How chemistry became biology

Foundations

The honest opening

This is the largest unanswered question in biology, and the answer is that nobody knows.

That sentence is doing real work, because the popular accounts do not say it. They tend to describe a lightning strike, a warm pond, and molecules assembling into the first cell, as if the sequence were established. It is not. What exists is a well-developed chemistry of the ingredients, several serious and incompatible hypotheses about the sequence, and a gap in the middle that no one has closed.

The gap is worth stating precisely, because it is not where most people think.

What is solved: the ingredients

Making the molecules of life turns out to be easy. This is the single most robust result in the field, and it has been demonstrated many times over.

In 1953 Stanley Miller, a graduate student working with Harold Urey, sealed a mixture of methane, ammonia, hydrogen and water into glassware and passed electrical sparks through it for a week. The water turned brown. In it were amino acids, the components of proteins, made from nothing but simple gases and energy.

The experiment was a sensation and it is still in every textbook. Two corrections have since been needed. First, the gas mixture is probably wrong: the early atmosphere was likely dominated by carbon dioxide and nitrogen rather than methane and ammonia, as lesson 4.5 described. Later experiments with revised mixtures still produce organic molecules, though less abundantly. Second, and more importantly, amino acids are not the hard part.

We know they are not, because they turn up without any help at all. Carbonaceous meteorites contain amino acids, nucleobases and sugars, formed in space with no planet involved. Lesson 3.5 followed that chemistry. The early Earth would have been supplied with building blocks whether or not it made its own.

What is not solved: the assembly

Building blocks are not a living thing, in the same way that a pile of bricks is not a house. The unsolved problem is the assembly, and specifically the arrival of a system that copies itself with variation, because that is the point at which natural selection can take over and the rest of this chapter follows.

The classic difficulty is a circle. In every cell alive today, DNA stores the instructions, proteins do the work, and each requires the other: DNA cannot be copied without protein enzymes, and proteins cannot be built without DNA instructions. Neither could have come first.

The best idea: RNA first

The way out is that a third molecule may have done both jobs.

RNA is chemically similar to DNA and can carry information the same way. In the early 1980s it was discovered that some RNA molecules also act as catalysts, speeding up reactions the way proteins do. Molecules that do this are called ribozymes, and they still exist: the ribosome, the machine that builds every protein in your body, has RNA at its catalytic core. That is arguably a fossil of an earlier world.

So the hypothesis, named the RNA world by Walter Gilbert in 1986, is that RNA came first, doing both jobs badly, and that DNA and proteins were later specialisations.

It is the strongest single idea in the field. It also has two serious problems. Making RNA nucleotides under plausible conditions was, for decades, considered close to impossible, because attaching the sugar to the base does not work well in water. And no one has made an RNA molecule that fully copies itself. Laboratory ribozymes can copy substantial stretches of RNA, which is real progress, and none can copy a sequence as long as itself.

The synthesis problem has eased considerably. In 2009 a Cambridge group found a route to activated pyrimidine nucleotides that sidesteps the sugar-plus-base step entirely, building the molecule through a different order of operations from simple feedstocks. Later work from the same group produced precursors of nucleotides, amino acids and lipids from one chemical network under one set of conditions, which is a striking result because it suggests the three classes of molecule need not have separate origins.

The rival: metabolism first

The other major family of hypotheses says the whole framing is backwards. Life is not fundamentally information, it is fundamentally a flow of energy, and self-sustaining chemical cycles came first, with heredity added later.

The most developed version puts the origin at alkaline hydrothermal vents, which are not the black smokers people picture. These are cooler, produce alkaline fluid, and precipitate porous mineral structures full of tiny interconnected compartments with thin walls.

The reason that is interesting is specific rather than poetic. Every living cell on Earth generates energy the same way: by pumping protons across a membrane and letting them flow back through a molecular turbine. It is a strange mechanism, and it is universal, which suggests it is very old. An alkaline vent in an acidic early ocean produces exactly that: a natural proton gradient across a thin barrier, for free, continuously, for tens of thousands of years.

On this account the first life did not invent its energy source. It grew inside one, and only later learned to make its own membranes and take the machinery with it.

The part that is easy

One piece of the puzzle turns out to be surprisingly cooperative. Fatty acids, which are simple molecules, spontaneously form hollow spheres in water. Those vesicles grow when more fatty acids are available, and divide when agitated. They can take up small molecules and retain larger ones.

So compartments are not the bottleneck. Something that looks superficially like a cell will assemble itself in a test tube in minutes. What it cannot do is inherit anything.

The things nobody can explain

Two facts about life have no accepted explanation, and both are worth knowing because they are the kind of clue that eventually breaks a problem open.

Handedness. Amino acids come in two mirror-image forms. Ordinary chemistry makes both in equal amounts. Life uses almost only the left-handed ones, and only right-handed sugars. Something selected one hand and locked it in. Proposals exist, including polarised light from certain stellar environments and amplification by crystallisation, and none is established.

The code. Why these twenty amino acids, and why these particular triplet assignments? There is structure in it, as lesson 5.1 noted, but no accepted account of how it arose.

How long there was

The window is genuinely uncertain, and it may be short.

Earth was assembled 4.567 billion years ago. Liquid water was present by around 4.4. The oldest widely accepted evidence for life is about 3.5 billion years old. That gives up to a billion years.

But two contested claims would compress it dramatically. Carbon inclusions in a zircon 4.1 billion years old have been reported as possibly biological. Structures in Canadian rocks at least 3.77 billion years old have been interpreted as microfossils in ancient vent deposits. If either holds, life appeared very soon after the planet cooled.

That would be significant. A fast origin suggests the transition is not wildly improbable, which bears directly on whether we should expect to find life elsewhere. It would also make it harder to explain, because there was less time for anything unlikely to happen.

What panspermia does and does not do

The idea that life arrived from space is sometimes offered as an answer. It is not one. Moving the origin to Mars or to a comet relocates the problem without solving it, since the same chemistry must still happen somewhere. It is a legitimate hypothesis about where, with no supporting evidence, and no hypothesis at all about how.

Deeper

Why the last common ancestor cannot help directly

Reconstructions of the last universal common ancestor indicate an organism with a genetic code, ribosomes, ATP-based energy currency and chemiosmotic coupling across a membrane. That is a sophisticated cell. Everything between the first replicating chemistry and that organism is invisible to comparative genomics, because the lineages that would have recorded it are extinct. The origin problem is therefore not a problem about the deepest branches of the tree; it is a problem about what happened before the tree.

The Powner synthesis, and why it mattered

Attempts to make ribonucleotides had failed at the glycosidic bond, the link between ribose and the nucleobase, which forms poorly in water. The 2009 route assembles the molecule from cyanamide, cyanoacetylene, glycolaldehyde, glyceraldehyde and inorganic phosphate through 2-aminooxazole, arriving at activated cytidine and, on ultraviolet irradiation, uridine, without ever forming free ribose or free base. Phosphate participates as a catalyst and buffer at multiple stages rather than being added at the end. The significance is methodological as much as chemical: it showed that assuming the biological disconnection points is what made the problem look intractable.

Chemiosmosis as an argument from universality

The proton-motive force is used by bacteria, archaea and eukaryotes, and the enzyme that harnesses it is structurally conserved across all three domains. Since horizontal transfer cannot easily explain a mechanism this deeply integrated with membrane architecture, its universality is usually read as indicating presence in the last common ancestor. The vent hypothesis takes the further step of arguing that a mechanism this awkward is more easily explained if it was inherited from a geological setting than if it was invented by a cell.

What a demonstration would require

For any hypothesis to be considered established, it would need to show, under conditions defensible as early-Earth-like and without investigator intervention between steps, a system that sustains itself, copies its own information, and undergoes heritable variation. No route has achieved this. Partial demonstrations exist for compartment formation, for nucleotide synthesis, for template-directed copying of short sequences, and for autocatalytic networks, but the joins between them are where the work remains.

Frontier

Convergence, slowly

The historical split between metabolism-first and replication-first has softened. Sutherland’s work in particular suggests that a single chemical network can generate precursors for all the major classes of biomolecule, which weakens the assumption that one class must have preceded the others. The current framing is less about which came first and more about which chemical environments can sustain the whole set simultaneously.

The Mars comparison

Mars had liquid water at the same time Earth did, and its surface has not been recycled by plate tectonics, so rocks from that era survive there in a way they do not here. Whether life ever began there is the single observation that would most constrain the question, because a second independent origin in the same solar system would imply the transition is easy, and a confirmed absence would imply the opposite.

Transcript and shot list

Target runtime about 8 minutes. Built as an honest inventory: what is solved, what is not, and the serious rival hypotheses with their unsolved problems. Shot types: ANIM is rendered from code with Manim, AI is a generated shot from Higgsfield with the shared style reference, REAL is agency imagery with credit, CAM is the narrator.

[00:00] S01 · ANIM · A single line of text on black: HOW LIFE BEGAN. Beneath it, after a pause: WE DO NOT KNOW.

This is the biggest unanswered question in biology, and the honest answer is that nobody knows. That sentence matters, because the popular version does not say it. You get a lightning strike, a warm pond, molecules assembling into a cell, as if the sequence were settled. It is not. What exists is very good chemistry of the ingredients, several serious and incompatible hypotheses about the sequence, and a gap in the middle that nobody has closed.

[00:35] S02 · ANIM · The gap drawn explicitly: simple molecules on the left, a replicating evolving cell on the right, a wide blank space between them labelled THE PROBLEM.

And the gap is not where most people think it is.

[00:50] S03 · REAL · The Miller-Urey apparatus, glassware and electrodes (Smithsonian or UC San Diego archive; confirm credit line). S04 · ANIM · Methane, ammonia, hydrogen and water sealed in; sparks; the water browning; amino acids appearing.

In 1953 a graduate student called Stanley Miller sealed methane, ammonia, hydrogen and water into glassware and passed sparks through it for a week. The water turned brown. In it were amino acids: the components of proteins, made from nothing but simple gases and energy. It was a sensation, and it is in every textbook. It also needs two corrections. The gas mixture was probably wrong, because the early atmosphere was more likely carbon dioxide and nitrogen. Later experiments with better mixtures still make organic molecules, just less of them. And more importantly: amino acids are not the hard part.

[01:35] S05 · ANIM · A carbonaceous meteorite cross-section; amino acids, nucleobases and sugars identified within it; a caption: made in space, no planet required.

We know they are not the hard part, because they turn up with no help at all. Carbonaceous meteorites contain amino acids, nucleobases and sugars, formed in space. Lesson 3.5 followed that chemistry. The early Earth got building blocks delivered whether it made its own or not.

[02:05] S06 · ANIM · A pile of bricks beside a house; then the real problem stated: a system that copies itself with variation.

Building blocks are not a living thing, the way a pile of bricks is not a house. The unsolved problem is the assembly, and specifically getting to something that copies itself with variation, because at that point natural selection takes over and the rest of this chapter follows.

[02:30] S07 · ANIM · The circular dependency: DNA needs proteins to be copied; proteins need DNA to be built; an arrow chasing its own tail. Then RNA appears in the middle, doing both jobs adequately.

And the classic difficulty is a circle. In every cell alive today, DNA holds the instructions and proteins do the work, and each needs the other. DNA cannot be copied without protein enzymes. Proteins cannot be built without DNA instructions. Neither could have come first. The way out is a third molecule that does both jobs. RNA carries information like DNA, and in the early 1980s it turned out that some RNA molecules also act as catalysts. They still exist: the ribosome, the machine that builds every protein in your body, has RNA at its catalytic core. That is arguably a fossil of an older world.

[03:15] S08 · ANIM · The RNA world hypothesis stated; then two problems raised: nucleotide synthesis fails at the sugar-base bond in water; and the longest self-copying ribozyme cannot copy a strand as long as itself.

So the hypothesis, named the RNA world in 1986, is that RNA came first, doing both jobs badly, and DNA and proteins were later specialisations. It is the strongest single idea in the field. It has two serious problems. For decades, making RNA nucleotides under plausible conditions looked close to impossible, because attaching the sugar to the base does not work in water. And nobody has made an RNA molecule that fully copies itself. Laboratory ribozymes copy substantial stretches, which is real progress, and none can copy a sequence as long as itself.

[04:00] S09 · ANIM · The 2009 route: simple feedstocks assembling through a different order of operations, never forming free ribose or free base; then a single network yielding nucleotide, amino acid and lipid precursors together.

The synthesis half has eased a lot. In 2009 a Cambridge group found a route that sidesteps the sugar-plus-base step entirely, building the molecule in a different order from simple starting materials. And later work from the same group produced precursors of nucleotides, amino acids and lipids from one chemical network under one set of conditions. That is a striking result, because it suggests the three classes of molecule may not need separate origins at all.

[04:40] S10 · AI into ANIM · An alkaline hydrothermal vent field on a dark seafloor, pale mineral chimneys; then a cross-section showing porous compartments with thin walls, alkaline fluid inside and acidic ocean outside, protons flowing across.

Now the rival, which says the whole framing is backwards. Life is not fundamentally information, it is fundamentally a flow of energy, and self-sustaining chemical cycles came first, with heredity added later. The most developed version puts the origin at alkaline hydrothermal vents. Not the black smokers you are picturing: these are cooler, alkaline, and they precipitate porous mineral structures full of tiny compartments with thin walls.

[05:20] S11 · ANIM · Every cell type shown pumping protons across a membrane and running them back through a rotary turbine; the same mechanism in bacteria, archaea and eukaryotes; then the vent supplying the same gradient for free.

And the reason that is interesting is specific rather than poetic. Every living cell makes energy the same way: pumping protons across a membrane and letting them flow back through a molecular turbine. It is a strange mechanism, and it is universal, which suggests it is very old. An alkaline vent sitting in an acidic early ocean gives you exactly that. A natural proton gradient, across a thin barrier, free, continuously, for tens of thousands of years. On this account the first life did not invent its power supply. It grew inside one, and only later learned to make its own membrane and take the machinery with it.

[06:05] S12 · ANIM · Fatty acids self-assembling into vesicles, growing, dividing. Then two unexplained facts: left- and right-handed amino acids, life using almost only one; and the genetic code table with an open question mark.

One piece is surprisingly cooperative. Fatty acids spontaneously form hollow spheres in water. They grow, they divide when agitated, they hold things in. So compartments are not the bottleneck. Something that superficially resembles a cell will assemble itself in a test tube in minutes. What it cannot do is inherit anything. And two facts have no accepted explanation at all. Handedness: amino acids come in mirror-image forms, ordinary chemistry makes both equally, and life uses almost only the left-handed ones. Something picked a hand and locked it in, and nobody knows what. And the code itself: why these twenty amino acids, why these assignments.

[07:00] S13 · ANIM · A timeline: Earth forms, water by 4.4 Ga, accepted life by 3.5, contested claims at 4.1 and 3.77 compressing the window.

How long was there? Earth was assembled four and a half billion years ago, had water by about four point four, and the oldest widely accepted evidence for life is three and a half. So up to a billion years. But two contested claims would compress that hard: carbon in a zircon four point one billion years old, and structures in Canadian rocks at least three point seven seven billion years old, both argued to be biological, both disputed. If either holds, life appeared very soon after the planet cooled. Which would cut two ways. A fast origin suggests it is not wildly improbable, which matters for whether we expect to find it elsewhere. It also makes it harder to explain, because there was less time for anything unlikely to happen.

[07:50] S14 · AI into ANIM · A Martian outcrop of ancient sediment; the calendar returns.

And the single observation that would settle most of this is not on Earth. Mars had water at the same time, and its surface has not been recycled by plate tectonics, so rocks from that era are still lying there. If life began twice in one solar system, the transition is easy. If it demonstrably did not, that is the other answer. Either way, the honest position today is the one we started with. We know the ingredients form readily. We know a cell can build itself a container. We do not know how the copying started. Next lesson: what that copying, once started, is able to do.

Shot list

Three AI shots, ten code-rendered shots, one archive image. Estimated Higgsfield use: 4 to 8 generations including retries.