The NPC-transporter, a ghost in the machine

Summary

Yeast FG nucleoporins are intrinsically-disordered proteins that contain cohesive molten globular regions and repulsive extended-coil regions. When placed along the central axis of the NPC, FG nups may self-assemble to create a novel transport channel that provides a series of docking sites for karyopherin-cargo complexes (Yamada et al., 2010).

The Nuclear Pore Complex (NPC) is a singular transport machine, both in its size and mechanism. This mammoth assembly is shaped like a wheel with eight spokes radiating from a central hub. In many eukaryotes, the spokes are connected by rings to form a spoke-ring complex, which resides in a pore created by a fusion of the inner and outer nuclear membranes. However, the nature of the central region of the NPC, which provides a gated conduit for cargo transport across the nuclear envelope, has remained mysterious. A recent paper by Michael Rexach and colleagues now provides new insights into the biophysical properties of intrinsically-disordered nuclear pore proteins, known as FG nups, which line the central transport region. In so doing, this work provides a plausible model for the central transport assembly in the NPC (Yamada et al., 2010).

Before describing this exciting new work some background information is in order. Given the large size of vertebrate NPCs, ~1200Å diameter and ~500Å in height, it is not surprising that this assembly is one of the largest known machines in eukaryotic cells. Thus, it was first seen in thin sections more than 50 years ago, and in the ensuing decades, the architecture and transport mechanism of the NPC have been studied extensively (see Strambio-De-Castillia et al., 2010). A large body of experimental data in the 1980s and 1990s showed that large cargo, including the Balbiani ring mRNP (Daneholt, 1997) and proteins, exemplified by colloidal gold particles coated with nucleoplasmin (Feldherr et al., 1984), both move through the center of the NPC. These diverse substrates are remarkably similar when visualized by EM as they transit the pore complex. In both cases, the cargo forms an extended configuration that spans the NPC and is restrained within a central region with a diameter of ~150-200Å. The first tomographic 3D reconstruction of a single, negatively-stained NPC revealed an elongated particle on the transport axis, described as a central plug (Unwin and Milligan, 1982). Subsequent studies showed a cylindrical “plug” in 3D reconstructions of S. cerevisiae and X. laevis NPCs (Yang et al., 1998; Akey and Radermacher, 1993), as well as in FEISEM images (Allen et al., 2000). Since cargo appeared to be caught in transit within this channel-like structure (Yang et al., 1998; Allen et al., 2000 and references therein), this feature was named the central transporter (Akey and Radermacher, 1993; Figures 1a, 1b).

Figure 1.

a- and b- Side and top views of the central transporter (pink) are shown within the Xenopus spoke-ring complex (blue; reproduced from Akey and Radermacher, 1993). c- and d- A detailed cartoon is shown of the new central transporter model, which is based on measured physical properties of FG nups and their relative positions along the vertical axis of the NPC (adapted from 1). Molten globule regions are shown as spheres and ellipses; extended-coils are indicated by wavy red lines. (left) A central cross-section is shown of the yeast NPC. (right) The model is shown as viewed from the cytoplasmic surface. (inset center): Various types of FG nups are shown diagrammatically (reproduced from Yamada et al., 2010).

NPCs provide a path for the flow of information between the cytoplasm and nucleus, which involves the transport of proteins and RNA complexes across the nuclear envelope. While the overall morphology of the NPC varies somewhat in different eukaryotes, NPCs share a set of ~30 nups which form the octagonal spoke-ring complex and the transport apparatus. Of particular importance is the FG nup family (for phenylalanine and glycine rich), which provide binding sites for transport factors, known as karyopherins or importins (Strambio-De-Castillia et al., 2010). The basic mechanism of nucleocytoplasmic transport utilizes a facilitated diffusion mechanism. Hence, transport involves a series of transient binding and release interactions between karyopherins and FG nups, which line the central path of the NPC. Current research has resulted in two models of translocation across the NPC, each of which is based on the unusual properties of intrinsically-disordered FG nups. In the Brownian affinity gating model, extended and flexible FG nups are thought to continuously sweep the transport axis of the NPC, creating a barrier to the diffusion of macromolecules. This barrier can be crossed more readily by cargo bound to transport factors that transiently bind to FG nups (Rout et al., 2003). In the gel phase model, regions of FG nups in extended-coil configurations are thought to form a gel-like matrix through FG-FG interactions. This in turn, creates a barrier for large macromolecules in the transport conduit. Cargo bound to transport factors may move through the gel due to transient binding, which causes a state change in the hydrogel, allowing transport complexes to diffuse inwards before another cycle of binding and release (Frey et al., 2006).

To further clarify the nature of the central transport region, Rexach and colleagues now describe a comprehensive analysis of the biophysical properties of yeast FG nups (Yamada et al., 2010). The data were then used to create a unifying model for the central transport region, which is consistent with known structural data and incorporates many properties of existing transport models. In brief, natively-unfolded FG nups contain distinct regions that behave quite differently. One subset of sequences form flexible and extended random-coils, while a second subset contain FG domains that form molten globules with a more defined hydrodynamic radius. Remarkably, these random-coils and molten globules can either be repulsive or attractive towards each other, depending upon whether they are strongly charged or relatively uncharged. Interestingly, the linear organization of random-coils and molten globules on each FG nup creates both longer and shorter molecules, described as trees and shrubs (Yamada et al., 2010). Models for each FG nup were then placed in their approximate locations within the NPC, where they line the central transport path (Figures 1c-1d, inset). This modeling step was made possible by a recent structure of the yeast NPC derived from biochemical and physical restraints (Alber et al., 2007).

In the new model, it is hypothesized that attractive properties of molten globular FG domains, when combined with the projected length of each FG nup, would naturally lead to self-assembly of rings in the close confines of the NPC. In this way, a series of octogonal rings would form on the central axis with each ring being formed by a particular type of FG molten globule. Importantly, inaccuracies in the current model of nup distribution within the NPC should not adversely affect the new hypothesis, but would alter the predicted position of a particular FG nup. Ring stacking along the central axis would then form a remarkable quaternary structure, a tube-like feature reminiscent of the central transporter in size, symmetry and topology, including the presence of a central pore (Figures 1c-1d). Unlike traditional channels constructed from folded proteins, the transporter might best be described as a quasi-ordered structure built from dynamic molten globules.

Importantly, the new model redefines the nature of the basic building blocks of the central transporter. Because of the unique properties of FG nups, the quaternary structure of the transporter may retain the necessary flexibility to allow repeated binding and release of karyopherin-cargo complexes, which lies at the heart of the facilitated diffusion mechanism. At the same time, the plasticity of a channel comprised of flexible building blocks may accommodate substrates of different sizes and shapes. This model also explains properties of the central transporter. For example, the transporter in EM studies is more weakly contrasted than the spoke region and is variable in appearance. The lower contrast of the transporter would arise from local plasticity of the FG nup molten globules and their lower average density. Depending upon the isolation procedure, the central transporter may have been “lost” from individual NPCs in some experiments due to proteolysis, mechanical disorder and/or partial disassembly. Indeed, a number of studies revealed NPCs without a central transporter, or with a blurred central feature, creating the perception that this cylindrical structure was a ghost in the machine!

How does the rediscovered NPC transporter accommodate current transport models? As so often happens, facets of current models may be germane to the new hypothesis. The observed plasticity of the central transporter and its mode of anchoring to the spoke-ring complex may entail dynamic aspects of the virtual gating model. In addition, flexible linkers that attach the transporter to the spokes could also participate in transport. At the same time, cohesive properties of FG nups that promote hydrogel formation may be manifest to some degree in ring assembly and give cohesiveness to the overall architecture of the central transporter. Significantly, the new transporter model can readily account for the fact that large cargos are restricted to a defined region on the central axis of the NPC during transit, which is consistent with computer modeling of FG nups within the spokes (Alber et al., 2007).

In summary, the current report from Rexach and colleagues now raises many questions concerning the nature and function of the central transporter. Are molten globular regions of FG nup rings cohesive enough to form a defined cylindrical channel through vertical ring stacking? Many intrinsically-disordered proteins become more ordered when they bind to a ligand: could FG molten globules become better folded when they interact with partners to form rings and stacks of rings? Do adjacent rings of FG nups cooperate during cargo translocation? Is the intrinsic flexibility and structural plasticity of FG nups harnessed to facilitate translocation within the transporter, perhaps through contraction and expansion of the rings? Are FG nups at the entrance/exit able to gate the central channel? Finally, do FG nups at the boundaries of the NPC adopt a more extended conformation that may guide karyopherin-cargo complexes down into the channel entrance? With so many questions remaining, it will be instructive to follow the re-emergent central transporter hypothesis and monitor its impact on studies of this remarkable transport process.