Endosymbiosis and the Complex Cell

It is not a part you grew — it is the one that got swallowed and not digested

More than a century ago, microscopists noticed that the green granules inside plant cells split in two the way bacteria do, rather than being assembled from parts by the cell. At the time this was just a curiosity. What turned it into a checkable lead were the traces that accumulated afterwards.

First, the membranes. The energy granules in your cells — mitochondria — are wrapped in two membranes, not one. When one cell engulfs a bacterium whole and fails to digest it, this is exactly what is left over: the inner layer is the bacterium's own skin, the outer one belongs to the hand that grabbed it. Second, that ring of DNA. Third, the protein-building machinery: every cell contains devices called ribosomes that string amino acids into proteins, and the ribosomes inside these granules are not the same model as the rest of the cell's — they are closer to the bacterial ones. Fourth, a granule cannot be built from scratch; a new one only ever comes from an existing one dividing.

The last piece was decisive. Once sequences could be read, that ring of DNA was compared gene by gene against bacteria of every kind, and mitochondria landed squarely inside a group called the alphaproteobacteria, while the green granules of plants — chloroplasts — landed inside the cyanobacteria. They are not bacteria-like. They have addresses on the bacterial family tree.

One sentence needs care here: nobody swallowed anybody "in order to" get energy. Swallowed by chance, undigested by chance, both parties still alive by chance — that combination must have occurred countless times and almost always came to nothing. A very few lineages left more descendants, and what we see today is that one. The three conditions from the natural-selection issue are not bypassed anywhere in this story. One more point: your mitochondria come almost entirely from your mother, since what the sperm carries is mostly kept out or destroyed. That is a way of holding down an internal conflict — mix two lineages of granules and you invite variants that simply copy themselves faster.

A paper rejected a dozen times was overturned by a new measurement, not a better argument

In 1967 a researcher publishing as Lynn Sagan — later known as Lynn Margulis — set out the whole case in a long paper. By later accounts, the manuscript had been turned down by more than a dozen journals before it appeared, some of them without a full reading.

Looking back, the rejections were not all bad temper. The evidence available then was mostly about form: it looks like a bacterium, it divides like a bacterium, its membrane is doubled. And looking like something always admits a second reading — a doubled membrane could be the cell's own membrane folded in around a pocket. The methods of the day could not decide it, and both sides were arguing over the same photographs.

What broke the deadlock was a new kind of measurement. Once DNA could be sequenced, the question became decidable for the first time: put the organelle's genes beside a large sample of bacterial genes and read off which branch it falls on — an answer that does not depend on what anything reminds you of. A dispute that had run for decades ended not because one side phrased it better, but because an observation appeared that could pull the two claims apart.

Something else is worth remembering. That paper made three claims. Mitochondria and chloroplasts went into the textbooks. The third — that the whip-like flagella on cell surfaces also came from a different bacterium — has never found support and is generally regarded as wrong. Same person, same paper, part of it standing and part of it falling: that is the normal case, not the exception.

So the swallowed party is identified. Who did the swallowing? The current answer is the archaea, a large group of single-celled organisms that look bacterial from outside but whose inner machinery is closer to ours. In 2015, genomes assembled from deep-sea sediment revealed a group of archaea carrying genes previously thought to belong only to complex cells. In 2020 a Japanese team actually grew one: more than a decade from sampling to a stable culture, growing so slowly that its numbers take a couple of weeks to double, and surviving only by handing its metabolic waste to partner organisms. The resulting picture is that complex cells are not a third branch standing beside bacteria and archaea, but a branch growing out of the archaea, plus one bacterium. This too is still being tested.

After the merger, the one taken in handed over almost the whole manual

The ring-shaped manual inside your mitochondria holds 37 genes. Its bacterial ancestors had several thousand. Where did the rest go? Some were genuinely lost — living inside somebody else's cell, many skills go unused, and nothing punishes them for breaking. But the more interesting fate is the other one: they moved into the cell nucleus. Of the proteins working inside your mitochondria today, more than a thousand kinds are built from genes in the nucleus and then shipped in.

The move is not an inference; somebody measured it directly. In 2003 a group did it in tobacco: they inserted into the chloroplast's manual a resistance gene that only works once it has landed in the nucleus, then raised offspring in bulk and counted how many showed the gene had made the trip. The answer was roughly one in sixteen thousand pollen grains. That is far more often than anyone expected — the migration is not an ancient event that finished long ago. It is still running.

The genes that stayed mostly govern one thing: watching the flow of electrons across the energy-producing membrane and adjusting to its state moment by moment. Allen's proposal is that such adjustments have to be made on the spot — send a signal to the nucleus and wait for instructions to come back, and the moment has passed; so the genes have to sit as close to the action as possible. Some evidence supports this and some does not; it is not settled.

A bigger dispute is about energy. Lane and Martin proposed in 2010 that a bacterium makes energy across its own outer skin, and skin area cannot keep pace with growing volume, so it cannot get large; a complex cell moves the energy membranes inside and adds hundreds of small control points, each carrying a scrap of manual, so that the energy available per gene rises by orders of magnitude (they put it around two hundred thousand fold), which is what makes a large genome affordable. In 2015 Lynch and Marinov ran a different accounting and objected that the gap does not exist at all, and that what really limits bacteria is population size and random drift, not energy. The argument is not over. For now this one goes down as unresolved and should not be used as a conclusion.