Every Living Cell Runs the Same Translation Machine

Lex Fridman opens with the tool at the center of Betül Kaçar's work: the phylogenetic, or evolutionary, tree of life, and what happens if you run it in reverse. Kaçar, an astrobiologist at the University of Wisconsin who trained as a chemist, calls those trees one of the most romantic and beautiful notions biology has produced. They are a picture of connectedness. Biologists love the graphic. The picture is also easy to get backward.

You do not start at a known root and grow branches. You start at the tips, with whatever is alive enough to sample today. The root is determined by what you placed on those tips. Reconstruction is reverse engineering: look at the leaves, then move upstream in time and guess at a trunk that is long dead. Depending on the question, the unit on a tip can be a gene, a protein, a species, or a group of species.

If you want to know whether a stretch of DNA was swapped between lineages, you build a tree for that event. If you want to resurrect ancient DNA, which is what her lab does, you build more than one. Kaçar's group compares trees from gene sequences, from protein sequences, and from species relationships. Where the three topologies agree, the ancestor they infer is a little less of a guess.

Fridman asks the question that sounds like a joke and is really a boundary condition: how realistic is Jurassic Park? Can ancient organisms be brought back? Kaçar redirects. By the time you reach dinosaurs and large mammals, the exciting part is already over. Eukaryotes and big bodies are late. The fun stuff, she says, is what people think is boring.

Two stretches of geologic time hold her attention. One is the origin of life and the look of the first cells. The other is those long intervals that seem to have produced no innovation loud enough to leave a record. Photosynthesis, eukaryotes, multicellularity, and the oxygenation of the planet all happened in single-celled organisms of a primitive appearance, and the record of them is poor. Those events are why the two of them can have a conversation at all.

Human understanding of past biological invention, she says, is very small. The materials are of two kinds, each limited. Geology offers imprints in rock. Biology offers the trees built from living sequences. You stitch the two together through a fog that thickens the farther back you try to see. Fridman likes the fog metaphor. Kaçar takes it: it is pretty foggy, and it gets foggier.

What makes biology fascinating, Kaçar says, is that it erases its own history. You work with a four-billion-year product, the genome, a dynamic chemical thing that will give you information only if you know where to look, because it has been answering the environment the whole time. Other historical sciences can visit a stratum or a ruin. The biologist's archive is the survivors, and the archive has been editing itself.

Even the bacteria we have today are not primitive. They are successful. Fridman apologizes for the insult. Kaçar, speaking as someone who works with them, feels obliged to defend a lineage that gets little shout-out when people picture life. She quotes Douglas Adams: humans do not like their ancestors and rarely invite them to dinner. Bacteria are in the dinner. They are in the gut. They helped themselves to an invitation.

They are older, more sophisticated, and more resilient than the large animals that usually get the starring role. Simplicity of appearance is not originality. A single cell that has been running, without a shutdown, through billions of years of hostile chemistry is not a low-resolution version of a whale. On the metric of staying alive, the winners have been the microbes all along.

Kaçar's lab studies a variety of bacteria and engineers them. The ideal partner is one you can actually rewrite. The choice follows the question. To connect biology with geology and to probe early innovations across billions of years, cyanobacteria are frequent guests. Their genomes can be engineered. Foreign DNA can be poked into their own, perturbing metabolic function on purpose.

For the most basic jobs of the cell, they use *E. coli*, the well-studied lab pet, a workhorse for cloning and for functions that are already mapped. They inject pretty much every bacterium they work with with foreign DNA. A third class is diazotrophs, nitrogen-fixing bugs such as *Azotobacter*. Into those living fixers the lab loads a nitrogen-fixing metabolism it has resurrected with computational and phylogenetic tools.

Asked how complicated these little organisms are, she answers that it depends how you define complication, and that none of them are primitive. Other groups have engineered *E. coli* to fix nitrogen. Her group does not. It uses nature's own fixer and then swaps in the ancient pathway the trees suggested. Complexity, for her, is not a compliment you save for animals.

Nitrogen in the air is a triple-bonded gas, abundant and almost unusable. Cells cannot drink it. It has to be converted to ammonia before it can enter downstream metabolism. That conversion is nitrogen fixation. Nitrogen is easy to underestimate and hard to do without. It is in ATP. It is in chlorophyll. It is among the elements biology uses in greatest bulk. Plenty of nitrogenous compounds exist on Earth; without this step they sit outside the door.

So far, Kaçar says, evolution has invented only one biological pathway to do the job. Carbon is different: microbes found seven or eight ways to fix it. Nitrogen is a singularity across geologic time, evolved around 2.7 billion years ago, probably less than three. The enzyme at the center is nitrogenase, working with electron donors, ATP, and a great deal of hydrogen to pry the triple bond apart. Life, she says, pushes this metabolism down to create fixed nitrogen. It is quite remarkable.

Fridman asks whether we would still have life as we know it without that step. She cannot think of it. No. It is essential. The two of them are talking because life found a way to fix nitrogen. The cocktail life actually drinks on this planet is five elements: carbon, hydrogen, oxygen, nitrogen, sulfur. Whether that is the only mix that could work, she does not know. It is how it happened here.

Abiotic routes exist. Lightning accumulates ammonia; a thunderstorm is a small fertilizer plant. Humans found another way about a hundred years ago, around World War I: the Haber-Bosch process, nitrogen and hydrogen forced into ammonia under high pressure. Urea for agriculture comes from that flow. The food comes from the fertilizer.

Then the number that should stop a reader. About fifty percent of the nitrogen in our bodies, Kaçar says, comes from that human conversion. A large share of the atoms in a modern person traveled a path opened by industry a century ago, not the enzymatic path evolved roughly 2.7 billion years back. We helped. We found a way to fix nitrogen for ourselves.

She is not impressed by the elegance of the copy. Humans replicated in the simplest way what nature had already done. Life manages the same chemistry with one enzyme, ATP, and hydrogen at ordinary temperatures. Industry uses brute pressure. Without the original invention there would have been no later evolutionary steps to feed. The industrial one is a recent overlay on a much older necessity.

Asked what machinery she is actually injecting DNA into, Kaçar starts with translation. It is a unique subsystem next to metabolism. People treat the cell as the building block of life. From a key perspective, she says, that is not quite right. One may argue that everything inside the cell serves translation. A paper she likes puts it more sharply: the entire cell is hopelessly addicted to this informatic, computing, biological, chemical system at its heart.

The conversion is from RNA to enzymes, a linear mRNA sequence into a protein that will later fold. That is the core processing center for information in life. It initiates, elongates, terminates, and recycles. It operates on discrete bits. It is a chemical decoding device she does not think you will find anywhere else in the cell. Fridman hears a factory that reads a schematic and builds a three-dimensional object. She offers a printer.

Every cell on Earth has this machine. The conserved core is those four steps. At the center sits an RNA that is both information and enzyme; around it, on the order of a hundred protein components that chew ATP or GTP and do their enzymatic jobs. It is, she says, the oldest computational device of life, likely the operating system that had to evolve for life to emerge, and in interesting ways more complicated than the computers we have now.

Kaçar unpacks translation into five overlapping descriptions. It is chemical: the compounds that make it up are chemicals, with rate constants and binding affinities that cannot be wished away. It is physical: it takes energy and moves real molecules through real space. It is informatic: what is processed are bits, four letters in sequence. It is computational: placing the system in discrete states as information is handled is itself computation. It is biological: imperfect replication yields variation and inheritance. The bugs are the features.

A desktop computer has the informatic and computational layers. It does not have the chemical, physical, and biological ones in the same sense, and it does not link information to function the way translation does. Sequence changes, shape changes, function changes, with no middleman. That tight coupling is, for Kaçar, the number one difference. You cannot have the machine without chemistry, and you cannot have cellular life without the machine.

Fridman asks whether chemistry made the machine or the machine merely picked among Earth's reactions to serve a purpose. She will not split them. What is interesting is that even when chemistry is imperfect, even when something binds as it should not, there is resilience. The system may not repair itself, but it moves on. Mistakes can be handled. That handling is the biology. The five components coexist. If the initiator of the computation is too fast, the next message cannot be delivered; if it is too slow, the process stalls. Each part has a margin of error outside which the whole assembly collapses.

Fridman offers a jazz ensemble: chemistry writes the notes, and there is room to improvise. Kaçar likes jazz. It is a party, everybody is invited, and they still have to operate together. The striking experimental fact is that if you remove the translation machinery from the cell and put it in a cell-free system, it works. In a test tube it will translate. It does not need the rest of the cell, so long as you feed it information.

That independence makes the system researchable. It also makes the dream of evolving a translation machine in the lab look less like fantasy, and Kaçar admits they have not done it. Origin-of-life chemists keep looking for a pot that generates the whole thing and have not found one, because it is far more than chemistry. You need biochemistry, networks, computation, information. Once you see the information-system view, she says, you cannot unsee it.

Single-pot chemistry, as she defines it, is the opposite of luck. You add all the recipes you know will lead to the final product. Origin-of-life chemists are chefs of a certain kind: chemicals in a container, a known environment, a rigid input-output map. Her own Ph.D. is in chemistry, though she does not do that kitchen work. Interested in origins anyway? Make chemist friends, she says. It should be mandatory. Chemists can already strip the machinery down, manipulate RNA, and show that a linear polymer can arrive at a processing center. Minimal information processing is not yet cellular translation.

DNA's four nucleotide letters are read in triplets, four times four times four, sixty-four possible states. There is only one codon that says start, and two or three that say stop. Life uses those sixty-four states, minus the start and stops, to encode twenty amino acids. Nature has on the order of five hundred amino acids to choose from. It narrowed the set to twenty. Nobody knows why. Kaçar will not pretend four billion years have explained it. We may wait another four billion. Early life may not even have used this same twenty; we would be fooling ourselves, she says, to claim we know.

Fridman pulls up a standard codon wheel, RNA letters on an inner shell, twenty amino acids on the outer. Kaçar lights up at the methionine. The start is always the same. All life starts with the same start. There is no other start code. Send AUG to the cell and translation knows that function is coming, a chain of information until a stop, marked on the chart in black squares, arrives. One way to begin. Three ways to halt.

Using this language, in theory, you can describe all life that has lived on Earth. In practice, she says, we are not that good at it yet. Genes are the four letters; proteins are the twenty amino acids, sometimes an easier alphabet for building trees. The table in the textbook does not change from bacterium to human. After four billion years of splitting, extinction, and recombination, that is the machine still running.

Using roughly a third of the possible states can look like a waste. Informatically, Kaçar says, the unused room is the point. The code is abundant and redundant. Degeneracy is implemented inside the translation machinery, which means you can make errors and the message still gets through. You can speak while missing some letters. Lysine, K, can be AAA or AAG; mess the third base and you still have lysine. Two-thirds of the unused states buy robustness.

Room for error is not only informatic. She agrees it is probably present in all five layers, computation included, because the informatic capacity is made possible together with the other components, and because the product yields a function, an enzyme. Colors on some codon charts even group chemically similar nucleotides, so a slip still lands on the same amino acid or a near neighbor. We do not know whether those neighbor spellings were locked in because they were once errors, or whether some deeper mechanism selected them.

Fridman asks how natural selection interplays with that resilience: which errors keep the function and which break it. Kaçar says we understand to some degree how an error at this core level hits the rest of the cell. We understand very little about the evolutionary mechanisms that selected the system. It is thought to be one of the hardest problems in biology, still the dark side, even though it is so essential. The language of life found ways to tolerate its own mistakes. How those ways were chosen remains mostly unread.

Kaçar closes the loop back to the tree. When linguists study vanished languages, they start with modern ones. Indo-European comparisons, a Slavic word for snow, trees of descent, an inferred ancestor. If many cultures needed a word for snow, you can imagine that it snowed. Biology is the same move. Understand the function of an ancient enzyme and you understand something about the world its owner lived in. Genes are words. Protein function is meaning. Environment is the weather that made the word worth having.

The codon wheel is life's alphabet, four letters rather than twenty-five. The phylogenetic tree is the text written with that alphabet over four billion years. Each living species is an essay the code produced under environmental pressure. Fridman, coming from computer science, cannot help seeing language models that encode not only text but images and actions in a similar compression. Kaçar keeps the analogy terrestrial: using this language we can, in one perspective, describe all life that has lived on Earth, and we are not yet good at the description.

As an astrobiologist she is after a harder split: which parts of this language were accidents of Earth, and which any life would be forced to invent. Nitrogen fixation, found once and never again, hints that some paths are chemically boxed in. Translation, coupling information to function across five layers of error tolerance, is hard to imagine replacing with something cheaper. She does not then declare the answer. We have one sample. The five-element cocktail is how it happened here. The only way to tell contingency from necessity is a second origin, and until then the honest position is the one she keeps returning to: we do not know.