Polyubiquitin chains serve a variety of physiological roles. Typically the chains are bound covalently to a protein substrate and in many cases target it for degradation by the 26S proteasome. However, several studies have demonstrated the existence of free polyubiquitin chains which are not linked to a specific substrate. Several physiological functions have been attributed to these chains, among them playing a role in signal transduction and serving as storage of ubiquitin for utilization under stress. In the present study, we have established a system for the detection of free ubiquitin chains and monitoring their level under changing conditions. Using this system, we show that UFD4 (ubiquitin fusion degradation 4), a HECT (homologous with E6-AP C-terminus) domain ubiquitin ligase, is involved in free chain generation. We also show that generation of these chains is stimulated in response to a variety of stresses, particularly those caused by DNA damage. However, it appears that the stress-induced synthesis of free chains is catalysed by a different ligase, HUL5 (HECT ubiquitin ligase 5), which is also a HECT domain E3.
- free ubiquitin chains
- homologous to E6-AP carboxy terminus ubiquitin ligase 5 (HUL5)
- methyl methanesulfonate (MMS)
- 26S proteasome
- ubiquitin fusion degradation 4 (UFD4)
Polymers of Ub (ubiquitin) are formed on various protein substrates in eukaryotic cells by the concerted action of three enzymes: the Ub-activating enzyme (E1), a Ub-carrier protein [E2; known also as UBC (Ub-conjugating enzyme)] and a Ub ligase (E3) which is the specific substrate-recognizing element of the system. There are two major classes of E3s: the HECT (homologous with E6-AP C-terminus) domain and RING domain E3s. The two families differ in their mechanism of substrate conjugation. Conjugated Ub can be removed by various DUBs (deubiquitinating enzymes) . The polyubiquitin chain formed may serve several physiological roles, such as targeting the tagged proteins for proteasomal degradation, regulating the activity of DNA repair mechanisms and activating transcriptional factors.
In the majority of cases polyubiquitin chains are anchored to a protein substrate. However, it was previously shown that free polyubiquitin chains can be formed both in vitro [8,9] and in cells . Although the function of the free chains has remained largely unknown, it was suggested that they can play a role in signal transduction  and serve as storage for a large amount of free monomeric Ub released under stress . It was also shown that free Ub chains can be conjugated directly to target substrates [13,14]. The potential source of free chains can vary: they can be released, intact or partially intact, from the target substrate in a reaction mediated by a DUB, or they can be synthesized de novo using monomeric Ub as a substrate.
Ub can be degraded via several mechanisms, one of them probably targets monomeric Ub which is degraded following its E3-mediated ubiquitination [3,15], thus generating a free Ub chain as a proteolytic intermediate. One such E3 ligase, TRIP12 (thyroid hormone receptor interactor 12), which can polymerize monomeric Ub, has been described , although additional enzymes may exist.
In the present study, we have established a yeast-based system for the generation and monitoring the formation of free Ub chains. We used the system to study their synthesis under basal and stressed conditions, and to identify two ligases that are involved in these distinct processes, UFD (Ub fusion degradation) 4 and HUL5 (HECT Ub ligase 5) respectively.
MATERIAL AND METHODS
The reagents for SDS/PAGE and the Bradford assay were from Bio-Rad Laboratories. Mouse anti-HA (haemagglutinin) antibody (16B12) was from Covance. Rabbit anti-Ub-conjugates antibody was described previously . Peroxidase-conjugated secondary antibodies were from Jackson ImmunoResearch Laboratories. Anti-HA affinity matrix and protease inhibitors (mixture) were from Roche Molecular Biochemicals. Anti-actin antibody was from Millipore. Anti-MyoD (myogenic differentiation) antibody was from Santa Cruz Biotechnology. Cycloheximide and zirconia beads were from Sigma. The reagents for ECL (enhanced chemiluminescence) were from Pierce. Restriction enzymes were from New England Biolabs. Real-time PCR master mix was from Thermo Scientific. Ni-NTA (Ni2+-nitrilotriacetate) agarose beads were from Qiagen. All of the other reagents were of high analytical quality.
For the detection of free Ub chains, we used the BY4741 yeast strain (Open Biosystems). For identification of the E3 ligases involved in the formation of free Ub chains, we used a library of viable E2(s)/E3(s) gene deletion strains (generated by Open Biosystems and selected by Dr Mark Hochstrasser, Yale University, New Haven, CT, U.S.A.), and a subset of a library of yeast strains with tetracycline-induced silencing of essential Ub system genes (Open Biosystems). The construction of a Δufd4Δhul5 strain was based on a Δufd4 strain (Open Biosystems, YSC1053 collection, record 4859), and was carried out by PCR-mediated replacement of HUL5 with a hul5Δ::LEU2 allele. For HUL5 overexpression and HUL5 rescue experiment, we used the W303 yeast strain . The construction of Δhul5 was carried out similarly to that of the double knockout strain.
All of the plasmids were constructed and manipulated using standard molecular biological techniques. For yeast transformation, cDNAs were subcloned into the following vectors: UbVV–HA (UbG75,76V–HA), lysine-less UbVV–HA (K0UbVV–HA), HA–UbVV, HA–K0UbVV, HA–Ub and MyoD into p416; UFD4 and UFD4C1450S into pRS313; and HUL5, HUL5C878A, FLAG–HUL5 and FLAG–HUL5C878A into pJS92. For the transient transfection of HEK (human embryonic kidney)-293 cells, UbVV–RGSHis6 (UbVV–His) or K0UbVV–His were subcloned into the pCAGGS vector .
Protein concentration measurement
Protein concentration was measured by the Bradford assay  using BSA as the standard.
Detection of proteins of interest was carried out following their resolution via SDS/PAGE (13.7% gel) and Western blotting using the appropriate antibodies.
Strains of Saccharomyces cerevisiae were transformed with plasmids coding for the proteins of interest, using the PEG/LiAc [poly(ethylene glycol)/lithium acetate] method as described previously .
A SYBR Green real-time PCR method was employed for the quantification of gene overexpression.
Immunoprecipitation of HA-tagged proteins from yeast
Yeast cells were transformed with p416 expressing the indicated HA-tagged proteins under the GAL promoter. A 10-ml yeast culture was grown overnight in YPR medium (1% yeast extract, 2% tryptone and 2% raffinose) at 30°C. Next, protein expression was induced by adding 2% galactose and the cells were incubated for an additional 2 h at 30°C. The cells were collected by centrifugation (3600 g for 2 min at 4°C), washed twice in PBS (150 mM NaCl, 3 mM KCl, 10 mM Na2HPO4 and 2 mM KH2PO4, pH 7.4) and frozen in liquid nitrogen. Lysis was performed by the addition of 400 μl of lysis buffer (50 mM Tris/HCl, pH 7.6, 150 mM NaCl, 0.1% Nonidet P40, 10 mM N-ethylmaleimide, 10 mM iodoacetamide and protease inhibitor mix) and 2 g of 0.5 mm zirconia beads. The tubes were shaken at 4°C for 10 min and the resulting lysates were cleared by centrifugation at 17000 g for 30 min at 4°C. Extract protein (3 mg) was incubated with 10 μl of packed anti-HA affinity matrix for 2 h at 4°C, and washed twice with lysis buffer. HA-tagged proteins were eluted by adding 15 μl of SDS/PAGE sample buffer (10% glycerol, 62.5 mM Tris/HCl, 2% SDS, 0.01 mg/ml Bromophenol Blue and 5% 2-mercaptoethanol, pH 6.8) and incubated at 95°C for 5 min.
Ni-NTA pull down of His-tagged proteins
HEK-293 cells were transfected with pCAGGS expressing the indicated His-tagged proteins. Proteins were precipitated using immobilized Ni-NTA as described previously . Briefly, 100% confluent cells from 35-mm plates were lysed in buffer A (6 M guanidine-HCl, 0.1 M each of Na2HPO4 and NaH2PO4, 0.01 M Tris/HCl, 5 mM imidazole and 10 mM 2-mercaptoethanol, pH 8.0). The extracts were incubated with 20 μl of packed Ni-NTA beads for 2 h at 4°C. The beads were washed with buffers A, B (8 M urea, 0.1 M each of Na2HPO4 and NaH2PO4, 0.01 M Tris/HCl, 10 mM imidazole and 10 mM 2-mercaptoethanol, pH 8.0) and C (8 M urea, 0.1 M each of Na2HPO4 and NaH2PO4, 0.01M Tris/HCl, 10 mM imidazole and 10 mM 2-mercaptoethanol, pH 6.3). The bound proteins were eluted with buffer D (300 mM imidazole, 0.15 M Tris/HCl, 30% glycerol, 0.72 M 2-mercaptoethanol and 5% SDS, pH 6.7).
Determination of protein stability in yeast cells
Yeast cells were transformed with cDNAs encoding either galactose-induced UbVV–HA or galactose-induced MyoD, and grown overnight in YPG medium (1% yeast extract, 2% tryptone and 2% galactose) at 30°C. Next, the culture was diluted to 5.0 D600nm and any further protein synthesis was inhibited by adding 0.1 mg/ml cycloheximide and 2% glucose. At each time point, a 10-ml sample was removed and free chains were isolated as described above. Proteins were resolved via SDS/PAGE (13.7% gel), blotted on to PVDF membrane and visualized using anti-Ub-conjugates, anti-HA or anti-MyoD antibodies. Actin was used as the loading control.
Stress induction in yeast
Yeast cells were transformed and induced for expression of UbVV–HA as described above. Prior to centrifugation, cultures were subjected to one of the following treatments: (i) heat treatment at 40°C for 1 h; (ii) exposure to 0.1% MMS (methyl methanesulfonate) for 2 h; (iii) exposure to 0.5 mM H2O2 for 1 h; or (iv) exposure to 60 μM CdCl2 for 1 h.
Free Ub chains are formed in cells
To demonstrate the formation of free polyubiquitin chains, we transformed yeast cells with UbVV–HA and subjected them to anti-HA immunoprecipitation followed by anti-Ub-conjugates immunoblot analysis. The G75,76V double mutation was introduced to prevent this Ub species from serving as a donor in the conjugation reaction (it can serve only as a Ub acceptor) and to prevent the cleavage of the HA tag by DUBs [17,18]. As shown in Figure 1(A) (lane 1), free polyubiquitin chains that are anchored to HA-tagged Ub were generated. The chains were probably based on internal lysine residue(s) in the UbVV–HA substrate, as its lysine-less counterpart could not serve as an acceptor (lane 2). To assess the formation of free chains in an additional biological context, we transfected HEK-293 cells with UbVV–His. As displayed in Figure 1(B), and in agree-ment with our results in yeast cells (Figure 2A), free Ub chains were generated. This generation was dependent on an internal lysine residue, probably Lys48, in the acceptor molecule.
The UFD4 ligase catalyses the formation of free Ub chains
To identify the conjugating enzymes involved in free chains formation, we used a yeast library of viable genomic deletions where in each strain a specific E2 or E3 gene was deleted (Table 1; selected from the Open Biosystems library). These strains were monitored for their ability to generate free chains. As can be seen in Figure 2(A), the level of free chains varied among the different strains. However, a consistent marked decrease was detected in the ΔUFD4 strain, accompanied by a parallel increase in the level of monomeric UbVV–HA. It should be noted that the screen was broadened to include a library of yeast strains with conditional silencing (tetracycline induced) of essential components of the Ub system. No additional factors involved in free Ub chain formation could be identified in this extended screen (results not shown).
Consistently, an increase in the level of free chains was observed in the ΔUBC8 E2-deleted strain, in conjunction with a parallel decrease in the level of the monomeric UbVV–HA (see the Discussion section).
UFD4 is a 167 kDa HECT-domain E3 ligase that was shown to target the E2 Ubc7  and the DNA alkyltransferase Mgt1 (O-6-methylguanine-DNA methyltransferase)  for proteasomal degradation. In addition, it was recently shown to interact with the N-end rule E3 UBR1, leading to enhanced activities of both the N-end rule and the UFD pathways . The respective human homologue, TRIP12, was shown to assemble free chains both in vitro and in cells .
As a component of the UFD pathway, it was reported that UFD4 recognizes specifically protein substrates with an N-terminal Ub fusion . Thus it may not be surprising that the UbVV–HA model substrate, where Ub is N-terminally fused to an HA peptide, is recognized by UFD4. However, we found that UFD4 can conjugate also Ub C-terminally fused to an HA tag (Figure 2B), suggesting that it recognizes mostly Ub itself.
It should be noted that deletion of UFD4 does not abolish completely the generation of free chains, generated either on N-terminally (Figure 2A, lane 2) or C-terminally (Figure 2B, lane 3) fused Ub, suggesting that other ligase(s) may also be involved in this reaction.
To further corroborate the involvement of UFD4 in the generation of free Ub chains, we carried out a rescue experiment. ΔUFD4 yeast cells were transformed with a cDNA encoding HA–UFD4 (pUFD4). As shown in Figure 2(C), pUFD4 restored, to a large extent, the WT (wild-type) phenotype (compare lane 3 with lane 2). In contrast, the catalytic inactive mutant pUFD4C1450S did not display a rescue effect (compare lane 4 with lane 3). It should be noted that the different HA–UFD4 constructs were equally expressed (results not shown). These results clearly show that UFD4 is involved in the formation of free Ub chains in a manner that requires the catalytic activity of its HECT domain.
UFD4 does not mediate degradation of Ub
To monitor the effect of ubiquitination of Ub on its stability, we followed UbVV–HA degradation in either the WT or ΔUFD4 yeast strains. As can be seen in Figure 3, Ub is stable in the presence or absence of UFD4 (lanes 1–4 and 5–8 respectively).
Ub is a relatively stable protein, probably because its entry into the proteasome 20S catalytic particle is inefficient [3,22]. The addition of a tail larger than 20 residues to the C-terminus of Ub renders it susceptible to rapid degradation without further ubiquitination . Therefore it has not been surprising to find that UbVV–HA is stable regardless of the presence or absence of UFD4. Thus it is possible that free Ub chains serve non-proteolytic function(s).
The level of free Ub chains is elevated during stress
Searching for possible physiological roles of free chains, we monitored their level following exposure of yeast cells to various types of stress (see the Materials and methods section). As can be seen in Figure 4, free chain formation was stimulated following heat shock (Figure 4Ai), DNA damage (Figure 4Aii) and oxidative stress (Figures 4Aiii and 4Aiv). The most prominent effect was observed following treatment with the DNA-alkylating agent MMS, which is known to induce DNA damage . To assess the specificity of the increase, it was important to examine the effect of MMS on the entire Ub conjugates pool. Towards this end we transformed yeast cells with HA–Ub (the Ub is from the WT species). The resulting conjugates were immunoprecipitated using anti-HA antibody and identified by blotting using anti-Ub-conjugates antibody. In addition, total Ub conjugates were detected by a simpler method, using anti-Ub-conjugates Western blotting of crude yeast extract. As shown in Figure 4(B), MMS did not have any effect on the cellular Ub pool, suggesting that its effect is specific for free Ub chains formation.
The Ub ligase HUL5 is involved in MMS-induced formation of free Ub chains
Searching for enzymes involved in the MMS induction of the synthesis of free chains, the obvious candidate to start with was UFD4. To check if UFD4 is indeed involved in stress-induced free chain formation, we assayed their induction in WT and ΔUFD4 strains following MMS treatment. As demonstrated in Figure 5(A), although the overall level of free chains was lower in ΔUFD4 yeast cells, a marked stress-induced increase could still be observed. This result indicates that there are additional factors, most probably Ub ligases, involved in free chain formation, and suggests the existence of distinct ligases that differentially regulate basal and stress-induced synthesis of these chains.
As UFD4 appears to regulate the basal synthesis of free chains, we searched for the enzyme(s) involved in induction of their formation under stress. To do so, we explored the yMGV (yeast microarray global viewer) database (http://transcriptome.ens.fr/ymgv/) for genes that are induced by MMS and are associated with the Ub system. Out of the 15 genes whose expression was found to increase following MMS treatment, the most prominent was HUL5, a HECT domain Ub ligase. As shown in Figure 5(B), the MMS-induced formation of free chains was significantly lower in the ΔHUL5 strain compared with its WT counterpart. In contrast, the basal level of free chains was unaffected by the deletion of HUL5. Specifically, MMS treatment of ΔHUL5 yeast cells less efficiently induced the formation of high molecular mass free chains. This observation is consistent with the characterization of HUL5 as a chain-elongating factor .
HUL5 is a 105 kDa HECT domain Ub ligase. It is associated with the proteasome and interacts with the DUB Ubp6 (Ub-specific protease 6) which is also associated with the proteasome . Additionally, it is involved in the ERAD (endoplasmic reticulum-associated degradation) pathway , and in the degradation of cytosolic misfolded proteins . Importantly, it has never been reported that HUL5 is associated with DNA-damage-response pathways.
To further corroborate that HUL5 is indeed involved in stress-induced free chain formation, we conducted a rescue experiment. ΔHUL5 yeast cells were transformed with cDNA coding for FLAG–HUL5, treated with MMS and the level of their free chains was assessed. As shown in Figure 5(Ci), HUL5 could rescue, though modestly, the induction of the formation of free chains following stress. In contrast, the catalytic-inactive HUL5C878A had no effect. The partial rescue may be due, for example, to the inability to regulate the level of the enzyme that may itself be regulated by the stress.
These results suggest a possible model for stress-induced formation of free Ub chains. According to this model MMS induces HUL5 expression, and the elevated level of HUL5 facilitates the formation of free Ub chains. If the induction would have been solely dependent on HUL5, the requirement for MMS could have been bypassed by overexpression of HUL5. To examine this hypothesis, WT yeast cells were transformed with HUL5 overexpressed under the CUP1 (copper metallothionein 1) promoter, and the level of free chains was monitored. As demonstrated in Figure 5(Cii), HUL5 overexpression did not affect the level of free chains. It should be noted that HUL5 transcription was stimulated 20-fold as determined by real-time PCR (results not shown).
As it is possible that both UFD4 and HUL5 can contribute to stress-induced free chain formation, it was important to check whether additional factors are also involved. Accordingly, we tested induced free chain formation in a strain that harbours genomic deletions of both UFD4 and HUL5. As can be seen in Figure 5(D), a residual induction could still be detected in the double mutant. However, its low level suggests that the two ligases are the key players in this process.
Free chains are not essential for yeast proliferation following MMS treatment
At that point we made an initial attempt to identify the physiological role of the free chains. We followed yeast growth in the presence and absence of MMS using a standard drop-test. MMS had no differential effect on either the ΔUFD or ΔHUL5 strain, compared with the WT strain (results not shown).
In the present study, we identified two HECT domain-containing E3 ligases, UFD4 and HUL5, that are involved in the formation of free Ub chains under basal and stressed conditions. The finding that the two enzymes belong to the HECT domain-containing family is intriguing, as the yeast proteome contains only five HECT domain proteins, compared with more than 25 RING and U-box domain-containing proteins. The specific recognition of one small molecule such as Ub by two different ligases may be surprising, and reflect differential expression, activation or availability of the ligases under different pathophysiological conditions. In searching for the E2 involved, we could not identify any particular enzyme that is involved in free chain formation (see for example Figure 2A). However, we noted that deletion of UBC8 leads to increased generation of free chains. It is possible that in the absence of UBC8, more free Ub that otherwise would have been used by UBC8, is available for the formation of free chains.
It is interesting to note that both UFD4 and HUL5 have been shown to interact with the proteasome [27,28]. This may imply a role for free chains in modulating proteasomal activity. In this context, it was previously demonstrated that free Ub chains are able to inhibit proteasomal degradation, probably by competing with substrate-bound chains for the proteasomal recognition elements .
As we have shown, MMS treatment stimulates significantly free chains generation (Figure 4Aii). Remarkably, we have found that exposure to MMS increases transcription of both UFD4 and HUL5 using real-time PCR (results not shown), and genomic assays have found a corresponding increase in HUL5 . Since both enzymes are proteasome associated, we speculated that their expression may be enhanced concomitantly by RPN4 (regulatory particle non-ATPase 4), a major stress-induced transcription factor that regulates proteasomal subunits expression . However, in an in silico search, we were unable to find any putative RPN4-binding sites in the proximity of the UFD4 and HUL5 loci. Thus the up-regulation of these genes may be facilitated by a different transcription factor, or by an uncharacterized RPN4-binding site.
It appears that the effect of MMS is more complicated than just inducing HUL5-mediated formation of free chains under stress. As shown in Figure 5(Cii), overexpression of HUL5 cannot mimic MMS in inducing free chain formation, suggesting that MMS exerts additional molecular effects. An example of such an effect can be protein methylation that may modulate enzymatic activity and specificity . Additionally, other factors that are MMS-induced can probably play a role as well. Therefore the overexpressed HUL5 molecules are stoichiometrically limited by the lack of these putative partners.
Taken together, the data from the present study uncover certain aspects of the biochemistry underlying the formation of free Ub chains. We suggest the involvement of free Ub chains in the DNA-damage response. As free Ub chain formation can also be regarded as ubiquitination of Ub, this process is probably important in the context of the autoregulation of the Ub system, and particularly in storing Ub and targeting it for degradation.
Ori Braten designed and performed experiments, analysed results and wrote the paper; Nitzan Shabek designed experiments and analysed results; Yelena Kravtsova-Ivantsiv provided valuable technical support and advice; Aaron Ciechanover conceived many of the ideas, accompanied and followed the entire study and, along with Ori Braten, wrote the paper.
This work was supported by the Dr Miriam and Sheldon Adelson Foundation for Medical Research (AMRF) [grant number 2009548], the Israel Science Foundation (ISF) [grant number 2015340], and the Deutsch–Israelische Projektkooperation (DIP) [grant number 2014378]. A.C. is an Israel Cancer Research Fund (ICRF) U.S.A. Professor [grant number 2015216].
We thank Dr Tommer Ravid (Hebrew University of Jerusalem, Jerusalem, Israel) and Dr Mark Hochstrasser (Yale University, New Haven, CT, U.S.A.) for their gifts of the yeast E3-deleted and E3-silenced libraries and the cDNAs encoding the different UFD4 species; Dr Daniel Kornitzer for the W303 yeast strain and the cDNAs encoding the different HUL5 species; and Dr Yifat Herman-Bachinsky for her technical assistance.
Abbreviations: DUB, deubiquitinating enzyme; HA, haemagglutinin; HECT, homologous with E6-AP C-terminus; HUL5, HECT ubiquitin ligase 5; HEK, human embryonic kidney; MMS, methyl methanesulfonate; MyoD, myogenic differentiation; Ni-NTA, Ni2+-nitrilotriacetate; RPN4, regulatory particle non-ATPase 4; TRIP12, thyroid hormone receptor interactor 12; Ub, ubiquitin; UBC, Ub-conjugating enzyme; UFD, Ub fusion degradation; WT, wild-type
- © The Authors Journal compilation © 2012 Biochemical Society