A tangled chain is not particularly useful for carrying out its usual tasks. Similarly, when proteins get tangled up, interacting with each other to form oligomers, some of them might function incorrectly, or even cause harm. Oligomerization is essential for the oncogenicity of the BCR–ABL oncoprotein, which is responsible for a wide range of human leukaemias. In the February issue of Nature Structural Biology, Peter Kim and colleagues describe the crystal structure for the N-terminal oligomerization domain of BCR–ABL, together with a new mode of oligomer formation. This N-terminal structure is a promising target for the design of oligomerization inhibitors that could disrupt the transforming activity of BCR–ABL.

The BCR–ABL oncogene is formed by the fusion of the BCR (breakpoint cluster region) gene on chromosome 22 with the proto-oncogene ABL (Abelson murine leukaemia viral oncogene homologue) on chromosome 9. The activity of cytoplasmic ABL (c-ABL), a tyrosine kinase, is normally under tight control. However, fusion of the BCR sequences constitutively activates ABL, which is an essential step for BCR–ABL transformation. The first specific kinase inhibitor to be developed, Gleevec, targets the ABL kinase domain and has been shown to be clinically effective in the treatment of chronic myelogenous leukaemia. The N-terminal BCR oligomerization domain is also essential for ABL activation.

The structure showed that BCR–ABL first forms dimers, which then dimerize to form tetramers. Initially, two monomers form a coiled–coil motif between C-terminal α-helices and swap N-terminal α-helices. Subsequently, dimers stack onto each other to form tetramers, the four α-helices lying approximately in the same plane. The use of both coiled–coil packing and domain swapping, each recognized as being important mechanisms of oligomerization, leads to a very stable structure. Clustering of kinase domains is thought to be sufficient to induce kinase activation. The crystal structure indicates that a tetramer might be even more efficient than a dimer at bringing domains together.

Protein–protein interactions are ubiquitous in many diseases and are likely to deliver a wealth of new sites for drug targeting. The few examples of drugs that inhibit such interactions show that this can be a successful strategy. Dimerized domains of anti- and pro-apoptotic proteins block apoptosis, but small-molecule inhibitors of dimerization have been shown to induce cell death. Inhibition of the human immunodeficiency virus 1 (HIV-1) gp41 fusion protein with various peptides provides a further proof of principle. Coiled–coils are well-studied protein–protein interface motifs, and the solved structure of the oligomerization domain of BCR–ABL provides a template for the rational design of inhibitors to disrupt oligomerization.