Human thymidylate synthase (hTS; EC 184.108.40.206) is one of a small group of proteasomal substrates whose intracellular degradation occurs in a ubiquitin-independent manner. Previous studies have shown that proteolytic breakdown of the hTS polypeptide is directed by an intrinsically disordered 27-residue domain at the N-terminal end of the molecule. This domain, in co-operation with an α-helix spanning amino acids 31–45, functions as a degron, in that it has the ability to destabilize a heterologous polypeptide to which it is attached. In the present study, we provide evidence indicating that it is the 26S isoform of the proteasome that is responsible for intracellular degradation of the hTS polypeptide. In addition, we have used targeted in vitro mutagenesis to show that an Arg–Arg motif at residues 10–11 is required for proteolysis, an observation that was confirmed by functional analysis of the TS N-terminus from other mammalian species. The effects of stabilizing mutations on hTS degradation are maintained when the enzyme is provided with an alternative means of proteasome association; thus such mutations perturb one or more post-docking steps in the degradation pathway. Surprisingly, deletion mutants missing large segments of the disordered domain still function as proteasomal substrates; however, degradation of such mutants occurs by a mechanism that is distinct from that for the wild-type protein. Taken together, our results provide information on the roles of specific subregions within the intrinsically disordered N-terminal domain of hTS in regulation of degradation, leading to a deeper understanding of mechanisms underlying the ubiquitin-independent proteasomal degradation pathway.
The proteasome is a large multi-subunit complex of several dozen proteins that is responsible for the bulk of protein degradation within the cell [1–3]. A number of structural isoforms of the proteasome have been described. All contain a common core complex known as the 20S particle, which consists of 28 subunits organized into four, stacked heptameric rings that form a chamber within which proteolysis takes place [4–6]. The entrance to this chamber is protected by a small pore capped by regulatory complexes that differ among proteasomal isoforms. In the 26S proteasome, which has been examined in most detail, the 20S complex is capped by the 19S regulatory particle (or PA700), which includes six ATPases organized into a hexameric ring that interacts directly with the outer heptameric ring of the 20S core [3–5]. The 19S particle exhibits a chaperone-like activity, and is responsible for recognition and unfolding of substrates as they ‘thread’ their way through the pore and into the proteolytic chamber [4,7,8]. In other isoforms, the 20S core is capped by distinct activators, including PA28 [7,8] (also called the 11S activator or REG) and PA200 ; the physiological roles of these proteasomal isoforms are only partially understood.
Typically, polyubiquitin chains covalently attached to the target substrate mediate the latter's recognition of, and docking to, the proteasome; this is followed by passage of the substrate into the proteasome's inner chamber [1–3]. While polyubiquitin attachment is a common mode of targeting substrates for proteasome-mediated degradation, it is not required for all proteins. A growing number of substrates have been shown to undergo degradation in the absence of ubiquitin modification (see [10,11] for reviews). These include the polyamine biosynthetic enzyme ODC (ornithine decarboxylase) [12–14], the proto-oncoprotein c-Fos [15,16], the cyclin-dependent kinase inhibitor p21Waf1/Cip1 [17,18], tumour suppressors p53  and Rb (retinoblastoma) , the F protein of hepatitis C virus , the transcriptional co-activator SRC-3  and others. Although the number of proteins identified as degraded by the ubiquitin-independent pathway is relatively small, recent proteomic studies have indicated that the pathway may be a significant contributor to the regulation of protein turnover within the cell . Thus it is likely that many more ubiquitin-independent substrates remain to be discovered.
Current knowledge of the mechanism of ubiquitin-independent protein degradation is limited. A series of elegant studies by P. Coffino and colleagues have shown that the short (i.e. <30 min) half-life of ODC is conferred by an intrinsically disordered 37-residue region at the C-terminal end of the polypeptide [12,13,24]. This region functions as a degron, in that it has the capacity to destabilize a heterologous protein to which it is fused [12,13,24]. The ODC degron competes with ubiquitin for proteasome recognition and binding . A Cys–Ala dipeptide located 22 residues from the C-terminal end mediates this binding, which is followed by the polypeptide's entry into the catalytic chamber and proteolysis in a C-to-N direction [13,14]. The particular proteasomal subunit to which ODC docks is not known. An accessory protein, termed antizyme, binds to ODC and increases the availability of its C-terminal end for proteasomal entry and initiation of degradation .
Studies of other ubiquitin-independent substrates bear out the critical role of disordered regions in interaction with, and degradation by, the proteasome. Degradation of tumour suppressor TP53 requires a disordered, proline-rich domain located near the N-terminal end of the protein . Ubiquitin-independent degradation of c-Fos is mediated by its disordered C-terminal end , which may interact with the PSMC5 subunit of the 19S complex [16,26]. The cyclin-dependent kinase inhibitor p21Waf1/Cip1 is a loosely structured protein whose degradation is mediated by the binding of the C-terminal end of the protein to the α7 subunit of the 20S core . Clearly, disordered regions play key roles in the degradation of ubiquitin-independent substrates, a feature shared with ubiquitin-dependent substrates as well .
TS (thymidylate synthase; EC 220.127.116.11) catalyses the reductive methylation of dUMP to form dTMP, and is essential for the de novo biosynthesis of dTTP during DNA replication and repair [29,30]. Studies in our laboratory have identified the human enzyme (denoted hTS) as a ubiquitin-independent proteasomal substrate. Degradation of the hTS polypeptide is governed by a 45-residue region at its N-terminal end, a region comprising a flexible, intrinsically disordered domain spanning the first ~27 amino acids, followed by an amphipathic α-helix that is 18 amino acids in length [31–33]. Similar to the C-terminus of ODC, the N-terminal domain of hTS functions as an independent degradation signal or degron .
Mature hTS has a proline residue at its N-terminal end, which arises through co-translational excision of the initiating methionine residue [32,34]. Amino acid substitutions that promote acetylation of the N-terminal residue, such as Pro2→Ala or Pro2→Leu, result in resistance to degradation, indicating that a free, unmodified amino group at the N-terminus is necessary for degradation [32,33]. Furthermore, a Gly–Ser dipeptide motif at residues 5–6 appears to act as a stabilizing element. Finally, one or more residues within the proline-rich region spanning amino acids 9–15, promotes degradation . The mechanisms by which these regions modulate the half-life of hTS have yet to be fully elucidated.
In the present study, we provide evidence that hTS degradation is mediated by the 26S isoform of the proteasome. In addition, we have carried out extensive mutational analysis of the enzyme's disordered N-terminal domain, focusing on its role in regulating degradation. We have identified two classes of mutant molecules, based upon their intracellular half-lives. One class, involving either blocking of the N-terminal end or loss/substitution of an Arg–Arg motif at residues 10–11, has a long half-life, indicating resistance to degradation. These mutants appear to be defective in one or more post-docking steps in the proteasomal degradation pathway. The other class, consisting of rather large deletions, has a short half-life, signifying susceptibility to degradation; interestingly, this class of molecules is degraded by a mechanism that is distinct from that for the wild-type enzyme. We thus conclude that sub-elements within the N-terminal unstructured region of hTS mediate specific steps in the degradation pathway that override what appears to be a distinct proteasomal mechanism.
All constructs for mutant analysis were generated using standard molecular biology techniques, and were verified directly by DNA sequencing. The parent plasmid for expression of hTS and its various mutant derivatives was pJZ205, which contains a full-length TS gene under the control of the SV40 (simian virus 40) promoter. Single amino acid substitutions were generated by the QuikChange® site-directed mutagenesis kit (Stratagene), whereas chimaeric proteins were constructed by one or two rounds of overlapping PCR. Ligation of hTS to PSMD4 was carried out using a mouse cDNA clone (OriGene Technologies). Details regarding all constructs, including primers used for mutagenesis, are available upon request from the corresponding author.
All cell lines were cultured at 37 °C in a humidified 5% CO2 atmosphere. Cell line RJK88.13, which is a TS-deficient derivative of V79 Chinese hamster lung cells , was maintained in DMEM (Dulbecco's modified Eagle's medium; Cellgro) containing 4.5 g/l glucose, and supplemented with 10% heat-inactivated FBS (fetal bovine serum; Cellgro) and 10 mM thymidine. The cell line HCT116 was maintained in RPMI medium (Cellgro) containing 4.5 g/l glucose, and supplemented with 10% heat-inactivated FBS.
Stable transfection, RNAi (RNA interference) and protein half-life analysis
RJK88.13 cells in medium containing 10 μM thymidine were transfected with the plasmid expression vectors indicated using Lipofectamine™ 2000 or LTX (Invitrogen) according to the manufacturer's instructions. Stable transfectants were selected in thymidine-free medium containing the nucleoside transport inhibitor dipyridamole (5 μM; Sigma–Aldrich). Transfectants were pooled and maintained in mass culture.
For siRNA (small interfering RNA)-mediated ‘knockdown’ of 19S complexes, HCT116 cells were transfected with 50 nM concentrations of each of four PSMC5-specific siRNAs (Dharmacon) using Lipofectamine™ 2000. After 48 h, half-lives were analysed following addition of 50 μg/ml CHX (cycloheximide; Sigma–Aldrich) for the times indicated. Where indicated, cells were pre-incubated with a freshly prepared solution of the proteasome inhibitor lactacystin (7.5 μM; Calbiochem) for 4 h, or with 100 nM FdUrd (5-fluoro-2′-deoxyuridine; with 10 μM folinic acid) for 24 h, before addition of CHX.
Protein extracts and immunoblotting
Cells were harvested by scraping, and lysed by sonication (3×10 s) in NET2 buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl and 0.05% Nonidet P40] containing 10 mM DTT (dithiothreitol), 2 mM 2-mercaptoethanol, 5 mM PMSF, 200 μg/ml aprotinin, 100 μg/ml pepstatin and 50 μg/ml leupeptin. Crude lysates were centrifuged at 15000 g for 1 h at 4 °C, and the protein concentrations in the resulting extracts were quantified using the Bio-Rad assay reagent with BSA as a standard. Immunoblotting was performed using standard techniques. hTS was detected with a monoclonal antibody provided by Dr Sondra Berger (Department of Pharmaceutical and Biomedical Science, University of South Carolina, Columbia, SC, U.S.A.). For detection of PSMC5, a monoclonal antibody (Santa Cruz Biotechnology, catalogue number sc-81388) was used. As an internal control for equal loading, blots were re-probed with an antibody against actin (Sigma–Aldrich, Clone AC-40). The antigen–antibody complexes were visualized using appropriate secondary antibodies with the ECL (enhanced chemiluminescence) kit (Amersham). Densitometry was carried out using ImageJ software maintained by the National Institutes of Health (http://rsb.info.nih.gov/ij/). A 2-fold dilution series of each extract was included on the blots for calibration and to correct for film exposure times. All values were normalized to actin concentrations on the same blots. Experiments were carried out at least twice, and as many as seven times in some cases; the data depicted in the Figures are representative of the outcomes.
Native PAGE and in-gel proteasome assays
Cell extracts (60 μg of protein) were supplemented with 0.5 mM ATP and 2.5 mM MgCl2 and separated on 4% acrylamide gels containing 0.5 mM ATP and 2.5 mM MgCl2 at 100 V for 3 h at 4 °C. Following electrophoresis, gels were incubated with Suc–Leu–Leu–Val–Tyr–AMC (50 μM, Suc is N-succinyl and AMC is 7-amino-4-methylcoumarin; BIOMOL, catalogue number PW8720) in developing buffer [50 mM Tris/HCl (pH 7.4), 5 mM MgCl2 and 1 mM ATP) for 30 min at 30 °C to assess the activity of the 26S proteasome.
Degradation of hTS is carried out by the 26S proteasome
Given the highly folded structure of hTS [36–38], it can be predicted that degradation of the protein requires the 26S isoform of the proteasome. However, this remains to be demonstrated experimentally. To directly assess the role of the 26S proteasome in hTS degradation, we determined whether interference with formation of the 19S regulatory complex would cause stabilization of the enzyme. This was performed by RNAi-mediated down-regulation of subunit PSMC5 (also called Sug1 or S8), an ATPase that is a component of the 19S particle, and is therefore essential for establishment of an intact 26S complex . Transient transfection of an siRNA specific for PSMC5 into the human colon tumour cell line HCT116 resulted in >90% reduction in PSMC5 protein levels (Figure 1A). No change in PSMC5 expression was observed with a non-specific siRNA, or in mock-transfected cells (Figure 1A). An in-gel assay showed that, as expected, the PSMC5-specific siRNA caused a reduction in the concentration of the 26S proteasome, which was associated with elevation in the level of the 20S proteasome (Figure 1B). Importantly, the half-life of hTS, which was measured by CHX chase, increased by approx. 3–4-fold in cells transfected with the PSMC5-specific siRNA, as compared with either mock-transfected cells or cells expressing a non-specific siRNA (Figures 1C and 1D). Thus interference with the formation of the 26S proteasome results in stabilization of the hTS polypeptide, indicating that this form of the proteasome is required for degradation. Consistent with this conclusion is the observation that the half-life of hTS is unaffected by down-regulation of the 11S activator (K. Barbour, personal communication), a proteasomal regulator that is involved in degradation of ubiquitin-independent substrates, but lacks unfolding activity .
Disruption of the 19S complex impairs degradation of hTS
Absence of the proline-rich region spanning residues 9–15 is responsible for the degradation-resistance of mouse TS
In contrast with the human enzyme, mouse TS (denoted mTS) has N-acetyl-methionine at its N-terminus, reflecting the fact that it is resistant to methionine excision, yet is subject to N-α-acetylation [32,40]. We recently reported that mTS is a stable molecule, as expected for a protein that is modified by N-α-acetylation . However, a Leu2→Pro mutant of mTS, which eliminates the susceptibility of the protein to N-α-acetylation, is also stable, indicating that a free, non-acetylated N-terminus is not sufficient for degradation . To show that the degradation-resistant phenotype of mTS is due to its N-terminal domain, we prepared a construct that encodes a protein (m–hTS), having the N-terminal region of mTS ligated to the folded domain of hTS (see Figure 2A). The m–hTS construct was stably transfected into the Chinese hamster lung cell line RJK88.13 , and the half-life of the encoded polypeptide was determined by CHX chase. As shown in Figure 2(B), m–hTS was quite stable to degradation as compared with wild-type hTS; such a high level of stability was maintained when proline was substituted for leucine at position 2 within the chimaera (mutant m–hTS/L2P; Figure 2C). These results confirm the idea that a free, non-acetylated N-terminal end is necessary, but not sufficient, for hTS degradation. One or more residues that are located elsewhere within the disordered region, and that differ between the human and mouse enzymes, are responsible for the differential ability of the mouse and human N-termini to promote degradation.
The region corresponding to residues 9–15 of hTS contains a destabilizing element
As noted previously , and as depicted in Figure 2(A), the primary sequences of the intrinsically disordered domains of hTS and mTS are significantly different. The human sequence is 27 amino acids in length, and contains two proline-rich segments: one at residues 9–15 and one at 24–27. In contrast, the mouse sequence is 21 amino acids in length, and has no proline-rich segments. The two N-terminal domains show high sequence variability, with only a few residues being conserved (Figure 2A). It is noteworthy that alignment of the two indicates that mTS lacks the region corresponding to residues 11–15 of hTS (Figure 2A). Previous studies showed that deletion of residues 9–15 within hTS (sequence P9RRPLPP15) results in a very stable molecule; furthermore, conversion of each amino acid within this region into alanine (i.e. P9RRPLPP15→AAAAAAA) also stabilized the molecule . The apparent deletion of this sequence in mTS raises the possibility that it contributes to the differences in half-lives between the human and mouse polypeptides.
To test this, we inserted residues 9–15 of hTS into the chimaeric m–hTS/L2P polypeptide described above. The resulting molecule, denoted m–hTS/L2P(9–15)ins (Figure 2D), exhibited a half-life of 10±2 h, which is considerably shorter than that for the parental m–hTS/L2P polypeptide (Figure 2C). Thus the region spanning residues 9–15 of hTS functions as a destabilizing element, suggesting that the absence of this region in mTS contributes to the latter's degradation-resistant phenotype.
The Arg–Arg motif at residues 10–11 is required for degradation
In order to define more precisely which amino acids within region 9–15 are critical for degradation, we examined additional mutants. The region is proline-rich (four out of the seven residues are proline), which is a feature that is known to be important in protein–protein interactions [41,42]. To test the role of proline residues in the function of region 9–15, we produced a mutant in which each proline residue was converted into alanine. This mutant, termed hTS/P(9,12,14,15)A, was found to have a half-life of 5±1 h, which is similar to wild-type hTS (Figure 3A), indicating that proline can be replaced by alanine without disturbing the polypeptide's susceptibility to proteolysis. Thus proline residues at positions 9, 12, 14 and 15 are dispensable for degradation.
The Arg–Arg motif at residues 10–11 promotes degradation of hTS
We next tested the role of the two arginine residues at positions 10 and 11. A mutant containing alanine in place of each of these two residues was generated. This mutant, termed hTS/R(10,11)A, exhibited a half-life of 11±1 h, which is ~2-fold more stable than the wild-type polypeptide (Figure 3B). A more pronounced effect was observed when the two arginine residues were converted into glutamate. This mutant, denoted hTS/R(10,11)E, exhibited a half-life of 23±3 h, representing profound stabilization relative to wild-type hTS (Figure 3C). This is consistent with the stable phenotype of mutants in which residues 9–15 were either deleted or converted into alanine , and indicates that the Arg–Arg motif at positions 10–11 promotes degradation of the hTS polypeptide.
To assess the importance of the arginine residues themselves as opposed to their positive charge, we generated and analysed a mutant in which they were replaced by lysine. This mutant, termed hTS/R(10,11)K, exhibited a half-life of 17±1 h, indicating it to be ~3-fold more stable than wild-type hTS (Figure 3D). Thus it is the presence of arginine rather than positive charge, that is necessary for maximal degradation of the TS polypeptide.
Loss of the Arg–Arg motif is responsible for variations in TS degron function among mammalian species
Although the primary sequence of the N-terminal domain of TS is hypervariable among mammalian species, several features of the region are conserved, including its disordered nature, its high proline content, and the occurrence of proline at the penultimate site . This, along with the fact that the domain from mTS is inactive in promoting degradation (see above), led us to measure degron activities for the N-terminal regions from other mammalian species. We determined the half-lives of chimaeric TS molecules in which the disordered N-terminal domains of TS from several phylogenetically distinct mammals were fused to the folded domain of hTS. The disordered domains from cow, rabbit, platypus and armadillo were used to create interspecies chimaeric polypeptides c–hTS, r–hTS, p–hTS and a–hTS respectively (Figure 4A). As seen in Figures 4(B)–4(E), all of the chimaeric molecules exhibited very stable phenotypes, with half-lives >24 h. Thus, similar to m–hTS, these N-terminal regions lack one or more essential elements required for degron activity when fused to hTS.
The Arg–Arg motif is a destabilizing element
The Arg–Arg dipeptide at residues 10–11 of hTS is either deleted or mutated in each of the four mammalian species examined (Figure 4A), providing a possible explanation for the inability of their N-terminal domains to efficiently promote degradation. To test this, we inserted Arg–Arg motifs into the interspecies chimaeric polypeptides, and determined their effects on stability. First, we introduced a Ser→Arg substitution at residue 9 of c–hTS, generating a construct, termed c–hTS/S9R, having Arg–Arg at positions 9–10. We also introduced arginine in place of proline at position 11 within c–hTS, resulting in a protein (c–hTS/P11R) containing Arg–Arg at residues 10–11. Both the c–hTS/S9R and c–hTS/P11R polypeptides were found to be considerably less stable than c–hTS (Figures 4F and 4G). Thus the presence of the Arg–Arg motif within the disordered N-terminal domain of cow TS makes the domain effective in promoting degradation, indicating that it is responsible for the inter-species difference in protein stability.
Further evidence in support of this conclusion was obtained with the r–hTS chimaera. A Pro–Arg–Arg tripeptide was inserted into r–hTS between amino acids 8 and 9, resulting in an Arg–Arg motif at residues 9–10. This mutant, termed r–hTS/(PRR)ins, was unstable relative to r–hTS (Figure 4H), indicating that the presence of an Arg–Arg motif promotes the ability of the N-terminal region from rabbit TS to mediate degradation.
Taken together, these results indicate that introducing an Arg–Arg motif into the N-terminal domains of the TS molecules from non-primate mammalian species converts these domains into active degrons. Thus the absence of this motif contributes significantly to the stable phenotype of TS from these species.
Stabilizing mutations act on post-docking steps in the proteasomal degradation pathway
Recent studies by Takeuchi et al.  showed that specific residues within the C-terminal degron of ODC mediate docking to the proteasome, and that the role of these residues can be by-passed by fusion to proteasomal subunit Rpn10, which provides an alternative docking capability. Thus the ability of Rpn10 to ‘rescue’ degradation in stabilized mutants provides an effective means of testing whether or not those mutants affect proteasome docking. We used such a strategy to determine whether mutational changes within the disordered domain of the hTS degron perturb docking, as opposed to one or more post-association steps, in the degradation pathway.
We generated chimaeric molecules in which proteasomal subunit PSMD4 (the mammalian homologue of Rpn10, also known as S5a) was ligated to the C-terminal end of wild-type hTS; PSMD4 was also ligated to the P2A and A(9–15) mutants, both of which are stable to degradation . The half-lives of the resulting chimaeric polypeptides were determined. Among the molecules lacking PSMD4, the P2A and A(9–15) mutants were, as expected, significantly more degradation-resistant than the wild-type protein (Figure 5A). Importantly, although all three PSMD4 chimaeras had shorter half-lives relative to their parental counterparts, the P2A and A(9–15) chimaeras were significantly more stable than the wild-type chimaera (Figure 5B). Pretreatment with lactacystin stabilized each of the chimaeric molecules, indicating that their degradation is proteasome-mediated (results not shown). Thus the stabilizing impacts of the P2A and A(9–15) substitutions are not ‘rescued’ by fusion to PSMD4, but are maintained in the presence of the alternative docking site. This indicates that the substitutions exert their stabilizing impacts via perturbation of one or steps subsequent to substrate binding to the proteasome.
Stabilizing mutations affect a post-association step in the proteasomal degradation pathway
Removal of large segments of the N-terminal domain leads to alterations in the mechanism of hTS degradation
Previously, we observed that mutant hTS molecules missing large segments of the N-terminal region, such as the first 13 or 23 residues, had half-lives that were similar to or shorter than that for the wild-type enzyme . To verify and extend these observations, we measured the intracellular stability of a mutant protein in which residues 7–29, comprising most of the disordered region, were deleted. We fully expected that this mutant, termed hTS/Δ(7–29), would be resistant to proteolysis. However, much to our surprise, it was quite susceptible to degradation, having a half-life of 2±0.6 h (Figure 6A). The molecule was stabilized by lactacystin, indicating that its degradation is indeed proteasomal (Figure 6B). Furthermore, as shown in Figure 6(C), it was also stabilized by exposure to FdUrd, an analogue that is metabolized to 5-fluoro-2′deoxyuridylic acid, which binds to and stabilizes the TS polypeptide . The markedly unstable nature of hTS/Δ(7–29) led us to examine the properties of this enzyme in more detail.
Effect of removal of residues 7–29 on hTS degradation
In previous studies we showed that modifying the N-terminal end of wild-type hTS by N-α-acetylation inhibits the degradation of the polypeptide, whereas introducing a His5 tag at the C-terminal end has little effect [32,33]. One interpretation of this observation is that N-α-acetylation blocks the ability of the disordered region of hTS to serve as the initiation point of degradation, and promote proteolytic breakdown from the N- towards the C-terminus [32,33]. To test whether this is the case for the Δ(7–29) deletion mutant, we blocked the N-terminal end of hTS/Δ(7–29) by substituting alanine in place of proline at residue 2, making the protein susceptible to N-α-acetylation [32,33]. This molecule, termed hTS/Δ(7–29)P2A, was found to have a half-life similar to that of its unmodified parent (Figure 7A). We then blocked the C-terminal end of hTS/Δ(7–29) with a His5 tag . This molecule, termed hTS/Δ(7–29)–His5, was much more stable than its parent (Figure 7A). Thus degradation of the hTS/Δ(7–29) mutant is inhibited by blocking its C-terminus, but not its N-terminus, which is opposite to what is observed for wild-type hTS, and indicates a fundamental difference between the two proteins in the mechanism of proteolysis.
Effect of large deletions within the N-terminal domain of hTS
To assess the generality of these results, additional deletion mutants were examined. Similar to our previous findings , mutants hTS/Δ(2–13) and hTS/Δ(2–23), which harbour deletions of residues 2–13 and 2–23 respectively, exhibited half-lives of 1±0.1 and 10±2 h respectively (Figures 7B and 7C). Both were stabilized by introducing a His5 tag at their C-termini (Figures 7B and 7C), again indicating a degradation mechanism that is similar to hTS/Δ(7–29), but distinct from that for the wild-type enzyme.
To test whether the altered degradation phenotypes of mutants hTS/Δ(7–29), hTS/Δ(2–13) and hTS/Δ(2–23) are due to simple shortening of the N-terminal domain, we constructed and analysed a mutant lacking residues 16–29. This mutant, referred to as hTS/Δ(16–29), has an N-terminal domain that is shorter than that of both hTS/Δ(7–29) and hTS/Δ(2–23). Blocking its C-terminus had little, if any, effect on its intracellular stability; in contrast, blocking its N-terminus resulted in significant stabilization (Figure 7D). Therefore the hTS/Δ(16–29) mutant is similar to the wild-type protein, indicating that the altered mechanism observed for the 7–29, 2–23 and 2–23 deletion mutants is not due to a simple reduction in the length of the N-terminal domain.
Overall, therefore, our results indicate that deletion of region 7–29, 2–13 or 2–23 within the N-terminal domain preserves susceptibility to degradation, but causes that degradation to occur by a different mechanism from wild-type hTS. It is interesting to note that these three deletions have in common the loss of residues 7–13, which overlaps with the P9RRPLPP15 region that was shown to be important to the degradation phenotype (see above).
Substrates targeted for proteosomal degradation must have a means of recognizing and docking to the protease complex. Once proteasome-bound, they must unfold, gain entry to the chamber formed by the 20S core and ‘thread’ their way through the chamber during proteolytic processing [4–6,13,28]. For many substrates, ubiquitin modification provides the docking capability, whereas an extended disordered domain supplies the post-association functions [1–3]. In the case of hTS, the first 45 N-terminal amino acids, spanning a 30-residue unstructured domain followed by an amphipathic α-helix, provides both of these elements (i.e. docking and post-docking functions) in the absence of a requirement for ubiquitinylation [31,33].
We have shown in the present study that the degradation signal of hTS targets the enzyme to the 26S isoform of the proteasome, as shown by inhibition of proteolysis upon interference with assembly of the 19S regulatory particle. This is not unexpected, since the highly structured character of hTS makes it likely that the ATPases within the 19S complex are essential for unraveling it and promoting its processive entry into the 20S core chamber.
We used targeted mutagenesis to demonstrate that the Arg–Arg dipeptide at residues 10–11 within the disordered N-terminus is a critical element required for maximal degradation of the protein. This conclusion is supported by results of analysis of TS from several mammalian species, including mouse, cow and rabbit, the N-terminal regions of which lack the Arg–Arg motif, and have weak, if any, degron activity. Introducing an Arg–Arg motif into these domains activates degron function (Figures 2 and 4). The finding that a dipeptide near the N-terminal end of hTS plays such a central role in regulating degradation is reminiscent of the situation with ODC. For the latter, mutation of a Cys–Ala sequence at amino acids 441–442 located ~22 residues from the C-terminal end impairs the degradation of the polypeptide [13,14]. However, the ODC and hTS dipeptide motifs act in mechanistically different manners. The Cys441 substitution within the ODC degron causes a defect in proteasome association, as indicated by ‘rescue’ of the substitution's effect following fusion to Rpn10, an alternative docking element . In contrast, a similar approach does not overcome the stabilizing effects of mutations within the hTS degron (see Figure 5), indicating that these substitutions affect one or more post-association steps, such as positioning for entry into the proteasomal chamber, or entry into the chamber itself.
It is likely that the ODC degron, and other unstructured regions that mediate proteasomal degradation, contains a specific subregion that, like the Arg–Arg dipeptide of hTS, mediates steps that follow substrate binding to the proteasome. Consistent with this model, proteasomal degradation of the enzyme UbcH10, which is ubiquitin-dependent, requires a direct interaction between its unstructured N-terminus and the proteasome, thereby triggering post-association steps leading to proteolysis .
Although we have not yet identified the site of hTS docking to the proteasome, it is likely that one or more subunits within the 19S regulatory complex, which has the ability to bind unstructured regions and promote unfolding of ordered domains [4,7,8], is involved. Clearly additional work, including development of an in vitro system that mimics the primary features of hTS degradation, will be required.
As described previously, and verified in the present study, removal of most of the N-terminal domain does not confer a stable phenotype to the hTS polypeptide, despite the function of the region as a degron. Deletion of residues 7–29, 2–13 or 2–23, each of which removes a large segment of the region, resulted in proteins with half-lives that are shorter than that for wild-type hTS. By examining the effects of alternately blocking the N- or C-terminus, we determined that degradation of these three mutants occurs by a mechanism that is distinct from the wild-type enzyme. These deletions, although spanning distinct segments of the N-terminal region, have in common the loss of residues 7–13, which overlaps the P9RRPLPP15 sequence discussed above. However, deletion of the latter region alone is not enough to elicit a change in the mechanism of degradation , indicating that loss of additional flanking residues are required. Apparently, large deletions that include the region spanning amino acids 9–15 disarm the normal, wild-type mode of degradation, which is inhibited by blocking the N-terminus of the protein; at the same time, a distinct mechanism that is inhibited by blocking the C-terminus is ‘unmasked’ or activated. This is similar to what was observed previously for the P303L mutant, whose degradation is inhibited by modifying the C-terminus, but not the N-terminus .
It is clear, therefore, that hTS undergoes proteasome-catalysed degradation by at least two mechanisms that are distinguished on the basis of modifying the N- or C-terminus. Although the biochemical nature of these mechanisms is not known, it is interesting to speculate that the two differ in the directionality of proteolysis. Although protein substrates may enter the proteasomal chamber via internal loops [45,46], it is common for degradation to be initiated by passage of one or the terminus of the polypeptide into the 20S chamber, after which proteolysis progresses linearly toward the distal end of the polypeptide [7,8]. The observation that degradation of wild-type hTS is inhibited by blocking its N-terminus, may reflect an N-to-C polarity. In contrast, deletion mutants del(7–29), del(2–13) and del(2–23), which are all stabilized by blocking their C-termini, may undergo degradation with a C-to-N polarity. Such a switch from N- to C-directed degradation may reflect an ‘unmasking’ of the C-terminal end, allowing it to drive proteasomal degradation of the molecule.
It is interesting to note that c-Fos, which is degraded by both ubiquitin-dependent and -independent means, has degradation signals at both the N- and C-termini [16,47]. The direction of c-Fos proteolysis has not been determined; however, it may be that the two signals drive degradation from opposite ends of the polypeptide.
In conclusion, the findings of the present study indicate that distinct sequence elements within the disordered N-terminal region of hTS mediate its degradation by the 26S proteasome. Both a free N-terminal end, as well as the region spanning residues 9–15 (in particular, the Arg–Arg dipeptide at residues 10–11), appear to be involved in post-association steps in the proteasomal pathway. Such steps might include positioning for entry into the proteolytic chamber, or entry itself. We surmise that one or more elements mediate the docking step, similar to what has been demonstrated for the ODC degron [13,14]. Identifying and characterizing those elements remains an issue for future study.
Finally, it should be emphasized that the half-life of hTS, as a determinant of the intracellular concentration of the enzyme, as well as the rate at which that concentration changes in response to repressing agents such as siRNAs and antisense oligonucleotides [48–50], is likely to be a critical factor in the role of the enzyme as a chemotherapeutic target. Thus further understanding of the mechanism by which that degradation occurs may result in novel strategies to destabilize the enzyme, and increase the efficacy of TS-targeted agents [29,30].
Sandra Melo and Franklin Berger conceived the project, designed the experiments and analysed results. Sandra Melo and Asami Yoshida carried out the experiments. Sandra Melo's contributions to the present study were in partial fulfilment for a Ph.D. degree in the Department of Chemistry and Biochemistry at the University of South Carolina, Columbia, SC, U.S.A.
We thank Dr Marj Peña and Ms Karen Barbour for advice and feedback. We also thank Ms Yang Yang Xing for technical help with some of the experiments.
This work was supported by the National Cancer Institute of the National Institutes of Health [grant number CA044013].
Present address: Stanford University School of Medicine, 269 Campus Drive CCSR 2140, Stanford, CA 94305, U.S.A.
Present address: Graduate School of Science and Technology, Nagasaki University, 1-14 Bunkyo, Nagasaki 852-8521, Japan