Functional Roles of the E3 Ubiquitin Ligase UBR5 in Cancer

Ubiquitination is one of the most widespread and frequent cellular posttranslational modifications (PTM) and is essential for normal cellular functions. Ubiquitination can regulate protein levels via degradation by the proteasome (1, 2), regulate protein–protein interactions, and modulating protein function and localization (3, 4). The small protein modifier Ubiquitin (Ub) forms an isopeptide bond with substrate proteins via acceptor lysine residues in a hierarchical enzymatic reaction catalyzed by an E3 Ub ligase (5). Much variety in Ub modification of substrate proteins arises from the ability of Ub to form either single conjugates (mono-ubiquitination) or chains (poly-ubiquitination) after repeated rounds of Ub attachment (6). Further complexity and diversity in Ub PTMs is generated by variation in Ub chain topology, depending on which Ub lysine residue is further ubiquitinated on the growing poly-Ub chain. Various poly-Ub chain topologies are known to effect different signaling outcomes via binding to specific adaptor proteins [reviewed elsewhere (4)]. Specificity of the Ubiquitin-Proteasome System (UPS) is largely determined by the approximately 617 (human) E3 Ub ligases (7), but significant pleiotropy and redundancy among E3 Ub ligases provides tissue specificity and fine-scale regulation of signaling outputs (6).

Correct functioning of the UPS is essential for many cellular processes such as cell-cycle progression and apoptosis. Consequently, the UPS is the subject of intense interest for its therapeutic potential in many diseases including cancer [reviewed elsewhere (8)]. Currently, only broad-acting proteasome inhibitors are in clinical use. Proteasome inhibitors bortezomib (Velcade) and carfilzomib (Kyprolis) have reached phase II clinical trials for use in mantle cell lymphoma (MCL) and multiple myeloma (9–12), although use is associated with peripheral neuropathy (12, 13). The E3 ligase modulator Lenalidomide (Revlimid, an analogue of Thalidomide), is also used in treatment of multiple myeloma but has associated resistance and toxicity (14).

The UPS targets a large range of proteins with important roles in cancer biology. Therapeutic benefit with reduced toxicity might be obtained by targeting more specific aspects of the UPS. For example, Ub-specific proteases, an extensive class of de-ubiquitinating enzymes, have promise as therapeutic targets (15). Of comparable interest and specificity are the E3 Ub ligases, especially as key proteins commonly associated with proliferation and cell-cycle arrest (such as p53, p27, Cyclins, and NF-κB) are specifically regulated by these enzymes (16).

One of the more intriguing members of the E3 ligase family is UBR5 [Ubiquitin protein ligase E3 component n-recognin 5, also known as EDD (E3 identified by Differential Display), EDD1, HHYD, KIAA0896, or DD5]. UBR5 is highly conserved (17) and is essential for mammalian development (18). UBR5 is rarely mutated in healthy somatic tissues, but is mutated and/or overexpressed in cancer (19, 20). UBR5 is known to modulate the DNA damage response (DDR) and transcription (see below for detailed discussion), and emerging roles for UBR5 in a number of biologic contexts have been identified using high-throughput, functional genomics screens. This review will summarize the characterized and emerging biologic roles of UBR5, its role in tumorigenesis, and its potential as a therapeutic target.

UBR5 is a homologue of hyperplastic discs (HYD), a Drosophila melanogaster tumor suppressor (21). Mammalian UBR5 is a HECT (homologous to E6-associated protein at the carboxy-terminus) E3 Ub ligase recognizing n-degrons (see Fig. 1; refs. 17, 22). The human UBR5 gene has 59 exons, encoding an approximately 10 kb mRNA and >300 kDa protein with widespread expression in various cell types. One splice variant (loss of nucleotides 884–901, removing amino-acid sequence VLLLPL) has been identified in 26/26 cancer cell lines investigated (19), although numerous other splice variants are predicted in genome databases. The structural/functional consequences of these variants are not known. UBR5 is highly conserved, but appears to be restricted to the metazoan lineage, and based on phylogenetic analysis is grouped in a distinct class of HECT ligases (Class IV; ref. 23). Class IV likely evolved from a common ancestral gene duplication and includes UBR5/EDD and three other families that are highly divergent at the sequence level.

HECT E3 Ub ligases have a well characterized mechanism, forming a thio-ester linkage between Ub and the substrate protein, a reaction highly dependent on a conserved cysteine within the HECT domain (24, 25). The catalytically active HECT domain of UBR5 is made up of N- and C-lobes separated by a linker sequence as determined by high-resolution crystallography (26). However, the HECT domain of UBR5 has some unique structural features compared with other HECT ligases. The C-lobe of the UBR5 HECT domain does not have a surface for noncovalent binding of Ub, and it is unclear if this exists in the N-lobe (a feature seen in other HECT E3 Ub ligases such as NEDD4, and thought to keep Ub in close proximity for building poly-Ub chains; ref. 26). Instead, UBR5 interacts with Ub via the UBA domain (27). UBR5 also has a zinc finger Ubiquitin Recognin Box (UBR) domain, thought to be involved in N-rule substrate recognition (28), two nuclear localization sequences (29), and an MLLE/PABC domain [homologous to the C-terminal region of Poly-Adenylation Binding Protein (30); Fig. 1]. The MLLE/PABC domain is thought to be a protein–protein interaction motif (e.g., known to bind substrates such as PAIP2; refs. 31, 32) and may regulate ubiquitin transfer catalyzed by the HECT domain (33).

Expression of UBR5 is responsive to progestin in human breast cancer cells, and shows varied expression in normal tissues (19). UBR5 is highly overexpressed in breast cancer at the mRNA level with increases in UBR5 protein also evident in many samples (19). This indicates that regulation of UBR5 is likely at the transcriptional level, though regulation of activity may also occur through PTM. Many phosphorylation sites have been identified in UBR5 (see Fig. 1; refs. 34–41), and UBR5 is known to be a phosphorylation target of ATM (42), CHK (43), and ERK kinases (44). However, this phosphorylation has not as yet been directly linked to UBR5 activity.

A number of studies have implicated UBR5 in various aspects of cancer biology and many of the molecular functions of UBR5 (e.g., DDR) are consistent with a role in cancer. Indeed, the Drosophila homologue Hyd was originally identified as a tumor suppressor gene (21). Human UBR5 was originally identified in a screen for progestin-regulated genes in breast cancer cells (17). Amplification of UBR5 has been reported in breast and ovarian cancer, mostly in the form of allelic imbalance resulting in increased UBR5 mRNA levels (19). UBR5 is located at genomic locus 8q22.3 (17), just downstream from MYC (8q22.4) but UBR5 amplification is independent of microsatellites examined in the more distal regions of the 8q arm (19). Provisional data from The Cancer Genome Atlas (TCGA) clearly show UBR5 amplification to be a common alteration in many cancer types (see Fig. 2A). UBR5 has been shown to mediate therapeutic resistance in ovarian cancer (20), likely through modulation of the DDR (see below). UBR5 was also identified in a transposon mutagenesis screen for cooperating mutations and pathways in pancreatic adenocarcinoma (45).

Perhaps the most convincing evidence yet for an integral role for UBR5 in cancer comes from a recent study showing nonsynonymous mutations of UBR5 in 18% of MCL cases. Most of these mutations affected the conserved cysteine of the HECT domain, which would be expected to disrupt E3 ligase activity (46). Unfortunately, the embryonic lethality observed in Ubr5-null mice has precluded their use in cancer models but the development of conditional knockout models should make these studies more feasible.

Functional mutations in UBR5 may be more informative for inferring the biologic role of UBR5 in cancer. The COSMIC database contains details of published mutations throughout many cancer types (47). Point mutations occur frequently throughout the UBR5 open reading frame in cancer, and frameshift mutations tend to occur toward the PABC/HECT region, which would likely result in loss of E3 Ub ligase activity (see Fig. 2B). Loss of E3 Ub ligase activity in cancer suggests that UBR5-mediated ubiquitination has a tumor suppressive role. There is a clear overall survival advantage for breast cancer patients with normal UBR5 expression (provisional TCGA data; Fig. 2C).

UBR5 has been shown to directly interact with numerous proteins implicated in a wide variety of cellular processes, including the cell cycle, transcriptional and translational machinery, and the DDR. A list of currently known interacting proteins is shown in Table 1. Many of these are key to understanding the role of UBR5 in cancer development and progression.

Identification of E3 Ub ligase substrates is notoriously technically challenging and emerging systematic approaches hold great promise in this respect. Regardless, a subset of UBR5-interacting proteins have been validated as targets of UBR5 E3 Ub ligase activity, although many other interactions appear independent of E3 Ub ligase function, indicating that UBR5-mediated ubiquitination is tightly regulated and is likely context specific. Known targets of UBR5 ligase activity include the DDR proteins TopBP1 (48), RNF168 (49), and ATMIN (42), transcription factors β-catenin (50), pregnane X receptor (PXR; ref. 51), and E6-AP (52), translational machinery CDK9 (53) and PAIP2 (54), the KATNA1 subunit of the cell cycle–related protein KATANIN (55), and the rate limiting enzyme in gluconeogenesis PEPCK (56).

The first descriptions of a developmental role for UBR5 came from mutants of the UBR5 Drosophila homologue, hyperplastic discs [Hyd, otherwise known as I(3)c43]. These display a range of developmental phenotypes, including imaginal disc abnormalities (57), larval lethality, sterility (21), premature photoreceptor differentiation (58), failed oogenesis (59), defective spermatogenesis (60), and plasma membrane gap junction deficiency (61). The severity of these phenotypes appears dependent on the relative level of HYD protein in mutant animals (21). Functional studies identified that Hyd negatively regulates expression of Hedgehog (Hh) and Decapentaplegic (Dpp) in eye and wing discs. However the mechanism is complex and context specific (58). Recent studies suggest that Hyd regulates differential outputs of Hh signaling by directing selective association of Cubitis interruptus (Ci) with different promoters (62). Further, Hyd can coordinate Hh signaling by regulating Hh and Ci expression in concert with Shaggy (Sgg), the Drosophila orthologue of GSK3β (63).

In mice, homozygous deletion of Ubr5 causes embryonic lethality around E10.5 due to failed yolk sac and allantoic vascular development and failed chorioallantoic fusion (18). Although the molecular mechanism underlying this phenotype remains to be described, a cluster of UBR5-regulated genes was recently associated with vascular development. Specifically, vascular endothelial cell expression of UBR5 is necessary for normal vessel formation through a repressive effect on the angiogenic factor ACVRL (64).

The involvement of UBR5 in regulation of hedgehog signaling has not been investigated in mammalian models; however, depletion of Indian hedgehog disrupts murine vascular yolk sac development (65), and paracrine hedgehog signaling is involved in tumor angiogenesis via modulation of VEGF (66). Drosophila studies showing a role for Hyd in controlling Hh and Dpp signaling may also be relevant in this context, as both Hh (65) and BMP (67) signaling are key regulators of yolk sac vascular development. It remains to be determined if the role of UBR5 in vascular development has functional relevance to tumor angiogenesis.

The importance of UBR5 in development is further reinforced by results from a functional genomics screen identifying UBR5 as crucial for embryonic stem cells (ESC) growth (68). UBR5 was subsequently shown to regulate ESC self-renewal via maintenance of the pluripotency factor OCT4 (69). A key outstanding issue in understanding the role of UBR5 in regulating stem cell biology and development is to determine the role of E3 ligase activity in mediating these functions. Aside from its high conservation across species, perhaps the strongest indication for the importance of UBR5 in development comes from a systematic survey of loss-of-function (LoF) variants in human protein-coding genes by the Exome Aggregation Consortium (ExAC). UBR5 is one of the most highly conserved genes in this analysis, which identified just four LoF variants in UBR5 at very low allele frequencies [less than 1/10,000; ExAC, Cambridge, MA (http://exac.broadinstitute.org), June 2015]. A UBR5 missense mutation (c.5720G>A) was recently identified in affected individuals of a Japanese pedigree with familial adult myoclonic epilepsy, but not in any nonaffected family members or unrelated healthy residents of the pedigree's community (70). This mutation encodes a substitution at Arg¹⁹⁰⁷ (Arg¹⁹⁰⁷His), which is highly conserved across species but does not occur in any known functional domains.

The UPS is intimately involved at every stage of transcriptional control and it is now evident that transcriptional control is a key aspect of UBR5 function (see Fig. 3). These stages include: initiation of RNA transcription by transcription factors and their associated co-activators, which recruit the transcriptional machinery (RNA polymerase II/general transcription factors) to gene promoters [reviewed elsewhere (71)]; RNA elongation, which requires chromatin remodeling through the removal and re-deposition of histone proteins; and RNA processing and nuclear export [reviewed elsewhere (72)].

UBR5 was first identified as a target gene of the progesterone receptor (PR), a member of the nuclear receptor family of ligand (e.g., steroid hormones, retinoic acid, thyroid hormone, etc.) induced DNA binding transcription factors (29). Interestingly, UBR5 enhanced the transcriptional activity of the PR itself (29). It is now evident that UBR5 has similar effects on several other nuclear receptors through a variety of mechanisms, which in some (but not all) cases involve the E3 ligase activity of UBR5, and which either directly target the receptor or alternatively other proteins in the receptor transcriptional complex.

A screen of potential transcriptional modulators found UBR5 was a negative regulator of estrogen receptor α (ERα) protein levels and activity. This action was dependent on ubiquitination and an intact UBR5 HECT domain, but ubiquitination of ER was independent of ligand occupancy. Depletion of UBR5 in MCF-7 breast cancer cells resulted in increased endogenous ER levels, an estradiol-induced increase in transcription of GREB1 and CCND1 (ER target genes), as well as increased cell proliferation (73). UBR5 is also a negative regulator of PXR (an orphan nuclear receptor) transcriptional activity. UBR5 together with dual-specificity tyrosine phosphorylation-regulated kinase 2 (DYRKK2) are found in a multiprotein complex with PXR. Phosphorylation of PXR by DYRKK2 facilitates its ubiquitination by UBR5, resulting in reduced PXR protein levels and activity (51).

In T-helper (T_H17) cells, production of the pro-inflammatory cytokine IL17 is regulated by the transcription factor RORγt, an orphan nuclear receptor. Levels of RORγt are controlled by its proteasomal degradation, promoted by UBR5-mediated ubiquitination (74). Activity of UBR5 is regulated upstream by T-cell receptor signaling (deactivation) and TGF-B signaling (activating) through unknown mechanisms and by association between UBR5 and deubiquitin enzyme A (DUBA), which stabilizes the UBR5 protein possibly by inhibiting its auto-ubiquitlyation and consequent proteasomal destruction.

E6-associated protein (E6-AP) is a HECT domain ubiquitin ligase, which acts a co-activator for several nuclear receptors including ER, PR, AR, GR, and RARα [reviewed elsewhere (75)] as well as targeting the receptor co-activator SRC-1. As is the case for the UBR5 co-activation of PR (29), E6-AP's activity as a co-activator is not dependent on its HECT domain ubiquitin ligase activity. E6-AP itself is a ubiquitination substrate for UBR5 (52) and lowered UBR5 expression results in greatly increased levels and half-life of E6-AP. These results point to a possible role for UBR5 in regulating nuclear receptor transcriptional activity via its effects on E6-AP activity. Consequently, UBR5 affects the proteolytic activity of the E6-AP/E6 complex, which targets p53 and other substrates for ubiquitination. UBR5 knockdown in human papillomavirus (HPV) infected cells not only strongly reduced p53 levels but in line with UBR5's possible tumor suppressor function decreased apoptosis in response to etoposide-induced DNA damage and overcame S-phase arrest (52).

UBR5 also regulates other nonnuclear receptor transcription factors. β-catenin is ubiquitinated by UBR5, resulting in its stabilization, increased nuclear localization, and transcriptional co-activation of Wnt target genes (50). Ubiquitination is dependent upon phosphorylation of β-catenin by GSK-3B. Similarly, silencing of UBR5 decreases β-catenin levels and inhibits β-catenin/TCF transcription. This is in contrast to the previously well-defined role of GSK-3B, which is to phosphorylate β-catenin when Wnt signaling is absent, resulting in ubiquitination by the SCF ubiquitin ligase complex and subsequent β-catenin degradation (76). UBR5 also binds and stabilizes the transcriptional co-activator Myocardin, upregulating smooth muscle gene transcription independent of E3 ligase function (77).

These findings represent both positive and negative regulation of genes associated with lineage-specific gene expression. Given the requirement for UBR5 in maintenance of pluripotency, this indicates that UBR5 may act as a molecular switch involved with coordinating differentiation of multiple cell lineages.

A number of studies have implicated UBR5 in regulation of the cellular DDR (see Fig. 3), which is critical for maintenance of genomic integrity and suppression of tumorigenesis. These roles are likely partly responsible for the observed effects of UBR5 in mediating chemoresistance in ovarian cancer (see above).

ATM signaling is required for efficient cell-cycle arrest following DNA damage (78) and proteomic screens have identified UBR5 as a phosphorylation substrate of ATM, CHK1, and CHK2 (43, 79, 80). The functional consequences of phosphorylation on UBR5 function are not clear but there are reciprocal effects on ATM and CHK2 activity. For example, upon DNA damage UBR5 relocates to the nucleus and is required for ATM-mediated phosphorylation of CHK2 (81). The interaction with CHK2 is mediated via phospho-dependent binding to its FHA domain and is required for the activation of downstream DDR checkpoints. UBR5 depletion compromises the maintenance of cell-cycle checkpoints, leading to polyploidy and cell death via mitotic catastrophe (82). UBR5 also regulates ATM signaling via ubiquitination of ATMIN following DNA damage (42) and mediates TOPBP1 ubiquitination (83) via an interaction with the BRCT domain of TOPBP1 (48). The BRCT domain is a common feature of DDR enzymes (84, 85), indicating UBR5 may interact with other DNA damage-related proteins (48).

UBR5 also acts to restrict recruitment of DNA repair machinery at the sites of double-stranded breaks (DSB) by indirectly regulating ubiquitination of phosphorylated H2AX (γ–H2AX). In concert with TRIP12, UBR5 is required to control the nuclear pool of RNF168 (49). RNF168 is a chromatin-associated E3 Ub ligase, which promotes genome stability by acting with RNF8 to ubiquitinate γ–H2AX at the site of DSBs, promoting recruitment of DNA repair machinery (86–88). In the absence of UBR5, excess RNF168 activity promotes ubiquitination of γ–H2AX beyond sites of damage, causing hyperaccumulation of ubiquitin-regulated genome caretakers such as 53BP1 and BRCA1.

Hence, UBR5 plays a key role in homeostasis of ubiquitin-mediated signaling after DNA damage, emphasizing the potential importance of dysregulated UBR5 expression and/or function in the evolution of cancer.

A number of large-scale functional genomics screens have identified putative novel roles for UBR5 in various biologic processes, including ciliogenesis, viral infection, apoptosis, and glucose homeostasis (see Fig. 3). Many of these remain to be validated but point to further roles for UBR5 and the UPS more generally in various aspects of cell biology and pathophysiology.

Primary cilia are microtubule-based organelles present on the apical surface of most vertebrate cell types (89), coordinating signaling pathways during development and in tissue homeostasis. Ciliopathy (defective cilia) is emerging as a major cause of various diseases and developmental disorders. UBR5 was identified in a whole-genome shRNA screen as a positive regulator of ciliogenesis (90), with UBR5 depletion causing decreased primary cilia formation. This screening hit was not further validated and the mechanism underlying this effect is yet to be determined, but the UPS is an established regulator of ciliogenesis (91).

A similar functional genomics approach was used to identify UBR5 (along with the transcription factor GRHL2) as a mediator of sensitivity to apoptosis. Depletion of UBR5 or GRHL2 increased sensitivity to Fas-ligand mediated apoptosis, and expression of both UBR5 and GRHL2 was slightly decreased in breast cancer cell lines less resistant to cell death (92). This indicates that UBR5 may play a role in suppressing apoptotic signaling.

The UPS is a well-established mediator of cellular response to viral infection and a number of studies have recently implicated UBR5 in regulation of viral infection. For example, UBR5 was identified in a siRNA screen for mediators of HIV replication (93). At the molecular level, UBR5 interacts with the HIV-1 VPR protein to regulate TERT activity (94) and also interacts with the HPV E6 protein and E6-AP (52).

Cross-talk between PTMs is emerging as an important regulatory mechanism in cell signaling (95, 96). An interesting example of this concept is demonstrated by the recent finding that acetylation of phosphoenolpyruvate caboxykinase (PEPCK) by P300 promotes its ubiquitination by UBR5, and subsequent proteasomal degradation (56). PEPCK catalyzes the conversion of oxaloacetate into phosphoenolpyruvate, a key step in gluconeogenesis. Phospho-dependent interactions involving UBR5 discussed above (e.g., CHK2 and DYRK2) are further examples of cross-talk between PTMs relevant to UBR5 function in cancer.

UBR5 is emerging as a key regulator of the UPS in cancer and development. However, it is unclear where UBR5 falls in the delineation between oncogene and tumor suppressor. In some cases, amplification and overexpression is clearly linked with chemoresistance and disease outcome (e.g., breast and ovarian cancer), whereas in other contexts, LoF mutations are clearly implicated in tumorigenesis (e.g., MCL). Development of better animal models allowing conditional modulation of UBR5 expression will be key to better understanding context-dependent roles of UBR5 in tumorigenesis and progression in various cancer types. These will be important not only in better understanding the biologic functions of UBR5 in cancer but in addressing the potential of UBR5 as therapeutic target. At this stage it is probably too early to define the potential utility of UBR5 as a therapeutic target in cancer.

UBR5 is a key regulator of cell signaling relevant to broad areas of cancer biology. However, despite insights from genetics, biochemistry, and cell biology, the precise mechanisms by which UBR5 contributes to tumor initiation and progression remain poorly defined. As an E3 ubiquitin ligase, many of the functions of UBR5 relevant to cancer are facilitated by its ubiquitylation of substrates. Many UBR5-interacting proteins have been identified and a number of these have been validated as substrates. However, there remains a clear need for systematic approaches to define the suite of UBR5 substrates in various contexts. This is a key knowledge gap in understanding the function of UBR5 and should be a priority for further research. It is also important to note that any development of UBR5 as a drug target will be heavily dependent on defining enzymatic substrates in a cancer context.

A clear outstanding question around UBR5 function in cancer is why the enzyme possesses many apparently divergent roles in multiple pathways in different tissue and disease contexts? Do these reflect discrete specific activities in different settings? An important consideration in this respect comes from the unique structure of UBR5, very high conservation of sequence, and restriction to the metazoan (i.e., multicellular) lineage. It is tempting to speculate that UBR5 evolved as a kind of signal integrator, coordinating extracellular signaling pathways (e.g., growth factor, hormonal, cilia) with regulation of cell-intrinsic pathways (e.g., proteostasis, DNA damage, etc.). The emergence of improved molecular tools, development of new model systems, and better definition of enzymatic substrates will provide a powerful platform for integrated approaches to define the function of UBR5 in normal cell biology and cancer pathology.

No potential conflicts of interest were disclosed.