Part Of: Biology sequence
Followup To: A Portrait of LUCA
Content Summary: 4000 words, 20 min read

Cutting the Knot

Abiotic chemistry is leaky. Sugars brown into tar. Nucleic acid hydrolyzes in warm water on a timescale of days to weeks. Practically anything soluble diffuses away. Whatever complexity chemistry builds is temporary. To achieve LUCA, chemistry needed a ratchet.

The natural place to look for that ratchet is LUCA’s own subsystems, but they depend on one another. Genes prescribe proteins, yet cannot be expressed without proteins to transcribe and repair them. Metabolism supplies the materials for both, yet runs on protein catalysts that genes encode. Asking which came first is really asking which subsystem supplied the first ratchet. 

This Gordian knot is a question of functional priority. There are two ways to cut it.

One answer sees the first ratchet in nucleic acid. A system that can copy information can preserve innovation. This is the genetics-first tradition.

The other sees it in metabolism. A network that captures free energy can maintain its own organization and generate the building blocks heredity would later need. This is the metabolism-first tradition. 

The RNA World

Let’s set metabolism aside for a moment. Among the three biopolymers of the central dogma – DNA, RNA, and modern coded proteins – which came first? On this smaller question, the scientific community has achieved consensus. Metabolism-first and genetic-first camps agree: RNA came first. 

RNA is the answer because it can do both jobs. RNA stores information in its nucleotide sequence, like DNA. But it also folds itself into structures that catalyze chemical reactions, like proteins. The discovery of self-splicing introns revealed that RNA can act as a catalyst (Kruger et al. 1982). Most modern biological catalysts are protein enzymes. RNA catalysts are called ribozymes.  The universal coenzymes hint at the same history. NAD, FAD, coenzyme A, and SAM are all built on nucleotide scaffolds, which White (1976) read as molecular fossils of an RNA-based metabolism.

The ribosome is the engine of the genetic code. Remarkably, we are able to reconstruct its development at the atomic resolution. The oldest part of the ribosome, its center known as the peptidyl transferase center (PTC), contains no proteins at all. Its catalytic heart is entirely RNA. The ribosome is a ribozyme (Nissen et al. 2000). 

Other genetic polymers could, in principle, have preceded RNA. One genetic system could also have replaced another while preserving the basic function of heredity, a process sometimes called genetic takeover (Hud et al. 2013). But for now, in the name of parsimony, we will assume the first genetic ratchet was RNA. 

Why is protein-first no longer an active research tradition? Amino acids and short peptides may well predate RNA. But RNA apparently precedes modern coded proteins. The RNA World constitutes major progress in cutting the knot. The RNA world is silent on the order of metabolism. This is the residue. 

The genetics-first solution advocates a strong version of the RNA World: genetic heredity precedes the emergence of metabolism. Natural selection then improves the chemical system around the replicator. No one questions the capacity of nucleic acids to build and retain complexity. The burden is to show how geochemistry alone can birth such a wondrous digital inheritance. 

The metabolism-first tradition holds that metabolism existed before RNA, and therefore before genetically encoded proteins. A non-enzymatic protometabolism must exist first. This school fully accepts digital inheritance. But it proposes that protometabolism served as the first ratchet. Geochemistry is thus less burdened. The burden here is twofold: to show that non-enzymatic protometabolism is viable, and to show that a mere reaction network can preserve innovation without genes. We take up the second question in a later post.

Let us explore the viability of specific genetics-first and metabolism-first theories. These are grounded in specific habitats. But before we search for the Geography of Genesis, we must first understand what success looks like. Designing such a rubric takes us back to first principles. We begin with thermodynamics.

Life as a Dissipative Structure

Free energy is the part of energy that can do work. The second law of thermodynamics says that energy tends to spread out. Hot objects cool, gases mix, and chemical gradients become weak. 

A closed system does not exchange matter with its surroundings. Without a continuing source of usable energy, its chemical differences decay toward equilibrium. As a system approaches equilibrium, free energy is dissipated, often as heat, and less remains available to do useful work. 

Life manifests tremendous complexity and order. In What Is Life?, Erwin Schrödinger explained that life is not a closed system. An open system can exchange energy and matter, so an external process can maintain nonequilibrium conditions. Life exists in an open system.  It consumes external energy and matter, then releases heat and waste. Life keeps local order while disorder increases in its surroundings. 

The chemist Ilya Prigogine developed the idea of dissipative structures. These structures import free energy from their environment. They maintain local order by using this energy and releasing heat and waste into their surroundings. When the flow stops, the structure usually weakens or disappears. 

Dissipative structures occur at many scales. Examples include convection cells, flames, chemical oscillations, hurricanes, and living organisms. For the origin environments we will consider, two energy sources matter most: sunlight and geological activity. 

Our first requirement is sustained free-energy flux. The Garden of Eden must feature continuous (or at least recurrent) throughput sufficient to maintain far-from-equilibrium chemistry.

Chemical Selectivity

Free energy flux can drive useful reactions. It can also drive thousands of useless ones. 

In organic chemistry, a small set of reactive molecules can combine in many different ways. Each new product can then react with the original molecules or with other products. The number of possible pathways can grow rapidly. Researchers call this a combinatorial explosion. 

Scientific experiments (e.g., Miller Urey) and natural phenomena (e.g., carbonaceous chondrites) both tend to produce tar: complex mixtures that are difficult to separate or even characterize. 

The formose reaction gives another example. Recall that ribonucleic acid (RNA) uses one particular sugar: ribose. Under alkaline conditions, formaldehyde produces dozens of different sugars and related compounds, while ribose appears as only one component of the diverse products. As the reaction continues, useful products can disappear into still more complex chemistry. 

Life is different. Biology uses an extraordinarily restricted region of chemical space. About 500 naturally occurring amino acids have been identified, but the standard genetic code uses only 20 of these to build proteins. The same organizing principle appears in metabolism. Smith and Morowitz (2004) identify eleven TCA-cycle carboxylic acids as a universal anabolic core, from which the major pathways of biosynthesis branch. In their formulation, lipids, sugars, amino acids, nucleotides, and other major classes of biomolecules are ultimately built outward from this small conserved core. 

A biochemical environment must channel significant flux toward its target products, while limiting competing reactions. This is a second requirement for Eden: chemical selectivity. 

For genetics-first, this requirement must be met before selection can start. A nucleic acid is a monument of chemical selectivity: one sugar among dozens, a few bases among many, the correct linkage among several. The environment must achieve that discrimination unaided. Genetics-first chemistry therefore requires prebiotic mechanisms that narrow the product distributions. 

Metabolism-first theories propose another possibility. Positive feedback reactions (aka autocatalysis) can deliver selectivity without a special environment. But it only works if the network reinforces a restricted set of productive reactions, rather than amplifying chemical clutter. We will explore this topic in more detail in a followup post.

Deriving the Rubric

We can say more. Chemical principles allow us to derive five additional requirements. 

1. Provenance. The required feedstocks and catalysts must arise or arrive naturally.
2. Reactant Concentration. Feedstocks must reach concentrations high enough for useful reactions.
3. Material Flux. Fresh material and free energy must continue to enter, while waste leaves and spent material is removed, reactivated, or recycled. 
4. Product Retention. Useful products must also remain close by long enough to interact again, rather than immediately dispersing into the environment. 
5. Net Production. Eden must not only produce these target products. It must produce them faster than it destroys them. We require net production.

Material flux and product retention are in creative tension. Eden must eject waste while selectively retaining useful products. 

The above concerns single reactions. But theories of abiogenesis rely on many processes.

Suppose one reaction requires acidic conditions, another alkaline conditions. Or one may require intense ultraviolet radiation, while another important product is rapidly destroyed by the same radiation. A theory may need to confine different reactions to different microenvironments. Such conflicts do not make a pathway impossible, but they do require that these microenvironments exchange products in a realistic way. This is the requirement for condition compatibility.

It is not enough to demonstrate each reaction separately in a laboratory. Products of one step must naturally reach the next without artificial purification, isolation, or reagent replacement.  This is the requirement for process continuity.

A favorable environment is of little use if it disappears before cumulative organization can develop. The environment must persist or recur often enough to provide continuity across time. Our last requirement is habitat persistence. 

These requirements are universal. But genetics-first and metabolism-first theories also generate idiosyncratic requirements. Let’s examine each in turn.

Building Nucleic Acid

Genetics-first demands a specific product. Geochemistry must synthesize a nucleotide. Every nucleotide brings together a sugar (ribose), a nitrogenous base, and a phosphate group into a single large molecule:

Carbon, hydrogen, and oxygen are readily available in many geological habitats. Nucleotides also require nitrogen and phosphorus, and these present new challenges for genetics-first models.

Most environmental nitrogen is molecular nitrogen, N₂. The strong triple bond in N₂ makes it unusually unreactive. Before nucleotide synthesis can begin, some process must supply fixed nitrogen in a reactive form. Nitrogen fixation is one way to make it more reactive (Mateo-Marti et al. 2019). Similarly, phosphorus is present in rocks, but much of it occurs in minerals with low solubility. In water containing calcium, dissolved phosphate is quickly locked away into minerals. Many prebiotic reactions, meanwhile, require phosphate concentrations far above those found in ordinary natural waters. This is the phosphate problem (Toner and Catling 2020). Together, these problems generate two requirements. A genetics-first theory requires bioavailable nitrogen and phosphate. 

Of course, individual nucleotides aren’t enough. They must come together as long strands of nucleic acid. 

RNA polymerization occurs through a condensation reaction. But condensation polymerization is disfavored in water. The equilibrium instead favors the reverse reaction, hydrolysis. Joining nucleotides into a chain therefore requires an input of free energy (Yadav et al. 2020). Monomer activation denotes the requirement to store or couple this free energy so that phosphodiester-bond formation can proceed.

Providing free energy is not enough. Polymerization also requires high efficiency. A 45-nucleotide polymer requires 44 bond-forming steps. At each step, the chain can degrade. Even modest inefficiency therefore compounds across the chain. Polymer elongation denotes the requirement for bond formation to be efficient enough to produce functional-length polymers.

Towards a Replicator

We’ve established a long RNA polymer. Next, it must copy itself. 

Natural selection cannot act on sequence information until polymers are long enough for different sequences to produce different functional structures. Very short oligomers have limited capacity to fold into complex three-dimensional forms.

We do not know a strict minimum length for a useful hereditary polymer. But Gianni et al. (2026) provide an important empirical benchmark. They discovered QT45, a polymerase ribozyme only 45 nucleotides long. QT45 can catalyze synthesis of both its own sequence and its complementary strand.

Recall that nucleotides attach to their complements via Watson-Crick pair bonds. These bound duplexes are very stable. This stability assists during the replication, but presents a problem afterwards. If the strands remain bound, neither is available as a template for the next generation. This is the strand inhibition problem. For replication to persist, the resulting duplex must separate without damaging the sequence. Template reset denotes this requirement. 

A catalyst that helps every molecule around it, freeloaders included, gains nothing from its own innovation. Some mechanism must couple products to their producers: what a system makes must preferentially benefit that system. Compartments are the obvious mechanism, but mineral pores, surfaces, and reaction-diffusion gradients are candidates too. As we will see in a later post, this requirement bites both camps, and selection will demand a sharper version of it.

So ends our discussion of genetics-first. Now we turn to metabolism-first theories.

The Twin Engines of Protometabolism

Genetics-first requires a specific deliverable: self-replicating RNA. Where is the analogue for metabolism-first? Which seeds accreted the fantastic complexity of modern metabolism?

Metabolic networks are not tangled webs. Their architecture is hierarchically modular: tightly linked modules nest within larger modules, all organized around a small connective core (Ravasz et al. 2002). Architectures like this arise when networks expand by attachment, each new pathway bolting onto machinery that already exists. And once a dozen downstream pathways depend on an upstream reaction, that reaction is frozen in place; changing it breaks everything built on top. Wimsatt (2007) calls this generative entrenchment. Entrenchment gives the network a readable stratigraphy. Deeply embedded core chemistry should tend to be older than dependent peripheral pathways. We will meet the same logic again when we explore the ribosome’s history from its onion-like layers. 

All living things rely on energy metabolism to fuel themselves and anabolism to build up complex organics. The differences reside where the carbon comes from. Autotrophs use carbon fixation to build biomass from inorganic carbon such as CO2. Heterotrophs obtain reduced carbon from the environment, and rely on catabolism to break down the complex organics in their diet. 

The modern biosphere rests on an autotrophic foundation. Nearly all carbon enters through the fixation of CO2. Life does not depend significantly on extraterrestrial, atmospherically or geologically produced organics. But on the Hadean Earth, impact delivery supplied organics at a far higher rate. Oparin and Haldane’s original vision of primordial soup presumed a heterotrophic origin. So which was it? Did the first life eat chondrites or fix carbon? 

Two converging lines of evidence suggest fixation. The first comes from genes and energetics. Weiss et al. (2016) traced 355 gene families back to LUCA and recovered the profile of an anaerobic, H2-dependent autotroph. Phylogenetic reconstructions are often contested. But Wimmer et al. (2021) computed the thermodynamics of LUCA’s ~400 core reactions and found that roughly 96% run exergonically under H2-rich anoxic conditions. The genes describe an autotroph, and the metabolism they encode fits an autotroph’s environment.

The second line of evidence applies our stratigraphy. Catabolism sits at the periphery. Major catabolic pathways characteristically converge on the universal anabolic core. Moreover, the substrates it digests are sugars, lipids, and proteins: the characteristic products of anabolism (Schönheit, Buckel & Martin 2016). The enzyme record concurs. The ancestral fructose-bisphosphate enzyme of archaea is locked into the gluconeogenic direction, marking glycolysis as a later retrofit (Say & Fuchs 2010). 

Taken together, these suggest that anabolism came first. The first life was likely autotrophic.

Carbon fixation places specific requirements on the environment. It requires both CO2 and a continuous electron supply. Candidate electron donors (reductants) include H2, Fe(II), and H2S. Reductant and carbon co-delivery denotes the requirement that electron donors and inorganic carbon arrive together, continuously, at rates sufficient to sustain fixation.

Anabolic metabolism synthesizes thousands of biological molecules. The stratigraphy points inward: the oldest chemistry of all should sit at the very center of the network. Smith and Morowitz (2004) identify it. Eleven carboxylic acids form a universal anabolic core.

With a handful of glycine-related exceptions, all biosynthesis radiates from five of these acids. Acetate seeds the lipids; pyruvate the sugars and the alanine-family amino acids; oxaloacetate the aspartate family, and through aspartate the pyrimidines; α-ketoglutarate the glutamate family; succinyl-CoA the pyrroles. (Purines are the untidy exception, assembled from glycine, aspartate, glutamine, and the same C1 chemistry owned by the WL pathway) These are the pillars of anabolic metabolism. 

The entire biosphere uses seven carbon fixation pathways (CBB, rTCA, WL, 3-HP, 3-HP/4-HB, DC/4-HB, and roTCA). Which of these was used by LUCA? 

You may recall learning about the tricarboxylic acid (TCA) cycle (aka Krebs cycle) at school. In modern organisms, this cycle is catabolic. For decades that was its only known direction. But Evans et al (1966) and Shiba et al (1985) discovered it also runs in the reverse direction. This rTCA cycle is an autotrophic engine of carbon fixation. The eleven carboxylic acids are precisely the universal anabolic core. It is a strong candidate for the seed of protometabolism. 

Several steps of any fixation pathway are endergonic (thermodynamically dysfavored) under ordinary conditions. Modern cells pay for these steps with ATP hydrolysis. A protometabolism has no ATP. The environment itself must pay. Thermodynamic drive denotes the requirement that environmental chemistry makes the overall pathway favorable without modern energy currencies like ATP.

We have identified the rTCA cycle by reading the network’s stratigraphy. Comparative biology provides another glimpse. Consider organisms that are anaerobic and chemoautotrophic. In archaea, these are the methanogens; in bacteria the acetogens. What do they have in common? 

Both methanogens and acetogens rely on the Wood-Ljungdahl (WL) pathway (aka acetyl-CoA pathway). WL is in fact the only carbon-fixation pathway found in both bacteria and archaea. Weiss et al (2016) place WL in LUCA itself. WL is also thermodynamically unique. Every other fixation pathway runs uphill and is financed with ATP. The Calvin cycle spends roughly nine ATP per pyruvate; rTCA is among the cheapest. WL runs downhill. Under anoxic conditions with H2 as the electron donor, the overall reaction is exergonic, releasing roughly 100 kJ/mol. Acetogens and methanogens exploit this: the same pathway that builds their biomass also charges their membranes and makes their ATP.  A cell running WL on a H2 supply occupies what Everett Shock called a “free lunch you are paid to eat.” WL is the only fixation chemistry that could have paid its own way. 

Finally, methanogens and acetogens also share a second piece of machinery: flavin-based electron bifurcation. The problem it solves is that reducing CO₂ requires electrons at potential lower than H₂ itself can supply. Bifurcation takes an electron pair from a mid-potential donor and splits it: one electron falls to a high-potential acceptor, and the energy released pays for lifting the other onto low-potential ferredoxin, the reductant that carbon fixation demands. Its distribution across both domains marks it as ancient (Buckel & Thauer 2013).

These two engines have complementary strengths and weaknesses. rTCA contains the entire anabolic core, and it is autocatalytic. But it runs uphill. WL has the right energetics: it is the only chemistry that runs downhill. But it is a linear, non-autocatalytic path, and hence cannot auto-amplify. Braakman and Smith (2013) place a fused WL-rTCA network at the base of the tree. For now, we will treat these fixation pathways as complementary halves of one machine.

Both WL and rTCA are now executed by enzymes. But buried inside many of these proteins are metal-sulfur cofactors, and these are often the catalytically active sites. Ferredoxin contains cubanes whose geometry resembles motifs found in  the mineral greigite (via its precursor mackinawite). The nickel-iron-sulfur cluster at the heart of CO dehydrogenase and acetyl-CoA synthase likewise resembles violarite. Zhao et al. (2022) quantified these cofactor-mineral correspondences. These structural correspondences hint at evolutionary continuity. Russell and Martin (2004) argue that these catalytic sites are mineral fossils, and the surrounding protein is later packaging around an ancient inorganic core. 

Metabolism-first envisions a non-enzymatic protometabolism. Geochemical continuity suggests the earliest catalysts were mineral. But modern enzymes accelerate reactions by factors up to 10^17. Without them, most reactions proceed at negligible rates. Kinetic accessibility denotes the experimental requirement that mineral, metal, or small-molecule catalysts sit in the flow path and drive the key reactions at meaningful rates. 

Vigorous flow solves one problem and creates another. The same throughput that delivers reductant and exports waste can wash a network’s working parts away faster than the chemistry regenerates them. A pathway turns on a characteristic timescale, and its intermediates must survive long enough to complete the loop. Intermediate retention denotes the requirement that residence times exceed network turnover times.

The deepest layer of the network has another striking property. If we abstract away the modern enzymes, cofactors, and carrier groups that now mediate it, the carbon skeletons of WL and rTCA contain only carbon, hydrogen and oxygen. These skeletons are strictly a CHO metabolism. Nitrogen is not needed to construct this central carbon framework. It seems to have entered later, converting pre-existing carbon skeletons into amino acids and the broader CHON metabolism (Braakman & Smith 2013).  Phosphate may show a similar pattern: Goldford et al. (2017) reconstructed a plausible ancient metabolic network with little dependence on phosphate, although Tian et al. (2019) contest how far that inference can be pushed.  This is compatible with de Duve (1991)’s hypothesis that thioesters (like acetyl-CoA) served as an energy currency predecessor of ATP. 

The resulting picture is one of elemental expansion: a CHO carbon-skeleton core, followed by nitrogen incorporation and increasingly phosphate-dependent chemistry. This ordering is naturally compatible with metabolism-first. If RNA came first, despite already requiring fixed nitrogen and phosphate, it becomes less obvious why the deepest metabolic layers should preserve a simpler elemental architecture. 

A Scorecard For Habitat

We have generated twenty requirements.

Universal (11)

• Sustained free-energy flux. Continuous or recurrent throughput maintaining far-from-equilibrium chemistry.
• Chemical selectivity. Flux channeled toward targets, against the combinatorial tar.
• Provenance. Feedstocks and catalysts arise or arrive naturally.
• Reactant concentration. Feedstocks reach reaction-viable concentrations.
• Material flux. Energy and material in, waste out, spent material recycled.
• Product retention. Useful products stay long enough to react again (in creative tension with material flux).
• Net production. Target made faster than destroyed.
• Condition compatibility. Incompatible reactions confined to microenvironments that exchange products realistically.
• Process continuity. Each chemical step connects in geochemically plausible way, without artificial laboratory techniques.
• Habitat persistence. Eden either outlasts, or recurs frequently.
• Localization. Products must preferentially benefit the systems that produce them, rather than diffusing to competitors. 

Genetics-first (5)

• Bioavailable nitrogen. Reactive C–N precursors.
• Bioavailable phosphate. Soluble, concentrated phosphate despite mineral lock-up.
• Monomer activation. Free energy stored or coupled so phosphodiester bonds can form.
• Polymer elongation. Bond formation is efficient and repeatable to functional lengths.
• Template reset. Duplexes separate non-destructively for the next round.

Metabolism-first (4)

• Reductant and carbon co-delivery. Electron donors and CO₂ arrive together, continuously.
• Thermodynamic drive. The environment pays for endergonic chemistry without ATP.
• Kinetic accessibility. Non-enzymatic catalysts must drive key reactions at meaningful rates.
• Intermediate retention. Residence times exceed network turnover times.

Next post, we will use this scorecard to grade specific hypotheses from both camps. See you then!

References

• Braakman & Smith (2013). The Compositional and Evolutionary Logic of Metabolism.
• Buckel & Thauer (2013). Energy Conservation via Electron-Bifurcating Ferredoxin Reduction and Proton/Na⁺-Translocating Ferredoxin Oxidation.
• Butlerov (1861). Formation Synthétique d’une Substance Sucrée.
• de Duve (1991). Blueprint for a Cell: The Nature and Origin of Life.
• Evans, Buchanan & Arnon (1966). A New Ferredoxin-Dependent Carbon Reduction Cycle in a Photosynthetic Bacterium.
• Gianni et al. (2026). A Small Polymerase Ribozyme That Can Synthesize Itself and Its Complementary Strand.
• Goldford, Hartman, Smith & Segrè (2017). Remnants of an Ancient Metabolism without Phosphate.
• Hud et al. (2013). The Origin of RNA and “My Grandfather’s Axe.”
• Kruger et al. (1982). Self-Splicing RNA: Autoexcision and Autocyclization of the Ribosomal RNA Intervening Sequence of Tetrahymena.
• Mateo-Marti et al. (2019). Pyrite-Induced UV-Photocatalytic Abiotic Nitrogen Fixation: Implications for Early Atmospheres and Life.
• Nissen et al. (2000). The Structural Basis of Ribosome Activity in Peptide Bond Synthesis.
• Pitsch et al. (1995). Mineral Induced Formation of Sugar Phosphates.
• Ravasz et al. (2002). Hierarchical Organization of Modularity in Metabolic Networks.
• Russell & Martin (2004). The Rocky Roots of the Acetyl-CoA Pathway.
• Say & Fuchs (2010). Fructose 1,6-Bisphosphate Aldolase/Phosphatase May Be an Ancestral Gluconeogenic Enzyme.
• Schönheit, Buckel & Martin (2016). On the Origin of Heterotrophy.
• Shiba et al. (1985). The CO₂ Assimilation via the Reductive Tricarboxylic Acid Cycle in an Obligately Autotrophic, Aerobic Hydrogen-Oxidizing Bacterium, Hydrogenobacter thermophilus.
• Smith & Morowitz (2004). Universality in Intermediary Metabolism.
• Tian et al. (2019). Phosphates as Energy Sources to Expand Metabolic Networks.
• Toner & Catling (2020). A Carbonate-Rich Lake Solution to the Phosphate Problem of the Origin of Life.
• Weiss et al. (2016). The Physiology and Habitat of the Last Universal Common Ancestor.
• White (1976). Coenzymes as Fossils of an Earlier Metabolic State.
• Wimmer et al. (2021). Energy at Origins: Favorable Thermodynamics of Biosynthetic Reactions in the Last Universal Common Ancestor.
• Wimsatt (2007). Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality.
• Yadav et al. (2020). Chemistry of Abiotic Nucleotide Synthesis.
• Zhao et al. (2022). Quantifying Mineral-Ligand Structural Similarities: Bridging the Geological World of Minerals with the Biological World of Enzymes