Integrins are transmembrane receptors composed of α and β subunits. Although most integrins contain β1, canonical activation mechanisms are based on studies of the platelet integrin, αIIbβ3. Its inactive conformation is characterized by the association of the αIIb transmembrane and cytosolic domain (TM/CT) with a tilted β3 TM/CT that leads to activation when disrupted. We show significant structural differences between β1 and β3 TM/CT in bicelles. Moreover, the ‘snorkeling’ lysine at the TM/CT interface of β subunits, previously proposed to regulate αIIbβ3 activation by ion pairing with nearby lipids, plays opposite roles in β1 and β3 integrin function and in neither case is responsible for TM tilt. A range of affinities from almost no interaction to the relatively high avidity that characterizes αIIbβ3 is seen between various α subunits and β1 TM/CTs. The αIIbβ3-based canonical model for the roles of the TM/CT in integrin activation and function clearly does not extend to all mammalian integrins. DOI: http://dx.doi.org/10.7554/eLife.18633.001, eLife digest Proteins called integrins span the membranes of most human cells, and help our cells to interact with their surroundings, enabling them to organise, communicate and to form a variety of structures. Cells in different parts of the body typically produce different integrins so that they can specifically connect with other cells and proteins in their local environment. There are many different kinds of integrin proteins found in cell membranes and they consist of one alpha and one beta subunit. Different integrin pairs can have different effects based on their environment and the other molecules that they encounter. Much of the research into how integrins work has involved one specific integrin found in platelets – cells in blood that aid clotting and wound repair. Yet, it is unknown if all integrins actually operate in the same way as the platelet integrin. Lu, Mathew, Chen et al. studied the part of integrins that are located inside cells (referred to as the cytoplasmic tail) and the part that crosses the membrane (the transmembrane domain). Three-dimensional structures of these parts of the proteins showed that they varied between different beta integrin proteins. Further experiments revealed that the strength of the association between different alpha and beta integrins also varied. Finally Lu, Mathew, Chen et al. demonstrated that components shared by several beta integrins actually have different purposes in different contexts. The diversity of structures and interactions within the group of integrin proteins suggests that integrins are likely to behave very differently in different cells. This means that platelet integrins cannot be used to fully understand the activity of all other types of integrin. More work is now needed to understand how the differences between integrins affect the roles that they fulfil and the molecules that they interact with. A deeper understanding of the differences between integrins could ultimately shape the development of strategies to specifically target them to treat a range of diseases – such as cancer and diseases in which there is a build-up of fibrous connective tissue. DOI: http://dx.doi.org/10.7554/eLife.18633.002