Life consists of multiple layers of complexity. When viewed in isolation, I suppose each layer is not impossible for human engineers to copy, although it would require an in-depth knowledge of chemistry, physics, engineering and coding to build the equivalent of, say, a nuclear pore complex or a ribosome.
But with each layer, the problems faced by nature are often very different. For example, to evolve a protein, a functional sequence must be found from a natural search of almost endless variations, or by smaller proteins being accidentally stitched together, or by genes being accidentally duplicated, or by domains and themes from existing genes being shuffled about somehow, or by some other method.
Then a potential new protein must acquire the equivalent of an address label, so it can be sent somewhere in or outside the cell. If a new gene is a duplicate of an old one, it will have the same label as the original. But how can it acquire a different address label?
This is a different type of problem compared with evolving a functional sequence, but the two problems need to be solved in conjunction with each other. We could refer to this as “parallel complexity.”
Just to begin the process of turning a gene into a protein, the gene also needs the right control sequences made up of “enhancer” and “promoter” regions, so that RNA polymerase machinery can latch on and transcribe the sequence. This, in itself, is a remarkably complex process. Enhancer regions can often be many thousands or even hundreds of thousands of nucleotides away from the gene, so the DNA strand is folded into a loop, to bring the enhancer and promoter regions together.
Then various transcription factors congregate near the binding site in high numbers, helping each other to latch on. Hundreds of mediator proteins join the party, forming into giant clusters, all helping RNA polymerase to do its job of copying the DNA strand.1 This is a huge co-operative effort just to produce one mRNA transcript that can be converted into a protein. It would be incredibly challenging for human engineers to replicate this process.
So far in this letter I have focused on what we might call “upward complexity,” where sequences in DNA and RNA give rise to proteins, and proteins form multi-protein machinery and structures such as nuclear pore complexes.
However, a different form of complexity that I have hardly touched on is the process of development. Multi-cellular organisms go through various stages of development before arriving at their mature form, which involves many temporary processes along the way.
This is another form of parallel complexity, where nature has to get multiple stages of complexity to work in sequence over time. Since we spent a whole chapter looking at the nuclear pore complex, let’s briefly look at how an NPC develops in the first place.
In early embryos, NPCs appear in some parallel stacked membrane sheets of the endoplasmic reticulum, or ER, which I have previously compared to a mail processing center. Biologists initially thought that the NPCs in the ER were storage compartments for nucleoporins, to be made available to the cell during early embryo development. But there was apparently no clear path for getting the whole NPC structure directly from the ER and into the nucleus. Researchers then looked at fruit flies, and discovered how it was done.2
During early development of the flies, parallel ER sheets containing NPC structures are highly interconnected in three dimensions; and the nuclear envelope, the membrane around the nucleus, contains openings. The openings and sheets line up, allowing the NPC structures from the ER to be inserted into the openings. Initially, the NPC structures are more like scaffolding, but once inserted, they can recruit other nucleoporins so they become fully working NPCs complete with transport controls. Without this mechanism, the development of the embryo would be slower. However, once the organism is more developed, other mechanisms take over the task of building NPCs.
Of course, this is just a minuscule fraction of the complexity that goes into turning one cell into an organism made up of trillions of cells. How does the shuffling of nucleotides come up with such intricate features? In evolutionary theory, new inventions usually come about because of things going wrong, such as the accidental addition, deletion or substitution of letters in the genome, the spluttering of copying mechanisms, cut and paste mechanisms going a bit crazy, parasites getting out of hand, large helpings of serendipity, and enormous amounts of trial and error by the population of organisms as a whole.
But how does the feature I have just described evolve in a cumulative manner? Either it works, or it doesn’t. If nature evolves holes in the nuclear envelope that aren’t filled, it will die, since the cellular transport system will no longer work. Of course, the magic wand of “co-evolution” is the standard answer given by theorists, but this doesn’t actually provide the details. To build such a feature, I would suggest, requires vast engineering knowledge, and can’t be achieved merely by tweaking nucleotides a bit at a time.
We have looked at parallel and upward complexity, but there is also what we might call “downward complexity,” and in many ways this is perhaps the most remarkable of all. For example, in a previous chapter I discussed the electron transport chain, which shuffles around electrons, drawing off some of their energy, which is then used to pump protons into a confined space. These protons are then forced through the tiny equivalent of a turbine, producing the power for our cells.
What makes this all so remarkable, quite apart from the ingenuity of the system, is that protons and electrons are subatomic particles, both much smaller than an atom. How did evolution, which is ultimately the shuffling of nucleotides consisting of dozens of atoms in fairly fixed configurations, combined with huge amounts of trial and error, manage to figure out how to manipulate subatomic particles in such an elegant way? There is a vast scale difference between an atom and an electron. Furthermore, the fairly strict chemistry of nucleotides doesn’t allow much room for experimentation.
Some have speculated that the process arose naturally around deep sea vents in the ocean, and then somehow genes took over later.3 In this idea, perhaps tiny crevices in the rocks acted like cell membranes, and a certain flow of chemicals through the rocks provided energy, and perhaps even formed a simple electron transport chain. In other words, nature supposedly “invented” the electron transport chain through the interaction of chemicals, and then organisms “adopted” it later.
But this is just another example of evolutionary storytelling, which waves an enormous magic wand over the critical details of how an organism encoded the whole process into nucleotide sequences, amino acids, proteins and protein complexes. It is, after all, complex machinery and proteins that do the electron shuffling and protein pumping in an actual cell, not deep sea vents.
The vital details that are missing from the evolutionary story also highlight why the electron transport chain couldn’t have evolved through cumulative selection. The chain requires several critical abilities. It needs to be able to receive donated electrons and then shuffle them through protein complexes. It needs to capture the released energy from the electrons. It needs to use that energy to pump protons into a confined space. Finally, an ATP synthase complex is needed, to take advantage of this accumulation of protons. Take away any of these abilities, the system fails and the organism dies.
These processes involve complex chemistry that is finely tuned. I accept there may be some room for flexibility. For example, it could be that not all four of the initial complexes are essential. Maybe one, two or possibly even three of them could be ditched. But if we wanted to design our own molecular process to shuffle electrons, and utilize their energy to pump protons into a confined space, there are probably only a limited number of ways to do it with a reasonably high degree of efficiency.
This is where chemists and engineers can help. They might find a simpler way, but even this would probably involve a lot of complexity. After all, we’re talking about pumping protons and shuffling electrons here, and these are much smaller than atoms! In other words, in all probability, the process couldn’t be dramatically simplified. At best, a slightly simpler version could perhaps be designed, but this still wouldn’t account for the evolution of the process we see in real cells.
Unlike proteins, whose exact functions can often be difficult to determine, chemistry is very specific. If the electron transport chain really did evolve from a sequence of simpler versions, then unlike most evolutionary stories, those simpler versions could be put together and tested in a lab. But as I have said, even the simplest version would still be complex, because the process needs to pump protons and shuffle electrons, showing that cumulative evolution couldn’t have happened.
In any case, the fatal point to the evolution of this process is that nature would have no way of easily tweaking it. Evolutionary changes take place in the DNA blueprint of an organism, at the nucleotide level. It’s a game of Nucleotide Shuffle. The mutation of a nucleotide can perhaps change an amino acid in a protein, which could change the chemical composition and structure of the protein, but there is no simple correspondence between changing a nucleotide that encodes a protein, and tweaking a chemical or chemical reaction at the atomic level. They are at two very different scales.
This is why saying that nature “invented” the electron transport chain and then an organism “adopts” it is equivalent to a fairy tale. An organism has no mechanism for taking a process happening around it, and turning it into multiple complex, specific sequences of nucleotides that become proteins, that in turn make up complexes that must be put together in sequence to even begin to build its own useful electron transport chain. Mutation followed by selection will not arrive at this, because it requires the invention of multiple specialized proteins that must be pieced together in very specific ways to create the chain.
To put the magnitude of the task into some perspective, nature must first evolve the proteins that make up at least one of the four complexes in the electron transport chain, which means it has to invent the ability to pump protons, shuffle electrons and utilize the released energy.
Perhaps, as is typical in evolutionary stories, nature invented them in some other context, and then managed to recruit them for another purpose, with multiple waves of the “translocation” or “recruitment” serendipity wand. Of course, the inventions would have had to be brought together all at once, because without all the abilities of an electron transport chain – receiving electrons, shuffling them around, and utilizing their energy to pump protons into a confined space – the chain would be useless.
Besides, all these activities would be pointless without something like an ATP synthase complex to benefit from the proton flow, so how do all of these complexes get fitted neatly together? Again, if the functions don’t all come together, the system won’t work. Cumulative improvements can’t happen to a process that already requires a fairly strict set of chemical processes to begin with, and it can’t be tweaked by shuffling nucleotides around.
There is variety in what can deliver electrons to an electron transport chain. For example, bacteria called “lithotrophs” (literally “eaters of rock”) can eat other chemicals to get its electrons. Since the chain is widely used across all domains of life, from an evolutionary perspective it must have been invented very early on.
It is also used in other contexts, such as in photosynthesis, the process of converting light into chemical energy. Clearly then, there is some flexibility in the system. The system seems to be somewhat modular – that is, the process is a discrete unit that can be used elsewhere in other contexts.
When we see complex systems being re-used like this, maybe we should ask whether modularity could actually be a design feature. Perhaps instead of serendipity wands copying and pasting entire systems and using them elsewhere, maybe less luck and more intelligence is really behind it, being able to do so precisely because the systems were designed to be modular in the first place; or at the very least, an intelligent hand would know precisely where to copy and paste.
Either way, I would suggest that evolution, as a game of Nucleotide Shuffle, can’t solve the problem of subatomic particle manipulation by itself. But what if a different game is being played, not at the divine level, but maybe at the cellular level?
1 See the article “What Does It Look Like to ‘Turn On’ a Gene?” by Alla Katsnelson, Casey Rentz and Knowable Magazine, posted at the-scientist.com on May 3, 2019. 2 Hampoelz et al, “Pre-assembled Nuclear Pores Insert into the Nuclear Envelope during Early Development”, Cell, 2016. 3 For example, see Chapter 1 of Life Ascending – The Ten Great Inventions of Evolution by Nick Lane; W. W. Norton & Company, 2009.