Genes VII

How does a nuclear pore accommodate the transit of material of varied sizes and characteristics in either direction? Nuclear pore complexes have a uniform appearance when examined by microscopy. The pores can be released from the nuclear envelope by detergent, and Figure 8.36 shows that they appear as annular structures, consisting of rosettes made of 8 spokes. Figure 8.37 shows a model for the pore based on three-dimensional reconstruction of electron microscopic images. It consists of an upper ring and a lower ring, connected by a lattice of 8 structures.

The basis for the 8-fold symmetry is explained in terms of individual components in the schematic view from above shown in Figure 8.38. This includes the central structure of Figure 8.37, and extends it with surrounding radial arms and an internal transporter. The outside of the pore complex as such consists of a ring of diameter ~120 nm. The ring itself consists of 8 subunits. The 8 radial arms outside the ring may be responsible for anchoring the pore complex in the nuclear envelope; they penetrate the membrane. The 8 interior spokes project from the ring, closing the opening to a diameter of ~48 nm. Within this region is the transporter, which contains a pore that approximates a cylinder <10 nm in diameter (Hinshaw et al., 1992).

The pore provides a passage across the outer and inner membranes of the nuclear envelope. As illustrated in Figure 8.39, the side view has two-fold symmetry about a horizontal axis in the plane of the nuclear envelope. There are matching annuli at the outer and inner membranes, comprising the surfaces that project into the cytosol and into the nucleus, and each is connected to the spokes, which form a central ring. (Only 2 of the 8 spokes are seen in this side view.) The spokes are symmetrical about the horizontal axis. The central pore projects for the distance across the envelope. Sometimes material can be seen within the pore, but it has been difficult to equate such sightings with the transport of particular material.

The size of the nuclear pore complex corresponds to a total mass ~50 10⁶ daltons (compare this with the 80S ribosome at 4 10⁶ daltons). We can identify the smallest repeating component by using the 8-fold symmetry as seen in cross-section (see Figure 8.38) and the 2-fold symmetry seen from the side (see Figure 8.39). This divides the scaffold into 16 identical units. Each of these units consists of ~30 different proteins, most often each present in 1-2 copies per unit (Rout el al., 2000). The central pore constitutes only a small part of the overall complex (for review see Forbes, 1992; Davis, 1995).

The ability of compounds to diffuse freely through the pores is limited by their size, and it is convenient to consider the material in three size classes:

• Molecules of <5000 daltons that are injected into the cytoplasm appear virtually instantaneously in the nucleus: we may conclude that the nuclear envelope is freely permeable to ions, nucleotides, and other small molecules.
• Proteins of 5-50 kD diffuse at a rate that is inversely related to their size, presumably determined by random contacts with, and passage through, the pore. It takes a few hours for the levels of an injected protein to equilibrate between cytoplasm and nucleus. We may conclude that small proteins can enter the nucleus by passive diffusion (but they may also be actively transported). The nuclear envelope in effect provides a mesh or molecular sieve that permits passage of material <50 kD.
• Proteins >50 kD in size do not enter the nucleus by passive diffusion; a mechanism of active transport must be required for their passage (Akey and Goldfarb, 1989; for review see Dingwall and Laskey, 1986)

The lattice-like structure of the nuclear pore suggests that different features could be responsible for active transport of large proteins versus passive transport of smaller proteins.

Eight channels are created by the open regions between the spokes, that is, around the periphery between the inner and outer annuli. These have a rigid structure. They are oval in cross-section, with a diameter of ~10 nm. We might speculate that this could provide a mesh to allow passive diffusion for small proteins. A globular protein of 50 kD in mass would have a diameter of ~5 nm if it were spherical. Presumably objects would need to be somewhat smaller than the mesh size in order to be able to pass through by diffusion.

If these channels allow free passage, material below the size limit will equilibrate in concentration between the nucleus and the cytoplasm (although as the material approaches the size of the mesh, equilibration is slow). Proteins that enter the nucleus in this way would be retained, and therefore accumulate in the nucleus, if they participate in some nuclear function. For example, if a protein becomes incorporated into a large structure, such as a chromosome, this removes it from the equilibrating pool, and thus pulls more protein into the nucleus.

The central pore is used to transport larger material. A protein requires a specific signal to pass through the central pore. Smaller proteins may be transported in this way (as well as through the peripheral channels), and larger proteins must use an active transport mechanism that overcomes the apparent size restriction of the pores. But how can a protein or ribonucleoprotein with a diameter exceeding that of the pore pass through it?

Transport through the pore has been characterized by using colloidal gold particles coated with a nuclear protein. When these particles are injected into the cytoplasm, they cluster at the nuclear pores, and then accumulate in the nucleus. This suggests that the pore structure can widen to accommodate objects of the size of the coated gold particles (~20 nm). Similar experiments have shown that gold particles coated with polynucleotides can be exported from the nucleus via pores (Feldherr et al., 1984).

The rigidity of the gold particle excludes the possibility that transport through the pore requires the protein to change into a conformation with a diameter physically smaller than the pore. We conclude that the nuclear pore has a "gating" mechanism that allows the interior to expand as material passes through. Pores engaged in transporting material appear to be opened to a diameter of ~20 nm, possibly by a mechanism akin to the iris of a camera lens. It is possible that two irises, one connected with the cytoplasmic ring and one connected with the nucleoplasmic ring, open in turn as material proceeds through the pore. Very large substrates, such as exported ribonucleoprotein particles, may have to change their conformation to conform with the limit of 20 nm.

The nuclear pore complex provides a structural framework that supports the proteins actually responsible for binding and transporting material into (or out of) the nucleus; it does not include all of the active components that are involved in binding and translocation. A major question about nuclear pores is whether they are all identical, or whether their uniform appearance disguises functional differences. We should like to know whether the same pores undertake transport into and out of the nucleus or whether different classes of pores exist. We have yet to determine to what extent the functions of the pore are intrinsic as opposed to being determined by the accessory factors with which it associates.

This section updated 4-22-2000