Furthermore, far from being a disadvantage, lack of specificity
can actually be an advantage. For example, Gleevec was developed
as an Abelson kinase inhibitor for the treatment of a specific
type of leukemia. However, it is also an effective treatment for
gastrointestinal stromal cancers because it inhibits the c-Kit
receptor and the platelet-derived growth factor (PDGF) receptor
tyrosine kinases, which are overexpressed or mutated in gastrointestinal
cancers (Demetri et al., 2006). In addition, the efficacy
of several anticancer drugs depends on their combined inhibition
of several different kinases, and these drugs may be less prone
to the development of drug resistance than ones that act on only
one specific kinase. Thus, some of the original prejudices against
protein kinases as drug targets have subsequently turned out to
have little substance.
The beauty of targeting protein kinases for therapeutics and
the basis for their popularity is that the same technologies and
small-molecule libraries can be used to develop inhibitors of
many types of protein kinases in almost every therapeutic
area. However, the vast amount of medicinal chemistry that
has been carried out in recent years has meant that novel patent
space is becoming quite difficult to find. Plus, there is a growing,
but probably unfounded, concern that the most important drug
targets in this area have been fully exploited. Therefore, the pharmaceutical
industry has begun to wonder where they may find
the next large set of drug targets that can be tackled in a manner
analogous to protein kinases. In this Perspective, we discuss the
premise that components of the ubiquitin system are prime
candidates for these new targets.
Ubiquitination More Versatile than Phosphorylation?
Ubiquitination is the covalent attachment of a small protein,
ubiquitin (8.5 kDa), to other proteins. In the first step, a thioester
bond is formed between the C-terminal carboxylate group of
ubiquitin and the thiol or sulfhydryl group of a cysteine residue
on an E1-activating enzyme. Next, the ubiquitin is transferred
to a cysteine on an E2-conjugating enzyme. In the third step,
the E2 interacts with an E3 ligase, and the ubiquitin is then transferred
from the E2 enzyme to substrates, which also interact with
the E3 ligase. This last step can occur directly, as in the RING E3
ligases, or it can occur indirectly with the ubiquitin first transferred
to a cysteine residue on the E3 ligase before being linked
to the substrate, as in the HECT family of E3 ligases. Chains of
ubiquitin are created by the same enzymatic process.
Similar to phosphorylation, ubiquitin can be linked covalently
to only one or several amino acid residues on the same protein
(Figure 1). However, in contrast to protein phosphorylation, ubiquitin
can also form polyubiquitin chains. Ubiquitin has seven
lysine residues and an a-amino group; thus eight different types
of polyubiquitin chains can form (and probably more because
chains with ‘‘mixed’’ linkages are also present in cells).
Even greater versatility is provided by ubiquitin-like proteins,
such as Nedd8, SUMO (1, 2, and 3), FAT10, and ISG15, which
are also attached covalently to proteins in processes called neddylation,
SUMOylation, tenylation, and ISGylation, respectively.
The formation of polyubiquitin chains and the existence of these
‘‘ubiquitin-like modifiers’’ make the ubiquitin system a more
complex and potentially more versatile control mechanism
than phosphorylation.