Department of Medical Oncology, University Hospital of Larissa, Larissa, Greece
To whom correspondence and reprint requests should be addressed: Ioannis A. Voutsadakis, M.D., Ph.D., Department of Medical Oncology, University Hospital of Larissa, Larissa 41110, Greece Tel: 30-2413502028 Fax: 30-2413502027 e-mail: ivoutsadakis@yahoo.com
January 2010
Ubiquitin: Gene and protein
Ubiquitin is a 8 kDa protein of 76 amino-acids that has taken its name from its ubiquitous presence in cells. Its covalent link to a target protein is a signal for different fates for this target protein. Ubiquitination (or ubiquitylation) refers to this covalent link and represents a signal analogous to phosphorylation. Ubiquitin gene exists in multiple copies in eukaryotic cells.
Ubiquitin protein contains seven lysine residues at positions 6, 11, 27, 29, 33, 48 and 63 through which it can be attached to the substrate or to one another. Ubiquitin is characterized by a β-grasp superfold termed ubiquiton in which a central α-helix is surrounded by four β-sheets (Welchman et al., 2005). This superfold defines also ubiquitin-like proteins such as SUMO (Small Ubiquitin-like modifier) and NEDD8 (Neuronal Precursor cell-expressed developmentally down-regulated protein 8). Other human ubiquitin-like proteins containing ubiquitons include ISG15 (Interferon stimulated gene 15), FAT10 (Human leukocyte antigen F-associated Transcript 10), FUB1 (Fan Ubiquitin-like protein 1) and URM1 (Ubiquitin Related Modifier 1). All ubiquitin-like proteins, although varying in amino-acid sequence, share with ubiquitin the common structure and the common biochemical mechanism of tagging through an isopeptide bond (Pickart, 2004; Pickart and Eddins, 2004).
Ubiquitination
The covalent attachment of ubiquitin to a target protein is a very well controlled process that is executed with the aid of three types of enzymes. A first type of enzyme called E1 or Ubiquitin activating enzyme, using energy from the conversion of ATP to ADP, binds ubiquitin and transfers it onto a second type of enzyme called E2 or Ubiquitin conjugating enzyme. E2-loaded ubiquitin is then attached, with the help of a third type of enzymes called E3 or Ubiquitin ligase, to the ε-amino group of a lysine residue on a target protein. Two E1 enzymes exist in the human genome called Ubiquitin-activating enzyme 1 (UBE1) and Ubiquitin-like modifier Activating enzyme 6 (UBA6) (Groettrup et al., 2008). UBA6 is also performing the activation function for fatylation, the addition of FAT10 to target proteins (Chiu et al., 2007). E2 enzymes are more abundant (about 30 to 40 exist in the human genome) and have a conserved 150 aminoacids central structure that includes four β sheets and four α-helices and surrounds the active cysteine residue. This cysteine accepts ubiquitin through the formation of thiol-ester bond with the final glysine residue of ubiquitin. A signature HPN (Histidine-proline-asparagine) sequence is found 7 to 8 amino-acids amino-terminal to this cysteine (Michelle et al., 2009). The formation of thiol-ester group requires ubiquitin to be activated, that is linked to the E1 enzyme, while free ubiquitin has very low affinity for E2 enzymes. E2-bound ubiquitin transfer to the target protein is facilitated by a ubiquitin ligase or E3 enzyme. There are about 600 E3 ligases in human genome. This step confers substrate specificity to the process of ubiquitination given that every E3 ligase can interact only with specific substrate proteins. Nevertheless this specificity is partial, as several substrates can interact with an E3 ligase while a specific protein undergoing ubiquitination can interact with several E3 ligases. In addition each E3 ligase can interact with several E2 enzymes and the reverse is also true given that each E2 can interact with several E3s (Van Wijk et al., 2009).
Figure 1: Enzymatic cascade of ubiquitination.
Two types of E3 ligases exist having a mechanistically different catalytic mode of action through which they perform ubiquitin ligation. RING (Really Interesting New Gene) type E3s act by bringing E2-bound ubiquitin in close proximity with the substrate protein in order for ubiquitin to be directly transferred to the substrate. In addition, RING E3s probably mediate a conformational change of bound E2 that facilitates ubiquitin transfer (Passmore and Barford, 2004). In contrast HECT (Homologous to Human Papilloma Virus E6 Carboxyterminal domain) type E3s possess an active cysteine residue that forms a thiol-ester bond with ubiquitin before it is transferred to the substrate. A third type of E3 ligases called U-box domain E3 ligases is considered by many as a sub-type of RING type E3s because U-box domain has a RING domain-like conformation and the mechanism of action is also by bridging E2-bound ubiquitin with the substrate, similarly to RING type E3s.
RING type E3s are by far more abundant than HECT E3s and comprise about 95% of human E3s (Li et al., 2008). RING domain has several cysteines and a histidine in its core structure which bind two zinc atoms. RING domains create the rigid platform that constitutes the surface for the Ubiquitin conjugating enzyme bound with ubiquitin binding. Some E3s are comprised of a single polypeptide that possesses both the RING E2-binding domain and the substrate binding domain, while other E3s are constituted by several distinct proteins, one of which is the RING domain E2-binding protein and another binds the substrate protein to be ubiquitinated (Deshaies and Joazeiro, 2009). The prototype of this latter group is the cullin-RING ubiquitin ligases (CRLs) comprised of a RING protein linked through a family of proteins called cullins to a substrate binding sub-unit (Bosu and Kipreos, 2008). In addition to transferring a first ubiquitin molecule to a substrate (chain initiation), RING E3 ligases perform chain elongation, the attachment of further ubiquitin molecules. These are distinct reactions and chain initiation is taking place in a much slower pace than the elongation step which, in many occasions, is completed 5 to 30 times quicker than the initiation (Petroski and Deshaies, 2005). In the U-box type E3 ligases the conserved cysteines and histidine of RING type ligases are replaced by charged and polar residues.
The other major type of E3s, HECT type has 28 members in human genome (Rotin and Kumar, 2009). All HECT ligases possess in the carboxy-terminal part of their molecule a HECT domain first identified and named by E3 ligase E6-AP (Human Papilloma Virus E6-Associated Protein), while their amino-terminal part is comprised of various other domains. HECT domain has two sub-domains, one of which binds the E2 Ubiquitin conjugating enzyme and the other binds the substrate protein.
Ubiquitination is a reversible process and there are specific de-ubiquitinating enzymes that reverse it. These enzymes recognize the isopeptide bond between the carboxyterminal glycine of a ubiquitin molecule and the ε-aminogroup of a lysine of another ubiquitin molecule or of a target protein. There are five families of de-ubiquitinating enzymes: the UBP (Ubiquitin-specific processing protease) family, the UCH (Ubiquitin Carboxy-terminal Hydrolase) family, the OTU (Ovarian Tumor related proteases) family, the ataxin/Josephin group having ataxin 3 as the only member and the JAMM (Jab1/MPN domain metalloenzyme)/MPN+ motif proteases (Amerik and Hochstrasser, 2004). The role of de-ubiquitinating enzymes is to maintain the ubiquitin pool in the cell and to perform proof-reading for proteins that had been inappropriately ubiquitinated. De-ubiquitinating enzyme Rpn11 of the JAMM/MPN+ family is part of the proteasome (see also next) and recycles ubiquitin from proteins that had been recognized and processed for degradation. The importance of de-ubiquitination is underlined by the fact that their dysfunction is associated with diverse diseases (Singhal et al., 2008).
Different types of ubiquitination and role
Although ubiquitination was initially identified as a signal that leads to proteasome degradation of the target protein, it has become since clear that attachment of ubiquitin can lead to different outcomes depending on the type of this attachment. All lysine residues of the ubiquitin molecules can be used for isopeptide bond formation and result to different outcomes. In addition another dimension of diversification is conferred by whether one ubiquitin molecule or a chain of ubiquitins is attached.
A chain of at least four ubiquitin molecules linked through lysine 48 is the signal for recognition of a target protein by the proteasome complex in order to be degraded. The entire structure that leads to degradation of a target protein by the proteasome is called a degron and is comprised of two parts, the first being the covalently attached ubiquitin tag and the second being an unstructured region of the target protein that is a pre-requisite for the delivery of the recognized and captured protein to the interior of the core proteasome particle where the enzymatic degradation activities reside (Schrader et al., 2009). Other lysines such as lysine 6 and 11 of the ubiquitin molecule can also serve as anchors of proteasome-recognized ubiquitin chains. Ubiquitin chains linked through lysine 63 regulate processes such as DNA repair, endocytosis and protein kinases activation (Hoeller et al., 2006). Proteasome degradation after lysine 63 poly-ubiquitination has been described in some instances to occur (Babu et al., 2005) but most often lysine 63 poly-ubiquitination leads to proteolysis through autophagy (Li and Ye, 2008). Lysine 63-linked chains differ significantly in their conformation from their lysine 48-linked counterparts. Lysine 63-linked chains undertake an open conformation with little contact between ubiquitins except for the covalent link (Varadan et al., 2004), although in solution lysine 63 chains can adopt a continuum of conformations in a dynamic manner (Datta et al., 2009). In contrast, lysine 48-linked chains have a more compact conformation with neighboring ubiquitins developing additional non-covalent links with each other. These differences between ubiquitin chains form the basis for divergent functions (Tenno et al., 2004) due to recognition specificity by different ubiquitin receptor proteins (Raasi et al., 2005).
The Ubiquitin-Proteasome System (UPS) plays an important role in DNA transcription. Co-activators bound to activated transcription factors recruit histone acetyltransferases such as CBP (CREB Binding Protein)/p300 and p/CAF (p300/CBP-associated Factor) and histone arginine methyltransferases such as CARM1 (Coactivator-associated Arginine Methyltransferase-1) and PRMT-1 (Protein Arginine Methyltransferase-1) (Jenster et al., 1997). These enzymes promote histone acetylation and methylation that opens nucleosomes in order for transcription complex to obtain access to transcription factor binding sequences in target promoters. The signal for histone methylation is provided by sequential histone mono-ubiquitination and de-ubiquitination (Zhang, 2003; Dover et al., 2002; Sun and Allis, 2002), a process in which the 19S regulatory part of the proteasome is also involved (Laribee et al., 2007; Ezhkova and Tansey, 2004). This process is important in transcription elongation and defines a point of regulation of transcription by the UPS. RING domain-containing E3 ligase hPIRH2 (human p53-induced ring-containing H2) binds transcription factors such as nuclear receptors and promotes suppression of histone deacetylase 1 (HDAC1) stabilizing histones in the acetylated state (Logan et al., 2006). Histone modifications are an intermediary state that promotes nucleosomal histone octamer dissociation from the promoter transcription initiation site and leave DNA naked for transcription machinery binding (Boeger et al., 2005; Boeger et al., 2004). In addition ubiquitination of co-repressors CtBP1/2 and NCoR/SMRT leads to their proteasome degradation releasing transcriptional repression in order for the transcription complex to bind DNA (Perissi et al., 2008). Many transcription factors such as nuclear receptors undergo ubiquitination after DNA binding (Gaughan et al., 2005; Ramamoorthy and Nawaz, 2008). In parallel a molecular complex called mediator is recruited and helps recruit, in its turn, RNA polymerase II to begin transcription (Vijayvargia et al., 2007). After a few rounds of transcription ubiquitin ligases have attached four ubiquitin molecules to transcription factor molecules which can now be recognized by the proteasome for degradation. Components of the general transcription machinery that possess E3 ligase activity collaborate in this ubiquitination (Conaway et al., 2002). Some transcription factors such as the AR (Androgen Receptor) are stabilized in a transient monoubiquitinated state by a protein called TSG101 (Tumor Susceptibility Gene 101), which later is displaced from the AR for poly-ubiquitination to take place (Burgdorf et al., 2004). Proteasomal degradation is a pre-requisite for the transcription process to continue because it frees the way for new transcription factor molecules to occupy the promoter as long as the signal that activates the transcription factor exists. In this way there is a strict time regulation of transcription.
Ubiquitination plays also significant role in DNA repair. Nucleotide Excision Repair (NER), one of the modes of DNA repair is activated when DNA damage, for example after UV light, is detected. NER requires ubiquitin-associated (UBA) domain of protein hHR23 in order to interact with ATP activity-possessing components of 19S proteasome (Reed and Gillette, 2007). This interaction does not result in proteasome degradation but promotes XPC (Xeroderma Pigmentosum Complementation group C) protein stabilization by preventing this protein from being poly-ubiquitinated and recognized by the proteasome for degradation (Raasi and Pickart, 2002). XPC mono-ubiquitination is at least temporarily promoted (as poly-ubiquitination is inhibited) and may serve as the signal for further factors involved in NER recruitment.
A role of ubiquitination exists in DNA damage tolerance pathway. In this instance, after DNA damage the protein PCNA (Proliferating Cell Nuclear Antigen) is mono-ubiquitinated and recruits trans-lesion synthesis polymerases that bypass DNA lesion allowing replication despite lesion existence. In contrast, PCNA lysine 63 poly-ubiquitination promotes recovering of stalled replication fork at sites of DNA damage in an error-free manner (Chiu et al., 2006). Other DNA repair pathways such as base excision repair (BER), mismatch repair (MMR) and Double Strand Break (DSB) repair involve both proteolytic and non-proteolytic ubiquitin regulation (Vlachostergios et al., 2009).
Mono-ubiquitination is a signal involved in receptor endocytosis and lysosomal sorting. Many receptor tyrosine kinases (RTKs) such as EGFR (Epidermal Growth Factor Receptor) and PDGFR (Platelet-Derived Growth Factor Receptor) undergo ligand-induced mono-ubiquitination. In this process ligand-induced phosphorylation of the receptor gives the signal for receptor ubiquitination. E3 ligase cbl facilitates receptor ubiquitination and is the major E3 ligase for this purpose (Hugland et al., 2003). Ubiquitinated receptors interact with ubiquitin-binding proteins of the endocytic pathway and are escorted through clathrin-coated pits to clathrin-coated vesicles, endosomes and finally lysosomes. In this travel, surface receptors are transferred to different ubiquitin-binding proteins. Mono-ubiquitination in multiple receptor sites (multiple mono-ubiquitination) has also been found to play a role in receptor endocytosis. Cbl E3 ligase mediates also multiple mono-ubiquitination. Multiple mono-ubiquitination is believed to stabilize interaction of receptors with ubiquitin receptors in order to enhance their transfer to lysosomes. Some ubiquitin receptors may also recognize only multi-ubiquitinated RTKs through multiple domain interactions.
The type of ubiquitination performed which, as discussed, will specify the fate of the target protein depends on the E2 and E3 enzymes that are involved. It appears that HECT type E3s due to their distinctive mode of action retain the decision of the ubiquitination type while RING type E3s are more promiscuous in the type of ubiquitination performed and depend on their E2 partner in each case to define ubiquitination type (Ikeda and Dikic, 2008). For example, HECT domain ligase E6-AP forms lysine 48 ubiquitin chains, while RING domain ligase BRCA1 can mono-ubiquitinate substrates when interacting with E2 enzymes UBCH6, UBE2E2, UBCM2 and UBE2W, forms lysine 63-linked chains when interacting with E2 MMS2-UBC13 and lysine 48-linked ubiquitin chains when interacting with E2 UBE2K (Christensen et al., 2007).
The proteasome
The whole proteasome structure is called 26S proteasome representing a complex of 2.5 MDa. It is localized in both the nucleus and the cytoplasm, near the endoplasmic reticulum and even in the centrosome (Fabunmi et al., 2000). 26S proteasome is comprised of two parts: The 19S regulatory particle (RP) and the 20S Core Particle (CP), comprised in their turn of several protein sub-units each. After attachment of at least four ubiquitin molecules the target protein is recognized by specific sub-units of 19S regulatory particle (RP) of the proteasome. 19S RP is a multi-protein structure that caps the two sides of the core particle (CP) of the proteasome. 19S (also known with the alternative name PA700) is made of two sub-complexes called the lid and the base and a total of 17 peptide molecules. Six of them possess ATPase activity while the 11 others are non-ATPases. The lid sub-complex is comprised of eight sub-units, six of which contain a PCI [Proteasome, COP9 signalosome and eIF3 (eukaryotic Initiation Factor 3)] domain mediating interactions between them. One of the other two sub-units, S13 in mammals and Rpn11 in yeast, is the metallopeptidase that performs de-ubiquitination of the substrates in order for ubiquitin molecules to be recycled. Both S13 and the eighth lid sub-unit contain a so called MPN (Mpr1p and Pad1p N-terminal regions) domain (Hanna and Finley, 2007). Nevertheless Rpn8 lacks key residues in the MPN domain and has no metallopeptidase activity.
Figure 2: Schematic representation of the proteasome multi-protein complex.
The 19S base sub-complex is made up of the six ATPases and three other peptides. ATPases belong to the AAA (ATPases Associated with various cellular Activities) family and are able to hydrolyze all four nucleotide triphosphates and to alter the conformation of protein, preventing aggregation. Thus, they function to prevent aggregation of proteasome substrate proteins before these proteins enter the Core Particle to be degraded. AAA ATPases have also functions independent of their membership in the proteasome structure notably in transcription and membranes fusion (Hanson and Whiteheart, 2005; Meyer, 2005). The three other peptides of the 19S base possess ubiquitin recognition domains that allow them to recognize poly-ubiquitin chains.
The core particle of the proteasome is a cylinder-shaped multi-unit structure with a hollow central chamber (Rechsteiner, 2005). Inside this chamber enzymatic degradation of target proteins takes place executed by three enzymatic activity-possessing subunits of the CP. CP consists of four seven-member rings that are stacked one on the other. The two peripheral rings are similar and are called α rings and the two central rings are also similar and are called β rings. Each of the seven sub-units of the α and β rings is distinct resulting in the CP to be comprised of two copies each of 14 distinct sub-units. Three of the seven sub-units of the β rings, β1, β2 and β5 possess the enzymatic activities of the proteasome, trypsin-like (post-basic residues cleavage) activity, chymotrypsin-like (post-hydrophobic residues cleavage) activity and post-glutamyl (caspase-like or post-acidic residues cleavage) activity respectively. Resulting fragments after proteasome degradation range in general between 4 and 14 amino-acids in length (Wolf and Hilt, 2004).
Carcinogenesis processes: The role of the UPS
Normal cells need to obtain six essential capabilities to become malignant (Hanahan and Weinberg, 2000): Self sufficiency in growth signals, insensitivity to anti-growth signals, inhibition of apoptosis, limitless replicative potential, angiogenesis potential and ability to invade and metastasize. UPS is involved in the regulation of all these processes as will be discussed briefly below.
Cell cycle machinery is in the heart of cell growth and the final destination of growth and anti-growth signals. Cell cycle is regulated in multiple levels by the UPS. Proteins called cyclins are associated with Cyclin-dependent kinases (CDKs) to activate their actions of phosphorylation of substrates for the cell to progress through the different phases of the cell cycle. In late G1 phase, cyclin D in collaboration with CDKs 4 and 6 phosphorylates and inactivates protein Rb. As a result transcription factor E2F is freed to transcribe genes necessary for the progression into the S phase. Transcription of Cyclin D is induced by the β-catenin/TCF4 transcription factor complex. β-catenin is regulated by the UPS through degradation after phosphorylation and ubiquitination with the aid of E3 ligase βTrCP (β-Transducin repeat Containing Protein). The stability of Cyclin D is also regulated directly by the UPS. Proteasome degradation keeps it in low levels through the cell cycle except for its up-regulation in late G1 (Kitagawa et al., 2009). Cyclins E1, E2 and A in collaboration with CDKs 1 and 2 get cell through S phase into G2 and Cyclin B functions in collaboration with CDK1 at G2 phase and is degraded by the proteasome at late mitosis (Vodermaier, 2004). CDKs are further regulated by CDK inhibitors such as p21 and p27, the stability of which are also determined by proteasome degradation (Carrano et al., 1999; Bornstein et al., 2003). The E3 ligase facilitating degradation of these CDK inhibitors is a RING finger type E3 with four sub-units. Of these sub-units F-box protein Skp2 (S-phase kinase protein 2) is the substrate recognition sub-unit. The same SCF type E3 is involved in the degradation of other cell cycle inhibitors such as the Rb family protein p130. In contrast, a SCF ligase with three identical sub-units but a different substrate recognition sub-unit called Fbxw7 (alternatively named hCDC4 or Archipelago) is involved in the proteasome degradation of proliferator promoting transcription factor c-myc as well as of cyclin E (Onoyama and Nakayama, 2008). Given their respective substrates E3 ligase sub-units Skp2 and Fbxw7 are acting as an oncogene the former and a tumor suppressor the latter (Shapira et al., 2005; Onoyama et al., 2007).
Another point of particular importance of cell cycle regulation by the UPS is at the anaphase phase of mitosis. At that point the chromosomes are aligned at the center of the cell and develop connections through the centromere with both poles of the mitotic spindle. When all chromosomes have completed their attachment to both poles the signal is given for each sister chromatid to begin moving to a pole, detached from the other sister chromatid. Up to that point sister chromatids are kept attached at the centromere with the action of proteins cohesins. When all chromosomes are attached, APC/C (Anaphase Promoting Complex/Cyclosome), an E3 ligase, ubiquitinates the protein securin which is degraded by the proteasome (Castro et al., 2003). Securin is an inhibitor of the enzyme separase, which, after securin destruction, is activated and cleaves cohesins allowing sister chromatids to be pulled to the two poles at the end of anaphase. In parallel APC/C promotes the destruction of Cyclin B allowing dephosphorylation and inactivation of CDK1, another prerequisite for progression from anaphase to telophase and completion of mitosis (Matyskiela et al., 2009).
Figure 3: Schematic representation of events leading to sister chromatids separation in anaphase.
Apoptosis is another process important in carcinogenesis that is regulated by the UPS. Many proteins of the cellular core apoptosis machinery are substrates of the proteasome. Bcl-2 family includes both pro-apoptotic and anti-apoptotic members and both categories contain members that are proteasome substrates. UPS regulates the balance between the pro-apoptotic and anti-apoptotic family members which in turn will determine ultimate cell fate after various stimuli (Yang and Yu, 2003).
IAPs (Inhibitors of Apoptosis) are a family of RING finger E3 ligases that inhibit apoptosis through ubiquitination and degradation of effectors of apoptosis, caspases. Apoptotic stimuli promote auto-ubiquitination of IAPs which leads to caspase stabilization in order to perform their apoptotic function (Vaux and Silke, 2005; Ni et al., 2005).
p53 is a transcription factor of importance for the induction of apoptosis after DNA damage and thus, it has been named "the guardian of the genome". p53 is regulated by the UPS through ubiquitination by several E3 ligases. Mdm2 (mouse double minute 2, also known as hdm2 in humans) is the first identified E3 ligase that ubiquitinates p53 for proteasomal degradation. In different stress conditions, p53 degradation is inhibited either through its phosphorylation that prevents interaction with mdm2 or through inhibition of mdm2 activity through interaction with inhibitor p14 ARF (Alternative Reading Frame, a name that this protein takes from the fact that it is transcribed from the same DNA sequence but with a different reading frame with the CDK inhibitor p16INK4a at chromosome 9p). p53 degradation is also prevented by de-ubiquitination by the enzyme HAUSP (Herpes virus-Associated Ubiquitin Specific Protease). Other E3 ligases have been found to ubiquitinate p53. In papilloma virus-infected cells, the HECT domain E3 ligase E6-AP (E6-Associated Protein) binds with viral protein E6 and promotes p53 degradation, an event that, together with degradation of tumor suppressor Rb, greatly contributes to viral oncogenesis. PIRH2 (p53-induced RING H2) is a RING type E3 ligase that promotes p53 ubiquitination independently of mdm2 and inhibits p53 transcription (Leng et al., 2003). Like mdm2, PIRH2 is a p53 target gene, this fact serving in both occasions as a negative feed-back loop. Another ubiquitin ligase ubiquitinating p53 is ARF-BP1/Mule (ARF-Binding Protein 1/Mcl1 ubiquitin ligase E3). This is a HECT domain ubiquitin ligase that, as its name implies, can be bound and inactivated by p14/ARF, in a manner analogous to mdm2 (Chen et al., 2005). ARF-BP1/Mule inactivation leads to promotion of apoptosis in both p53-dependent and -independent ways implying that the ligase has other apoptosis promoting substrates besides p53. In addition it ubiquitinates and promotes degradation of an anti-apoptotic protein, the bcl2 family member Mcl1 (Zhong et al., 2005). ARF-BP1/Mule possesses a BH3 (Bcl2 homology 3) domain through which it interacts with Mcl1. As a result of having both p53 and Mcl1 as a substrate, ARF-BP1/Mule can promote or impede apoptosis under different conditions (Shmueli and Oren, 2005). Finally, COP1 (Constitutively Photomorphogenic 1), a RING domain E3 ligase, is also promoting p53 degradation (Dornan et al., 2004).
Figure 4: Regulation of p53 by ubiquitination. Ubiquitin ligases involved in p53 ubiquitination are depicted.
The existence of multiple pathways regulating p53 stability and degradation by the UPS allow both a strict control of its function and a versatility of its activation and inhibition. Nevertheless the UPS constitutes a vital component of all pathways.
In order for a cell to obtain limitless replicative potential, it needs to neutralize the mechanism that shortens telomeres with each successive division and limits the total number of cell cycles that it can successfully undergo. A protein called TRF1 (Telomeric Repeat binding Factor 1, alternatively called PIN2- Protein Interacting with NIMA 2) binds telomeres and prevents access of telomerase, thus physiologically preventing telomere length maintenance through the action of telomerase. In this way, in normal cells, telomere length is decreased with each successive cell cycle. Casein kinase 2 phosphorylates TRF1 and promotes its binding to telomeres (Kim et al., 2008). In contrast in neoplastic cells, TRF1 is ADP-ribosylated by a poly(ADP-ribose) polymerase (PARP), tankyrase and dissociates from telomeres (Smith and de Lange, 2000). Dissociated TRF1 is then ubiquitinated with the mediation of F-box family E3 ligase Fbx4 and degraded by the proteasome (Chang et al., 2003; Lee et al., 2006). This degradation allows telomerase to access the telomere and perform telomere length maintenance contributing to limitless replicative potential avoiding chromosome erosion that would lead to apoptosis.
Figure 5: The role of UPS in telomere maintenance in cancer. In normal cells (left), telomerase access to telomeres is prevented by protein TRF1 and telomeres are shortened with each cell division. In neoplastic cells (right), after ADP-ribosylation, TRF1 is displaced from the telomere and is ubiquitinated and degraded by the proteasome. As a result, telomerase can access telomeres and prevent their shortening.
Angiogenesis is a crucial process in carcinogenesis and is regulated by the UPS in multiple levels. For example, the α sub-units of transcription factor HIF-1 (Hypoxia Inducible Factor-1) is kept suppressed under normoxic conditions by proteasome degradation (Corn, 2007). This degradation requires the action of oxygen sensing prolyl-hydroxylases (PHDs) that hydroxylate HIFα in two proline residues (Pro402 and Pro564) of a so-called oxygen-dependent degradation domain (ODD). Proline hydroxylation gives the signal for HIFα ubiquitination with the help of E3 ligase complex consisting of VHL (Von Hippel Lindau) protein, elongin B, elongin C, Cul2 and Rbx1. Ubiquitination is followed by HIFα proteasome degradation (Koh et al., 2008). In contrast, in hypoxia, prolyl-hydroxylases are inactive and HIFα remains hypo-hydroxylated and is stabilized in order to perform, in collaboration with constitutively present factor HIF-1β, a transcription program which induces dozens of genes among which genes important for angiogenesis, such as VEGF, are included. VHL protein constituent of HIFα's E3 ligase is mutated in Von Hippel Lindau syndrome which encompasses increased frequency of renal cell carcinomas (RCCs), retinal and central nervous system tumors as well as in sporadic RCCs, leading to constitutively active HIFα in this malignancy. PHDs are themselves proteasome substrates and their ubiquitination is mediated by E3 ligases Siah1 and 2 (Seven in absentia Homolog 1 and 2) (Nakayama and Ronai, 2004).
Figure 6: Regulation of transcription factor HIF-1 by the UPS. In normoxia (left), HIF1 is hydroxyprolinated and ubiquitinated with the aid of E3 ligase VHL to be degraded by the proteasome. In hypoxia (right), hydroxylation of HIF-1 is inhibited and the transcription factor is stabilized in order to perform its transcription program.
Several other control points of angiogenesis by the UPS exist and include, as another example, direct HIF transcription regulation by proteasome-dependent and -independent functions of ubiquitin and regulation of the intra-cellular signal emanating from VEGFR (the receptor of VEGF).
Invasive and metastatic potential is another characteristic of the neoplastic cell and is also regulated by the UPS. Activation of several receptor tyrosine kinases such as EGFR, PDGFR and GDNFR (Glial cell line-Derived Neurotrophic Factor Receptor, also known as ret) favour invasion. These receptor proteins are proteasome substrates (Gur et al., 2004; Pierchala et al., 2006; Kim et al., 2008; Baron and Schwartz, 2000) and the same is true for intracellular proteins that take part in signal transduction such as Akt and ERK (Adachi et al., 2003; Mikalsen et al., 2005) as well as transcription factors that are final effectors of the pathways.
Lysophosphatidic acid (LPA) receptors are also an example of invasion and motility promoting receptors. They are seven domain membrane spanning G-protein-coupled receptors. GBM (Glioblastoma multiforme), a central nervous system malignancy characterized by a propensity of tissue invasion, expresses high levels of LPA receptors which are stimulated by LPA derived from lysophosphatidylcholine through the action of autotaxin, an enzyme with lysophospholipase D activity also produced and secreted by GBM cells (Kishi et al., 2006; Hoelzinger et al., 2005). Autotaxin gene is under the control of transcription factor β-catenin which, as already mentioned, is proteasome-regulated (Kenny et al., 2005).
UPS role in cancer: The example of colorectal carcinogenesis
Colorectal cancer develops along two major pathways. In the first pathway which takes place in about 85% of sporadic colorectal cancer patients as well as in patients with the hereditary syndrome Familial Adenomatous Polyposis (FAP), there is a sequence of molecular events leading stepwise from hyperplasia to adenoma to carcinoma. These cases have the characteristic of chromosomal instability. The remaining 15% of sporadic cases share molecular pathogenesis with another hereditary syndrome, hereditary non-polyposis colorectal cancer (HNPCC). In these cases there are mutations of genes involved in mismatch repair (MMR) of DNA such as MSH2, MLH1 and PMS2 leading to microsatellite instability (Voutsadakis, 2007).
The FAP type sequence begins with mutations in the gene encoding for APC (Adenomatous Polyposis Coli) protein. This is a protein taking part in a complex together with scaffolding proteins axin and conductin and kinases GSK3β (Glycogen Synthase Kinase 3β) and CKII (Casein Kinase II) that facilitates phosphorylation of transcription factor β-catenin, leading afterwards to ubiquitination with the help of E3 ligase TrCP and proteasomal degradation of β-catenin (Ilyas, 2005). If APC acquires debilitating mutations in both alleles as it happens in about 85% of sporadic colorectal carcinomas, or has already a germline mutation in one allele and acquires a mutation in the other allele as it happens in FAP syndrome, β-catenin cannot be ubiquitinated and degraded by the proteasome and thus, it remains constitutively active to perform a proliferation program leading to formation of lesions called aberrant crypt foci (ACF) in the colon. These are the first lesions in this sequence of colorectal carcinogenesis. Subsequently, activating mutations in the oncogene k-ras promote progression of ACF to adenoma. These mutations activate proliferation programs normally emanating from receptor tyrosine kinases without the need for receptor activation. Both pathways down-stream of activated k-ras, the Raf/MAPKs pathway and the PI3K/akt pathway have members that are regulated by ubiquitination and proteasome degradation, while additional intersection of k-ras-activated pathways and the UPS exist at the level of transcription factors activated by MAPKs such as AP-1 (Activated Protein 1) given that transcription is a process that requires ubiquitin in both a proteasome-dependent and -independent manner (Voutsadakis, 2008).
Next step of colorectal carcinogenesis, the transition from adenoma to carcinoma requires accumulation of additional lesions such as p53 and Smad4 (also known as DPC4- Deleted in Pancreatic Carcinoma 4) mutations. These are also pathways that are UPS regulated. As already discussed in a previous section, stability of p53 is regulated by multiple E3 ligases. Smad4 is part of the TGFβ (Transforming Growth Factor β) signalling cascade and is also a proteasome substrate for proteolytic regulation.
The other sequence of colorectal carcinogenesis involving lesions in MMR genes is also UPS-regulated in multiple levels (Hernandez-Pigeon et al., 2004; Hernandez-Pigeon et al., 2005). As mentioned in a previous paragraph, DNA repair processes in general are UPS regulated.
It becomes clear from the above discussion that several important lesions and pathways involved in both the FAP type sequence and the MMR sequence of colorectal carcinogenesis are UPS regulated. Given this multitude of regulations in oncogenesis by UPS, there are multiple opportunities for therapeutic interventions. Paradoxically this same multitude may diminish the probability that a single intervention affecting UPS regulation would be beneficial in a wide range of cancers with diverse molecular lesions. In contrast there is the need for identification of sub-sets of cancer types with specific molecular lesions that will be particularly sensitive to a specific therapeutic intervention affecting the UPS. Therapeutic success with proteasome inhibitor bortezomib in multiple myeloma creates a hope that inhibition of UPS can be a valid target in other types of malignancies and specific sub-sets of tumor locations. Further clinical trials based on the rational of solid pre-clinical data are needed to identify them.