In PI-103 purchase breast tissue, both LMO4 or DEAF1 are thought to play roles in cell proliferation and ductal side-branching. Their abilities to increase proliferation of mammary cells mark both proteins as potential contributors to breast tumour growth and metastasis. LMO4 is present in all human breast tumour subtypes, with.50% of primary tumours showing increased levels of expression, with a high level of nuclear LMO4 expression being associated with poor patient survival. Forced overexpression of LMO4 causes mammary epithelial cells to proliferate ex vivo, increases mammary cell populations in a transgenic mouse model, and promotes cell invasion and tumour formation in human cell lines. Although LMO4 contains little more than two proteinbinding LIM domains, it can affect gene expression by modulating transcriptional events, presumably by recruiting transcription factors, including DEAF1. LMO4 and DEAF1 are co-expressed in breast tissue and were shown to interact in mammalian two-hybrid assays. Given the potential functional significance of this interaction in breast cancer, we sought to understand how LMO4 and DEAF1 might cooperate to regulate cell proliferation. In this work, we used a combination of yeast two-hybrid, biophysical and cell-based assays to identify a tightly defined LMO4-binding region of DEAF1. This region contains a specific LMO4-interaction domain and the majority of a coiled coil domain encompassing the nuclear export signal of DEAF1. Further, we show that LMO4 can regulate the subcellular localisation of a DEAF1 construct incorporating the new LMO4-binding region. Together these results support the idea that high levels of LMO4 in the nucleus, which is a hallmark of sporadic breast cancers, may upset the delicate balance between interactions with partner proteins such as DEAF1. Our data indicate that the DEAF1 coiled coil forms a tetramer in vitro, and contributes to a bipartite LMO4-binding motif in yeast two-hybrid assays. The native tetrameric coiled coil can be replaced by a non-native dimeric coiled coil with only a moderate loss of apparent affinity in this assay. Our current model for binding is that DEAF1404–438 makes direct contacts with LMO4 in a manner similar to other well characterised LMO and LIM-homeodomain binding domains and DEAF1457–479 either stabilises the construct, or provides an appropriate self-association state for the interaction with LMO4. The presence of an NES in the coiled coil domain is not uncommon; NESs can be found in leucine rich segments of proteins, including coiled coil domains, located proximal to disordered regions. Leucine rich NESs from at least two different proteins bind the exportin protein CRM1 as helices. Conserved leucine residues that form the hydrophobic core of the coiled coil are critical for recognition by the exportin protein. The DEAF1 coiled coil sequence resembles a typical NES. In this scenario the DEAF NES would only become available to exportins either by movement of the helix containing the NES, or monomerisation of individual helices.