It can be hard to tell, at first, when a cell is on the verge of self-destruction.

It appears to be going about its usual business, transcribing genes and making proteins. The powerhouse organelles called mitochondria are dutifully churning out energy. But then a mitochondrion receives a signal, and its typically placid proteins join forces to form a death machine.

They slice through the cell with breathtaking thoroughness. In a matter of hours, all that the cell had built lies in ruins. A few bubbles of membrane are all that remains.

“It’s really amazing how fast, how organized it is,” said Aurora Nedelcu, an evolutionary biologist at the University of New Brunswick who has studied the process in algae.

Apoptosis, as this process is known, seems as unlikely as it is violent. And yet some cells undergo this devastating but predictable series of steps to kill themselves on purpose. When biologists first observed it, they were shocked to find self-induced death among living, striving organisms. And although it turned out that apoptosis is a vital creative force for many multicellular creatures, to a given cell it is utterly ruinous. How could a behavior that results in a cell’s sudden death evolve, let alone persist?

The tools for apoptosis, molecular biologists have found, are curiously widespread. And as they have sought to understand its molecular process and origins, they’ve found something even more surprising: Apoptosis can be traced back to ancient forms of programmed cell death undertaken by single-celled organisms — even bacteria — that seem to have evolved it as a social behavior.

The findings of one study, published last fall, suggest that the last common ancestor of yeast and humans — the first eukaryote, or cell bearing a nucleus and mitochondria — already had the tools necessary to end itself some 2 billion years ago. And other research, including a key paper published last May, indicates that when that organism was alive, programmed cell death of some kind was already millions of years old.

Some researchers believe that the origins of apoptosis practiced in our cells might be traced to the mitochondrion, which is curiously central to the process. Others, however, suspect that the origins of cell death may lie in a long-ago bargain between our ancestors and bacteria. Whatever the route, the new research surfaces tantalizing evidence that programmed cell death may be older than anyone realized, and more universal. Why is life so haunted by death?

When Death Is the Plan

In the late 1950s, the cell biologist Richard Lockshin grew fascinated by what happens to tissues an organism no longer needs. He was working in the Harvard University lab of insect expert Carroll Williams, who had acquired 20,000 silkworm cocoons from Asia; by the time they arrived at the lab, their metamorphosis had begun. Inside each cocoon, the silkworm’s cells were dying so the creature could become a silk moth. Lockshin went on to document targeted tissue death inside their bodies, which he dubbed “programmed cell death.”

At around the same time, the Australian pathologist John Kerr was turning an electron microscope on the cells of rat embryos to make a similar discovery. As the embryo developed, new cells were being added to the body plan. However, cells were dying, too. It wasn’t an accident, and it wasn’t the result of an injury. This death, which he called “apoptosis,” was “an active, inherently controlled phenomenon,” Kerr wrote. In the rat embryos, death was the plan.

Researchers observing this kind of death eventually arrived at a reasonable explanation for it. During development, a globe of rapidly dividing cells becomes something with wings and antennae, or fingers and toes. Along the way, some of those cells have to get out of the way of the rest. Even in adults, programmed cell death made scientific sense. Unhealthy cells — such as those that accrue DNA damage — must be able to eliminate themselves from a multicellular body, lest they cause additional destruction to the cells around them. Researchers also found that failures of apoptosis could lead to disease, which was also fitting. In cancer, a cell that should have died — a cell whose DNA has so many mistakes that it should have removed itself — does not. In autoimmune and other diseases, cells that should not die do, and vice versa: Cells that should die don’t.

Experts assumed, though, that this skill was unique to multicellular organisms, which had bodies made of many cells for which other cells could die. What possible good could a single-celled organism reap from its own death? Evolution could hardly favor a behavior that removed its carrier from the gene pool.

“It didn’t seem to make sense why anything would actively kill itself,” said Pierre Durand, an evolutionary biologist at the University of the Witwatersrand in South Africa.

But as scientists sketched these death protocols in greater detail, some began to realize that single-celled eukaryotes had similar tools and abilities. In 1997, a team of researchers led by the biochemist Kai-Uwe Fröhlich reported yeast cells methodically dismantling themselves — the first known instance of a “unicellular lower eukaryote” having the basic machinery of programmed cell death. Soon, single-celled algae, protists and other fungi joined the ranks of creatures known for self-induced death.

As biologists tried to understand how organisms could have evolved this ability, they were forced to grapple with another question: If programmed cell death didn’t appear with multicellularity, then where did it come from?

The Tools for the Job

Here is what happens when a eukaryotic cell dooms itself to die.

First, there comes a signal that the end has come. If it’s from outside the cell — if the surrounding cells have marked their neighbor for death — the signal arrives at the cell’s surface and binds a death receptor, which jump-starts apoptosis. If the signal comes from inside the cell — if the reason for death is damage to the genome, for instance — then the process starts with the mitochondria turning against their host cell.

In either case, specialized enzymes soon leap into action. Some apoptotic factors, such as caspases in animals, can activate each other in a cascade of startling swiftness that becomes a swarm and cuts the cell’s structures to ribbons. After that, the cell’s fate is sealed.

“There are many roads to cell death,” said L. Aravind Iyer, an evolutionary biologist at the National Center for Biotechnology Information, who goes by L. Aravind. They all end with apoptotic enzymes, and with fragments of protein and DNA where a cell used to be.

Apoptosis is so tightly controlled, and so widely practiced, that it’s hard not to wonder where its mechanisms originated — both the pieces that make up the machine, which must have come first, and the ways they work together. That curiosity is what drove Szymon Kaczanowski and Urszula Zielenkiewicz of the Polish Academy of Sciences to a recent set of experiments. They wanted to know whether apoptotic proteins from one eukaryote would function if plugged into the apoptotic machine of a distant relative. If the process still worked, they figured, then the enzymes’ functions — the way they slice and dice DNA or activate other parts of the machinery — must have been largely conserved over long periods of time.

The team engineered yeast chimeras that had apoptotic enzymes from across the eukaryotic world: from mustard plants, slime molds, humans and the parasite that causes leishmaniasis. Then, the researchers induced apoptosis. They saw that many of these chimeras were able to execute themselves regardless of the proteins’ origins. What’s more, “the different hallmarks of apoptosis are frequently maintained,” Kaczanowski said, including DNA breakage and condensation of chromatin in the nucleus.

They wondered, too, whether bacterial proteins could stand in for eukaryotic ones. When they subbed in analogue protein genes from a handful of bacteria, the team observed programmed death in some chimeras, but not all. That suggested that the tools for self-induced death predated even the eukaryotes, the researchers concluded.

Not everyone agrees with their interpretation. Some of these proteins, especially those that cut DNA and proteins, are dangerous for the cell, Aravind said; a cell might die simply because of the damage, rather than because of an apoptotic process.

Still, Kaczanowski and Zielenkiewicz believe that what they are seeing is true programmed cell death. And one of their speculations about why bacterial genes might work in eukaryotes connects with an idea that’s been bandied about by biologists for decades.

The theory involves the mitochondrion — an organelle that was once a free-living bacterium. It is the cell’s energy producer. It also crops up again and again in apoptosis pathways. Guido Kroemer, who studies the role of mitochondria in apoptosis, dubbed them “the suicide organelles.”

“Many call it,” Nedelcu said, “the central executioner of cell death.”

An Inside Job?

The mitochondrion is a pretty little thing under the microscope, a neat lozenge containing a labyrinth of membranes. It breaks sugars down to generate ATP, a molecule whose energy powers nearly every cellular process. We don’t know precisely how it wound up within us: The original bacterium might have been the prey of our single-celled ancestor and then escaped digestion by means still mysterious. It might have been a neighbor cell, sharing resources with our ancestor until their fates were so intertwined that their bodies became one.

Whatever its origins, the mitochondrion has its own small genome, left over from its days of independence. But many of its genes have moved to the host’s genome. In 2002, Aravind and Eugene Koonin wrote a landmark paper considering the idea that eukaryotes may have gotten some of their apoptosis genes from the mitochondrion. This little remnant of a bacterium might be the source of some tools eukaryotic cells use to kill themselves.

The genes for apoptosis reminded Kaczanowski and Zielenkiewicz of an arms race between a predator and its prey. In their new paper, they speculated that they might be holdovers from the tools evolved by a prey organism, presumably the original mitochondrial bacterium, to defend itself.

Maybe, once caught inside our ancient ancestor, apoptotic proteins became a way for the mitochondrion to stress the host into changing its behavior, goes a hypothesis collected by Durand and Grant Ramsey, a philosopher of science, in a review they published last June. Or maybe they are the remnants of a way the mitochondrion ensured that the host could not get rid of it — a poison for which only the mitochondria possessed the antidote. Somewhere along the way, the process was captured or transformed by the host, and a variant evolved into apoptosis proper.

The search for answers about the origin of eukaryotic apoptosis seems to be drawing researchers deeper into the bacterial world. In fact, some wonder whether the answers may lie in why single-celled organisms take their own lives. If some form of programmed cell death is older than multicellular life — older even than eukaryotes — then perhaps understanding why it happens in organisms with no bodies to benefit and no mitochondria to speed the process can explain how this all got started.

For the Good of Some Whole

Here’s one reason a single-celled organism might choose to die: to help its neighbors.

In the 2000s, when Durand was a postdoctoral researcher at the University of Arizona, he discovered something intriguing during an experiment with single-celled eukaryotic algae. When he fed algae the remains of their kin who had died by programmed cell death, the living cells flourished. But when he fed them the remains of kin killed violently, the algae’s growth slowed.

Programmed cell death appeared to create usable resources from dead parts. However, this process could only benefit relatives of the dead algae, he found. “It was actually harmful to those of a different species,” Durand said. In 2022, another research group confirmed the finding in another algae.

The results possibly explain how cell death can evolve in single-celled creatures. If an organism is surrounded by kin, then its death can provide nutrition and therefore further its relatives’ survival. That creates an opening for natural selection to select for the tools for self-induced death.

Bacteria, too, are single-celled, and may live among their kin. Can they also die for some greater good? There are hints that under the right conditions, bacteria infected with a virus may kill themselves to arrest the spread of disease. These revelations have reshaped how researchers think about programmed cell death, and Aravind recently discovered another piece of the puzzle.

It involves protein regions called NACHT domains, which appear in some animal apoptosis proteins. NACHT domains also exist in bacteria. In fact, in the wild, the microbes that have the most NACHT domains sometimes partake of what looks very much like multicellular living, Aravind said. They grow in colonies, which makes them especially vulnerable to contagion and especially likely to benefit from each other’s self-sacrifice.

Aravind’s colleague Aaron Whiteley and his lab at University of Colorado and his lab equipped E. coli with NACHT domains and grew them in test tubes. Then they infected the cells with viruses. Strikingly, they found that NACHT-bearing proteins were required to trigger a form of programmed cell death, with infected cells killing themselves so swiftly that the viruses were unable to replicate. Their sacrifice could protect others around them from infection, Aravind said.

These preserved domains tell a story of apoptotic origins, according to Aravind. “You already had a premade cell-death apparatus that was there in certain bacteria,” he said. Then, at some point, some lineages of eukaryotic cells picked up this toolkit, which eventually endowed cells in multicellular organisms with a way to die for the greater good.

He no longer believes the evidence points to the mitochondrion as the only bacterial source of apoptosis proteins. The mitochondrion is the primary bacterial leftover still living within most eukaryotic cells, and 25 years ago it was the logical candidate for these mysterious genes, he said. In the years since, however, something else has become clear: The mitochondrion probably wasn’t alone.

The Bacteria in Us

Eukaryotic genomes, researchers have gradually realized, bear many traces of bacterial genes, remnants of a silent parade of other creatures that left their marks on us. They may have been symbionts, like the mitochondrion, that popped in and out of various eukaryotic lineages, leaving genes behind. “We should now realize that this situation probably continued all through eukaryotic evolution,” Aravind said.

Genes involved in apoptosis may have come from former symbiotic partners that have since departed. Or they may be the results of horizontal gene transfer — a process once thought to be rare and now considered relatively widespread — where genes can hop from one organism to another through processes that are still being worked out. Packages of useful genes can leap between kingdoms of life and persist in new organisms if the benefits are great enough.

One of those benefits, strangely enough, seems to be programmed self-destruction.

All of this is important because it brings into focus the tangled reality underlying the glib phrase “survival of the fittest.” Evolution works in surprising ways, and genes have many purposes. Yet what is becoming clearer is that some sort of primitive collectivity — and, with it, organized self-sacrifices by living things — went on for possibly billions of years before multicellular life arose. Perhaps, as scientists continue to piece together the origins of cell death, we will find a broader notion of what death, and life, are for.