Before we can build a nuclear pore complex (or “NPC”), let’s take a closer look at its structure, so we can understand the problem from an engineering perspective. To help you visualize these things, think of the nuclear envelope, the membrane of the cell’s nucleus, as a giant ball, into which we poke hundreds of holes spaced out evenly around it. We then fill the holes with NPCs, which act like channels to filter what gets in and out of the ball.
A typical human NPC is made up of about 1,000 proteins called “nucleoporins,” although there are only about 30 different types of nucleoporin. Some of them act as scaffolding, to keep the membrane of the nucleus curved and stable.
Around the central transport channel of an NPC are 8 identical spokes, made from 16 almost identical half-spokes. This means the channel is in the shape of an octagon. Research has shown that 8 spokes is the most efficient way of helping to transport very large molecules in and out of the nucleus.1
If we were to cut an NPC open from top to bottom, we would see a central ring sitting between two outer rings, one facing down toward the nucleus, and one facing up toward the main compartment of the cell. The two outer rings act as a structural scaffold.
If we were to slice the NPC through the middle like a burger bap, the two halves of the bap would be mirror images of each other, which is why the inner core is called the “symmetric core.”
On the side of the NPC facing the nucleus, nucleoporins attach to the symmetric core and form what looks a little like a basket, so biologists call it the “nuclear basket.” It interfaces with the transcription machinery in the cell nucleus. On the side facing out, a different set of nucleoporins form filaments which assist transport through the NPC. These are referred to as “cytoplasmic filaments.”
The central channel has a barrier made up of several nucleoporins that contain stretches of repeated amino acid sequences, called “FG repeats.” This creates a kind of mesh that helps control what enters and leaves the nucleus.
Proteins called “karyopherins” recognize molecules that need to get in. When a molecule has an address label called a “nuclear localization signal,” it is helped through the NPC by a set of karyopherins that attach to the molecule and bind to the mesh. Energy for transport comes from what is known as the “Ran gradient.”
mRNA strands are produced in the nucleus, but the ribosomes needed to translate them into proteins are located in the cytoplasm, which is outside the nucleus. This means an mRNA strand needs to pass through an NPC. To indicate where it needs to go, the strand has an address label called a “nuclear export signal.”
While being transcribed in the nucleus, other proteins bind to the mRNA strand, to form what is called a “ribonucleoprotein complex” (or “mRNP”), that prepares the sequence for export. This is tagged with a “nuclear export factor,” a small protein that interacts with the mesh inside an NPC, helping the sequence through. It also serves as a one-way ticket, because it is removed once the mRNP has made it to the other side, preventing it from going back into the NPC.
An interesting question we could ask here is: why are mRNA strands coding for proteins that are destined for the nucleus first of all exported, only for the proteins to be imported once they have been made? Wouldn’t it be simpler just to make these proteins within the nucleus? Ribosomes are located outside the nucleus, so proteins couldn’t be made without this step. But why aren’t they also located in the nucleus, so at least some proteins could be made there?
I suspect this is for safety and quality control reasons. From a safety perspective, if proteins were made in the nucleus, there is a risk that faulty ones could damage or clog up the equipment that is critical for the healthy functioning of the cell.
Therefore, it makes sense that processes within the nucleus focus on storing and maintaining the information in chromosomes, and in transcribing some of this information in the form of mRNA strands, while more potentially dangerous functions, such as producing energy and building protein machinery, are done outside of the nucleus.
The export and import process can also provide quality control checks, to help ensure that mRNP sequences are in good shape before they are exported from the nucleus, and to make sure that only the right proteins are allowed in. Evidence indicates that the nuclear basket, which an mRNP needs to pass through to get into an NPC, can act as a hub for mRNA export quality control.2
In a living cell, nuclear pore complexes interact with other parts of the cell. However, for the moment, let’s ignore this and think about what would be needed just to build one stand-alone NPC, one “hole” in the nuclear envelope.
First of all, we would need the raw materials, namely, the correct 1,000 or so nucleoporins. To help visualize one of these proteins, first picture a long balloon used by entertainers at children’s parties, the kind that can be shaped into a man or a dog. Each balloon would be the equivalent of one amino acid, out of which proteins are built.
Now imagine several hundred of these balloons taped together to form one huge strand. The entire strand would visually represent one protein, one nucleoporin, and this would still be just one tiny piece of a nuclear pore complex. We would need about 1,000 of these strands of balloons to build an NPC. As a quick side note, I don’t recommend trying to build a nuclear pore complex at children’s parties.
In the microbiological world, proteins aren’t as well-behaved as strings of long balloons. Different segments have properties that make them behave in certain ways. For example, some segments hate water. They are called “hydrophobic.” Put them in water, and they will try to stay away from it, perhaps by hiding within other segments that love water, known as “hydrophilic.” These and other properties, such as electrical charge and polarity, are part of the reason why some proteins can fold into shape almost by themselves, and why certain shapes are rigid enough to be used by cells as components for molecular machines.
In other words, building an NPC is far more complicated than simply twisting lengthy strings of long balloons around each other. We would also need to simulate the cellular environment in which an NPC is built, and the reactions to that environment by each amino acid in the protein sequence. I guess it could be done by smart human engineers, but it would be quite a feat, and would probably need to be done on a computer, in a virtual environment that simulated the conditions in a cell.
In addition, each piece would have to be fitted together in the right order, like a three-dimensional jigsaw puzzle in which certain pieces were repulsed by their environment, and some were attracted to it. In other words, for someone to build the equivalent of a stand-alone nuclear pore complex, it would be a formidable task requiring a lot of chemical and engineering knowledge.
Now, what is the story of how the nuclear envelope and NPC came about, according to evolutionary theorists? Nobody can be sure, because they weren’t around to observe them evolving, but the story goes something like this:
Prokaryotic cells already had the ability to build a membrane, so perhaps another membrane accidentally mutated inside of a cell around the DNA molecule, forming what would eventually become the nucleus; or perhaps the development of an inner membrane was more gradual, evolving as an extension of the cell’s own membrane. To begin with, it wasn’t sufficient to shield the DNA molecule, but it somehow gave the cell a small survival advantage as it evolved.
Either way, a core scaffold co-evolved to support this new membrane, made up of proteins used elsewhere in the cell for membrane binding and tethering. Natural selection would then test out many different scaffold structures, and the one that worked best incorporated simple pores, which would gradually evolve into pores with basic gating mechanisms.
Since there is a symmetry to the NPC as we know it, multiple duplication events took place, where the structure of certain parts of the scaffold were duplicated, resulting in the symmetry we see today. Gradually the structure was improved on through natural selection, to include a more selective transportation system, until we arrive at our current NPC structure.
Stories like this perhaps sound convincing at first. Proteins are recruited from elsewhere. Functions get drafted in from one part of a cell to another. Sequences of code get duplicated. Throw in natural selection and vast amounts of time, and it almost sounds like the evolution of an NPC was inevitable.
But let’s look at the details more carefully. First of all, the evolutionary story requires a complete or partial membrane to appear around the DNA molecule in one particular ancestral cell. For this to happen consistently in its offspring, some kind of control code is needed that regulates the construction of this new membrane from generation to generation, and calls up specific proteins in a particular order. But how can the control code organize the construction of this, and direct proteins to the right places, without new address labels for each protein?
Second, scaffolding is needed, to support the curvature of the membrane. Perhaps scaffold proteins were brought along by whatever events caused the new membrane to evolve, but how would these proteins get to their new places of work without new address labels? Indeed, how does the cell define the location of a place that didn’t need to be defined before?
Third, the new membrane must be constructed carefully enough to allow transport in and out, so that the cell can not only continue to survive, but actually have a survival or reproduction advantage. If the new membrane lets nothing through, the cell dies, because the DNA molecule contains the blueprint for the cell’s machinery. The membrane would effectively strangle the cell. On the other hand, if the new membrane let’s everything through, evolution would likely discard it, unless it has some kind of useful function.
What about the nucleoporins that make up the NPC? Where could they have come from? Proteins are usually made up of one or more domains that fold into specific shapes. Research has shown that nucleoporins share a handful of similar domains, so biologists assume they evolved by the duplication and shuffling around of these domains from a smaller set of domains.3 Why the original domains would exist on their own at first isn’t clear, and how whole sequences of code get shuffled around is also not clear; presumably a lot of the right type of copying and pasting would be needed.
Due to the symmetry of the NPC, multiple duplication events are assumed to have taken place during its supposed evolution. For example, nature could start with one half-spoke, and duplicate it seven more times to get the eight half-spokes. But that begs the question: what good is half a spoke on its own, before nature accidentally copied it another seven times? Evolutionary theory suggests it must have had a function, but these functions are rarely even explained, let alone scientifically tested.
To get eight whole spokes, nature could also duplicate all of the eight half-spokes at the same time. But this wouldn’t arrange them into an octagon shape. If it did, this would imply nature knew that it was aiming for an eight-sided shape from the beginning, because the angles for each piece would need to be planned for in advance.
The point here is, duplication doesn’t really explain the specific design of the NPC. Even after these alleged duplication events, nature still has to assemble the correct nucleoporins in the right order and shape.
Of course, according to evolutionary theory, nature doesn’t have to figure out how to assemble 1,000 complex parts all in one go. It doesn’t even need to start with 1,000 parts. It only has to make occasional tweaks, bring in or remove parts once in a while, and then test out slightly different designs through natural selection. We could think of evolution as a game of constant tinkering, played out over an almost endless series of rounds. The best tweaked design in one round goes on to compete in the next round.
But how exactly does nature get to play around with the design of an evolving NPC? The shuffling of nucleotides in a nucleoporin gene isn’t really the answer, because nucleoporins share similar domains, with re-used fold structures. Furthermore, there are only about 30 choices of nucleoporin, but they are used multiple times in the NPC structure. Therefore, if the gene for one nucleoporin mutates, this may have a dramatic impact on the overall structure of the NPC.
The repeated use of only about 30 or so nucleoporins could actually be a design feature, meant to restrain changes caused by mutations. Too many mutations in any one nucleoporin would distort the shape of the NPC, perhaps even breaking it and thus killing the cell.
In theory, the way nature could test out alternative designs is by shuffling and tweaking parts of the blueprint that control how the NPC is put together consistently. But there doesn’t seem to be a mechanism in the cell that allows pieces of the NPC to be shuffled around experimentally, like switching pieces in a jigsaw puzzle. There is only the serendipity wand of translocation; but this is often just a storytelling device rather than an actual mechanism. It’s true that pieces of the genome can be translocated at times, but this usually happens in a strictly regulated manner by mechanisms following specific markers, pathways and blueprints.
Even if we allow for perhaps a virus or some other mechanism to do the job, this would still involve incredible serendipity. To test just one slightly different NPC design, an entire nucleoporin sequence would have to be cut or copied, and pasted into just the right place in the blueprint so the tweaked design could be made in the first place.
In other words, nature doesn’t have an easy way of altering the design. And a similar problem exists for all of the complex structures found in the cell. Most proteins work alongside other proteins, and some of them form larger machines, called protein “complexes,” which are often made from several, dozens or even hundreds of proteins. The nuclear pore complex is just one example.
What brings these proteins together in the right order? Clearly some kind of master blueprint needs to be in control of the process, so protein complexes can be built consistently and reliably, and so their designs can be passed down through the generations.
But how can these blueprints be improved on over time? According to classical evolution theory, they are tweaked through the game of Nucleotide Shuffle, the mutation of letters in DNA sequences.
From an evolutionary point of view, a cell is merely a vessel for the genome stored in a DNA molecule. A slightly mutated genome might be passed on to the offspring of an organism, and the offspring’s cells are then built from it. If a mutation happens to give the offspring a survival or reproduction advantage, the mutated genome is more likely to be preserved than competing genomes. Or perhaps the organism just gets lucky.
Either way, nature doesn’t have an easy way to piece together new protein complexes or try out alternative design plans. Evolutionary theorists talk about proteins and functions being “recruited” from elsewhere, but there is no cellular recruitment agency hiring proteins such as Alice and Bob who want to change careers. In other words, “recruitment” is just another serendipity wand, or another form of the “translocation” wand.
Besides, in the case of the NPC, there is yet another layer of remarkable complexity that has an enormous impact on its supposed evolution. Not only does it need to be assembled correctly, it also needs to be disassembled quickly when the cell goes through the process of division. This is critical, so the genome can be duplicated and passed on to daughter cells. In other words, while nature is supposedly evolving the NPC, it also has to figure out how to tweak both its assembly and disassembly at the same time.
The inner ring complex of the NPC is almost identical across all eukaryotic life forms, so it is assumed to be the core that holds the rest of it together. However, there is some flexibility in the size and shape of the outer ring complex, across different forms of life. Some also have wider or smaller pores, depending on the needs of the organism.
The way the NPC is put together, with its scaffold of similar building blocks, allows for this flexibility in the first place. It’s almost as if nature had incredible foresight, by putting together a structure that could be adapted to suit the needs of a variety of creatures. But evolution has no foresight; so once again, in the evolutionary paradigm, this was all just incredible luck.
In the last few chapters, I have presented arguments implying that the features of eukaryotic cells pose a difficult challenge for evolution. However, in the next chapter I will present an intriguing line of evidence that suggests these features were deliberately designed after all.
1 Wolf, Mofrad, “On the Octagonal Structure of the Nuclear Pore Complex: Insights from Coarse-Grained Models”, Biophysical Journal, 2008. 2 For more information, refer to the article “The Great Escape: mRNA Export through the Nuclear Pore Complex” by Paola De Magistris, International Journal of Molecular Sciences, 2021. See the “Remodeling at the Nuclear Basket” section in particular. 3 Devos et al, “Simple fold composition and modular architecture of the nuclear pore complex”, PNAS, 2006. See “Evolution of the NPC” section in particular.