The Role of Fatty Acid Vesicles in the Origin of Polymer Function

Abstract

The three signature structures of cells are membranes, proteins, and nucleic acids. These structures differ markedly in their composition, so how did they first come together in one unit? And how were peptides and oligonucleotides with functions benefiting the unit selected from random sequences? I review evidence for the following scheme: The first membranes were composed of fatty acids that self-assembled in shallow bodies of fresh water into dynamic, metastable vesicles. The vesicles encapsulated peptides and oligonucleotides during cycles of dehydration and rehydration. Alternatively, polymer formation from membrane-associated monomers yielded peptide and oligonucleotide-containing vesicles. In either case, the polymer-bearing vesicles then became enriched for specific polymers due to the dynamic character of fatty acid membranes. Vesicles bearing peptides that increased vesicle stability and growth would have increased in frequency. Vesicles bearing oligonucleotides that increased the concentration of beneficial peptides would have been further favored. Complementary oligonucleotides could have stabilized peptides and reduced their diffusion out of the vesicles. They could also have directed de novo, templated peptide synthesis, which would have opened the path to the generation of novel peptides.

Keywords

Origin of life—Prebiotic—Self-assembly—Polymer—Vesicle—Membrane

1. Introduction

The fundamental unit of life is the cell. The three signature structures of cells are membranes, proteins, and nucleic acids.1 Together, they draw energy and material from the environment and use it to maintain and reproduce the unit. These structures differ markedly in their composition, that is, in their building blocks and the linkage between building blocks. So how did they come together in one unit? Furthermore, how were polymers with a function that benefited the unit enriched relative to random prebiotic polymers?

Section 2 reviews how prebiotic membranes composed of fatty acids assemble spontaneously into dynamic compartments called vesicles that can grow, divide, shrink, or collapse. Section 3.1 reviews how such vesicles can encapsulate preformed polymers, and Section 3.2 describes a possible parallel process in which the polymer-building blocks assembled with membranes prior to polymerization. These sections set the stage for addressing the harder question of why vesicles that bore specific polymers would have accumulated. What functions benefiting the unit did these polymers carry out? Section 4 reviews work which shows how specific peptides increase membrane stability or growth. Section 5 addresses how specific oligonucleotides could stabilize these peptides and increase their production. This section includes a proposal for the origin of oligonucleotide-templated peptide synthesis.

This scheme differs from most other perspectives on the origin of cells. In my scheme, functional polymers emerged due to their selection by the unit for positive effects on stability and growth, rather than for effects on self-replication or metabolic reactions. The effects on stability and growth would have led to compositional evolution of a vesicle population, prior to the origin of polymer replication. The oligonucleotide-templated peptide synthesis I will describe would have led to the transition to replication-based evolution. Only after that transition could the peptide catalysts required for energy harvesting and metabolism have arisen (Orgel, 1998).

2. Cellular Organization Most Likely Originated with Fatty Acid Vesicles Formed in Shallow Bodies of Fresh Water

The initial driver of biological organization was probably the hydrophobic effect, because it drives the spontaneous formation of membranes from amphiphilic molecules (Woolf, 2015; Graham, 2023; Tanford, 1978). The aggregation of the hydrophobic tails increases entropy by freeing the movement of water molecules in the solution, and the hydrophilic heads’ favorable interaction with water orients the tails (Fig. 1).

FIG. 1.

Prebiotic amphiphiles form membrane-bound compartments called vesicles. An increase in entropy drives hydrophobic portions of the molecules to sequester from water, thereby forming small clusters called micelles and larger structures called vesicles that contain an aqueous lumen. Not shown: the amphiphiles can aggregate directly to form vesicles rather than passing through the micellar phase. Environmental conditions affect the equilibria among the different phases, as discussed in the text. Decanoic acid is a prebiotic fatty acid with sufficiently hydrophobic and hydrophilic ends to form vesicles. The micrograph is from a sample prepared by dissolving decanoic acid in a solution of NaOH and then titrating to pH 7.65 with HCl (reprinted from Black et al., 2013).

Fatty acids, the hydrophobic component of phospholipids, were likely the first membrane-building blocks. They are found in meteorites and also could have been made on early Earth by the Fischer–Tropsch reaction (Yuen and Kvenvolden, 1973; Lawless and Yuen, 1979; Naraoka et al., 1999; Cohen et al., 2023). Fatty acids of eight or more carbons spontaneously form vesicles, bilayered membranes that enclose an aqueous space (Deamer et al., 2002; Morigaki and Walde, 2007) (Fig. 1).

Mixtures with other prebiotic compounds increase the robustness of fatty acid vesicles. Including long-chain alcohols substantially lowers the concentration required for vesicle formation (Apel et al., 2002) and extends the pH range over which the vesicles are stable. And mixing the fatty acids with either long-chain alcohols or certain amino acids stabilizes the vesicles against disruption by divalent cations (Monnard et al., 2002; Chen et al., 2005; Cornell et al., 2019; Bonfio et al., 2020).

The most likely location for the formation of the first fatty acid–based vesicles would have been shallow bodies of fresh water. The high concentrations of both sodium and divalent cations in the oceans would have been challenging even with the additional components noted above. Shallow rather than deep bodies of water would have been favored because the concentration of fatty acids would have been higher there. For example, calculations indicate that a concentration sufficient for vesicle formation would have been released into a shallow lake hit by a large meteorite on which the Fischer–Tropsch reaction took place during infall (Cohen et al., 2023). Periodic dehydration of the lake would have increased the concentration and thereby made vesicle formation even more likely. Experiments have demonstrated that fatty acid vesicles do form and persist in water from hot springs and soda lakes (lakes which are high in carbonate salts) (Milshteyn et al., 2018; Cohen et al., 2024a).

Three properties of fatty acid vesicles make them particularly suitable for the first cell membranes: (1) They can passively take up essential nutrients such as nucleotides, that is, without evolved transport systems (Mansy et al., 2008); (2) they can grow by incorporating additional fatty acid molecules, either from added micelles or from other vesicles (Lowe et al., 2025; Todd et al., 2022; Walde et al., 1994; Berclaz et al., 2001; Hanczyc et al., 2003; Chen and Szostak, 2004; Adamala and Szostak, 2013); and (3) they can divide without losing their contents (Hanczyc et al., 2003; Zhu and Szostak, 2009).

There are certainly other candidates for the first protocell compartments, including vesicles composed of long-chain alkyl amines or alkyl phosphates (Maurer and Monnard, 2011), coacervates (Keating et al., 2021), and micelles (Kahana et al., 2023). It is also possible that phospholipids emerged in the prebiotic environment (Gibard et al., 2018). Nonetheless, fatty acid vesicles stand out, in sum, because of their prebiotic plausibility, experimentally established cell-like characteristics, direct path to modern lipids, and—as described in the remainder of the perspective—the rationale they provide for the evolution of peptides and oligonucleotides.

3. The Colocalization of Fatty Acid Vesicles with Polymers Most Likely Occurred in Shallow Bodies of Water Subject to Periodic Dehydration

3.1. Encapsulation of preformed polymers

The next question is how fatty acid vesicles with encapsulated polymers could have arisen. The origin of the polymers is beyond the scope of this review; it is generally believed to have involved condensing agents or dehydration (Livio and Szostak, 2024; Frenkel-Pinter et al., 2020a) (Fig. 2A).

FIG. 2.

The three signature components of cells could have colocalized by several mechanisms, including (A) dehydration–rehydration of preformed oligomers mixed with vesicles and (B) formation of oligomers from monomers associated with vesicles. The formation of oligomers depicted in (A) would have been due to dehydration and/or activation of the building blocks by condensing agents. The building blocks of polypeptides are amino acids, which form abiotically. The prebiotic synthesis of the building blocks of oligonucleotides (i.e., nucleotides) is poorly understood; it could have proceeded, for example, as described by Powner et al. (2009), rather than by joining a sugar and base. The binding of monomers to vesicles depicted in (B) would have been spontaneous; the nonbinding of “other compounds,” represented by the stars, illustrates the selective potential of vesicles. Condensing agents, in addition to dehydration, could have contributed to vesicle-associated oligomer formation from the monomers, since at least one such agent works in vesicles (Bonfio et al., 2020).

A variety of procedures generate encapsulated peptides or oligonucleotides during the formation or reformation of fatty acid vesicles. Formation of vesicles from fatty acid micelles has been shown to result in the encapsulation of polymers present in the micelle solution. For example, induction of vesicle formation by the addition of decanol to a solution of decanoic acid micelles (Fig. 1) and catalase resulted in encapsulation of the enzyme (Apel et al., 2002). Combination of an alkaline preparation of oleic acid micelles with a neutral solution of a peptide, followed by vortexing and tumbling, also yielded peptide-containing vesicles (Adamala and Szostak, 2013). RNA has been encapsulated in myristoleic acid/glycerol monomyristoleate vesicles by vortexing and repeated freeze-thaw cycles (Chen et al., 2004). Another reported mechanism for oligonucleotide encapsulation is the uptake of RNA-coated mineral particles. Following the finding that certain clays stimulate vesicle formation, investigators observed that polyadenosine coated on clay particles was encapsulated as myristoleic acid vesicles formed (Hanczyc et al., 2003).

Whether these laboratory procedures had prebiotic analogs is unclear. In a dynamic environment, changes in pH, temperature, and pressure may have been sufficient to generate cycles of membrane formation and disaggregation (or transition to an oil phase) (Rubio-Sánchez et al., 2021; Mayer et al., 2018). Fatty acids generally revert to micelles above pH 8 (Namani and Walde, 2005) and may form oils under acidic conditions and at elevated temperatures. Restoration of more moderate conditions can result in re-encapsulation of contents, including oligonucleotides (Rubio-Sánchez et al., 2021).

Another mechanism for encapsulation of solutes in fatty acid vesicles is dehydration followed by rehydration (Fig. 2A). In an early experiment, decanoic acid vesicles were mixed with salmon testis DNA that had been sonicated to yield fragments under 1000 bp long (Apel et al., 2002). The mixture was dehydrated and then rehydrated on a microscope slide. Staining with acridine orange showed the DNA concentrated in vesicles. The experiment employed a substantial mass ratio of DNA to decanoic acid (1:2), and the pH and temperature were not specified. A similar result was obtained in a later experiment from this laboratory, which employed lauric acid and its glycerol monoester instead of decanoic acid, and yeast ribosomal RNA as the polymer. The components were dissolved in water collected from a hot spring at pH 3.3, again in a 1:2 mass ratio for the nucleic acid and amphiphile (Milshteyn et al., 2018). A recent article from a different group also found encapsulation of yeast RNA upon dehydration followed by rehydration. These investigators added the RNA to a solution of decanoic acid and glycerol monodecanoate at a nucleic acid:amphiphile ratio of about 1:50, both at neutral and mildly acidic pH. Interestingly, thin-walled, unilamellar vesicles, which resemble modern membranes more closely than do multilamellar vesicles, encapsulated the RNA more efficiently than the highly multilamellar vesicles in the preparation (Steller et al., 2022).

In addition to encapsulation, another way preformed polymers and vesicles could have coalesced as a unit is by binding of the polymers to the membrane. Evidence for the binding of hydrophobic peptides is discussed in Section 4. However, binding would not have provided a mechanism for colocalization of vesicles with diverse polymers.

Overall, encapsulation due to dehydration followed by rehydration is probably the most likely way preformed polymers would have been localized with vesicles, and this conclusion suggests that shallow bodies of water subject to periodic dehydration were where vesicles and polymers coalesced as a unit.

3.2. Polymerization of monomers associated with vesicles

Another mechanism for colocalization with vesicles is the formation of the polymers from monomers associated with the membrane (Fig. 2B). The monomers could have bound to the membrane, which would thereby have concentrated them, and the binding could have been selective, which would thereby help explain why only certain monomers were incorporated into vesicle-associated polymers. The increase in concentration would increase reaction rates, and the reduced water activity would facilitate the dehydration required for the formation of peptide and phosphodiester bonds (Grochmal et al., 2015). The membrane could have provided conformational constraints favorable for polymerization, which is not thermodynamically favorable in an aqueous solution. Condensing agents or cycles of drying and wetting could have further driven membrane-associated polymer formation.

The plausibility of this scenario is supported by the finding that heterocyclic bases, sugars, and amino acids do bind to fatty acid micelles and vesicles and do so selectively (Black et al., 2013; Cornell et al., 2019; Xue et al., 2020). Since these compounds also stabilize the vesicles (Black et al., 2013; Cornell et al., 2019), a population of vesicles would become enriched over time for those that bore polymer-building blocks.

The formation of polymers from membrane-associated monomers has been demonstrated in several systems. Drydown of decanoic acid membranes with serine ethyl ester increased the initial rate of serine–serine formation, compared with drydown with a shorter fatty acid that does not form bilayers (Cohen et al., 2022). Previous work, albeit with nonbiological, highly hydrophobic amino acids, also showed that fatty acid vesicles can induce peptide formation (Adamala and Szostak, 2013; Murillo-Sanchez et al., 2016). These two articles argue that the hydrophobic core of the bilayer could lower the pKa of the amines (favorable for peptide bond formation) and reduce hydrolysis, while the carboxylate headgroups could mediate acid–base catalysis. Finally, vesicles can enable the formation of peptides from amino acids that diffuse in from the environment, under conditions in which the polymerization does not occur in the environment due to dilution (Kwiatkowski et al., 2021).

Regarding oligonucleotides, drydown of mononucleotides with membranes (albeit composed of phospholipids rather than fatty acids) leads to the formation of RNA-like polymers 25–100 nucleotides long (Rajamani et al., 2008). The length increases as the number of dehydration–rehydration cycles increases. Subsequent work showed that, in the dehydrated state, the membrane aligns the mononucleotides favorably for phosphodiester bond formation (Toppozini et al., 2013).

In sum, both encapsulation of environmentally available polymers and formation of polymers from monomers associated with membranes would have been most likely in shallow bodies of water that could periodically dry down.

4. Origin of Function for Peptides: Increased Stability and Growth of Fatty Acid Vesicles

The harder question that remains is how functional polymers were selected, that is, how they were enriched in a population of vesicles relative to the total pool of random sequences. The key to answering this question is the metastable character of fatty acid vesicles: their capacity to grow and shrink and their susceptibility to collapse in the presence of various cations. As a result of these characteristics, vesicles bearing molecules that increase stability or growth would have increased in abundance, while those bearing molecules that decrease stability or growth would have decreased in abundance (Fig. 3).

FIG. 3.

The first function of peptides was to increase the growth and stability of vesicles. In the study depicted, decanoic acid vesicles, without (A) or with (B) an added peptide, were exposed to decanoic acid micelles or a high concentration of Na⁺. The effect on vesicle size or extent of flocculation was then determined as described in Cohen et al. (2024b). Dipeptides increased vesicle diameter by up to 2-fold. The micrographs are reprinted with permission from Cohen et al. (2024b), ©Copyright 2024 American Chemical Society.

In support of this scheme, investigators have demonstrated the evolution of a vesicle-stabilizing peptide (Mayer et al., 2018). They generated random peptides from 12 amino acids by placing the monomers under the pressure and temperature conditions of a deep tectonic fault zone. They included two amphiphiles, octadecylamine and octadecanoic acid, and cycled the conditions to alternately generate supercritical or gaseous CO₂. Vesicles formed as the system transitioned from the supercritical to the gaseous state. After multiple cycles, they found an enrichment for amphiphilic peptides, which presumably had integrated into the vesicle bilayer and were thereby protected from hydrolysis (which occurs in the system in addition to synthesis). One of the amphiphilic peptides was analyzed in detail and found to stabilize the vesicles against leakiness induced by high temperatures. A caveat regarding this work is that neither the amphiphiles nor some of the amino acids employed are considered prebiotic. Nonetheless, these results strongly confirm that vesicles can preferentially accumulate specific peptides that have a stabilizing effect in a fluctuating environment. Another, possibly more relevant fluctuating environment would have been shallow pools subject to cycles of dehydration and rehydration.

Prebiotically plausible peptides that stabilize fatty acid vesicles have also been identified. Fatty acid vesicles flocculate (clump together) in the presence of high NaCl concentrations, which prevents them from growing and dividing. In one study, 3 of 15 dipeptides tested reduced salt-induced flocculation of decanoic acid vesicles (Cohen et al., 2024b). All three of the effective dipeptides contained leucine at the amino terminus, and replacing leucine with even the structurally similar isoleucine eliminated the effect. In a molecular simulation, one of the effective dipeptides, Leucine–Leucine, stably docked to the membrane via insertion of the two hydrophobic side chains into the core of the bilayer, while an ineffective dipeptide, Alanine–Alanine, did not dock to the membrane. This finding implies that beneficial, leucine-containing peptides could accumulate on the vesicles, while other peptides would not.

Stabilization of a fatty acid bilayer by amyloid-forming peptides has also been demonstrated (Bomba et al., 2018). (Valine-ornithine)₄-NH₂ forms structured aggregates with decanoic acid, in which the carboxylates of the fatty acid bilayer interact with amines from the peptide. This interaction increases the thermal stability of the bilayer. The bilayer, in turn, enables the aggregation of the peptide and thus stabilizes it against a proteolytic enzyme.

Coacervates composed, in part, of a charged peptide may provide a further example of vesicle stabilization that involves peptides. Fatty acid bilayers can form on a coacervate base, and the resulting interaction can influence the membrane’s properties. Thus far, it has been shown that with a coacervate composed of a synthetic polyamine and adenosine diphosphate, a supported oleic acid bilayer is relatively resistant to disruption by Mg²⁺ (Lee et al., 2024).

Peptides that have a negative effect on stability have also been identified (Cohen et al., 2024b). Both of the heterochiral forms of Leucine–Leucine (L-D and D-L) increased vesicle leakiness and, at high pH, reduced vesicle formation relative to the two homochiral forms (both residues L or both D). Thus, over time, a population of vesicles would be expected to become enriched for homochiral versus heterochiral Leucine–Leucine.

In addition to having differential effects on vesicle stability, peptides can also differentially affect vesicle growth. In a recent study with decanoic acid vesicles, 3 dipeptides out of 15 tested significantly increased the rate of growth induced by the addition of micelles (Todd et al., 2022). As seen in the stabilization study described above, all of the effective dipeptides contained leucine at the amino terminus (the set tested was different from the set used in the stability study, but two of the three growth-promoting dipeptides were effective stabilizers as well). Subsequent work found that a leucine-containing dipeptide can increase the final radius of the vesicles, as well as initial growth rate, when multiple rounds of micelles are added (Cohen et al., 2024b). Since larger vesicles would be more prone to division, a population of vesicles would over time become enriched for those with growth-promoting peptides.

The frequency of vesicles that contain a growth-promoting peptide would increase for an additional reason: they can actually cause the shrinkage of other vesicles, which could lead to their dissolution. Oleic acid vesicles that bear the hydrophobic dipeptide Ac-Phe-Leu-NH₂ increase in diameter when mixed with vesicles that lack the dipeptide, while the diameters of the latter decrease (Adamala and Szostak, 2013). Apparently, the peptide integrates into the membrane and increases the rate of free fatty acid incorporation (recall that fatty acid vesicles exist in equilibrium with free molecules, as depicted in Fig. 1).

In addition to effects on the stability and growth of membranes, peptides may also alter membrane permeability. As noted above, heterochiral Leucine–Leucine increases the leakiness of decanoic acid vesicles, relative to homochiral Leucine–Leucine. In this case, the increased permeability was probably nonspecific, since a dye (calcein) was released (Cohen et al., 2024b). Longer peptides, however, could form channels with specificity that improve vesicle viability (Pohorille et al., 2003).

While all the beneficial peptides described above contained hydrophobic amino acids, consistent with interaction with the core of the lipid bilayer, stabilizing or growth-promoting peptides can also contain charged residues. The effective peptide in the CO₂ system contained a lysine, and in the salt-induced flocculation system, preliminary results show that Leucine–Arginine reduces flocculation at least as well as Leucine–Leucine. More speculatively, peptides that contain negatively charged amino acids could also interact with a fatty acid membrane, since such peptides do complex with another negative structure, RNA, in the presence of cations (Giacobelli et al., 2022). The significance of peptides that can interact with both the membrane and oligonucleotides is discussed in the following section.

In sum, multiple observations suggest that the first functions of peptides may have been to increase the stability and growth of fatty acid vesicles. Again, these effects would have led to compositional evolution of a vesicle population, as vesicles harboring beneficial peptides increased in abundance relative to other vesicles. The vesicles in turn would have impacted the repertoire of peptides by stabilizing those that associated with the membrane against hydrolysis.

While any beneficial peptide would probably have to be present in multiple copies to have a significant impact, a collection of similar single-copy peptides could have had an additive effect. Still, in the long run, vesicles with multiple copies of a given peptide must have emerged. This transition could have occurred in two ways: (1) some peptides could have accumulated on vesicles due to binding the membrane, as suggested above for Leucine–Leucine, and (2) oligonucleotides could have increased the retention of specific peptides and ultimately could have templated their synthesis—as I will discuss in the next section.

5. Origin of Function for Oligonucleotides: Increasing Peptide Concentration in Vesicles

Unlike peptides, oligonucleotides have not been reported to stabilize fatty acid vesicles. Nucleic acid can increase vesicle growth by increasing osmotic pressure on the membrane and thereby facilitating the incorporation of additional fatty acid molecules (Chen et al., 2004), but this effect is not sequence dependent. So how could specific oligonucleotides contribute to the stability or growth of vesicles indirectly?

The key to explaining the first function of oligonucleotides is their ability to increase the peptide content of vesicles, because some of the resulting peptides would have increased fatty acid vesicle stability or growth as described in Section 4.

I will review work showing that oligonucleotides could increase the peptide content of vesicles in four ways (Figs. 4 and 5): (1) stabilizing peptides against hydrolysis, (2) impeding diffusion of peptides out of vesicles, (3) binding amino acids and activating them for the formation of peptide bonds, and—most significantly—(4) aligning activated amino acids for the formation of specific peptides.

FIG. 4.

The first function of oligonucleotides was increasing the peptide content of vesicles. By complexing with a peptide, an oligonucleotide could stabilize it and prevent its diffusion out of the vesicle. Oligonucleotides can also increase the synthesis of peptides, as described in the text and Figure 5.

FIG. 5.

The complementary strand to a peptide-stabilizing oligonucleotide could direct formation of a copy of the peptide due to base-pairing with unjoined amino acid-bearing oligos. See text for explanation of acylation of oligonucleotides by amino acids.

Mutual stabilization of protopeptides and oligonucleotides has been demonstrated. Depsipeptides (which include both ester and amide bonds) increase the melting temperature of oligonucleotides, and the oligonucleotides reduce hydrolysis of the depsipeptides (Frenkel-Pinter et al., 2020b). Depsipeptides containing amino acids found in modern proteins were more effective stabilizers than those containing nonproteinaceous amino acids, which indicates the selective potential of such stabilizing interactions. The frequency of peptides and oligonucleotides that formed mutually stabilizing complexes would have increased over multiple rounds of wetting and drying. Again, if the peptide in such a complex stabilized or increased the growth of vesicles, the vesicle population would become enriched for that oligonucleotide–peptide pair. And if the membrane of the vesicle stabilized the oligonucleotide–peptide pair, there would have been a positive feedback loop.

Some peptides readily diffuse through a fatty acid membrane (Kwiatkowski et al., 2021), whereas oligonucleotides generally do not (Mansy et al, 2008). So complexing of a diffusible peptide with an oligonucleotide would reduce loss of the peptide to the environment. In support of the plausibility of this phenomenon, Arg-Arg-Arg-Phe-Phe-Phe added to the outside of vesicles penetrated the membrane and stably associated with internal RNA (Kamat et al., 2015). Thus, peptides with beneficial effects on the membrane due to hydrophobic interactions could be retained in vesicles via interaction of their positively charged residues with an oligonucleotide. This interaction could in turn have stabilized the oligonucleotide.

Turning to effects of oligonucleotides on peptide formation: The ability of even unjoined nucleotides to increase peptide synthesis, by esterification of amino acids, was demonstrated by Orgel’s group over 40 years ago (Weber and Orgel, 1978; Weber and Orgel, 1979). They synthesized the 2′(3′)-glycyl ester of adenosine-5′(O-methylphosphate) and mixed it with one of several amino acids. Some dipeptides, with amino-terminal glycine, formed in all cases, most notably with the addition of serine: 36% of the serine was incorporated into Glycine–Serine. Without an added amino acid, Gly-Gly formed, and an oligonucleotide, poly(U), increased the rate of Gly-Gly formation 7-fold. Presumably, the poly(U) acted as a template that aligned the acylated adenosine-5′(O-methylphosphate) (Weber and Orgel, 1980). The rate of peptide formation was still low relative to hydrolysis of the ester, and most of the dipeptides were cyclized rather than linear. Nonetheless, these findings are significant because the prebiotic environment could have been more conducive to peptide formation and elongation. For example, a higher pH, as would have been likely in soda lakes, would have favored nucleophilic attack by amines (Rodriguez-Garcia et al., 2015).

The glycine-acylated nucleotide used in these studies was synthesized by nonprebiotic methods, but more recent work shows that acylation can occur by mixing a nucleotide with an amino acid in the presence of the prebiotic compounds methyl isonitrile and dicyanoimidazole (Liu et al., 2020). A caveat regarding this work is that it required a high concentration of dicyanoimidazole (100 mM) and relatively low pH (4–5). Dehydration in the prebiotic environment could have contributed to acylation.

In another system in which nucleotides stimulated the production of peptides, the investigators used a condensing agent–catalyst combination that worked at neutral pH and 0°C. Under these conditions, a mixture of adenosine and uridine monophosphates with glycine and phenylalanine yielded peptides in phosphoramidate linkage to the 5′ end of oligonucleotides that also formed during the incubation (Jauker et al., 2015). The condensing agent, ethyl(dimethylaminopropyl)carbodiimide, and catalyst, 1-ethylimidazole, have not been suggested to be prebiotic and were used at high concentrations, but some product was also obtained using cyanamide as the condensing agent.

Other investigators have started with oligonucleotides. Yarus and Illangasekare showed that a 29-nucleotide-long RNA can self-aminoacylate at the 3′ end when provided with phenylalanine activated by a phosphoester linkage to adenosine monophosphate (Illangasekare and Yarus, 1999). The acylated product promotes extension to a dipeptide due to the increased nucleophilicity of the amine, which reacts with the ester of another acylated RNA. It was subsequently found that simply a 5-nucleotide-long RNA can similarly aminoacylate the 3′ end of a complementary oligonucleotide, leading again to the formation of peptides via reaction of the product with a remaining phosphoester-activated amino acid (Turk et al., 2011).

Carell and coworkers have established a more complex system based on two complementary RNA strands, one with a modified nucleoside found in tRNA at the 5′ end and the other with a different nucleoside specific to tRNA at the 3′ end (Müller et al., 2022; Singer et al., 2023). The duplex RNA effects the peptide bond formation between amino acids attached to these ends. The amino acid at the 5′ end is attached via a urea linkage under somewhat elaborate but possibly prebiotic conditions, and its carboxyl group is activated by plausibly prebiotic reagents. The amine of the 3′-bound amino acid then attacks the activated carboxyl group; this results in a peptide bond that links the two oligos via a hairpin loop. Finally, the urea linkage is cleaved at an elevated temperature and low pH, which leaves the peptide attached to the 3′ end of the other oligonucleotide.

All the systems described thus far would probably generate only randomly ordered peptides.

However, an oligonucleotide could have at least preferentially increased the level of beneficial peptides by preferentially activating leucine. Oligonucleotides as short as 20 nucleotides can noncovalently bind specific amino acids (Yarus, 2017). This binding could position the amino acid for acylation of a 3′ hydroxyl (in conjunction with, e.g., methyl isonitrile and dicyanoimidazole as noted above). The amine of a free amino acid or of a peptide could then attack the ester to yield a new peptide with the previously oligonucleotide-bound amino acid at the amino terminus. An oligonucleotide that preferentially bound and then was acylated by leucine would benefit a vesicle because several beneficial peptides have an amino-terminal leucine, as described in Section 4. Again, Orgel found that serine was by far the most effective amino acid at attacking an acylated nucleotide (Weber and Orgel, 1979). If the nucleotide was acylated by leucine, the predominant dipeptide produced would be Leucine–Serine, one of the peptides that reduces the salt-induced flocculation of decanoic acid vesicles.

Still, there is no apparent way any of these oligonucleotide-based mechanisms for increasing peptide synthesis would yield peptides with specific sequences.

The key to the generation of specific peptides is alignment by a template. How, then, could oligonucleotides have aligned activated amino acids? A possible clue is the observation in (1) above that some oligonucleotides would have been enriched in vesicles due to their stabilizing interactions with beneficial peptides. Some of these interactions would have been between the oligonucleotide and the amino acid sidechains of the peptide, as occurs in RNA–protein binding today (Corley et al., 2020) and in the binding between amino acids and aptamers (single-stranded RNAs) that Yarus identified. Critically, Yarus also found aptamers that bind a specific dipeptide (Turk-Macleod et al., 2012). This finding suggests that an oligonucleotide can bring two amino acids sufficiently close together to form a peptide bond.

Further evidence in support of this concept was provided by a study of aptamers that bind to a peptide from the HIV Tat protein (Harada et al., 2014). The investigators found one aptamer that binds to two adjacent arginine residues in the peptide via interaction with the arginine side chains. Structural analysis indicated two distinct arginine-binding sites in the aptamer, and tests with variant aptamers showed that close proximity of the two sites was required for significant binding. Most telling in support of Yarus’s hypothesis, the rate of activated arginine addition to a peptide ending in arginine was increased when the peptide was complexed with the aptamer.

Templating of longer peptides could have been effected by longer oligonucleotides that contained multiple amino acid-binding sites, as would be present in oligonucleotides that bound to and stabilized beneficial peptides in vesicles.

An intriguing variation of this scheme could have led to a primitive version of the modern translation mechanism, along the lines described by Ma (2010). In this variation, adaptor oligonucleotides––self-acylated as discussed above––would have base-paired with the complementary strand to the stabilizing oligonucleotide (Fig. 5). The basis for this idea is Yarus’s finding that amino acid-binding aptamers are enriched in anticodons for the amino acid. A peptide-stabilizing oligonucleotide would therefore be a string of anticodons (albeit with intervening sequences), and its complementary strand would be a string of codons. The complementary strand could thus serve as a template for the alignment of unjoined, acylated oligonucleotides.

From start to finish, this scheme would work roughly as follows (see Fig. 5 for an illustration): 1.

Vesicles would have been enriched for oligonucleotides that bound to and stabilized beneficial peptides.

The nucleotide sequences that bound the amino acid sidechains of the peptide would likely resemble the sequences that bind the corresponding unjoined amino acids (which again can be as short as 15–20 nucleotides).

According to Yarus’s findings, these binding sites would have been enriched in anticodons for the bound amino acid.

During transient dissociation of the stabilizing oligonucleotide from the peptide, the complementary strand could form (Livio and Szostak, 2024), and it would contain codons for the amino acids.

Anticodons in free amino acid-binding oligonucleotides would base-pair with the codons in the complementary strand. This could occur as the bound amino acid is transferred to the 3′ end of the binding oligonucleotide.

Peptides would then form from the aligned, activated amino acids. As noted above, Orgel showed that adding poly-uridine to a solution of acylated adenosine nucleotide increases peptide formation.

This scheme is essentially the same as Ma’s elaboration of Yarus’s “direct RNA template” hypothesis (Yarus, 2017), with the important addition that the first indirect templates were complementary to oligonucleotides that stabilized beneficial peptides. Unlike Yarus or Ma, I therefore provide a rationale for why oligonucleotides that served as templates for the synthesis of beneficial peptides would have been enriched in protocells.

To be sure, this scheme raises several questions. For example, why would the adaptor oligonucleotides (those that bound unjoined amino acids) have been present in vesicles? One answer is that they would have helped retain the amino acids they bound, which can themselves be beneficial for vesicles (Cornell et al., 2019). And, as Ma notes, there would have been two ready sources for these oligonucleotides: partial hydrolysis of peptide-binding oligonucleotides and incomplete replication of these oligonucleotides. How would the anticodon-containing, adaptor oligonucleotides both bind the amino acid and base-pair with the codon-containing oligonucleotide template? The structure of the adaptor oligonucleotides would have to open as the amino acid was transferred to the 3′ end. This shift could be induced by the formation of a phospho-ester of the amino acid, which would activate it for acylation of the oligonucleotide (Liu et al., 2020). As noted above, self-acylation is plausible, since a 29-nucleotide-long RNA has been shown to self-acylate, and even a 5-nucleotide-long RNA can acylate the 3′ end of another oligonucleotide. How could the nucleotide polymerization required for the formation of the complementary strand occur? This is a problem for any origin of life scheme and is the subject of ongoing research (Livio and Szostak, 2024). A possible explanation is that a peptide that accumulated in protocells due to its stabilizing interaction with the membrane also increased nucleotide polymerization.

In sum, stabilization of beneficial peptides and increasing their synthesis were probably the original functions of oligonucleotides. The key point is that oligonucleotides would have been selected based on their ability to increase the concentration of peptides favorable for vesicle stability or growth.

6. Conclusion

Understanding the origin of cells requires explaining how their three signature components coalesced as a functional unit. I divide the task of explaining the origin of these units into two questions: How did membranes, peptides, and oligonucleotides come together, and how were polymers with a function benefiting the unit selected from random sequences?

I conclude that the coalescence of the three components can be explained by cycles of dehydration and rehydration in shallow bodies of fresh water. This is where fatty acid vesicles, peptides, and oligonucleotides would have been most likely to form, due to the concentrating effect of dehydration, and where the process of dehydration–rehydration would have driven encapsulation of the polymers by the vesicles. Such locations would also have been favorable because of the relatively low concentration of vesicle-disrupting divalent cations.

The key to understanding the origin of polymer selection and function is mutual stabilization of the three components. Specific peptides can increase the stability or growth of fatty acid vesicles. Oligonucleotides, in turn, can stabilize peptides against hydrolysis, increase their retention in the vesicles, and increase their de novo synthesis. The frequency of vesicles that bore such peptides and oligonucleotides would increase over time due to the dynamic properties of fatty acid membranes. In addition, peptides that were stabilized by the membranes would be favored. As noted throughout this perspective, cycles of drying and wetting would have driven the selection for mutually stabilizing interactions. Others have noted the importance of mutual stabilization as a driver of prereplication chemical evolution (Edri et al., 2023), and the concept that function arose from the interaction of diverse components (Ruiz-Mirazo and Moreno, 2024).

The emergence of oligonucleotide-templated peptide synthesis, possibly by the mechanism that I describe, would have initiated Darwinian evolution. This transition would have led to the diversity of peptides required for metabolism and energy harvesting.

The scheme presented in this review entails three major issues that require further research. First, we must demonstrate that a population of fatty acid vesicles does actually become enriched over time for those bearing stabilizing and growth-promoting peptides and oligonucleotides. Second, further work is required to determine whether oligonucleotides that bind amino acids can both self-acylate and base-pair with a primitive messenger RNA as proposed by Ma. Third, a fundamental outstanding question is how oligonucleotide replication arose, since it would have been required for sustained templated peptide synthesis; any given oligonucleotide template would have been hydrolyzed over time despite stabilizing interactions with a peptide. Nonenzymatic template-directed primer extension has been demonstrated by Szostak’s group, but it is not robust (Livio and Szostak, 2024). Perhaps peptides or oligonucleotides enriched in vesicles due to their stabilizing or growth-promoting effects also increased the rate of nucleotide polymerization.

Footnotes

Acknowledgments

This review was stimulated in part by 4 years of conversation with the late Frank Harold, who unrelentingly focused on the need to explain the origin of biological organization and function. The author also acknowledges Sarah Keller for providing office space and the entire Keller group for discussions on membrane dynamics and for advice on the figures. He thanks John Doedens and Ann Hobson for comments on the article.

Author’s Contributions

R.A.B. conceptualized and wrote this review.

Author Disclosure Statement

The author declares no competing financial interest.

Funding Information

The author received no funding to support this review.

Associate Editor: Lewis Dartnell

1

Polysaccharides, too, are a signature of extant cells, but are not generally considered likely in the prebiotic environment (Sutton et al., ).

References

Adamala

, Szostak

. Competition between model protocells driven by an encapsulated catalyst. Nat. Chem. 2013;5(6):495–501; doi: 10.1038/nchem.1650

Apel

, Deamer

, Mautner

. Self-assembled vesicles of monocarboxylic acids and alcohols: Conditions for stability and for the encapsulation of biopolymers. Biochim Biophys Acta 2002;1559(1):1–9; doi: 10.1016/S0005-2736(01)00400-X

Berclaz

, Muller

, Walde

, et al. Growth and transformation of vesicles studied by ferritin labeling and cryotransmission electron microscopy. J Phys Chem B 2001;105(5):1056–1064; doi: 10.1021/jp001298i

Black

, Blosser

, Stottrup

, et al. Nucleobases bind to and stabilize aggregates of a prebiotic amphiphile, providing a viable mechanism for the emergence of protocells. Proc Natl Acad Sci U S A 2013;110(33):13272–13276; doi: 10.1073/pnas.1300963110

Bomba

, Kwiatkowski

, Sanchez-Ferrer

, et al. Cooperative induction of ordered peptide and fatty acid aggregates. Biophys J 2018;115(12):2336–2347; doi: 10.1016/j.bpj.2018.10.031

Bonfio

, Sutherland

, Russell

, et al. Activation chemistry drives the emergence of functionalised protocells. Chem Sci 2020;11(39):10688–10697; doi: 10.1039/D0SC04506C

Chen

, Roberts

, Szostak

. The emergence of competition between model protocells. Science 2004;305(5689):1474–1476; doi: 10.1126/science.1100757

Chen

, Salehi-Ashtiani

, Szostak

. RNA catalysis in model protocell vesicles. J Am Chem Soc 2005;127(38):13213–13219; doi: 10.1021/ja051784p

Chen

, Szostak

. A kinetic study of the growth of fatty acid vesicles. Biophys J 2004;87(2):988–998; doi: 10.1529/biophysj.104.039875

10.

Cohen

, Kessenich

, Hazra

, et al. Prebiotic membranes and micelles do not inhibit peptide formation during dehydration. Chembiochem 2022;23(3):e2021006; doi: 10.1002/cbic.202100614

11.

Cohen

, Todd

, Ding

, et al. Natural soda lakes provide compatible conditions for RNA and membrane function that could have enabled the origin of life. PNAS Nexus 2024a;3(3):1–9; doi: 10.1093/pnasnexus/pgae084

12.

Cohen

, Todd

, Maibaum

, et al. Stabilization of prebiotic vesicles by peptides depends on sequence and chirality: A mechanism for selection of protocell-associated peptides. Langmuir 2024b;40(17):8971–8980; doi: 10.1021/acs.langmuir.4c00150

13.

Cohen

, Todd

, Wogan

, et al. Plausible sources of membrane-forming fatty acids on the early earth: A review of the literature and an estimation of amounts. ACS Earth Space Chem 2023;7(1):11–27; doi: 10.1021/acsearthspacechem.2c00168

14.

Corley

, Burns

, Yeo

. How RNA binding proteins interact with RNA: Molecules and mechanisms. Mol Cell 2020;78(1):9–29; doi: 10.1016/j.molcel.2020.03.011

15.

Cornell

, Black

, Xue

, et al. Prebiotic amino acids bind to and stabilize prebiotic fatty acid membranes. Proc Natl Acad Sci USA 2019;116(35):17239–17244; doi: 10.1073/pnas.1900275116

16.

Deamer

, Dworkin

, Sandford

, et al. The first cell membranes. Astrobiology 2002;2(4):371–381; doi: 10.1089/153110702762470482

17.

Edri

, Fisher

, Menor-Salvan

, et al. Assembly-driven protection from hydrolysis as key selective force during chemical evolution. FEBS Lett 2023;597(23):2879–2896; doi: 10.1002/1873-3468.14766

18.

Frenkel-Pinter

, Samanta

, Ashkenasy

, et al. Prebiotic peptides: Molecular hubs in the origin of life. Chem Rev 2020a;120(11):4707–4765; doi: 10.1021/acs.chemrev.9b00664

19.

Frenkel-Pinter

, Petrov

, Krishnamurthy

, et al. Mutually stabilizing interactions between proto-peptides and RNA. Nat Commun 2020b;11(1):3137; doi: 10.1038/s41467-020-16891-5

20.

Giacobelli

, Fujishima

, Lepsık

, et al. In vitro evolution reveals noncationic protein–RNA interaction mediated by metal ions. Mol Biol Evol 2022;39(3):msac032; doi: 10.1093/molbev/msac032

21.

Gibard

, Bhowmik

, Karki

, et al. Phosphorylation, oligomerization and self-assembly in water under potential prebiotic conditions. Nat Chem 2018;10(2):212–217; doi: 10.1038/nchem.2878

22.

Graham

. Molecular Storms. Springer: Cham, Switzerland; 2023.

23.

Grochmal

, Prout

, Makin-Taylor

, et al. Modulation of reactivity in the cavity of liposomes promotes the formation of peptide bonds. J Am Chem Soc 2015;137(38):12269–12275; doi: 10.1021/jacs.5b06207

24.

Hanczyc

, Fujikawa

, Szostak

. Experimental models of primitive cellular compartments: Encapsulation, growth, and division. Science 2003;302(5645):618–622; doi: 10.1126/science.1089904

25.

Harada

, Aoyama

, Matsugami

, et al. RNA-directed amino acid coupling as a model reaction for primitive coded translation. Chembiochem 2014;15(6):794–798; doi: 10.1002/cbic.201400029

26.

Illangasekare

, Yarus

A tiny RNA that catalyzes both aminoacyl-RNA and peptidyl-RNA synthesis. RNA 1999;5(11):1482–1489; doi: 10.1017/s1355838299991264

27.

Jauker

, Griesser

, Richert

. Spontaneous formation of RNA strands, peptidyl RNA, and cofactors. Angew Chem Int Ed Engl 2015;54(48):14564–14569; doi: 10.1002/anie.201506593

28.

Kahana

, Segev

, Lancet

. Attractor dynamics drives self-reproduction in protobiological catalytic networks. Cell Rep Phys Sci 2023;4(5):101384; doi: 10.1016/j.xcrp.2023.101384

29.

Kamat

, Tobe

, Hill

, et al. Electrostatic localization of RNA to protocell membranes by cationic hydrophobic peptides. Angew Chem Int Ed Engl 2015;54(40):11735–11739; doi: 10.1002/anie.201505742

30.

Keating

, Martin

, Santore

. Editorial overview: Coacervates and membraneless organelles. Curr Opin Colloid Interface Sci 2021;56:101527; doi: 10.1016/j.cocis.2021.101527

31.

Kwiatkowski

, Bomba

, Afanasyev

, et al. Prebiotic peptide synthesis and spontaneous amyloid formation inside a proto-cellular compartment. Angew Chem Int Ed Engl 2021;60(10):5561–5568; doi: 10.1002/anie.202015352

32.

Lawless

, Yuen

. Quantification of monocarboxylic acids in the Murchison carbonaceous meteorite. Nature 1979;282(5737):396–398.

33.

Lee

, Cakmak

, Booth

, et al. Hybrid protocells based on coacervate-templated fatty acid vesicles combine improved membrane stability with functional interior protocytoplasm. Small 2024;20(52):e2406671; doi: 10.1002/smll.202406671

34.

Liu

, Wu

L-F

, Xu

, et al. Harnessing chemical energy for the activation and joining of prebiotic building blocks. Nat Chem 2020;12(11):1023–1028; doi: 10.1038/s41557-020-00564-3

35.

Livio

, Szostak

. Is Earth Exceptional? Basic Books: New York; 2024.

36.

Lowe

, Kaushik

, Wang

. Natural size variation amongst protocells leads to survival and growth under hypoosmotic conditions. Small 2025;21(3):e2406241; doi: 10.1002/smll.202406241

37.

. The scenario on the origin of translation in the RNA world: In principle of replication parsimony. Biol Direct 2010;5:65; doi: 10.1186/1745-6150-5-65

38.

Mansy

, Schrum

, Krishnamurthy

, et al. Template-directed synthesis of a genetic polymer in a model protocell. Nature 2008;454(7200):122–125; doi: 10.1038/nature07018

39.

Maurer

, Monnard

P-A

. Primitive membrane formation, characteristics and roles in the emergent properties of a protocell. Entropy 2011;13(2):466–484; doi: 10.3390/e13020466

40.

Mayer

, Schreiber

, Davila

, et al. Molecular evolution in a peptide-vesicle system. Life (Basel) 2018;8(2):16; doi: 10.3390/life8020016

41.

Milshteyn

, Damer

, Havig

, et al. Amphiphilic compounds assemble into membranous vesicles in hydrothermal hot spring water but not in seawater. Life (Basel) 2018;8(2):11–25; doi: 10.3390/life8020011

42.

Monnard

, Apel

, Kanavarioti

, et al. Influence of ionic inorganic solutes on self-assembly and polymerization processes related to early forms of life: Implications for a prebiotic aqueous medium. Astrobiology 2002;2(2):139–152; doi: 10.1089/15311070260192237

43.

Morigaki

, Walde

. Fatty acid vesicles. Curr Opin Colloid Interface Sci 2007;12(2):75–80; doi: 10.1016/j.cocis.2007.05.005

44.

Müller

, Escobar

, Xu

, et al. A prebiotically plausible scenario of an RNA–peptide world. Nature 2022;605(7909):279–284; doi: 10.1038/s41586-022-04676-3

45.

Murillo-Sanchez

, Beaufils

, Gonzalez Manas

, et al. Fatty acids’ double role in the prebiotic formation of a hydrophobic dipeptide. Chem Sci 2016;7(5):3406–3413; doi: 10.1039/C5SC04796J

46.

Namani

, Walde

. From decanoate micelles to decanoic acid/dodecylbenzenesulfonate vesicles. Langmuir 2005;21(14):6210–6219; doi: 10.1021/la047028z

47.

Naraoka

, Shimoyama

, Harada

. Molecular distribution of monocarboxylic acids in Asuka carbonaceous chondrites from Antarctica. Orig Life Evol Biosph 1999;29(2):187–201; doi: 10.1023/a:1006547127028

48.

Orgel

. The origin of life—a review of facts and speculations. Trends Biochem Sci 1998;23(12):491–495; doi: 10.1016/S0968-0004(98)01300-0

49.

Pohorille

, Wilson

, Chipot

. Membrane peptides and their role in protobiological evolution. Orig Life Evol Biosph 2003;33(2):173–197; doi: 10.1023/a:1024627726231

50.

Powner

, Gerland

, Sutherland

. Synthesis of activated pyrimidine ribonucleotides in prebiotically plausible conditions. Nature 2009;459(7244):239–242; doi: 10.1038/nature08013

51.

Rajamani

, Vlassov

, Benner

, et al. Lipid-assisted synthesis of RNA-like polymers from mononucleotides. Orig Life Evol Biosph 2008;38(1):57–74; doi: 10.1007/s11084-007-9113-2

52.

Rodriguez-Garcia

, Surman

, Cooper

GJT

, et al. Formation of oligopeptides in high yield under simple programmable conditions. Nat Commun 2015;6:8385–8392; doi: 10.1038/ncomms9385

53.

Rubio-Sánchez

, O’Flaherty

, Wang

, et al. Thermally driven membrane phase transitions enable content reshuffling in primitive cells. J Am Chem Soc 2021;143(40):16589–16598; doi: 10.1021/jacs.1c06595

54.

Ruiz-Mirazo

, Moreno

. On the evolutionary development of biological organization from complex prebiotic chemistry. In Organization in Biology. ( Mossio

., ed.) Springer: Cham, Switzerland, 2024; pp 187–218.

55.

Singer

, Müller

, Węgrzyn

, et al. Loading of amino acids onto RNA in a putative RNA-peptide world. Angew Chem Int Ed Engl 2023;62(21):e202302360; doi: 10.1002/anie.202302360

56.

Steller

, Van Kranendonk

, Wang

. Dehydration enhances prebiotic lipid remodeling and vesicle formation in acidic environments. ACS Cent Sci 2022;8(1):132–139; doi: 10.1021/acscentsci.1c01365

57.

Sutton

, Pulletikurti

, Lin

, et al. Abiotic aldol reactions of formaldehyde with ketoses and aldoses—implications for the prebiotic synthesis of sugars by the formose reaction. Chem 2025;11(11):102553; doi: 10.1016/j.chempr.2025.102553

58.

Tanford

. The hydrophobic effect and the organization of living matter. Science 1978;200(4345):1012–1018; doi: 10.1126/science.653353

59.

Todd

, Cohen

, Catling

, et al. Growth of prebiotically plausible fatty acid vesicles proceeds in the presence of prebiotic amino acids, dipeptides, sugars, and nucleic acid components. Langmuir 2022;38(49):15106–15112; doi: 10.1021/acs.langmuir.2c02118

60.

Toppozini

, Dies

, Deamer

, et al. Adenosine monophosphate forms ordered arrays in multilamellar lipid matrices: Insights into assembly of nucleic acid for primitive life. PLoS One 2013;8(5):e62810; doi: 10.1371/journal.pone.0062810

61.

Turk

, Illangasekare

, Yarus

. Catalyzed and spontaneous reactions on ribozyme ribose. J Am Chem Soc 2011;133(15):6044–6050; doi: 10.1021/ja200275h

62.

Turk-Macleod

, Puthenvedu

, Majerfeld

, et al. The plausibility of RNA-templated peptides: Simultaneous RNA affinity for adjacent peptide side chains. J Mol Evol 2012;74(3–4):217–225; doi: 10.1007/s00239-012-9501-8

63.

Walde

, Wick

, Fresta

, et al. Autopoietic self-reproduction of fatty acid vesicles. J Am Chem Soc 1994;116(26):11649–11654; doi: 10.1021/ja00105a004

64.

Weber

, Orgel

. Poly(U)-directed peptide-bond formation from the 2′(3′)-glycyl esters of adenosine derivatives. J Mol Evol 1980;16(1):1–10.

65.

Weber

, Orgel

. The formation of dipeptides from amino acids and the 2′(3′)-glycyl ester of an adenylate. J Mol Evol 1979;13(3):185–192.

66.

Weber

, Orgel

. The formation of peptides from the 2′(3′)-glycyl ester of a nucleotide. J Mol Evol 1978;11(3):189–198.

67.

Woolf

. A hypothesis about the origin of biology. Orig Life Evol Biosph 2015;45(1–2):257–274; doi: 10.1007/s11084-015-9426-5

68.

Xue

, Black

, Cornell

, et al. A step toward molecular evolution of RNA: Ribose binds to prebiotic fatty acid membranes, and nucleosides bind better than individual bases do. Chembiochem 2020;21(19):2764–2767; doi: 10.1002/cbic.202000260

69.

Yarus

. The genetic code and RNA-amino acid affinities. Life (Basel) 2017;7(2):13; doi: 10.3390/life7020013

70.

Yuen

, Kvenvolden

. Monocarboxylic acids in Murray and Murchison carbonaceous meteorites. Nature 1973;246(5431):301–303; doi: 10.1038/246301a0

71.

Zhu

, Szostak

. Coupled growth and division of model protocell membranes. J Am Chem Soc 2009;131(15):5705–5713; doi: 10.1021/ja900919c