Protein ubiquitylation and protein phosphorylation as cellular regulators and as systems for focused drug discovery
Protein phosphorylation
Discovered in the mid 1950's as a mechanism for the regulation of glycogen metabolism, it became clear many years later that this protein modification regulates almost all aspects of cell life. In protein phosphorylation, phosphate becomes attached covalently to serine, threonine and tyrosine residues by enzymes known as protein kinases, and is removed by a distinct class of enzymes, termed protein phosphatases. Phosphorylation of a protein can alter its conformation and ability to function in almost any conceivable way, for example by switching the biological activity of the protein on or off, by altering its stability or its interaction with other proteins, or by causing it to move from one cellular compartment to another. It is the simplicity, flexibility and reversibility of protein phosphorylation, coupled with the availability of ATP as a phosphoryl donor, which explains why this covalent modification has been adopted so widely to regulate cell functions.
Protein ubiquitylation and ubiquitin-binding proteins
Discovered in the late 1970s as a mechanism for marking proteins for proteolytic destruction, it is now obvious that protein ubiquitylation regulates as many aspects of cell life as phosphorylation. In protein ubiquitylation, the C-terminus of a small, but very abundant 8 kDa cellular protein, termed ubiquitin, becomes attached covalently to lysine residues in other proteins, a process that requires an E1 activating enzyme, an E2 conjugating enzyme and an E3 ligase. The ubiquitin is removed from proteins by isopeptidases, termed deubiquitylases or DUBs. Reversible protein ubiquitylation is analogous to reversible phosphorylation in-as-much as a particular molecule (ubiquitin or phosphate) is attached to or removed from proteins to reversibly alter their function. However, ubiquitylation is an even more versatile control mechanism because, in contrast to phosphoryation, additional molecules of ubiquitin can be linked covalently to the first ubiquitin attached to a protein to produce polyubiquitin chains and polyubiquitylated proteins. Moreover, at least eight types of polyubiquitin chain can be formed in which any one of the seven lysine residues or the ?-amino group of ubiquitin can be linked covalently to the preceding ubiquitin in the chain. The different types of polyubiquitin chain adopt different conformations, which enables them to interact with distinct polyubiquitin binding proteins, and so regulate different cell functions. For example, polyubiquitin chains formed by linkage to Lys48 of ubiquitin (termed Lys48-linked polyubiquitin chains) interact with a particular component of the proteasome and thereby target proteins carrying such polyubiquitin chains for proteolytic destruction. In contrast, Lys63-linked or linear polyubiquitin chains can act as scaffolds for the recruitment of signalling complexes that regulate the innate immune system or endocytosis, while Lys11-linked polyubiquitylation is emerging as an important mechanism for regulating mitosis, and Lys29/Lys33-linked polyubiquitylation prevents the activation of a particular subfamily of protein kinases. It has become obvious that cells contain a bewildering array of ubiquitin- and polyubiquitin-binding proteins and that the characterisation of these proteins and elucidation of their physiological roles will be crucial for our understanding of how the ubiquitin system controls cell functions at the molecular level.
Ubiquitin-like modifiers
Further complexity in the ubiquitin system arises from the presence of a number of ubiquitin-like modifiers (ULMs), such as SUMO, NEDD8, FAT10 and ISG15 that can also be attached to proteins, in processes known at SUMOylation, NEDDylation, FAT10ylation, ISGylation and so on. Since ULMs, like ubiquitin, become linked covalently to lysine residues on proteins, this creates the potential for such modifications to prevent protein ubiquitylation (or vice versa). Moreover, some E3 ubiquitin ligases can only ubiquitylate proteins that are SUMOylated, while NEDDylation of the Cullin-RING E3 activates its ubiquitin ligase function. Thus, the different types of ubiquitin-like modification do not operate in isolation from one another, but are inextricably interlinked.
The interplay between protein ubiquitylation and protein phosphorylation
In recent years it has also become obvious that protein phosphorylation and protein ubiquitylation are not distinct and separate control mechanisms, and that the interplay between these systems is extensive. For example, a number of E3 ubiquitin ligases are activated or inactivated by phosphorylation. Conversely, the formation of Lys63-linked or linear polyubiquitin chains is critical for the activation of protein kinases that are important for the control of the innate immune system, while other protein kinases are inactivated by ubiquitylation. Moreover, some E3 ligases, such as members of the SCF-Cullin family, can only recognise and ubiquitylate their protein targets if these targets are phosphorylated. I am convinced that understanding the interplay between protein phosphorylation and protein ubiquitylation will become an increasingly important aspect of the study of cell regulation over the next decade.
Drug discovery in the ubiquitin system; lessons from protein phosphorylation
Abnormal levels of protein phosphorylation are a cause or consequence of many diseases and, in recent years, protein kinases have become the pharmaceutical industry's most important class of drug target, accounting for perhaps a third of R&D expenditure and over half of cancer drug discovery. Ten drugs that target protein kinases have been approved for clinical use in cancer and hundreds more are undergoing clinical trials. This situation is remarkable given that 15 years ago the pharmaceutical industry had little interest in protein kinases as drug targets. The situation changed with the advent of the first low nanomolar inhibitors of protein kinases that were relatively selective, followed by the demonstration in 1998 of the remarkable efficacy of Glivec. This drug, which inhibits the Bcr-Abl tyrosine kinase, transformed a rapidly fatal form of leukaemia into a manageable disease, and had a huge impact on the perception of protein kinases as drug targets.
Velcade (also called Bortezomib) was the first drug to be developed by targeting a component of the ubiquitin system. Approved for clinical use in 2007, this proteasome inhibitor is having a major impact in the treatment of multiple myeloma, a B cell lymphoma. Several other proteasome inhibitors and a drug that targets the E1 activating enzyme for NEDDylation have also entered clinical trials as anti-cancer agents. This raises the question of whether the ubiquitin system will eventually furnish the pharmaceutical industry with as many drug targets as protein kinases already have. Interestingly, the number of proteins involved in reversible phosphorylation and reversible ubiquitylation is comparable. There are over 500 protein kinases encoded by the human genome compared to 10 E1 activating enzymes, 45 E2 conjugating enzymes and around 600 E3 ligases, and 140 protein phosphatases compared to nearly 100 DUBs. Moreover, a number of diseases have been identified that are caused by mutations in E3 ligases or DUBs. I have no idea whether the ubiquitin system will become as important as protein kinases for the pharmaceutical industry. However, I am confident that in ten years time the ubiquitin system will be far more important for the pharmaceutical industry than it is at present.
by Professor Sir Philip Cohen FRS, FRSE
Director of the UK Medical Research Council Protein Phosphorylation Unit
Director of the Scottish Institute for Cell Signalling and Protein Ubiquitylation Unit
Co-Director of the Division of Signal Transduction Therapy
Dundee, February 2010
