|Home | About | Journals | Submit | Contact Us | Français|
Cell division depends upon the coordinated action of positive and negative regulatory factors that ensure high fidelity replication of the genome and its equivalent separation into daughter cells following cytokinesis. The role of positive factors such as the cyclin-dependent kinases in promoting cell division is firmly established, as is the function of CDK inhibitors and phosphatases that antagonize CDKs. In addition to these, regulated protein destruction is now appreciated as essential for temporal regulation of cell cycle transitions. Protein degradation serves as an irreversible switch that ensures temporally regulated cell cycle transitions. Signal-dependent regulation of protein degradation is best understood with regard to the 26S proteasome. Proteins are directed to this machine subsequent to enzymatic transfer of a highly conserved small polypeptide, ubiquitin. The focus of this review is the regulatory molecules that direct the regulated attachment of ubiquitin, polyubiquitylation, to proteins destined for degradation as cells transition through the G1 phase into S phase. During the past decade, it has become increasingly apparent that these molecules are critical mediators of normal cell proliferation, and as such, they are frequently deregulated in human cancers.
Mitotic cell division, which regulates equal distribution of genetic material to daughter cells, serves as the basis of cell proliferation and as such is critical for tissue homeostasis in multicellular organisms. The mitotic cell cycle can be divided into 2 business phases, S phase, during which the genome is duplicated, and mitosis, when the cell divides and all material, including the cell’s genetic blueprint, is equally divided between 2 daughter cells. These 2 phases are separated by gap phases: the G1 phase, separating mitosis from S phase, and G2, separating S phase from mitosis. Given the cellular complexity of higher eukaryotes, the regulation of the mitotic cell cycle is not simply a matter of duplicating the genome and acquiring sufficient protein/lipid mass to produce 2 daughter cells. Rather, this process must be coordinated with the surrounding microenvironment, which determines the availability of growth factors and nutrients; critically, most if not all of these signals are sensed during G1 phase.
Cells enter the cell cycle, G1 phase, and progress to S phase in response to growth factors and nutrients. Progression through G1 phase is facilitated by the D-type cyclins (D1, D2, D3), which form active holoenzymes with cyclin-dependent kinases 4 or 6 (CDK4, CDK6), and cyclin E, which activates CDK2.1 Although the D-type cyclin-dependent kinases are the first to be detected following stimulation of quiescent cells with mitogenic growth factors, both D-type and E-type cyclin-dependent kinases regulate G1 phase progression.2-4 The D-type cyclin-dependent kinase minimally performs 2 functions critical for G1 phase progression. The first function involves the phosphorylation-dependent inactivation of the growth inhibitory properties of the retinoblastoma (Rb) transcriptional repressor protein.5 Subsequent phosphorylation by the cyclin E–dependent kinase results in the release of members of the E2F family of transcriptional activators from Rb.5 The second function is the noncatalytic titration of p21/p27 inhibitors away for CDK2, thereby facilitating subsequent activation of cyclin E/CDK2 complexes.
While mammalian cells encode 3 distinct D-type cyclins, which provide both overlapping and distinct functions,6-9 deregulated accumulation of cyclin D1 is most frequently associated with cancer initiation/progression and hence is the best characterized. Cyclin D1 accumulation represents one of the key mitogen-regulated events required for G1 phase progression.10-12 Ras-mediated pathways play an integral role in the mitogen-dependent induction of cyclin D1 and in the inhibition of ubiquitin-dependent cyclin D1 destruction. The mechanisms associated with overexpression of cyclin D1 in human cancer are varied and include chromosomal translocation, gene amplification, and reduced protein degradation.13-23 The latter mechanism had been predicted for many years due to the frequent overexpression of cyclin D1 in carcinomas of the breast and esophagus without any associated alterations in the CCDN1 locus. The recent identification of the ubiquitylation machinery that directly targets cyclin D1 to the proteasome has provided direct evidence for this mechanism of overexpression.
In addition to cyclin D1, accumulation of numerous G1 phase regulators is determined by ubiquitin-dependent proteolysis. Among these is another key G1 cyclin, cyclin E. Cyclin E is overexpressed in carcinomas of the breast, and its overexpression is associated with poor prognosis. The overexpression of both cyclin D1 and cyclin E contributes to the unrestricted growth of cancers via their capacity to increase activity of their cognate CDK partners. In addition to positive cell cycle regulators, accumulation of negative regulators or CDK inhibitors (CKIs) is also determined through the ubiquitylation pathways. Unlike the cyclins, which are frequently stabilized in cancer, CKI destruction is accelerated. The reduction in a particular CKI triggers unrestricted cyclin/CDK function and deregulated cell cycle progression. As discussed below, the progressive accumulation and rapid destruction of each key regulatory protein through ubiquitin-mediated destruction contribute to the coordinated cell cycle transitions that are key to homeostatic growth; the loss of the key, nonreversible off switch (proteolysis) is frequently associated with malignant outgrowth.
Protein ubiquitylation is catalyzed in a stepwise manner through the sequential action of 3 classes of enzymes: E1, E2, and E3. Substrate specificity for ubiquitylation of critical regulators is determined by the E3 ubiquitin ligases.24-28 Ubiquitylation of a majority of the G1 regulatory proteins is regulated by SCF (Skp–Cullin 1–F-box) E3 ligases. SCF is a multicomponent E3 ligase and the best characterized of the cullin-based E3 ligases. Cullin-based E3 ligases typically contain a substrate-specific recognition factor, which is linked to the catalytic E2-E1 enzymes via the scaffolding property of the cullin. Cullin 1 (Cul1) was originally identified as CDC53 in budding yeast29; it was subsequently shown to serve as a scaffold for the SCF family of E3 ligases. Collectively, there are 7 cullin family proteins, cullin 1, 2, 3, 4A, 4B, 5, and 7, all of which appear to function in the context of protein ubiquitylation. As discussed below, cullins 1, 3, 4A, and 4B contribute to ordered cell cycle progression.
Cul1 represents the best characterized member of the cullin family. It functions as a core component of the SCF E3 ligase; in this multimeric enzyme, Cul1 bridges the key enzymatic components of the ligase (the small ring finger protein Rbx1/Roc1, an E2 and the E1) to one of 68 human or 74 murine F-box proteins, which in turn function as the substrate-specific recognition factors. The bridging of such a large number of substrate-targeting proteins to Cul1 is achieved by the adaptor protein, Skp1. Skp1 binds directly and specifically to the conserved F-box domain, thereby recruiting the “scaffolded enzyme” to the F-box–associated substrate (Fig. 1). The F-box domain, as initially identified in cyclin F and CDC4,30 contains approximately 45 residues in length and makes direct contact with Skp1. The crystal structure of one such complex reveals that Cul1 is concave such that it brings the Rbx1 and E2 components into close proximity with the substrate that is docked with the F-box protein, thereby permitting ubiquitin transfer (Fig. 1).31
The other cullins serve analogous functions but utilize unique specificity components. Cul2 and Cul5 utilize a Skp1-like factor, elongin C, as a bridge for substrate specificity factors harboring BC/SOCs box domains.32 Of note, Cul2 serves its scaffolding function in the context of the Von Hippel-Lindau tumor suppressor.32 Cul3 binds directly to substrate-specific recognition proteins that contain a BTB (Broad complex, Tram track, and Bric-a-brac domain).33,34 Cul4A and Cul4B are recruited to their substrate adaptors, the DCAFs (DDB1-Cul4–associated factor) by the DDB1 protein.35 Cul7, like Cul1, utilizes Skp1 to recruit Fbw8, which again serves a substrate-binding function.36 Collectively, this highly modular E3 ligase can regulate the stability of a vast array of proteins in a highly ordered and signal-dependent fashion.
The coordinated accumulation of G1 phase regulators depends upon the concerted timing of gene expression with protein degradation. It is now apparent that ubiquitin-dependent destruction of many, if not all, key regulatory proteins is governed by one or more E3 ligases. The SCF family of E3 ligases makes the largest contribution to this process and as such will be discussed in the context of individual substrate-ligase relationships in the greatest detail.
Substrate recognition by the F-box proteins is conferred by their C-terminal domains. The F-box proteins have been subdivided into 3 groups based upon their substrate-binding domains: Fbxw, Fbxl, or Fbxo. The Fbxw family utilizes a C-terminal WD40 repeat domain, and prototypical members of the Fbxw family include Fbw7 (CDC4); the Fbxl subgroup is characterized by a C-terminal leucine zipper domain, and an example of this subgroup is Skp2 (Fbxl1).37 The Fbxo subgroup generally lacks recognizable protein:protein interaction motifs that would be expected to mediate substrate recognition.38 However, as more of the members of this group are characterized and substrate-binding regions functionally defined and/or crystallized, unique domains will perhaps be identified, allowing the further subdivision of this group of F-box family members.
Fbxw7 was initially identified as CDC4 in the screen for budding yeast mutants that exhibited defects in cell cycle progression.39 Work in budding yeast revealed that CDC4 is indeed the specificity component of an SCF E3 ligase that regulates the accumulation of the cell cycle inhibitor, Sic1.27,40,41 Work in several laboratories facilitated the identification of orthologs in a variety of multicellular organisms including Caenorhabditis elegans (SEL-10),42 Drosophila (archipelago-AGO),43 and mammalian cells (Fbxw7).44,45 All orthologs share relative homology and structure, although it should be noted that yeast CDC4 is unique in its function as a negative regulator of a cell cycle inhibitor, while all orthologous Fbxw7 proteins share in their capacity to limit accumulation of proteins that facilitate rather than inhibit cell proliferation. Critical substrates of Fbxw7 include cyclin E, c-myc, Notch1, and SREBP1c.46 Among these, both cyclin E and c-myc contribute to regulated G1 phase progression in response to mitogenic growth factors.
Analogous with most F-box proteins, Fbxw7 can be divided into 2 and perhaps 3 conserved domains. The F-box domain is located near its N-terminus followed by 8 WD40 repeats that constitute the substrate-binding domain. Based upon structural analysis, these repeats are predicted to form an 8-bladed propeller structure containing multiple points of direct substrate interaction. Further analysis defined critical arginine residues within the propeller blades that make direct contact with phosphorylated substrates.47 The critical nature of these arginines was confirmed through analysis of Fbxw7 mutants identified in human cancers.
In addition to these 2 domains, a subset of F-box family members also contains a dimerization or D-domain prior to the F-box. While incompletely defined, dimerization depends on conserved hydrophobic amino acids in this domain. While the propensity of a number of F-box proteins to dimerize is an emerging theme, how dimerization contributes to function remains poorly defined. Dimerization of Fbxw7 is not essential for substrate recognition, but ubiquitin transfer to bound substrates is absolutely dependent upon dimerization.48 Similar to the latter observation, work on Fbxo4 suggests a role for dimerization in ubiquitin transfer rather than substrate binding.23
Substrate recognition by Fbxw7 is dependent upon prior phosphorylation of substrates. Strikingly, GSK3β has been implicated as the major kinase for both cyclin E and c-myc. The major phosphodegron in cyclin E lies at the C-terminus and is nucleated by a series of phosphorylation events initiated at Thr380. Thr380 was initially defined as a site of autophosphorylation by CDK2.49 It later became apparent that GSK3β also targets this residue.50 This latter work revealed a complex chain of events wherein Thr-380 phosphorylation by GSK3β functions as a priming site necessary for autophosphorylation of Ser-384 by cyclin E/CDK2; p-Ser384 then serves as phosphorylated residue that docks with the WD40 repeats of Fbxw7. Phosphorylation of cyclin E at Thr62 also generates a Fbxw7 phosphodegron.44,51 The precise role of N-terminal versus C-terminal degrons in the regulation of cyclin E remains to be determined.
Fbxw7 recognition of c-myc is also mediated by GSK3β-dependent substrate phosphorylation.52,53 GSK3β phosphorylation of myc at Thr58 generates an Fbxw7 phosphodegron. In contrast to cyclin E, the GSK3β-mediated event is preceded by MAPK-dependent phosphorylation of Ser62, and this phosphorylation event functions as a priming event necessary for GSK3β recognition. Ser62 phosphorylation also appears to have a stabilizing effect on myc such that specific dephosphorylation of this residue is necessary in order for Fbxw7-dependent degradation of myc.
While there is currently no evidence suggesting Fbxw7 function is regulated by signal input, alternative splicing of Fbxw7 generates 3 isoforms (α, β, γ) with unique N-termini.54 Although it is not evident that the N-termini of the respective isoforms should alter substrate specificity directly, the isoforms exhibit distinct patterns of subcellular localization, and this may have a significant impact upon substrate regulation. For example, Fbxw7γ localizes to the nucleolus,55 a site of ribosomal transcription, which is in turn directed in large part by c-myc. It is therefore not surprising that Fbxw7γ regulates c-myc ubiquitylation in this compartment.
The high frequency of inactivating Fbxw7 mutations in primary human cancer provides the best evidence for a critical biological function for Fbxw7 as an inhibitor of proproliferative proteins. Work from a number of laboratories has collectively revealed mutations at a frequency approaching 6%.46 The highest frequency appears to be in T-All, where it is inactivated via hemizygous mutations in nearly 30% of the primary tumors. In many tumors, mutations are missense and target critical arginine residues within the WD40 repeats that make direct contact with substrates. However, both missense and nonsense mutations have been identified in colon cancer.
Skp2 (Fbxl1) was initially identified as a cyclin A coprecipitating protein, along with Skp1, where it was coined S phase kinase-associated protein 2.56 The initial characterization suggested a potential role for Skp2/Skp1 to directly activate cyclin A/CDK2. Because of the noted interactions, regulation was postulated to be direct. However, subsequent studies demonstrating a role for Skp2-Skp1-Cul1 in the ubiquitylation of target proteins suggest that the ability of Skp2 to increase catalytic function of cyclin A/CDK2 is a reflection of Skp2-dependent ubiquitylation of key CDK inhibitors: p21Cip1 and p27Kip1.
As with all F-box proteins, Skp2 harbors an N-terminal F-box domain necessary for recruitment of Skp1. The C-terminus of Skp2 consists of 10 leucine-rich repeats (LRR). Like the WD40 repeat, the LRR is a protein:protein interaction domain utilized by 21 F-box family members.57 Structurally, unlike the WD40-based propeller, the LRR consists of a β strand and an α helix that forms a concave substrate-binding surface lined with the β strands.58
SCFSkp2 regulates ubiquitylation of a number of substrates, many of which function in G1 progression, including p21Cip1,59 p27Kip1,60 p130 (Rb2),61,62 and c-myc.63,64 Unlike Fbxw7, Skp2-dependent recognition of select substrates requires the utilization of a small protein cofactor, Cks1. Cks1 and related proteins were identified as genetic suppressors of CDK mutations in yeast.65,66 Skp2-dependent ubiquitylation of p27 was shown to be dependent upon Cks1; subsequent structural studies revealed that Cks1 binds to Skp2 within the concave LRR domain and therein coordinately and directly mediates association with Thr187-phosphorylated p27Kip1.67-69 While Cks1 appears to be superfluous for Skp2-dependent targeting of its other substrates, a potential role for additional substrate specificity cofactors should be considered.
Like its close relative p27Kip1, p21Cip1 can also be ubiquitylated by SCFSkp2.59 However, this does not require Cks1. In addition to the regulation of Cip1/Kip family ubiquitylation, SCF-Skp2 also targets the retinoblastoma family member p130.61,62 The degradation of these cell cycle regulators by Skp2 is likely key to coordinated cell proliferation. Consistent with this notion, Skp2 is frequently overexpressed in human cancers such as lymphoma and oral squamous carcinoma70,71; critically, Skp2 overexpression generally correlates with decreased p27Kip1 levels.71 Consistent with the Skp2 functioning as an oncogenic protein, transgenic expression of Skp2 in lymphocytes is associated with malignant lymphoma, and in vitro studies have revealed that Skp2 overexpression in concert with mutant Ras generates a neoplastic phenotype.71
The inverse relationship between Skp2 and p27Kip1 in human cancers suggests that p27Kip1 is a key substrate. It is worth noting that while Skp2 overexpression is associated with lymphoma, p27Kip1-deficient mice are not predisposed to lymphoma, suggesting the existence of additional biologically relevant targets. Among these is the c-myc oncoprotein63,64; like many key regulatory proteins, c-myc accumulation and activity are regulated through multiple, partially overlapping pathways. Skp2 was indeed initially found to regulate c-myc ubiquitylation. Strikingly, unlike Fbxw7, which appears to only mediate c-myc stability, Skp2-dependent regulation not only influences c-myc stability but also increases c-myc–dependent transcriptional activation properties.63,64 The relationship between Fbxw7 and Skp2 with regard to c-myc ubiquitylation remains to be established. Possible scenarios include subcellular access such that nucleolar c-myc is uniquely dependent upon Fbxw7γ, while Skp2 regulates nuclear c-myc stability. Alternatively, Skp2 could generate monoubiquitylated c-myc, which increases its function as an activator of transcription upon which Fbxw7 could polyubiquitylate c-myc, triggering proteasomal targeting.
Fbxo4 belongs to the large subgroup of F-box proteins that do not contain a recognizable protein:protein interaction domain within the C-terminus, hence, Fbx only (Fbxo). Although initially an orphan F-box protein (lack of a known substrate), work suggested that Fbxo4 functions in concert with a specificity cofactor αB crystallin.72 αB crystallin is considered a small heat shock protein and functions as a molecular chaperone. Coexpression of Fbxo4 with αB crystallin was noted to increase the level of ubiquitin-conjugated proteins within the cell. Shortly after this work, 2 substrates of Fbxo4 were identified: cyclin D173 and Trf1.74 The latter is a component of the shelterin complex, which regulates telomere homeostasis.75 Cyclin D1 is an established oncogene whose overexpression in cancer has been attributed to decreased proteolysis and is a key regulator of growth factor–mediated G1 progression.76
Given the absence of a characterized substrate-binding domain, our understanding of how Fbxo4 recognizes substrates has lagged behind that of Fbxws and Fbxls. As for cyclin D1, binding is mediated by phosphorylation of a conserved Threonine residue, Thr-286, in the C-terminus of cyclin D177 and is dependent upon αB crystallin (Fig. 2).73 In contrast, Fbxo4 binding to Trf1 is phosphorylation and αB crystallin independent.74,78,79 Recent structural studies on Fbxo4 have provided insights into the ubiquitylation of TRF1 and suggest that Fbxo4 binds to TRF1 through an atypical small GTPase-like fold.74,78 However, because Fbxo4 regulates its 2 substrates, cyclin D1 and TRF1, in distinct manners, further structural studies are needed to elucidate the mechanisms by which Fbxo4 recognizes phosphodependent substrates such as cyclin D1. Identification of additional substrates should provide greater insights into the mechanism of substrate recognition by Fbx4.
Until recently, it was thought that SCF ligases are constitutively active and ligase activity is regulated at the level of substrate phosphorylation. It is now clear that SCF ligases undergo dimerization mediated by a conserved N-terminal dimerization domain (D domain).48,78 Although dimerization appears to be a characteristic of SCF ligases, how dimerization regulates ligase activity is poorly understood. Recent work from our laboratory revealed that Fbx4 dimerizes through conserved D domain in its N-terminus and dimerization of Fbx4 is regulated by growth factor–dependent phosphorylation.23 During S phase of the cell cycle, accelerated cyclin D1 turnover induces phosphorylation of Fbx4 on Serine 12 (S12), thereby promoting dimerization and ligase activation. Consistent with this, mutation of S12 to alanine abolished dimerization and subsequent Fbx4 activation, leading to impaired cyclin D1 ubiquitylation and degradation.23 The importance of S12 phosphorylation is emphasized in the identification of cancers, wherein this residue is targeted by missense mutations.23 Interestingly, phosphorylation of Fbx4 at S12 is catalyzed by GSK3β, the same kinase that mediates phosphorylation of cyclin D1 and promotes its nuclear export.80 While it remains unclear how GSK3β itself is regulated such that it coordinates phosphorylation of both ligase (Fbxo4) and its substrate (cyclin D1), it clearly functions as the focal point necessary to direct substrate to the cytoplasmic ligase and also mark the degron within cyclin D1 with a phosphate group in order to facilitate Fbxo4–cyclin D1 binding.
Ubiquitin-mediated proteolysis plays a critical role in the maintenance and regulation of cellular homeostasis. On the other hand, deregulated proteolysis contributes to aberrant cellular growth and tumorigenesis. Recent studies strongly suggest that Fbxo4 has tumor suppressor properties, as loss-of-function mutations are associated with tumorigenesis. Fbxo4 functions as a negative regulator of cyclin D1, and loss of Fbxo4 expression is closely associated with cyclin D1 nuclear accumulation. Consistent with Fbxo4-antagonizing oncogenic properties of cyclin D1, Fbxo4-resistant cyclin D1 mutants have increased oncogenic potential in vitro and in vivo.81 In addition, activating mutations in cyclin D1 are mutually exclusive with inactivating mutations in Fbxo4.23
Cyclin D1 overexpression has been implicated in the pathogenesis of various cancers, and altered expression can be attributed to the deregulation of its proteolysis in many of these. While mutation of Fbx4 has been documented in esophageal cancers, there is also evidence for loss of αB crystallin in breast cancer–derived cell lines with a corresponding increase in cyclin D1 levels.73 Additional work is necessary to assess Fbxo4 status in other cancers associated with cyclin D1 overexpression. In esophageal carcinomas, it is worth noting that the cancer-associated mutations were found in the N-terminal domain of Fbxo4 and directly impair its dimerization and activation. Thus, unlike Fbxw7, where the majority of mutations directly impair substrate binding, mutations in Fbxo4 impair ligase activation.
In addition to Fbxo4, Fbxw8 has also been implicated in the regulation of cyclin D1 turnover.82 Biochemical characterization demonstrates that like Fbxo4, Fbxw8 binds to cyclin D1 via pThr-286 and can directly conjugate ubiquitin. Unlike Fbxo4, there is no evidence for disruption of Fbxw8 in human cancer, and work from knockout mice reveals that loss of Fbxw8 paradoxically reduces cell multiplication.83,84 Thus, the contribution of Fbxw8 to cyclin D1 regulation needs further examination.
The SCF E3 ligases play an essential role in the temporal regulation of G1/S phase transitions and in cell proliferation control in general, yet they are not the sole contributing entities. We will next consider individual E3 ligases that have been identified as critical regulators of individual cell cycle regulators. Literature regarding the regulation of cyclin E degradation is dominated by interests in Fbxw7, which directs polyubiquitylation of the Thr380/384–phosphorylated protein that is by definition associated with CDK2, given the role of autophosphorylation in the generation of the phosphodegron. However, monomeric cyclin E also undergoes ubiquitin-mediated degradation, and degradation of monomeric cyclin E does not require phosphorylation, suggesting an SCF-independent mode of ubiquitylation. Indeed, destruction of monomeric cyclin E is regulated via a distinct cullin complex coordinated by Cul3.85 A Cul3–cyclin E interaction was first suggested through a yeast 2-hybrid wherein cyclin E was used as bait. Cul3-deficient mice were subsequently generated to address the functional significance of the complex. While Cul3 is essential for embryonic development, available tissues exhibited overexpression of cyclin E and increased numbers of cells in S phase, a phenotype associated with cyclin E overexpression. Additional work utilizing cells derived from mice with a conditional Cul3 allele suggests that Cul3-dependent control of cyclin E is critical for maintaining cells in a quiescent state.86 How Cul3 targets cyclin E remains to be determined. Cul3, like Cul1, utilizes substrate specificity factors; whereas Cul1 uses F-box–containing proteins, Cul3 depends upon BTB domain–containing proteins.33,34 A vast number of BTB domain–containing proteins exist in the genomes of many species, yet only a limited number have been demonstrated to link Cul3 with a specific substrate, and at the present time, there are no obvious candidates that might link Cul3 with cyclin E.
Much of the initial efforts to unravel the mechanisms regulating p27Kip1 stability focused on the role for Thr-187 phosphorylation–directed recognition by SCF-Skp2. However, 2 critical pieces of evidence suggested the existence of a second regulatory E3 ligase. First was the generation of a mouse wherein a nonphosphorylatable Kip1 allele, p27-T187A, was knocked into the endogenous locus.87 Because of the large body of evidence for CDK2-dependent regulation of p27 stability, in concert with observations that p27Kip1 levels decline rapidly at the G1/S boundary, coinciding with CDK2 activation, it was anticipated that the p27-T187A mutant would be stable throughout the cell cycle. Analysis of cells from these mice revealed just the opposite; p27T187A proteolysis remained intact during G1 phase.87 In fact, it was only the S and G2 phase pools of p27Kip1 that were dependent upon Thr187 phosphorylation and hence upon SCFSkp2. A second piece of evidence implicating a distinct E3 ligase was the observation that p27Kip1 still underwent ubiquitin-dependent proteolysis in Skp2−/− cells during G1 phase.88 Furthermore, this alternative ligase activity was restricted to cytoplasmic fractions, whereas Skp2 is predominantly nuclear. These observations suggested the existence of a sequential pathway wherein a cytoplasmic ligase regulates p27Kip1 during G1, following CRM1-mediated p27Kip1 nuclear export; subsequent to S phase entry, regulation is handed off to SCFSkp2 in order to maintain levels of p27Kip1 below an inhibitory threshold.
The existence of the cytoplasmic p27Kip1 E3 was confirmed by the identification of the KPC (Kip1 ubiquitination–promoting complex).89 KPC was demonstrated to bind and directly mediate ubiquitylation of p27Kip1 in a phosphorylation-independent manner. KPC is a dimer composed of KPC1 and KPC2 components. KPC1 contains a Ring finger domain, while KPC2 contains a ubiquitin-like domain (UBL) and 2 ubiquitin-associated domains (UBA). Initial work with this ligase revealed that the KPC1 subunit alone could coordinate ubiquitin conjugation in a manner dependent upon its Ring finger domain.89,90 However, in vivo, both KPC1 and KPC2 function is required as the UBL domain of KPC2 mediates binding with KPC1, while the UBA domains in concert with the UBL mediate 26S proteasome binding.
Protein elimination via polyubiquitin-directed, 26S proteasome degradation is now widely accepted as a key regulatory process necessary for both normal and tumor cell proliferation and survival. Its intrinsic nature as a rapid, irreversible switch is a key aspect necessary to initiate progression through various cell fates. The importance of this regulatory process can be highlighted by considering the frequency with which key tumor-associated processes result from deregulation of proteolytic processes (Table 1). These observations have led to considerable interest by the pharmaceutical industry, with the hope of developing small molecules that can disrupt key cancer-associated proteolytic events (e.g., p27Kip1 destruction) or increase degradation of key growth-promoting proteins (cyclin D or E). In addition, 26S proteasome inhibitors such as bortezomib have shown striking success in cancers such as multiple myeloma. The ongoing efforts to dissect key regulatory processes in the ubiquitin-proteasome pathway will play a significant role in the development of new therapeutic strategies and provide key insights into normal as well as malignant cell growth.
The authors thank Shivani Sethi for editorial assistance and Eric Lee for critical reading of this paper.
This work was supported by grants from the National Institutes of Health (NIH) (CA93237) and a Leukemia & Lymphoma Scholar award (J.A.D.).
The author(s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.