Choose your partners: dimerization in eukaryotic transcription factors Grigoris D. Amoutzias 1, 2 , David L. Robertson 3 , Yves Van de Peer 1, 2 and Stephen G. Oliver 4 1 Department of Plant Systems Biology, VIB, Ghent University, Technologiepark 927, B-9052 Ghent, Belgium 2 Department of Molecular Genetics, Ghent University, Technologiepark 927, B-9052 Ghent, Belgium 3 Faculty of Life Sciences, University of Manchester, Manchester, M13 9PT, UK 4 Department of Biochemistry, University of Cambridge, Sanger Building, 80 Tennis Court Road, Cambridge CB2 1GA, UK In many eukaryotic transcription factor gene families, proteins require a physical interaction with an identical molecule or with another molecule within the same family to form a functional dimer and bind DNA. Depend- ing on the choice of partner and the cellular context, each dimer triggers a sequence of regulatory events that lead to a particular cellular fate, for example, proliferation or differentiation. Recent syntheses of genomic and func- tional data reveal that partner choice is not random; instead, dimerization specificities, which are strongly linked to the evolution of the protein family, apply. Our focus is on understanding these interaction specifi- cities, their functional consequences and how they evolved. This knowledge is essential for understanding gene regulation and designing a new generation of drugs. Dimerization in signal transduction Dimerization, in a biological system, is defined as the formation of a functional protein complex composed of two subunits [1]. In general dimeric interactions in signal transduction pathways are not very stable. Rather, they are dynamic and act as reversible switches in the process of information flow [2]. Although dimerization is observed in many signal transduction and regulatory gene families [1,3], this review focuses on eukaryotic transcription factors (TFs). There are several families of TFs (e.g. the ETS-domain family, named for the avian E26 transform- ation-specific retrovirus) that form heterotypic dimers (see Glossary), but here we will discuss homotypic dimer- ization (i.e. dimerization among members of the same TF family). These family members can homodimerize or heterodimerize: two monomers might first dimerize in solution and then bind DNA (dimer pathway) or, Review Glossary 2R event: The two rounds of whole-genome duplication that probably happened in the common ancestor of vertebrates. A gene subfamily consisting of one gene would duplicate to two and then to four genes. bHLH: Basic-region helix–loop–helix. This is the largest family of dimerizing TFs in humans, with 118 members. Many are key regulators of developmental processes, differentiation or the cell cycle. Bilateria: Animals with a bilateral symmetry; that is, they have a front and a back end, as well as an upside and downside. They have bodies that develop from three different germ layers: the endoderm, mesoderm and ectoderm. Mollusks, worms, insects and vertebrates belong to this group. bZIP: Basic region leucine zipper. This is the second-largest family of dimerizing TFs in humans. They are encoded by 51 genes, many of which are well-studied oncogenes. Choanoflagellates: Flagellate unicellular eukaryotes. They are considered to be the closest living relatives of the metazoa. The last unicellular ancestors of animals are thought to have resembled modern choanoflagellates. Cnidaria: Animals with radial symmetry and two germ layers, endoderm and ectoderm. Corals, sea anemones and jellyfish are some of the animals that belong to this group. Enhanceosome: A complex of TFs and other proteins that assemble and bind cooperatively to the enhancer region of a gene. Eumetazoa: Animals with true tissues organized into germ layers, and an embryo that goes through a gastrula stage. Cnidaria and bilateria belong to this group, although sponges do not. HD-ZIP: Homeodomain leucine zipper. The homeodomain is a helix–turn–helix DNA-binding domain, consisting of three alpha helices. The third helix confers DNA-binding specificity. Heterodimer: A dimer formed between two non-identical proteins. Heterotypic dimer: A dimer formed between two proteins derived from different gene families. Homodimer: A dimer formed between two identical protein molecules. Homotypic dimer: A dimer formed between two proteins derived from the same gene family. Hub: A highly connected (promiscuous) protein in the protein–protein interaction network. MADS box: Named after the four originally identified members: MCM1, AGAMOUS, DEFICIENS and SRF1. This family expanded significantly in plants, with 107 genes in Arabidopsis thaliana. Many of these genes encode regulators of organ development in plants. Metazoa: Animals. Eumetazoa and sponges belong to this group. NFAT: Nuclear factor of activated T cells. It is closely related to the NF-kB family of TFs. NF-kB: Nuclear factor-kappa B. This is a very well-studied gene family that is involved in the regulation of the immune system, among other processes. NR: Nuclear receptor. This is the third-largest family of dimerizing TFs in humans, with 48 members. They constitute major pharmaceutical drug targets for a wide range of diseases. Opisthokonta: Monophyletic group of eukaryotes, including both the metazo- an and fungal lineages, together with the phylum choanozoa. Their flagellate cells propel themselves with a single posterior flagellum. Paralogs: Genes or proteins of the same phylogenetic group in a given organism that are very similar in sequence. They originate from a common gene by one or more gene-duplication events. PAS domain: Named after three proteins in which it occurs: Per, period circadian protein; Arnt, Ah receptor nuclear translocator protein; and Sim, single-minded protein. Peripheral member: A poorly connected protein in the protein–protein interaction network. Protein array: A high-throughput technology that spots whole proteins or protein domains onto glass slides at high density. It is analogous to a DNA microarray but is used for assaying protein–protein interactions. Scale-free network: A network whose distribution of connectivity (number of proteins with K interactions versus K interactions) decays in a power-law fashion. It has a few hubs and many peripheral members. Its special topology is linked to robustness and fast information flow. STAT: Signal transducers and activators of transcription. This family has attracted a lot of attention because of its involvement in cancer and its potential as a target for therapeutics. Usually, it is connected to one of the hubs of the network. Vertebrates: Animals with backbones or spinal columns, for example, fishes, frogs, reptiles and mammals. Corresponding author: Oliver, S.G. (steve.oliver@mole.bio.cam.ac.uk). 220 0968-0004/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2008.02.002 Available online 9 April 2008