Protocells and Membranes

Fatty acids form vesicles by themselves, feed, grow, divide and compete without any genome. The unsolved problem is stated by Szostak's own laboratory: RNA copying needs 50 to 200 millimolar magnesium and a fatty-acid membrane falls apart above about 4. No protocell yet both divides and copies its contents across generations.

Draft:  September 2026

"It also is not clear how to achieve multiple generations of nucleic acid replication within the context of a protocell.  If and only if all of these challenges can be met will it be possible to synthesize an evolving protocell." — Gerald Joyce and Jack Szostak, 2018.1

"There is currently no plausible path for the emergence of a self-replicating protocell, because prevalent formulations of model protocells are built with fatty acid vesicles that cannot withstand the concentrations of Mg2+ needed for the function and replication of nucleic acids." — the opening sentence of a 2023 paper by the people trying to fix it.2

Summary

Of all the parts of the origin-of-life problem, this is the one where the laboratory work is most impressive.  A cell needs a bag, and bags turn out to be easy.

  • Membranes assemble themselves. Above about eight carbons, single-chain fatty acids "form micelles as concentrations increase above a critical value," and "at chain lengths of eight carbons and higher, bilayers begin to appear in the form of membranous vesicles."3 No template, no catalyst, no chemist.
  • They grow, divide and compete. Fed with fresh micelles, multilamellar vesicles stretch into threads and divide under gentle shear "into multiple daughter vesicles without loss of internal contents," through at least three cycles.4 Osmotically swollen vesicles steal membrane from relaxed neighbours, which is Darwinian competition with no genome doing any work.
  • Clay does three jobs at once. Montmorillonite "accelerates the spontaneous conversion of fatty acid micelles into vesicles," gets encapsulated in them, and carries adsorbed RNA inside.5 It is also the mineral that polymerises nucleotides.
  • The unsolved problem is a hundredfold mismatch, and Szostak's own laboratory found it. RNA copying needs 50 to 200 millimolar magnesium; fatty-acid membranes "become unstable in ≥4 mM Mg2+."1 Adamala and Szostak state it directly: at the concentrations RNA needs, magnesium "destroys vesicles."6
  • Every workaround costs something, and the researchers name the cost. Citrate works, but "it is unlikely that sufficient citrate was present in the prebiotic environment."1 Amino acids help by driving the vesicles multilamellar.7 Soda lakes cap magnesium near 1 millimolar, at the price of pH 10.8
  • One workaround reaches the required range and it is worth conceding. Coacervate-supported membranes stay "intact and impermeable to externally-added solutes" at 25 and 50 millimolar magnesium.9 The price is a synthetic polymer with no prebiotic source, and nobody has run replication inside one.
  • Nobody has a protocell that both divides and copies its genes. That is not a creationist claim. It is Joyce and Szostak's own summary, quoted above.

The verdict.  Compartments form, grow, divide and compete without help, and that should be conceded loudly.  What no laboratory has built is a compartment whose contents are copied and inherited across generations, and the best-understood reason is a hundredfold mismatch between the chemistry of the genes and the chemistry of the bag.

What the laboratory has actually built

Four results, each of which removes a claim that circulates in apologetics.

Vesicles form by themselves.  David Deamer's review gives the chemistry:  short-chain amphiphiles form micelles above a critical concentration, and at eight carbons and above they start forming bilayer vesicles.3  Nobody has to fold them.  This is ordinary physical chemistry, driven by the fact that a molecule with a water-loving head and a water-hating tail has nowhere else to go.

They can be made in one step, with cargo, on a mineral that also does the genetics.  Martin Hanczyc, Shelly Fujikawa and Szostak showed in 2003 that montmorillonite clay "accelerates the spontaneous conversion of fatty acid micelles into vesicles," that clay particles and RNA adsorbed to clay are encapsulated in the process, and that the resulting vesicles "can grow by incorporating fatty acid supplied as micelles and can divide without dilution of their contents by extrusion through small pores."5  Growth, division and a catalytic surface inside, in one experiment.

They grow and divide while keeping their contents.  Ting Zhu and Szostak fed multilamellar vesicles with micelles.  Surface area grows faster than volume, so the spheres stretch into threads, and then "modest shear forces are then sufficient to cause the thread-like vesicles to divide into multiple daughter vesicles without loss of internal contents."4  The shear stress needed is 0.35 dynes per square centimetre, which is less than a stream.  Vesicle number rises exponentially across at least three cycles.

They compete, and the competition is Darwinian without a genome.  RNA inside a vesicle exerts osmotic pressure on the membrane, which drives uptake of more membrane material from relaxed neighbours, which shrink.  More RNA inside means a faster-growing bag.  That is selection at the level of the compartment, driven by the contents, before anything is copied.

None of that needs a designer's hand in the sense apologetics usually means, and an argument that treats compartment formation as an unsolved problem is twenty years out of date.

The magnesium mismatch

Here is the specific, quantitative reason the programme has stalled, and it comes from the laboratory that built it.

Non-enzymatic RNA copying needs magnesium, at 50 to 200 millimolar.  Katarzyna Adamala and Szostak print the collision in one sentence:  RNA copying reactions "require high (50–200 mM) concentrations of Mg2+ (or other divalent) ions, but Mg2+ at such concentrations destroys vesicles by causing fatty acid precipitation."6  Joyce and Szostak put the membrane side of it at "≥4 mM Mg2+," and call it "an apparent incompatibility between the two systems."1

The problem is older than that.  In 2005 Irene Chen and Szostak measured how fragile these membranes are:  decanoic acid vesicles "aggregate at moderate concentrations of monovalent ions (∼0.2 M NaCl), and divalent cations cause aggregation and precipitation even at quite low concentrations (<1.5 mM MgCl2)."10  Seawater is 0.5 molar sodium chloride and about 50 millimolar magnesium.

So the compartment that forms most easily is destroyed by the ocean, and destroyed a hundred times over by the chemistry its contents need.

The workarounds, with their prices

Fix What it buys What it costs
Citrate chelation RNA copying inside an intact vesicle "unlikely that sufficient citrate was present in the prebiotic environment"1
Glycerol monoesters tolerance from 0.5 to 4 mM a factor of eight against a shortfall of a hundred
Amino acids survival at 10 mM magnesium works by driving the vesicles multilamellar7
Soda lakes magnesium capped near 1 mM by carbonate precipitation pH 10 and about a molar of sodium8
Coacervate-supported membranes intact at 25 and 50 mM a synthetic polycation with no prebiotic source9
Cyclophospholipids growth, division and primer extension still needs the chelate; a few nucleotides, once2

Two of those rows deserve more.

Amino acids are the most prebiotically comfortable fix.  Caitlin Cornell's group found that serine and glycine, but not leucine, protect decanoic acid vesicles against 10 millimolar magnesium, and that the protection "persists after mixing and dilution to lower overall concentrations."7  Ten millimolar is real progress and still far short of fifty.

And one fix reaches the required range, which cuts against this article's own argument.  Jessica Lee and colleagues grew a lipid membrane around a coacervate core and tested it at 25 and 50 millimolar magnesium:  "while fatty acid and blended membrane vesicles are disrupted by the addition of Mg2+, the corresponding coacervate-supported membranes remain intact and impermeable to externally-added solutes."9  Fifty millimolar is the bottom of the range RNA copying needs.  The honest caveat is the polycation, poly(allylamine hydrochloride), a laboratory polymer with no prebiotic source, and the paper claims none.  A 2026 follow-up found that mixtures of prebiotically plausible single-chain amphiphiles outperform pure ones, giving "higher yields, smaller sizes, and greater subcompartment formation."11

So the hundredfold mismatch is a statement about fatty-acid vesicles, not about compartments in general.  That is the accurate form, and this article states it that way rather than the stronger form that would be easier to argue.

Where membranes will and will not form

The membrane question is also the sharpest weapon in the argument between the two mainstream camps about where life started.

Deamer against submarine vents, and it is a membrane argument.  "Self-assembly of membranes composed of amphiphiles has not been demonstrated in hydrothermal vents, either in situ or in laboratory simulations.  In contrast, hydrothermal fields are supplied by what is essentially distilled water with ionic solutes in millimolar ranges that are more conducive to the self-assembly of amphiphiles into vesicles."12  The experiment behind it tested amphiphiles in water drawn straight from Yellowstone springs and in seawater.  "None of the amphiphiles that were tested formed stable vesicles in sea water because the high ionic concentrations disrupted membrane stability."13

Lane's laboratory answered it in 2019 and the answer is good.  Mixtures of twelve fatty acids and alcohols from ten to fifteen carbons form vesicles at 600 millimolar sodium chloride, 50 millimolar magnesium and 10 millimolar calcium, at 70 °C.  The critical bilayer concentration falls thirtyfold, from 39 millimolar for pure decanoic acid to 1.3 millimolar for the mixture.14  Damer and Deamer cite the result and do not dispute it;  they answer that a prebiotic source of such mixtures remains uncertain.

There is a catch neither camp stresses, and this article flags it as an inference rather than a finding.  The Lane result runs at pH 11 to 12, with "very few" vesicles below pH 9.  Alkaline hydroxide complexation is what takes the divalent cations out of play, so the salt tolerance and the RNA-hostile pH are the same fact.  RNA does not survive pH 11 to 12.  No paper states this objection, and it should be checked before anyone leans on it.

Neither camp is entitled to say the other's compartment is impossible.  That is the honest summary and it is more useful than either camp's press release.

Where the lipids come from

A quieter problem, and it is quantitative.

Fatty acids shorter than about eight carbons do not form bilayers at all.  Meteoritic straight-chain fatty acids fall off exponentially with chain length and stop around twelve carbons, so the ones long enough to work are present at roughly 104 to 105 parts per billion.  Work that through:  dissolving a kilogram of Murchison-grade material into a litre of water yields something like 0.6 millimolar of vesicle-capable fatty acid, against a critical bilayer concentration of 39 millimolar for decanoic acid.  That arithmetic is this project's own, on published numbers, and it is flagged as such.  Two orders of magnitude of evaporative concentration are needed before a vesicle exists.

And the leading prebiotic synthesis network does not supply the missing piece.  Sutherland's cyanosulfidic chemistry produces glycerol, which is the headgroup.  It does not make acyl chains.  A network described as producing "lipid precursors" has produced the part that is not the problem.

The membraneless alternative, and its mirror-image problem

If membranes are fragile, perhaps the first compartment had no membrane.  Coacervates are dense liquid droplets that form when oppositely charged polymers meet, and they concentrate RNA for free.

The problem was measured rather than asserted.  RNA moves between droplets so fast that no droplet keeps its own genes:  exchange times of about a minute for a 6-nucleotide RNA and five minutes for a 49-mer.  Szostak's group state the consequence in their own abstract:  "the rapid exchange of RNA between these compartments, along with their rapid fusion, both within minutes, means that individual droplets would be unable to maintain their separate genetic identities.  Hence, Darwinian evolution would not be possible."15

The best fix is a good one and it is very recent.  Rainwater rinsing crosslinks the droplet surface, and after it "shorter RNA sequences of 6 to 8 nt exchange within a timescale of minutes, longer RNA sequences of 35 nt or higher remained compartmentalized for days."15  Three limits the authors state themselves:  the polycation is synthetic, the droplets destabilise above roughly 1.3 millimolar salt, and the boundary is a crosslinked skin rather than a membrane, so short RNAs still leak in minutes.

What no protocell has done

The best published systems each get part of the way, and the gaps are complementary.

System Grows Divides Copies its contents
Zhu & Szostak 2009 yes yes, three cycles no, the RNA is inert cargo
Adamala & Szostak 2013 yes no yes, with citrate
Toparlak et al. 2023 yes yes a primer extended by a few nucleotides, on a defined template, once

The field's own diagnosis is the opening sentence of that 2023 paper:  "There is currently no plausible path for the emergence of a self-replicating protocell."2  And Joyce and Szostak's summary, which is the load-bearing citation of this article:  "It also is not clear how to achieve multiple generations of nucleic acid replication within the context of a protocell.  If and only if all of these challenges can be met will it be possible to synthesize an evolving protocell."1

The claims, audited

Claim Verdict
"Cell membranes could not form on their own" Dead. Fatty acids above eight carbons make bilayers unaided
"A membrane could not grow or divide without machinery" Dead. Growth on feeding and division under gentle shear, contents retained
"There is no competition without genes" Dead. Osmotically swollen vesicles take membrane from their neighbours
"Magnesium destroys the membranes RNA needs" Sound, in its narrow form. A hundredfold mismatch, and Szostak's laboratory found it
"Every workaround is artificial" Overstated. Amino acids and soda lakes are natural; the coacervate polymer is not
"No protocell copies its own contents across generations" Sound. Stated by Joyce and Szostak, and by the 2023 paper trying to fix it
"Prebiotic chemistry supplies the lipids" Not established. The leading network makes the headgroup, not the chains

Objections

1. The magnesium problem has been solved

Partly, and this article says so before arguing.  A coacervate-supported membrane survives 50 millimolar magnesium, which is the bottom of the range RNA copying needs.9  What has not been done is running replication inside one, and the polymer holding it together is a modern laboratory reagent.  The correct statement is that the mismatch is a fact about fatty-acid vesicles rather than about compartments as such.

2. Soda lakes solve it more plausibly

They are the most geologically honest fix on the table:  carbonate precipitation caps magnesium near one millimolar, and RNA polymerisation runs at comparable rates in lake water and in laboratory buffer.8  The cost is pH 10 and about a molar of sodium, and monovalent salt at that level aggregates fatty-acid vesicles on its own.  A partial answer, not a complete one.

3. These vesicles are nothing like a cell

True, and nobody claims otherwise.  A fatty-acid vesicle has no proteins, no transport machinery, no metabolism and no genes.  What it demonstrates is narrower and still worth having:  that the boundary of a cell is not a thing that needs explaining separately.  The hard part was never the bag.

4. The experiments always supply pure fatty acids

Fair, and the chain-length arithmetic above is the strongest form of it.  The meteoritic supply of the right lengths is thin, and the leading synthesis network does not make acyl chains at all.  This is the least-discussed problem in the area and probably the most durable.

5. Coacervates avoid the membrane problem

They trade it for a worse one.  Without a boundary, the contents mix in minutes, and the authors say that rules out Darwinian evolution.15  The rain-crosslinked skin fixes it for longer RNAs and not for short ones.

6. This is an argument from what has not been done yet

It is, and the article says so plainly.  The field is thirty years old, it is well funded, and the people running it publish their negative results.  What makes this a report rather than a prediction is that the specific obstacle is named and measured:  a hundredfold concentration mismatch, not a vague appeal to complexity.  If a protocell divides and copies its own genome next year, this section is wrong and will be rewritten.

What this article claims, and what it does not

It claims that fatty-acid vesicles form, feed, grow, divide and compete without help, that this should be conceded, and that clay does several of the required jobs at once.  It claims that RNA chemistry and fatty-acid membranes are incompatible by about two orders of magnitude in magnesium, that every workaround carries a stated cost, that no protocell both divides and copies its contents across generations, and that the prebiotic supply of the right-length fatty acids is thin.

It does not claim that compartments are impossible or that the magnesium problem is unsolvable;  one published system already survives the required concentration.  The pH objection to the Lane result is this project's own inference from published figures and is marked as unconfirmed.  The meteoritic fatty-acid arithmetic is likewise this project's own, on other people's numbers.  Those gaps are listed in the literature review.

Sources

  1. 1.

    Joyce, and Szostak.  “Protocells and RNA Self-Replication.”  Cold Spring Harbor Perspectives in Biology 10:a034801.  2018.  A review of protocells and how RNA could replicate inside them.  The authors argue that forming a protocell is not the hard part: the real problem is replication, meaning growth and division of the membrane together with copying of the encapsulated genetic material.  The article uses it for the unsolved problem, the magnesium figure and the citrate concession.

    1. 1a.

      “However, nonenzymatic copying of RNA templates requires very high concentrations of divalent cations, typically Mg2+, and most ribozymes require high concentrations of Mg2+ for optimal activity.  In contrast, fatty acid–based membranes become unstable in ≥4 mM Mg2+, indicating an apparent incompatibility between the two systems.” ↩ ↩

    2. 1b.

      “However, it is unlikely that sufficient citrate was present in the prebiotic environment, and thus alternatives must be sought.” ↩ ↩

    3. 1c.

      “It also is not clear how to achieve multiple generations of nucleic acid replication within the context of a protocell.  If and only if all of these challenges can be met will it be possible to synthesize an evolving protocell.” ↩ ↩

  2. 2.

    Toparlak, Ö. Duhan et al.  “Cyclophospholipids Enable a Protocellular Life Cycle.”  ACS Nano 17:23772–23783.  2023.  The paper's opening sentence is that diagnosis:  “There is currently no plausible path for the emergence of a self-replicating protocell, because prevalent formulations of model protocells are built with fatty acid vesicles that cannot withstand the concentrations of Mg2+ needed for the function and replication of nucleic acids.  Although prebiotic chelates increase the survivability of fatty acid vesicles, the resulting model protocells are incapable of growth and division.”  Their system grows, divides and extends a primer by a handful of nucleotides, on a defined template, in one direction, once, and it still uses citrate.  A 2023 laboratory study of protocells made from cyclophospholipids mixed with fatty acids, tested in magnesium held by citrate.  The protocells grew and divided, kept their encapsulated nucleic acids through the division, and extended an RNA primer.  The article uses it as the clearest recent attempt to close the protocell life cycle, and it opens with the field's own diagnosis. ↩ ↩ ↩

  3. 3.

    Deamer, David W. et al.  “The First Cell Membranes.”  Astrobiology 2(4):371–381.  2002.  Figure 2:  “form micelles as concentrations increase above a critical value”.  “at chain lengths of eight carbons and higher, bilayers begin to appear in the form of membranous vesicles”.  A review of how amphiphilic compounds, which have a water-loving head and a water-hating tail, could have reached early Earth from space and self-assembled into membranes.  Laboratory simulations show that such vesicles readily encapsulate functional macromolecules, including nucleic acids and polymerases.  Two companion papers give the concentration thresholds usually quoted alongside this review: Apel, Deamer and Mautner, BBA 1559 (2002): 1-9, and Monnard et al., Astrobiology 2 (2002): 139-152.  The article uses the review to show that fatty acids assemble into membranes without a template or catalyst. ↩ ↩

  4. 4.

    Zhu, Ting F., and Jack W. Szostak.  “Coupled Growth and Division of Model Protocell Membranes.”  Journal of the American Chemical Society 131:5705–5713.  2009.  Section Abstract:  “modest shear forces are then sufficient to cause the thread-like vesicles to divide into multiple daughter vesicles without loss of internal contents”.  The RNA inside is inert cargo; nothing in this system copies it.  The competition result the article refers to is in Chen, Irene A., Richard W. Roberts, and Jack W. Szostak, "The Emergence of Competition Between Model Protocells," Science 305 (2004): 1474-1476.  A laboratory study of large multilamellar fatty-acid vesicles, about 4 micrometres across, fed with fatty-acid micelles.  The vesicles stretch into long threads when fed, and under a shear stress of 0.35 dynes/cm² the threads divide without losing their contents through at least three cycles, while the vesicle number rises exponentially.  The article uses it as the clearest case of a membrane that grows and divides while keeping its contents.  Mirror:  PMC ↩ ↩

  5. 5.

    Hanczyc, Martin M. et al.  “Experimental Models of Primitive Cellular Compartments: Encapsulation, Growth, and Division.”  Science 302:618–622.  2003.  Section Abstract:  “accelerates the spontaneous conversion of fatty acid micelles into vesicles”.  “can grow by incorporating fatty acid supplied as micelles and can divide without dilution of their contents by extrusion through small pores.”  Szostak's own laboratory later supplied the limit: Zhu and Szostak's 2009 paper, which the article cites, says that this division by extrusion.  “results in significant loss of protocell contents during each division cycle.”  A laboratory study of how the clay montmorillonite affects fatty-acid vesicles.  The clay speeds up vesicle formation by about a factor of 100, and the vesicles that form can grow and divide without dilution of their contents.  The article uses it because one experiment shows growth, division and a catalytic surface working together. ↩ ↩

  6. 6.

    Adamala, Katarzyna, and Jack W. Szostak.  “Nonenzymatic Template-Directed RNA Synthesis Inside Model Protocells.”  Science 342:1098–1100.  2013.  The passage, in full:  “However, the conditions required for RNA replication chemistry and fatty acid vesicle integrity have appeared to be fundamentally incompatible (5) (Fig S1).  Both ribozyme catalyzed and non-enzymatic RNA copying reactions require high (50–200 mM) concentrations of Mg2+ (or other divalent) ions (6), but Mg2+ at such concentrations destroys vesicles by causing fatty acid precipitation.”  A laboratory study that screened small molecules for ones that protect fatty-acid vesicles from magnesium, then tested whether RNA copying still works inside the protected vesicles.  It found that citrate protected oleate vesicles at 10 mM magnesium and let RNA copying go on more slowly: at 50 mM magnesium the primer-extension rate fell from 1.4 to 0.67 per hour, while EDTA and NTA stopped it entirely.  The article uses it because the authors' own laboratory states the conflict between RNA copying and fatty-acid membranes in a single sentence.  Mirror:  PMC ↩ ↩

  7. 7.

    Cornell, Caitlin E. et al.  “Prebiotic amino acids bind to and stabilize prebiotic fatty acid membranes.”  Proceedings of the National Academy of Sciences 116:17239–17244.  2019.  The mechanism is a change of lamellarity, not chelation.  10 mM magnesium alone.  “converts individual ∼10-μm multilamellar vesicles into punctate structures,”.  and serine or glycine prevents that.  Leucine, which has a hydrophobic side chain, does not work.  Note what has not been tested: as Cohen et al. point out, RNA assembly inside these particular vesicles has not been investigated.  A laboratory study of whether amino acids bind to decanoic acid vesicles and protect them from salt and magnesium, using cryo-electron microscopy, fluorescence microscopy, NMR spectroscopy and filtration assays.  It found that several amino acids bind to the membranes and that a subset stabilizes them against salt and magnesium.  The article uses it as the most prebiotically comfortable fix for the magnesium problem. ↩ ↩ ↩

  8. 8.

    Cohen, Zachary R. et al.  “Natural soda lakes provide compatible conditions for RNA and membrane function that could have enabled the origin of life.”  PNAS Nexus 3:pgae084.  2024.  The abstract states the problem and the test:  Section Abstract:  “Subsequent assembly of RNA and membranes into cells is a long-standing problem because RNA function requires divalent cations, e.g. Mg2+, but Mg2+ disrupts fatty acid membranes.  The low solubility of Mg-containing carbonates limits soda lakes to moderate Mg2+ concentrations (∼1 mM), so we investigated whether both RNAs and membranes function within these lakes…  In the laboratory, nonenzymatic, RNA-templated polymerization of 2-aminoimidazole-activated ribonucleotides occurred at comparable rates in lake water and standard laboratory conditions (50 mM MgCl2, pH 8).”  The real samples were from Last Chance Lake and Goodenough Lake, British Columbia, November 2021:  “∼1 M Na+ and ∼1 mM Mg2+ at pH 10.”  The prices are the authors'.  The ligase ribozyme needed the pH dropped to 9.  The authors report that.  “a high concentration of monovalent cations… leads to lower yields of ribozyme-catalyzed ligation”.  and that at 1 M Na+ the amphiphiles survive as flocculated aggregates rather than free vesicles.  The same paper supplies two of the article's other statements.  On citrate:  “Chelation of divalent cations by excess citrate preserves fatty acid vesicles and enables internal RNA reactions, but high concentrations of citrate on the early Earth are unlikely,”.  On the amino-acid workaround:  “the assembly of RNA has not been investigated inside these vesicles.”  A laboratory and field study of water from two soda lakes, in which the authors tested whether RNA copying and fatty-acid membranes both work.  Soda lakes are rich in sodium and carbonate, and carbonate precipitation keeps their magnesium near 1 millimolar.  The article uses it as the most geologically honest fix on the table, and the authors name its costs. ↩ ↩ ↩

  9. 9.

    Lee, Jessica et al.  “Hybrid Protocells Based on Coacervate-Templated Fatty Acid Vesicles Combine Improved Membrane Stability with Functional Interior Protocytoplasm.”  Small 20:e2406671.  2024.  Section Abstract:  “while fatty acid and blended membrane vesicles are disrupted by the addition of Mg2+, the corresponding coacervate-supported membranes remain intact and impermeable to externally-added solutes.”  The cost is in the paper's own methods.  The coacervate is made from poly(allylamine hydrochloride) and adenosine diphosphate, and poly(allylamine hydrochloride) is a laboratory polymer.  A laboratory study that wrapped a fatty-acid membrane, or a mix of fatty acid and phospholipid, around a droplet of a synthetic polymer and ADP.  It tested the membranes at 25 and 50 millimolar magnesium: the coacervate-supported membranes stayed intact while plain fatty-acid vesicles were disrupted, which makes it the single strongest published answer to the magnesium argument.  It belongs in any honest treatment of that argument.  The article cites it as a fix that reaches the required range, which cuts against its own argument. ↩ ↩ ↩ ↩

  10. 10.

    Chen, Irene A. et al.  “RNA Catalysis in Model Protocell Vesicles.”  Journal of the American Chemical Society 127:13213–13219.  2005.  The quoted vesicles are made of decanoic acid:  “aggregate at moderate concentrations of monovalent ions (∼0.2 M NaCl), and divalent cations cause aggregation and precipitation even at quite low concentrations (<1.5 mM MgCl2, or 1:65 ratio of cation to lipid).”  For scale, modern seawater is about 0.5 M sodium chloride and about 53 mM magnesium.  A laboratory study of vesicles made from myristoleic acid and its glycerol ester, which stayed intact in magnesium, let magnesium pass through the membrane within minutes, and let a hammerhead ribozyme inside them cut itself when magnesium was added.  This is the earliest clear statement of the incompatibility, and it is from the laboratory that built the protocell programme.  The article uses it because the authors' laboratory states how fragile fatty-acid membranes are in divalent ions. ↩

  11. 11.

    Joshi, Manesh Prakash et al.  “Increasing the Compositional Heterogeneity of Single-Chain Amphiphile Membranes Supported by Coacervate Cores Alters Stability and Properties of the Hybrid Protocells.”  Langmuir 42:13582–13595.  2026.  Section Abstract:  “higher yields, smaller sizes, and greater subcompartment formation”.  A 2026 laboratory study that built hybrid protocells, with a membrane of mixed prebiotic amphiphiles around a coacervate droplet, and compared them with protocells whose membranes used one kind of amphiphile.  The mixed membranes gave higher yields and smaller protocells, and they stayed intact under acidic conditions.  The paper does not test magnesium, so the article cites it only for the heterogeneity result, and the magnesium finding belongs to Lee et al. 2024 alone. ↩

  12. 12.

    Deamer, David, and Bruce Damer.  “Can Life Begin on Enceladus? A Perspective from Hydrothermal Chemistry.”  Astrobiology 17:834–839.  2017.  The sentence quoted in the article, with the clause that follows it:  “Self-assembly of membranes composed of amphiphiles has not been demonstrated in hydrothermal vents, either in situ or in laboratory simulations.  In contrast, hydrothermal fields are supplied by what is essentially distilled water with ionic solutes in millimolar ranges that are more conducive to the self-assembly of amphiphiles into vesicles.”  A second Deamer objection in the same paper is independent of ionic strength and harder for the vent camp to answer:  “Because hydrothermal vents cannot undergo such [dehydration-rehydration] cycles, how protocells with encapsulated polymers could assemble is an open question.”  The supporting experiment is Milshteyn, Daniel, Bruce Damer, Jeff Havig, and David Deamer, "Amphiphilic Compounds Assemble into Membranous Vesicles in Hydrothermal Hot Spring Water but Not in Seawater," Life 8, no. 2 (2018): 11.  A perspective article, not an experiment, that compares deep-sea vents with hot-spring fields as sites for the origin of life.  Its membrane objection to vents is that amphiphiles have not been shown to self-assemble into membranes there.  The article uses it as Deamer's own case against submarine vents. ↩

  13. 13.

    Milshteyn, Daniel et al.  “Amphiphilic Compounds Assemble into Membranous Vesicles in Hydrothermal Hot Spring Water but Not in Seawater.”  Life 8:11.  2018.  The amphiphiles were dispersed in water.  “drawn directly from hot springs in Yellowstone National Park at two pH ranges, and the results were compared with sea water under the same conditions.”  The paper's summary of the seawater result:  Section Abstract:  “None of the amphiphiles that were tested formed stable vesicles in sea water because the high ionic concentrations disrupted membrane stability.”  Read it against Jordan et al. 2019, which got vesicles at seawater ionic strength from a twelve-component mixture rather than from the pure acids tested here.  The two results are compatible, and the difference is the composition of the amphiphile pool.  A laboratory study that dispersed 10- and 12-carbon monocarboxylic acids, and a mixture of the 12-carbon acid with its monoglyceride, in water from Yellowstone hot springs and in seawater.  The pure acids formed vesicles in the spring water, the mixture assembled into relatively stable membranes at both acidic and alkaline pH, and none of the compounds formed stable vesicles in seawater.  The article cites it as the experiment behind Deamer's flat statement about vents, and it is a real measurement rather than an argument from principle. ↩

  14. 14.

    Jordan, Sean F. et al.  “Promotion of protocell self-assembly from mixed amphiphiles at the origin of life.”  Nature Ecology & Evolution 3:1705–1714.  2019.  Damer and Deamer cite the result and do not dispute it.  Their answer, in Damer, Bruce, and David Deamer, "The Hot Spring Hypothesis for an Origin of Life," Astrobiology 20 (2020): 429-452, is that a plausible prebiotic source of such mixtures.  “remains uncertain.”  One catch neither camp stresses: the result needs pH 11-12, and Jordan et al. report.  “very few”.  vesicles below pH 9, and RNA does not survive pH 11-12.  The alkaline hydroxide complexation that removes the divalent cations and the RNA-hostile pH are the same fact.  A laboratory study of vesicles made from a mixture of twelve fatty acids and alcohols with 10 to 15 carbons, which form vesicles at 600 mM NaCl, 50 mM Mg2+ and 10 mM Ca2+, pH 11-12, 70 °C, which is modern seawater ionic strength under alkaline vent chemistry.  The critical bilayer concentration falls thirtyfold, from 39 mM for pure decanoic acid to 1.3 mM for the mixture.  The article uses it as the answer to the claim that salt and divalent ions rule out vesicles at hydrothermal vents. ↩

  15. 15.

    Agrawal, Aman et al.  “Did the exposure of coacervate droplets to rain make them the first stable protocells?”  Science Advances 10:eadn9657.  2024.  The problem they answer is stated in their own abstract:  Section Abstract:  “the rapid exchange of RNA between these compartments, along with their rapid fusion, both within minutes, means that individual droplets would be unable to maintain their separate genetic identities.  Hence, Darwinian evolution would not be possible.”  A laboratory study of membraneless droplets (coacervates) made from poly(diallyldimethylammonium chloride), a positively charged polymer, and ATP, which the authors sheared in distilled water so that their surfaces cross-linked.  The droplets stayed unfused for months and held RNA for days, depending on the RNA's length, and the authors' "should apply" to natural polycations is a promissory note, not a result.  The article uses it for the best recent fix to RNA leaking between droplets. ↩ ↩

    1. 15a.

      The baseline exchange times are Agrawal et al.'s own, measured by fluorescence recovery after photobleaching and reported in their introduction to the problem; they credit the underlying difficulty to Jia, Hentrich and Szostak, Origins of Life and Evolution of Biospheres 44 (2014): 1-12. ↩