The well-known version of this story is that chlorophyll and haemoglobin look almost alike, which is true and is the least interesting thing about it. The pathway is more specific than that. Every naturally occurring tetrapyrrole — the ring-shaped molecules that hold a metal at their centre — descends from a single precursor, 5-aminolevulinic acid, through conserved steps to the first cyclic intermediate, uroporphyrinogen III. That compound is the last common ancestor of the entire family.
Current Conditions
From uroporphyrinogen III the pathway forks. One route leads to the haems and the chlorophylls; the other, via a compound called precorrin-2, leads to sirohaem, haem d1, coenzyme F430 and cobalamin — vitamin B12. On the first route there is a second fork, and it is the one worth holding onto: both branches are identical all the way to protoporphyrin IX, where iron insertion gives haem and magnesium insertion gives chlorophyll. So the molecular difference between a leaf and your blood is a metal-selection step at the end of a long shared production line — four metals across the family, with cobalt for B12 and nickel for F430, the cofactor of the enzyme that produces methane.
The Stronger Version
Pigment resemblance is where most accounts stop, and stopping there misses the better fact. Chlorophyll and haemoglobin do different jobs — one captures photons, the other carries oxygen — so their similarity is one of building material rather than function. The functional identity is a layer below, in the machinery that actually generates energy.
In your mitochondria sits Complex III, the cytochrome bc₁ complex. In a chloroplast sits the cytochrome b₆f complex. These are not two solutions to a similar problem. They are members of one superfamily, widely distributed across photosynthetic and respiratory electron transport chains in both eukaryotes and prokaryotes. Their protein structures and the charge-transfer events at their cores are highly similar; where they differ is in subunits and cofactors out at the periphery, in proportion to distance from the central redox site. The core is conserved. The trim was redecorated.
Both do the same work: take electrons from a quinone in the membrane, hand them to a water-soluble carrier, and use the energy released to push protons across the membrane. Both operate by the same mechanism, the Q-cycle proposed by Peter Mitchell in 1975. And both build a proton gradient that is subsequently spent by ATP synthase — the same enzyme, in both.
Which is the sentence this piece exists for. A tree and a human both charge a proton gradient across a membrane and discharge it through the same turbine. What differs is the energy source that does the charging: sunlight there, oxidation of food here. Not two systems. One system, two supply lines.
Why the Question Is Slightly Wrong — and What Most People Get Backwards
Here the story acquires its most misunderstood turn, and it is worth being blunt about the misconception, because it is close to universal: animals did not descend from plants. The shared machinery is not evidence that we came out of the plant lineage. Almost everyone who hears the chlorophyll-and-haem story concludes exactly that, and the truth is both stranger and more elegant.
Plants are not our ancestors. They are our cousins. Roughly one and a half to two billion years ago there lived a single-celled organism — neither plant nor animal, and not remotely resembling either — from which both lineages descend. Everything after that is divergence: one branch eventually produced beeches, the other eventually produced you. Reading a tree as an earlier stage of yourself gets the family tree wrong by ninety degrees. It is a sibling, not a parent.
And that ancestor was already a merger. Long before it split into anything, a cell had swallowed a bacterium and failed to digest it. The bacterium stayed, kept its own genome, kept doing what it had always done — burning fuel with oxygen — and became the mitochondrion. Every eukaryote alive carries the descendants of that single event.
Which produces the fact that reframes everything: a tree has mitochondria too, and respires exactly as you do. A beech burns sugar with oxygen, day and night, in Complex III complexes that are homologous to yours by direct shared descent — through that common ancestor, not through anything botanical. Photosynthesis is not the plant’s alternative to respiration; it is an addition on top of it. The tree runs both. It builds sugar in the light and burns it around the clock, which is why a plant in a sealed dark box suffocates just as you would.
The chloroplast came later, and only in one branch. A second capture, a different bacterium — a cyanobacterium, already capable of photosynthesis — swallowed by an ancestor of the plants alone, long after the split. That is the actual asymmetry between you and the beech: not that it is older or more primitive, but that its lineage took in a second lodger and yours did not.
Now the sentence that started this section can be stated precisely. The last common ancestor of your Complex III and a beech tree’s b₆f complex is not an ancestor of you and the beech. It is a bacterium. Two different bacterial lineages were taken in at two different moments in history, and the family resemblance between the two energy complexes is their family resemblance, carried into two hosts that had nothing to do with it. You and the tree are not parent and child, and not even really the parties to the relationship. You are both landlords who took in tenants from the same very old family.
And the second correction is sharper. In photosynthetic prokaryotes the same quinol oxidoreductases can operate in both directions within one membrane: respiration in the dark, photosynthesis in the light. The literature has called such complexes metabolic hybrids. There were never two systems that later converged. There was one system, run one way or the other depending on what was available — and the tidy division into plants that photosynthesise and animals that respire is a later specialisation between organisms, not a partition inside the chemistry.
The last common ancestor of your Complex III and a beech tree’s is not an ancestor of you and the beech. It is a bacterium. You are both landlords to the same tenant.
The Branch That Is Not Ours
Now the part where the family portrait acquires a gap, and it runs against intuition twice.
First, the sequence. Reading the pathways against the branches of life that carry them, the order of appearance runs: F430 and cobalamin first, then sirohaem, then protohaem, and chlorophyll last. The oldest branch of the family is vitamin B12; the youngest is the one everybody assumes is primordial. Highly varied life existed long before nature arrived at protoporphyrin, the building block chlorophyll requires. Earlier authors proposed different orders — haem before chlorophyll in 1965, cobalamin before haem before chlorophyll in 1970 — so the direction of travel has been argued for sixty years and the current reading is a reading, not a measurement.
Second, the possession. Humans build haem and bilins. We do not build cobalamin, and neither does any other animal, nor any plant nor fungus. B12 is synthesised exclusively by microbes. The oldest branch of a family whose other branches we run in every cell is one we have to acquire from outside — which makes vitamin B12 the only vitamin whose absence is, at bottom, a missing biosynthetic pathway rather than a dietary preference.
Absorbed Upstream, Produced Downstream
Which sets up the geometry that makes this more than a curiosity. Roughly a fifth of gut bacteria can synthesise B12, and more than four fifths require it. Your colon is therefore a functioning production site. And human absorption of B12 is restricted to intrinsic-factor-dependent uptake in the terminal ileum — upstream of where the microbial production happens. The receptors sit before the factory.
The textbook conclusion follows: microbial synthesis inside your own gut is nutritionally inaccessible, and the vitamin must come from food. Ruminants solved the same problem by moving the fermentation chamber forward, ahead of absorption. Many non-ruminant herbivores solved it by eating their own faeces. We got neither, and human faeces contain appreciable quantities of B12-like material that is unavailable to the non-coprophagic individual.
Two things complicate that tidy story, and both belong in any honest telling.
Most of what is down there is the wrong molecule. Cobalamin represents less than 2% of total corrinoid content in faeces. The rest are analogues — same family, wrong side chains, inactive in humans. So even if the geometry were favourable, the material would be overwhelmingly unusable. It is not simply the right compound in the wrong place; it is mostly the wrong compound in the wrong place.
And the geometry itself is under revision. A 2023 tracer study did the direct experiment: labelled cyanocobalamin instilled straight into the ascending colon, 5 µg, uptake measured in plasma. Result: 7% ± 5% over four hours, against 63% ± 10% for an oral dose. The oral curve also showed a late absorption peak accounting for 12% of total absorption after an average lag of 8.7 hours — consistent with something happening well past the ileum. The authors derived a minimum daily requirement of 1 µg against a daily excretion rate of 0.7 µg, and proposed colonic absorption as one reason clinical deficiency among vegetarians is rarer than predicted.
Hold the size of that result correctly. Seven percent with wide error bars, in a small study, using a clean synthetic tracer placed directly where you want it — not the mixed, mostly-analogue material a real colon contains. It does not overturn the recommendation to obtain B12 from diet or supplements. What it does is convert an absolute into a quantity: the door is not sealed, it is narrow, and how much traffic passes through it under real conditions is now a measurable question rather than a settled one.
What the Whole Picture Says
The scaffold is old and was never redesigned — only decorated and re-purposed. One ring, four metals, jobs as different as capturing sunlight and producing methane. Evolution could not replace the macrocycle, so it varied the centre and the periphery. The construction rule is the same one that governs the anatomy of brains: extensible, not revisable, because you cannot take a running system offline to start over. Finding it here, in biochemistry, is a strong hint the rule is general rather than a story about nervous systems.
And the gap is not a defect. It is tempting to read the B12 situation as bad engineering — a factory installed past the loading dock. But selection does not price a capacity that costs nothing to lack. In a lineage that reliably ate animal tissue, the cobalt pathway was redundant and its loss was invisible. It became visible only when diets changed on a timescale of decades, against an anatomy that changes on a timescale of millions of years. The deficiency is not a design error; it is a design assumption that stopped holding.
Two Honesties
First, the evolutionary ordering is the softest part of this piece. It is inferred from pathway structure and taxonomic distribution rather than observed, and competing sequences have been proposed since the 1960s. The homology of the complexes and the shared branch points are solid; the chronology is a well-motivated reading. Related and equally soft: the frequently repeated claim that respiration evolved out of photosynthesis is a hypothesis about direction, not a settled result — what is established is the family relationship and the dual function, not which came first.
Second, the B12 story is one where an honest reader should notice who benefits from each version. “Your gut makes it, so supplements are unnecessary” is wrong and has consequences, because the material is overwhelmingly analogue and the absorption is small. “It is anatomically impossible to absorb any of it” is now also too strong, and the study that weakened it was funded to ask an inconvenient question about a population with low intake. Both convenient camps are wrong in the same way, and it is the twin-tribes formation again: they want a categorical answer where the finding is a quantity with error bars.
What to Actually Take From This
The distance between a human and a tree is one of the most confidently misjudged quantities in ordinary thinking — and the one place the shared machinery genuinely fails us is the one place a supplement industry has grown up.
The specialisation is shallower than the labels. Plant and animal are not opposite designs; they are one design with different energy inputs, running homologous complexes into the same turbine. Worth carrying as a general suspicion: categories that feel fundamental often turn out to be settings on a shared mechanism, and the interesting question is what varies rather than what is opposed.
Old scaffolds get re-purposed, never replaced. One ring and four metals produce photosynthesis, oxygen transport, sulphite reduction and methane generation. Nothing was redesigned, because a running system cannot be taken offline. Wherever you find a component doing wildly unrelated jobs across a system, expect an old core that could not be removed rather than a clever unification.
Watch the difference between a closed door and a narrow one. The B12 case is a small lesson in reading claims: microbial production in your colon does not feed you, the absolute version of that statement is now under revision, and both convenient conclusions are wrong. Obtain B12 from diet or supplements. But notice that the textbook certainty became a measured quantity the moment someone ran the direct experiment — which is what usually happens to categorical claims that nobody had tested.
Instrument Check — Worth Your Attention
Study — Kurpad et al., colonic B12 absorption by ¹³C-cyanocobalamin kinetics, Am J Clin Nutr 118(6), 2023. The direct experiment on a textbook certainty: labelled cyanocobalamin instilled into the ascending colon, plasma appearance modelled, 7% ± 5% recovered against 63% ± 10% orally. Read it for the design rather than the headline — small n, wide error, synthetic tracer — and for the derived minimum requirement of 1 µg/day, which is a more useful number than most of what circulates.
Read — “Biosynthesis of the modified tetrapyrroles — the pigments of life,” J Biol Chem (2020). The whole family in one review: the branch points, the metals, the ring contraction that makes B12 the odd one out. The single figure showing every pigment arranged around uroporphyrinogen III is the most efficient thing on this subject — the family resemblance is visible before a word of the text is read. Pair it with the prokaryotic haem-pathway review for the evolutionary ordering and its caveats.
Follow — Down to the Metal, meant literally here. The case for dropping below the packaging on any problem, applied to a question where the bottom layer really is a metal atom in a ring. Together with The Physics of Conspiracies it supplies both halves of what this piece needed: go down far enough to see the mechanism, and refuse the two convenient camps when the finding is a quantity.
Flight Log — Dispatch From Altitude
Every aircraft carries fuel it can never use. Certification is explicit about it: some quantity in the tanks cannot reliably reach the pumps under all attitudes and conditions, and that quantity is determined by test, declared, and subtracted from the usable total. It is real fuel: it has mass, it is carried on every sector, it is paid for. And it is unavailable, not because anything is broken but because of where it sits.
That is the B12 situation exactly. Real vitamin, produced continuously, physically inside the system, separated from the uptake point by geometry rather than by failure. And the comparison carries the correction this piece insisted on: unusable fuel is not a philosophical category but a measured number, established by test. Which is what happened when a group stopped asserting that the colon absorbs nothing and instilled a labelled dose into it: the answer came back small, uncertain and non-zero — a better kind of fact than the certainty it replaced.
Which leaves the thought worth carrying off the flight deck. The temptation, in aviation as in biology, is to read a system by its outputs — thrust here, sunlight in a leaf, oxygen in blood — and conclude that different outputs mean different machines. Go down a level and the difference shrinks to a supply line and a metal atom, while the core turns out to be one thing that was never redesigned. The kinship between you and the tree outside is neither sentimental nor metaphorical — it is a shared component list, held by two organisms that took in the same lodger long ago and never asked it to leave.