated with more rapid degradation.
The NLS1 mutation identified above did not result in complete absence of MyoD from the nucleus. It may be that MyoD undergoes nucleocytoplasmic shuttling and that the mutagenesis of NLS1 causes MyoD to enter the nucleus more slowly than wild-type MyoD, or it may be that the rate at which MyoD exits the nucleus is accelerated when NLS1 is mutated. To test if MyoD NLS1 undergoes nucleocytoplasmic shuttling, 16 -24 h after transfection of HeLa cells with MyoD NLS1 , LMB was added, and MyoD localization was visualized via immunofluorescence. As seen in Fig. 4, incubation with LMB led to a dramatic accumulation of MyoD NLS1 within the nucleus (17-93% nuclear). This suggests that the MyoD NLS1 mutant cycles between the cytoplasm and the nucleus, although it does not indicate whether MyoD NLS1 is entering the nucleus more slowly or exiting the nucleus more quickly than wild-type MyoD. Localization of this mutant to the nucleus in the presence of LMB was associated with a decrease in the half-life of MyoD NLS1 (t1 ⁄2 ϭ 1.0 Ϯ 0.1 h) compared with that seen in the absence of LMB (1.3 Ϯ 0.3 h; Fig. 4).
The ability of MyoD to cycle in and out of the nucleus suggests either the presence of an intrinsic NES or of association with a chaperone (32) that supports MyoD nuclear export. Typically, NES are regions high in leucine content. Comparison of the MyoD sequence to other known NES within various proteins, such as protein kinase A inhibitor (33), cyclin B1 (34), human immunodeficiency virus-1 Rev protein (35), or human Cdc25C (36), suggested a candidate NES within the N terminus region of MyoD. Thus, using the MyoD NLS1 mutant as a template, Leu-3, Leu-4, Leu-8, Ile-11, and Leu-13 were mutated to alanines (MyoD NLS1NES ), and the construct was transfected into HeLa cells. A marked increase in nuclear localization of MyoD (17-42% nuclear) was observed (Fig. 5) consistent with the notion that amino acid sequence 3-13 contains a nuclear export signal. Two other regions high in leucine content were mutated on the MyoD NLS1 background (L268A, L269A, and L270A; L248A, L251A, and I254A) to identify other potential NES within MyoD; however, no others were identified.
Having identified the major NLS and NES signals in MyoD, we next sought to determine which degradation pathway (N terminus-dependent or lysine-dependent) predominates in the nuclear and cytoplasmic compartments. To localize the various mutants to the cytoplasm or the nucleus, we prepared NLS1 or NES mutations on N-terminal-blocked MyoD and lysine-less MyoD.
The N-terminal-blocked species of MyoD are degraded via the lysine-dependent pathway. The N-terminal-blocked NLS1 mutant (N-terminal-blocked MyoD NLS1 ) was degraded with a shorter t1 ⁄2 (1.6 Ϯ 0.3 h) compared with the N-terminal-blocked wild-type MyoD (2.7 Ϯ 0.6 h), whereas the N-terminal-blocked NES mutant (N-terminal-blocked MyoD NES ) was degraded with a t1 ⁄2 (2.8 Ϯ 0.5 h) similar to the N-terminal-blocked wildtype MyoD (Fig. 7A). As expected, N-terminal-blocked wildtype MyoD and N-terminal-blocked MyoD NES were nuclear, whereas N-terminal-blocked MyoD NLS1 was predominately cytoplasmic. The more rapid degradation of the N-terminalblocked species localized to the cytoplasm (N-terminal-blocked MyoD NLS1 ) suggests that the lysine-dependent pathway may be more active for MyoD in the cytoplasm than in the nucleus.
The lysine-less species of MyoD are degraded via the N terminus-dependent pathway. Whereas lysine-less wild-type MyoD and lysine-less MyoD with the NES mutation (lysineless MyoD NES ) were both nuclear as expected, lysine-less MyoD with the NLS1 mutation (lysine-less MyoD NLS1 ) was only 40% cytoplasmic. This result is somewhat unexpected because the NLS1 mutation on either the wild-type MyoD or N-terminalblocked MyoD background yielded species with 83 or 75% cytoplasmic localization, respectively (Figs. 3, 4, and 7A). Lysineless MyoD NLS1 had a shorter t1 ⁄2 (1.5 Ϯ 0.4 h) than lysine-less wild-type MyoD (2.6 Ϯ 0.6 h). Lysine-less MyoD NES was degraded at a substantially slower rate (t1 ⁄2 ϭ 5.8 Ϯ 0.9 h). The slower degradation of the lysine-less species localized to the nucleus (lysine-less MyoD and lysine-less MyoD NES ) suggests that the N terminus-dependent pathway is less active in the nucleus.

DISCUSSION
The ubiquitin-proteasome system exists in both the cytoplasm and the nucleus and is responsible for the degradation of many short-lived cellular proteins. Ubiquitination of the target protein can occur on an ⑀-amino group of an internal lysine or on the N terminus of the protein tagged for destruction. MyoD, a basic helix-loop-helix transcription factor, is ubiquitinated and degraded by the ubiquitin-proteasome system in the nu-cleus (11). Ubiquitination of the protein occurs preferentially on the N terminus, leading to N terminus-dependent degradation. Blocking of the N terminus with a 6ϫ Myc tag, via methylation, or through carbamylation of the ␣-amino group leads to preferential degradation by the lysine-dependent pathway (8). Using NLS and NES mutants of N-terminal-blocked and lysine-less MyoD, we determined that the lysine-dependent pathway is more active in the cytoplasm, whereas both degradation pathways have equal activity in the nucleus.
This determination was based on t1 ⁄2 calculations of the different N-terminal-blocked and lysine-less mutants, and although the differences in the t1 ⁄2 may be small, they are likely to be biologically significant because a 2-4-fold change in the half-life of a protein may underlie a significant change in a protein steady-state level and hence in a biological function. It should also be noted that only the subcellular localization and t1 ⁄2 were assayed and that kinetic determinations of MyoD nucleocytoplasmic shuttling were not evaluated. Furthermore, it is certainly possible, if not likely, that these changes have multiple effects on MyoD, affecting such aspects as DNA binding, acetylation, and/or dimerization. Although the results presented herein are specific to MyoD, future studies will focus on FIG. 7. Localization and t1 ⁄2 of N-terminal-blocked and lysineless MyoD mutants and their NLS and NES mutants. The same NLS1 and NES mutations were prepared on N-terminal-blocked and lysine-less MyoD. The constructs were transfected into HeLa cells and were visualized by immunofluorescence. A, N-terminal-blocked MyoD NLS1 was localized 75% to the cytoplasm, and N-terminal-blocked MyoD NES was localized 96% to the nucleus. The t1 ⁄2 of N-terminalblocked MyoD and its NLS1 and NES mutants were found to be 2.7 Ϯ 0.6, 1.6 Ϯ 0.3, and 2.8 Ϯ 0.5 h, respectively. B, lysine-less MyoD and lysine-less MyoD NES were 100% nuclear, whereas lysine-less MyoD NLS1 was 40% cytoplasmic. The half-lives of the lysine-less MyoD and its NLS1 and NES mutants were 2.6 Ϯ 0.6, 1.5 Ϯ 0.4, and 5.8 Ϯ 0.9 h, respectively. these pathways of cellular localization and degradation and on how the pathways relate to other proteins degraded by the ubiquitin-proteasome system.
To determine the ability of the ubiquitin-proteasome system to degrade MyoD in the cytoplasm, the NLS of MyoD was mutated in an attempt to inhibit nuclear localization. NLS are characterized by single or bipartite regions of basic amino acids. In the present study NLS mutants of MyoD were prepared by mutagenesis of clusters of basic amino acids (Lys and Arg) to alanines. Among the six regions evaluated, MyoD NLS1 and MyoD NLS3 were found to significantly reduce nuclear localization of MyoD (17 and 21% nuclear, respectively). Mutation of NLS1 and NLS3 together did not significantly inhibit nuclear import to a greater degree than that seen with either single mutation alone. Therefore, it is likely that MyoD possesses one NLS signal (amino acid region 102-124) and that NLS1 and NLS3 function as subdomains of this sequence. These results differ from those previously reported by Vandromme et al. (31) in which they identify two NLS regions (amino acid regions 100 -112 and 130 -135) that appear to function independently in that deletion of both regions is required to inhibit nuclear localization of MyoD. We find that substitution mutagenesis of NLS1 (amino acid region 102-112) alone is sufficient to inhibit import and that NLS2 (amino acid region 133-134) is not a nuclear targeting sequence. Furthermore, MyoD NLS1ϩ2 was no more cytoplasmic than MyoD NLS1 alone (Fig. 3).
Using MyoD NLS1 as a background, we identified a single NES signal within the N terminus of MyoD (Fig. 5). Although MyoD NLS1NES was 42% nuclear compared with 17% nuclear for MyoD NLS1 , there was still a considerable amount of MyoD in the cytoplasm. Treatment of MyoD NLS1NES with LMB caused the protein to be greater than 88% nuclear (data not shown), suggesting that mutagenesis of the NES does not fully inhibit nuclear export. One reason for this may be that MyoD associates with cofactors that facilitate nuclear export regardless of the mutated NES, or it may be that a separate NES, consisting of residues other than leucine, exists.
MyoD NES had a slightly longer t1 ⁄2 than wild-type MyoD (1.05 versus 0.8 h; Fig. 6). One potential reason for this finding may be that we had introduced modifications to the protein at the N terminus, which appears to be its main ubiquitination site (8). Another possibility is that the nuclear degradation machinery is saturated when MyoD is exclusively nuclear. Consistent with this latter possibility, we found that the t1 ⁄2 for the degradation of wild-type MyoD in the presence of LMB is somewhat slower than that seen without the inhibitor (legend to Fig. 2).
To localize N-terminal-blocked MyoD and lysine-less MyoD to the cytoplasm or to the nucleus, we mutated their NLS and NES. The t1 ⁄2 of N-terminal-blocked MyoD was significantly longer than that of wild-type MyoD (2.7 Ϯ 0.6 h and 0.8 Ϯ 0.1 h, respectively), which is in agreement with previously published data ( Fig. 7A; Ref. 8). Cytoplasmically localized N-terminalblocked MyoD (N-terminal-blocked MyoD NLS1 ) was degraded with a shorter half-life than either wild-type N-terminalblocked MyoD or N-terminal-blocked MyoD NES , both of which localized to the nucleus (Fig. 7A). This supports the notion that the lysine-dependent pathway is more active in the cytoplasm than in the nucleus.
The locus of activity of the N terminus-dependent pathway is less clear. Mutation of the NLS within lysine-less MyoD did not yield a construct with predominant cytoplasmic localization. Although some cytoplasmic localization was seen, lysine-less MyoD NLS1 was ϳ60% nuclear (Fig. 7B). The reason or reasons for this finding are unclear. One reason may be that mutagenesis of lysines to arginines may affect DNA binding as lysine 146 has been shown to make contacts with the bound DNA duplex (37). The reduced ability of MyoD to induce myogenesis as a result of mutagenesis of lysine 124 (38) may also contribute to the nuclear localization and shorter half-life of lysineless MyoD NLS1 . Activation and the DNA-binding ability of MyoD are also regulated by acetylation of lysines 99, 102, and 104 (39,40), which are mutated in the MyoD lysine-less mutant. In addition, lysine-less MyoD NLS1 was degraded more rapidly than wild-type lysine-less MyoD, perhaps consistent with this notion. Lysine-less MyoD NES was nuclear with a half-life considerably longer than that of wild-type lysine-less MyoD (5.8 Ϯ 0.9 h versus 2.6 Ϯ 0.6 h). This finding likely results from the lack of recognition sites for the ubiquitinproteasome system, as all of the lysines had been mutated to arginines and had been mutated within MyoD NES . The N terminus (amino acid region 3-13) had also been mutagenized, leaving no known site for recognition and ubiquitination, as alterations of amino acids with the protein alter its threedimensional structure and hence its potential for recognition by the ubiquitin system.
A scheme for the pathways of MyoD degradation is illustrated in Fig. 8. MyoD is synthesized in the cytoplasm and rapidly targeted to the nucleus, where it exists both uncomplexed and bound to DNA. Wild-type MyoD can be tagged for ubiquitination on internal lysines or at its N terminus. Modification of either of these recognition sites decreases the degradation rate of MyoD. Wild-type MyoD was degraded more rapidly in the nucleus than in the cytoplasm. N-terminalblocked MyoD, however, was degraded more rapidly in the cytoplasm. This suggests that the lysine-dependent pathway of degradation is more active in the cytoplasm than in the nucleus. Because the half-lives of wild-type MyoD NLS1 and Nterminal-blocked MyoD NLS1 were similar, it may be that cyto- FIG. 8. A model of ubiquitin-proteasome degradation of MyoD within the cell. MyoD is synthesized in the cytoplasm and rapidly directed via an NLS to the nucleus, where it resides both free and complexed with DNA. An NES sequence allows MyoD to shuttle from nucleus to cytoplasm. In addition, another minor LMB-sensitive (LMB Sens) pathway supports nuclear export. In the nucleus, where the lysine-dependent and N terminus-dependent pathways are equally active, ubiquitin proteasome degradation occurs more rapidly than in the cytoplasm. The lysine-less-dependent pathway was more active than the N terminus-dependent pathway within the cytoplasm.
plasmic MyoD is degraded only through a lysine-dependent pathway, whereas both pathways function equally well in the nucleus. Equivalent activities of the two pathways within the nucleus are also supported by the similar half-lives of wild-type N-terminal-blocked MyoD and wild-type lysine-less MyoD.
At present it appears that many more proteins are tagged for ubiquitin-dependent degradation on internal lysines. In addition, it currently seems that the majority of "nuclear" proteins require nuclear export prior to degradation. Consistent with these observations, the lysine-dependent degradation pathway appears to be more active within the cytoplasm than in the nucleus. This raises the possibility that nuclear proteins that are degraded in the nucleus undergo degradation via an N terminus-dependent pathway, as is the case with MyoD.