The Nuclear Envelope and Nuclear Transport

Every protein in the nucleus was made in the cytoplasm. Proteins are made on ribosomes in the cytoplasm, so histones, polymerases and transcription factors all have to be brought in, while messenger RNAs and ribosomal subunits made inside have to be sent out. The border is busy in both directions at once. How does a nucleus collect the right proteins, keep them, and make traffic run the right way?

The idea: a gated border with a one-way rule

Think of a border with open gates. Anyone small can stroll across. Anything large needs a passport and a licensed courier who reads it. The couriers do not push anyone through: they walk the same open gates. What makes traffic one-way is a rule about where couriers put their loads down, and that rule depends on a chemical that is abundant on one side of the border only.

In the cell the gates are nuclear pore complexes, the passport is a short amino-acid signal, the couriers are transport receptors, and the chemical is Ran, a small GTPase, bound to GTP. Ran works like the GTPase switches in that lesson: on with GTP, off with GDP. The rest of this lesson builds each part and then looks at the experiments that found them.

The envelope, lamina and pore

The nuclear envelope is two lipid bilayers, an inner and an outer nuclear membrane, described as about 20 to 40 nm apart in animal cells. The outer membrane is continuous with the endoplasmic reticulum. So the narrow perinuclear space between the two membranes is ER lumen (a lumen is the space enclosed by an organelle’s membrane), topologically on the same side as the inside of the ER, as in that lesson. The inner and outer membranes meet at the pores, where the bilayer bends round to line a water-filled channel.

Under the inner membrane of animal cells lies the lamina, a meshwork of proteins called lamins that gives the nucleus mechanical stability and bridges the envelope and chromatin. Lamins belong to the intermediate-filament family, but in cryo-electron tomograms of extracted vimentin-null mouse fibroblasts they formed filaments only about 3.5 nm thick. Mutations in LMNA, the gene for A-type lamins, cause human diseases called laminopathies; a mechanically weaker nucleus is one proposed mechanism, and why one gene gives such varied diseases is still argued. Budding yeast has no lamins.

Each nuclear pore complex is built from roughly 30 different proteins, called nucleoporins, each present in multiple copies. Its scaffold has eightfold rotational symmetry. Lining the central channel are nucleoporins with long, disordered stretches rich in phenylalanine-glycine repeats, the FG repeats. Measurements of passive entry of inert probes into digitonin-permeabilized HeLa cells supported a model in which FG domains form the permeability barrier, with a fitted mesh radius of about 2.6 nm. How the FG domains do this, as a cohesive gel or as a brush of waving chains, is still debated. Nor is the scaffold rigid: in fission yeast, pores constricted reversibly during energy depletion and under hypertonic stress.

Worked problem: how soft is the barrier?

Textbooks once gave a sharp limit, often 40 kDa, above which proteins could not diffuse through the pore. Timney and colleagues tested this in 2016 in live budding yeast. They expressed fluorescent reporters of different size, bleached the cytoplasmic signal, and followed nuclear fluorescence as it returned to equilibrium. Model the excess nuclear signal above equilibrium as a single exponential:

x(t) / x(0) = exp(-t / τ)

Their normalized time constants were τ = 15 s for 27 kDa GFP and τ = 413 s for a 67 kDa reporter. Assume these values apply to the cell and ignore their uncertainty.

1. Half-times. Setting x/x(0) = 0.5 gives t½ = τ × ln 2. For GFP, 15 × 0.693 = 10.4 s. For the 67 kDa reporter, 413 × 0.693 = 286 s.
2. After one minute. For GFP, 60/15 = 4, and exp(-4) = 0.018, so 1.8% of the excess is left. For the large reporter, 60/413 = 0.145, and exp(-0.145) = 0.865, so 86.5% is left.
3. Does the large reporter cross at all? At 826 s, 826/413 = 2, and exp(-2) = 0.135. Only 13.5% of the excess remains, so it does cross, slowly.
4. What the ratio says. 413/15 = 27.5. A 2.5-fold rise in mass (67/27 = 2.48) slowed exchange about 27-fold. Permeability falls steeply but continuously with size: a soft barrier, not a cutoff.

These are fractions of excess signal in a yeast nucleus, not the time one molecule spends in one pore. For comparison, single NLS-carrying molecules were seen to interact with pores for about 10 ms on average in permeabilized HeLa cells.

Signals, receptors and the Ran gradient

A nuclear localization signal (NLS) is a short stretch of a protein that marks it for import. Classical NLSs are basic. A monopartite NLS has one cluster of basic residues, such as Pro-Lys-Lys-Lys-Arg-Lys-Val from SV40 large T antigen. A bipartite NLS has two basic clusters separated by a spacer, as in nucleoplasmin. Unlike the ER signal sequences in that lesson, an NLS is not cut off, so a protein can be imported again. The signal also depends on context: the same sequence works better in some proteins than in others.

In the classical pathway, importin α binds the NLS and importin β binds importin α. Importin β docks the complex at the pore and moves it through by binding FG repeats as it goes.

Now the one-way rule. RCC1, Ran’s exchange factor, sits on chromatin and loads GTP onto Ran, so RanGTP is made in the nucleus. RanGAP1 in the cytoplasm, helped by RanBP1, stimulates Ran to hydrolyse its GTP, so cytoplasmic RanGTP stays low. NTF2 carries RanGDP back into the nucleus to be recharged. The receptors read this gradient:

• Import. Importins load cargo where RanGTP is low. In the nucleus RanGTP binds importin β and the cargo is released. Importin β returns with RanGTP, and the exportin CAS returns importin α, also with RanGTP.
• Export. An exportin binds its cargo only together with RanGTP. CRM1, also called exportin 1, carries proteins with a leucine-rich nuclear export signal (NES), a short pattern of spaced hydrophobic residues, often leucines. In the cytoplasm RanGAP-stimulated hydrolysis breaks the complex apart.

Where is the energy spent? Not on pushing. In an in vitro assay, transportin, another importin, carried its cargo across pores with no Ran or nucleotide present; RanGTP, but not its hydrolysis, was needed to release the cargo and send the receptor back, and GTP hydrolysis appeared necessary only to reset the receptor for further rounds. Passage is best described as facilitated diffusion. GTP hydrolysis builds and resets the gradient that decides where loads are picked up and dropped.

Not all export uses this machinery. Most messenger RNA leaves through the receptor NXF1 (also called TAP), largely independent of Ran, and export of spliced mRNA proceeded without Ran in a Xenopus oocyte assay. What happens to RNA before it leaves belongs to that lesson. Cells also regulate import itself, which is how many transcription factors are switched on: see that lesson.

The experiments that found them

Entry is not retention. In 1982 Dingwall, Sharnick and Laskey clipped the tails off nucleoplasmin, a pentameric nuclear protein of Xenopus oocytes, by partial proteolysis. Cores injected into oocyte cytoplasm stayed out, but a pentamer with one remaining tail entered. Cores injected into the nucleus stayed there. Entry information was in the tail.

The route is the pore. In 1984 Feldherr, Kallenbach and Schultz injected nucleoplasmin-coated gold particles, about 5 to 20 nm across including the coat, into oocytes. By electron microscopy, they concluded the gold crossed through pores. Gold coated with transport-deficient nucleoplasmin or with polyvinylpyrrolidone was largely excluded.

A signal you can move. In 1984 Kalderon, Roberts, Richardson and Smith transferred the SV40 large-T signal region to β-galactosidase and pyruvate kinase, which then went to the nucleus; changing lysine 128 to threonine abolished this. Their abstract does not name the recipient cells.

Receptors and gradient. In 1995 Görlich, Laskey and colleagues reported that importin α recognizes the NLS and importin β docks at the pore. In 1997 Izaurralde, Mattaj, Görlich and colleagues collapsed the Ran asymmetry in oocyte nuclei with RanGAP1, RanBP1 or a Ran mutant unable to hold GTP stably, and major import and export pathways were blocked. The same year Fornerod, Ohno, Yoshida and Mattaj showed in oocytes that extra CRM1 stimulated export of Rev and U snRNAs, that leptomycin B inhibited it, and that CRM1 binds NES cargo and RanGTP cooperatively in vitro.

Design problem

Protein Z is cytoplasmic at steady state in cultured cells. Within an hour of adding leptomycin B it is nuclear. Interpret the result and design experiments to test your interpretation.