'Truly nut-like', as their name origin suggests, eukaryotic cells divide and conquer the intricacies of cellular physiology. A fine example is the spatial segregation of transcription and translation, with DNA kept snugly in the nucleus and RNA and proteins shuffling in and out of it to safeguard the readout of genetic mandates. This comes at the price of elaborate logistics, where nuclear pore complexes (NPCs) act as the sole molecular bouncers at the nuclear gates. With each NPC containing multiple copies of ∼30 nucleoporins that total nearly 1,000 subunits, the question is: how do you study such a massive structure?

“X-ray crystallographers have provided an almost complete catalog of high-resolution structures that build the NPC,” says Martin Beck from the European Molecular Biology Laboratory in Heidelberg. “However, the nuclear pore is so huge that obviously you cannot crystallize it entirely, so you need to combine different methods that integrate the different resolution levels in a meaningful way.”

To fit the available parts list in a bigger picture, Beck and colleagues resorted to a combination of state-of-the-art electron tomography (ET) and cross-linking mass spectrometry (XL-MS). Starting with human tissue culture cells, they successfully isolated a ten-protein NPC scaffold subcomplex, called hNup107, and subjected it to single-particle analysis by ET. The real challenges were to then assign the resulting low-resolution model to its exact location in the NPC and to fill in the resolved electron density with available higher-resolution structures.

Although cryo-ET has been the method of choice for studying the global NPC architecture, previous studies have been limited to resolutions coarser than 60 Å and thus have not permitted modeling of individual components. And as the high copy number of subcomplexes allows for both inter- and intrasubcomplex contacts, orthogonal techniques such as interaction screening methods have been themselves plagued by ambiguities. As a result, at least three starkly different architectural models for the NPC scaffold have been proposed.

Model of the NPC scaffold integrating different resolutions. Image adapted from data courtesy of M. Beck. Credit: Katie Vicari

What typically kill cryo-ET resolution are both the limited throughput and the very low signal-to-noise ratio of the data. With the latest generation of microscopes, detectors and automated data acquisition, however, “the data sets become larger; thus, the resolution goes up,” explains Beck. Nevertheless, he adds that although secondary-structure cryo-ET resolution is now possible in principle, high intrinsic dynamics of the NPC add an extra layer of complexity.

Working on isolated nuclear envelopes with cutting-edge instrumentation and a sophisticated set of analytic tools, Beck and colleagues managed to get a hold of NPC dynamics and to solve the NPC scaffold structure at 32-Å resolution. This allowed them not only to unambiguously assign the location of the predominant hNup107 subcomplex in the electron density map, but also to pinpoint the location of additional subcomplexes in the scaffold. The researchers used XL-MS to validate the resolved structure. To discriminate between inter- and intrasubcomplex contacts, they analyzed the isolated hNup107 complex in parallel to whole NPCs. Combining the resulting distance restraints with high-resolution structures offered an almost complete interpretation of the observed electron density for the hNup107 scaffold. Careful analysis of subcomplex arrangement and structure dynamics led to an architectural model that offers both a consensus to previous studies and insight into cargo transport mechanics.

“There are still other subcomplexes—which need to be looked at, [not only] in isolation but also in their exact position in the NPC,” says Beck. We are looking forward to a sequel.