It's hard for crystallographers to determine the structures of all the macromolecular assemblies that cell biologists want to get a close look at. But researchers can piece together precise molecular structures by considering other readily available data, Fernandez-Martinez et al. suggest.
Their new work builds on their previous findings that highlighted the value of low-resolution data, which most labs can obtain fairly easily but rarely use to infer molecular structures. For example, with a technique called domain deletion mapping, which involves pruning particular proteins and then determining which interactions remain and which are disrupted, researchers can deduce the connections between proteins and their orientations within a complex.
Fernandez-Martinez et al. applied the method to the Nup84 complex, which contains seven proteins. Sixteen copies of the complex form the outer rings of the yeast nuclear pore complex (NPC). Electron microscopy and domain deletion mapping allowed the team to obtain spatial restraints, or structural limitations on the complex's architecture. Using a computer algorithm, they could then build possible structures for the Nup84 complex that satisfied these restraints.
Although the analysis confirmed some previous findings, such as the complex's overall arrangement and “Y” shape, it also revealed new features and linked particular functions to specific structures. For example, the findings suggest that three proteins at the top of the “Y,” Nup85, Nup120, and Seh1, are particularly important for linking the complex to the rest of the NPC, whereas Nup120 and a related protein, Nup133, are crucial for stabilizing the NPC's interactions with the nuclear envelope.