Significance statement
While the formation of membranes is fundamental to cellular organization, it is not an evolutionary process in the traditional sense but a result of the physicochemical properties of molecules and their interactions. Evolutionary processes subsequently shaped the diversity and complexity of life, including the structures and functions of cellular membranes. Due to the chemical complexity of modern phospholipids, fatty acids have been proposed as the fundamental building blocks of simple membranous vesicles and protocells because they can be abiotically synthesized through basic Earth chemistry. We show that simple fatty acid membranes do not form in the absence of salt and suggest seawater tide pools and hydrothermal springs as the most promising prebiotic environments for the formation of basic membranous structures.
Introduction
Primitive cells (so-called protocells) played a crucial role in the transition from the prebiotic chemical world to the biological world (Schrum et al., Reference Schrum, Zhu and Szostak2010). This transition likely involved the encapsulation of genetic material (Deamer and Barchfeld, Reference Deamer and Barchfeld1982) and other molecules present in the environment. Warm little ponds (WLPs) have been proposed as one of the earliest environments for the origin of life, as speculated by Charles Darwin regarding the emergence of proteins (Follmann and Brownson, Reference Follmann and Brownson2009). WLPs undergo cycles of dehydration and rehydration, which promote the polymerization of RNA, making them a plausible environment for the emergence of life (Follmann and Brownson, Reference Follmann and Brownson2009; Da Silva et al., Reference Da Silva, Maurel and Deamer2015). They are also an ideal environment for salt-induced peptide formation, which is a proposed prebiotic pathway for the formation of biologically relevant oligomers (Rode, Reference Rode1999). It is hypothesized that multilamellar membrane structures could have formed within these WLPs through wet–dry cycles, with material encapsulated between the layers (Deamer and Barchfeld, Reference Deamer and Barchfeld1982). Vesicular encapsulation of RNA, which can act as an enzymatic-like catalyst and carry information capable of replication, could provide all the necessary components for the emergence of early protocells (Deamer and Barchfeld, Reference Deamer and Barchfeld1982; Szostak et al., Reference Szostak, Bartel and Luisi2001).
The importance of membranous structures and protocells is described in the ‘Lipid world’ hypothesis proposed by Segre et al., who suggested that spontaneous self-assembly of amphiphiles into supramolecular structures is a crucial component of the origins of life due to improved reaction rates, selective properties and potential for simple homeostatic systems (Segrè et al., Reference Segrè, Ben-Eli, Deamer and Lancet2001). While modern cell membranes are typically composed of phospholipids, they are not suitable candidates for components of the initial membrane boundary structures due to their complexity and the need for advanced biochemistry for fabrication. Simple amphiphilic molecules are more likely candidates for the initial cell membranes (Deamer and Dworkin, Reference Deamer and Dworkin2005).
Biological membranes are essential components of cells, separating the interior of cells from their external environment and compartmentalizing different cellular functions. The basic structure of membranes involves the organization of lipids, proteins and other molecules. This arrangement is driven by the hydrophobic and hydrophilic properties of these molecules. The formation of biological membranes is not typically considered an evolutionary process in the same sense as the evolution of species over generations. Instead, membrane formation is a fundamental aspect of cellular biology and is more closely related to the self-organization of molecules driven by physicochemical forces, such as surface tension (Hatta, Reference Hatta2011), entropy of aggregation (Budin et al., Reference Budin, Debnath and Szostak2012) and thermodynamic properties (Lombard, Reference Lombard2014).
While the evolution of life involves changes in species over time through natural selection, the basic principles governing membrane formation have likely been conserved across the evolutionary history of living organisms. The emergence of the first cells involved the spontaneous formation of lipid bilayers. Over time, the complexity of membranes and their associated functions evolved as organisms diversified.
Modern cell membranes typically contain hundreds of lipid species (van Meer et al., Reference van Meer, Voelker and Feigenson2008; Sampaio et al., Reference Sampaio, Gerl, Klose, Ejsing, Beug, Simos and Shevchenko2011; Klose et al., Reference Klose, Surma and Simons2013; Harayama and Riezman, Reference Harayama and Riezman2018). The diverse structure, distribution and composition of lipids impact the physical properties of membranes. Biological effects include the ability to withstand extremes in temperature and pressure, although the full biological effect of these variations is unknown (Harayama and Riezman, Reference Harayama and Riezman2018). Modern cell membranes are typically composed of phospholipids, which can spontaneously form lipid bilayers in an aqueous environment (Bangham et al., Reference Bangham, Standish and Watkins1965). Although it has been proven that these lipids can be synthesized abiotically, it is unlikely that all of the necessary components for phospholipid membranes, let alone complex modern cell membranes, would be abundant in a prebiotic environment (Hargreaves et al., Reference Hargreaves, Mulvihill and Deamer1977; Monnard and Deamer, Reference Monnard and Deamer2002; Pohorille and Deamer, Reference Pohorille and Deamer2009; Ruiz-Mirazo et al., Reference Ruiz-Mirazo, Briones and de la Escosura2014).
Additionally, primitive membranes would not have the same complex biosynthetic machinery and structure as modern bilayer cell membranes, so the origin of protocell membranes is typically attributed to spontaneous assembly of amphiphilic molecules (Deamer and Barchfeld, Reference Deamer and Barchfeld1982; Hanczyc et al., Reference Hanczyc, Fujikawa and Szostak2003; Maurer and Nguyem, Reference Maurer and Nguyem2015). These molecules can assemble into micelles, monolayers and bilayers (as highlighted in Fig. 1), to create early cell-like vesicles (Deamer et al., Reference Deamer, Dworkin, Sandford, Bernstein and Allamandola2002).
The fatty acids in this study were found to form distinct phases, depending on temperature and salinity. (a) Spherical micelles with average diameter ranging from 2 to 4 nm. (b) Larger spherical micelles began to group together with increasing salinity. (c) Example of cylindrical micelle formation which preceded bilayer formation. (d) Small bilayer patches began to form at elevated salt concentrations coexisting with cylindrical micelles. (e) Stable bilayer membranes with average thicknesses of approximately 1–1.5 nm formed for all fatty acids at the highest salt concentrations. Lauric acid (C12) is cyan, myristic acid (C14) is violet and palmitic acid (C16) is mauve. Water molecules and salt ions are not shown for clarity.
Fatty acids, or monocarboxylic acids, have been proposed as an alternative constituent of the first protocell membranes (Gebicki and Hicks, Reference Gebicki and Hicks1973; Hargreaves and Deamer, Reference Hargreaves and Deamer1978; Walde et al., Reference Walde, Wick, Fresta, Manngone and Luisi1994; Schrum et al., Reference Schrum, Zhu and Szostak2010). Fatty acids consist of carbon chains with a methyl group at one end of the chain and a carboxyl group at the other end (Rustan and Drevon, Reference Rustan and Drevon2005). Fatty acids have a simpler chemical composition than phospholipids, and prebiotic chemistry could more easily permit synthesis of the necessary fatty acids and amphiphilic compounds required for membrane formation (Schrum et al., Reference Schrum, Zhu and Szostak2010). This chemical simplicity could also evolve towards modern phospholipid membranes through phosphorylation and addition of a second acyl chain to form phosphatic acid, which is a simple phospholipid (Schrum et al., Reference Schrum, Zhu and Szostak2010). Fatty acids also have desirable membrane transport permeability properties resulting from the lipid dynamics associated with vesicles (Mansy, Reference Mansy2010). The permeability and the membrane growth associated with fatty acid membranes make them an ideal primitive membrane (Schrum et al., Reference Schrum, Zhu and Szostak2010). They are permeable to small polar molecules and some charged molecules such as nucleotides, which suggests that cellular membranes could reasonably support the uptake of nutrients without modern transport mechanisms (Mansy et al., Reference Mansy, Schrum, Krishnamurthy, Tobé, Treco and Szostak2008).
There is evidence that the necessary amphiphilic compounds were abundant in early Earth (Deamer et al., Reference Deamer, Dworkin, Sandford, Bernstein and Allamandola2002; Lazcano, Reference Lazcano2010). Amphiphilic molecules isolated from the Murchison carbonaceous chondrite meteorite were found to be composed of varying acyl chain length and head group chemistry (Lawless and Yuen, Reference Lawless and Yuen1979; Deamer, Reference Deamer1985; Deamer and Pashley, Reference Deamer and Pashley1989). Other carbonaceous chondrites were also found to contain low-molecular weight fatty acids, suggesting that these simple amphiphiles may have been delivered to primitive Earth as a source of organics required to form simple membranes (Naraoka et al., Reference Naraoka, Shimoyama and Harada1999). These findings matched simulated synthesis with UV photolysis of interstellar and precometary ice analogues (Dworkin et al., Reference Dworkin, Deamer, Sandford and Allamandola2001) and with Fischer–Tropsch reactions (McCollom et al., Reference McCollom, Ritter and Simoneit1999; Rushdi and Simoneit, Reference Rushdi and Simoneit2001). This suggests that meteoritic delivery of fatty acids to the early Earth could support self-assembly of membranes (Deamer and Pashley, Reference Deamer and Pashley1989).
While the exact chemical conditions at the origin of life remain uncertain, some key factors and conditions typically considered include the existence of energy sources such as ultraviolet (UV) radiation, lightning or geothermal heat; the availability of simple organic molecules; the presence of water, minerals and catalysts; different pH levels; and potentially reducing environments. In this study, we focus on the effect of salinity on the formation of prebiotic membranes made of simple fatty acid molecules. To describe the early environments that would have allowed the formation of fatty acid membranes, we conducted all-atom molecular dynamics (MD) computer simulations using 12, 14 and 16-carbon fatty acids (referred to as C12, C14 and C16), deprotonated fatty acids at various temperatures and salinity levels ranging from freshwater (no-salt conditions) to saline solutions, seawater, seawater tide pools, salty tide pools and up to Dead Sea salt levels. All simulations were performed at a neutral pH, but the use of deprotonated fatty acids mimics a basic environment. Our findings indicate that the presence of salt is the primary factor driving bilayer formation, with stable bilayers only forming at the highest salt concentrations for the short-chained fatty acid C12.
Results
MD simulations were performed using GROMACS and the CHARMM36 force field to investigate the behaviour of deprotonated C12, C14 and C16 fatty acids (lauric acid, myristic acid and palmitic acid). The simulations were conducted at temperatures both below and above the melting temperatures of the respective fatty acid. Detailed information about the simulation setup can be found in the ‘Materials and methods’ section. Six different salt conditions were simulated to mimic the range of potential salinities in aquatic bodies: no additional salt ions, 0.15 M NaCl, 0.5 M NaCl, 1 M NaCl, 3 M Nacl and 5 M NaCl. The 0.15 M NaCl concentration corresponds to physiologically relevant salt levels that resembles the conditions under which cells naturally operate, while 0.5 M is typical for present day seawater. Seawater tide pools can show elevated salt concentrations between 1 and 3 M due to evaporation. We also included a high salt concentration of 5 M NaCl comparable to the Dead Sea. Our investigation focused on studying the influence of fatty acid chain length, temperature and salinity on structure formation. The simulations started with an initial bilayer configuration, and if the bilayer was not stable, it typically dissolved within 30 ns, leading to the formation of spherical or cylindrical micelles, as shown in Fig. 1. Given the relatively fast structural dynamics observed, all simulations were run for a total simulation time of 50 ns.
Salinity is the main driver of bilayer formation
The complete phase diagram for all fatty acids at all temperatures and salt levels is shown in Fig. 2. Results are shown as a function of salt concentration for each fatty acid (C12, C14, C16) below and above the respective chain melting transition T m. At 0 M NaCl, none of the fatty acids was found to form stable membranes. All of the simulations yielded spherical micelles throughout the simulation boxes. When simulations were run below the chain melting temperature T m, C12 and C14 were found to form cylindrical micelles at physiological salt concentrations of 0.15 M, while C16 was forming a stable bilayer in Fig. 2(a). Notably, higher salinity was required for the formation of membranes above T m, which was highlighted by the salinity required for all fatty acids to form a stable membrane (3 M below T m, and 5 M above T m). Based on these observations, we propose that the phase sequence leading to more stable states progresses from spherical micelles to cylindrical micelles and eventually to bilayers. Transitions to more organized phases occurred when (1) the salt concentration was increased, and (2) longer-chain fatty acids were used. A higher salt concentration was needed at higher temperatures.
Phase diagrams for C12, C14 and C16. The fatty acids are represented along the vertical axis, and salinity increases along the horizontal axis. The red zones denote unstable phases, while blue indicates increased aggregation (such as cylindrical micelles), and green depicts stable bilayer membranes. (a) All simulations were performed below the respective fatty acid melting temperatures. C16 is the first fatty acid to form a stable membrane at 0.15 M NaCl. As salinity increases, the shorter-chain fatty acids begin to form stable membranes. (b) All simulations were performed above the fatty acid melting temperatures. Compared to (a), a higher salinity is required to promote fatty acid aggregation, and more simulations were observed in the intermediary phase. All simulations formed stable bilayer membranes at 5.0 M NaCl. Lauric acid (C12) is cyan, myristic acid (C14) is violet and palmitic acid (C16) is mauve. Water and salt molecules are not shown.
The presence of salt leads to more densely packed and better ordered bilayer membranes
To understand the effect of salt on the properties of the fatty acid bilayers, density profiles and thickness of stable membranes as a function of salt concentration were determined from the simulations, as described in ‘Materials and methods’ section. Figure 3(a) shows the density profiles of C16 bilayers as an example. The distinct peaks in density correspond to the position of the heavier oxygen atoms in the carboxylic acid head groups and were used to determine the position of the head group region.
(a) Mass density profiles along the perpendicular z-direction of C16 bilayers at 5 M NaCl, where stable bilayers formed. The two maxima in the mass density correspond to the location of the heavier oxygen atoms in the fatty acid head groups, as shown in the overlaid snapshot of the simulation output for C16 below T m at 5 M NaCl with water and salt not shown. (b) Bilayer thickness for the stable membrane-forming simulations for C12, C14 and C16 below T m which increased with increasing salinity. (c) Thickness of all stable membranes for C12, C14 and C16 above T m. Similarly to (b), there is a slight increase with increasing salinity. Numeric values for all parameters are given in Tables 1 and 2.
As shown in Fig. 3(b) and (c), the thickness varied from approximately 0.28 to 1.35 nm. There was a significant increase in thickness with increasing salt concentration at temperatures below the chain melting temperature T m, shown in Fig. 3(b). The thickness increase was less pronounced at temperatures above T m in Fig. 3(c).
The deuterium order parameter, S CD was used to examine the stability of the simulations based on the ordering of the lipid acyl chain tails. S CD calculates the degree with which lipids are oriented and ordered based on the experimental procedure of measuring carbon–deuterium bond parameters with NMR spectroscopy. The order parameter is dependent on both the orientation of C–H bonds and the disorder of the system. Low S CD corresponds to disordered fatty acids in the membrane core. Longer chains can decrease S CD due to increased alignment of the carbon bonds, and low temperatures typically correspond to higher S CD due to lipids being packed together (Piggot et al., Reference Piggot, Allison, Sessions and Essex2017; Loschwitz et al., Reference Loschwitz, Olubiyi, Hub, Strodel and Poojari2020). S CD was computed as described in ‘Materials and methods’ section. The S CD profiles and average values for C16 simulations are shown in Fig. 4(a). C16 was the only fatty acid that formed a bilayer also at a lower salinity of 0.15 M NaCl and therefore allows a comparison between salt concentrations. Those S CD profiles show values between 0.1 and 0.25, pointing at tightly packed and ordered fatty acids. S CD profiles in micelles are significantly lower, with values of less than about 0.05. Average values for S CD for all samples are plotted in Fig. 4(b) for C16 at temperatures below and above the chain melting transition. For both temperatures, increasing salinity led to an increase in the deuterium order parameter.
(a) Deuterium order parameter exemplary shown for C16 at different salt conditions and temperatures below and above the chain melting temperature. C16 formed a bilayer above salt concentrations of 0.15 M for low-T and above 0.5 M salt at high-T, as shown in the phase diagram in Fig. 2. Values for the order parameter in the bilayer phase were found to be significantly higher than values in the micelle phase. (b) Average values for the deuterium order parameter determined from the curves in part (a) for C16. For both temperatures, S CD was found to significantly increase with increasing salt concentration. (c) Average deuterium order parameter of C12. (d) Average deuterium order parameter for C14. Error bars correspond to one standard deviation. Numeric values for all parameters are given in Tables 1 and 2.
The area per fatty acid (area per lipid, APL) was determined for the simulations, and are listed in Tables 1 (for temperatures below T m) and 2 (for above T m). Values are also plotted in Fig. 5(a) and (b). The areas formed three distinct regimes: areas of about 0.85 nm2 are typically found in loosely packed spherical micelles. More tightly packed areas of ~0.65 nm2 indicate the formation of cylindrical micelles, while small areas around 0.4 nm2 correspond to stable bilayers. The data points in Fig. 5(a) and (b) for temperatures below and above T m follow the phase behaviour in the phase diagram in Fig. 2. The transitions from spherical micelles to cylindrical micelles and eventually to stable bilayers are accompanied with decreases in the areas per fatty acid, respectively.
Area per fatty acid (APL), average deuterium order parameter (S CD) and membrane thickness for the last 20 ns of all simulations below T m
Thickness values are only included for the stable membranes which had two distinct peaks corresponding to each bilayer. The high APL values are indicative of more loosely packed structures, such as spherical micelles, whereas the low APL values are related to closely packed fatty acids in stable membranes. S CD and thickness correspond to the data shown in Figs 3 and 4.
Area per fatty acid (APL), average deuterium order parameter (S CD) and membrane thickness for the last 20 ns of all simulations above T m
Area per fatty acid as a function of NaCl salt concentration. The areas fall into three distinct regimes: areas of about 0.85 nm2 are found in loosely packed spherical micelles. More tightly packed areas of ~0.65 nm2 indicate the formation of cylindrical micelles, while small areas around 0.4 nm2 correspond to stable bilayers. (a) Area for C16, C14, C12, below the respective chain melting temperature T m. Formation of spherical micelles, cylindrical micelles and bilayers, with increasing salt concentration, in agreement with the phase diagram in Fig. 2. (b) Area per fatty acid above T m.
We note that the areas in Tables 1 and 2 are in good agreement with Höltje et al. (Reference Höltje, Förster, Brandt, Engels, von Rybinski and Höltje2001), who report areas per fatty acid of ~0.35 nm2 in fluid bilayers. We also note that areas of ~0.22 nm2 were reported in this work and also earlier X-ray diffraction work in crystalline bilayers (Luzzati and Tardieu, Reference Luzzati and Tardieu1974). Areas per fatty acid in simulations of mixed oleic acid/oleate bilayers also report areas of 0.30–0.35 nm2, in good agreement with our values for fluid fatty acid bilayers (Han, Reference Han2013).
As membrane stability is typically correlated with lower APL values, these findings confirm the phase sequence in the phase diagram in Fig. 2, where the APL corroborates the increased stability with high salinity and low temperatures. These observations can be summarized as follows: increasing salt concentration leads to (1) an increase in bilayer thickness accompanied by (2) an increase in the deuterium order parameter S CD, and (3) a decreased area per lipid. These findings are indicative of a more dense packing of the fatty acid tails in the presence of salt, with suppressed fluctuations.
Discussion
The formation of vesicular structures by oleic acid (C18) was first demonstrated through the mechanical agitation of evaporated long-chain fatty acids in buffer solution (Gebicki and Hicks, Reference Gebicki and Hicks1973, Reference Gebicki and Hicks1976). Subsequently, it was found that fatty acids can spontaneously self-assemble into vesicles in aqueous environments, similar to phospholipids (Gebicki and Hicks, Reference Gebicki and Hicks1973, Reference Gebicki and Hicks1976; Hargreaves and Deamer, Reference Hargreaves and Deamer1978; Apel et al., Reference Apel, Deamer and Mautner2002; Monnard et al., Reference Monnard, Apel, Kanavarioti and Deamer2002). The stability of these vesicles was reported to be influenced by factors such as fatty acid chain length and composition, temperature, concentration and salinity (Apel et al., Reference Apel, Deamer and Mautner2002; Deamer et al., Reference Deamer, Dworkin, Sandford, Bernstein and Allamandola2002; Monnard et al., Reference Monnard, Apel, Kanavarioti and Deamer2002). Fatty acid bilayer membranes exhibit pH sensitivity due to the stabilization provided by van der Waals interactions and hydrogen bonding (Rosano et al., Reference Rosano, Christodoulou and Feinstein1969; Haines, Reference Haines1983). The specific phases and structures formed through amphiphilic self-assembly depend on the packing behaviour of surfactants, which is governed by their soft matter characteristics and thermodynamic properties (Schäfer et al., Reference Schäfer, Kolli, Christensen, Bore, Diezemann, Gauss, Milano, Lund and Cascella2020).
Hargreaves and Deamer demonstrated that the presence of uncharged amphiphiles is necessary for the formation of membranes from ionic soaps and detergents. They observed the formation of liposomes from single-chain fatty acid amphiphiles and connected these liposomes to biomembranes through the visualization of permeability using phase-contrast microscopy (Hargreaves and Deamer, Reference Hargreaves and Deamer1978). The fatty acid vesicles were found to be fluid and stable above the melting temperature. It was proposed that pH is linked to the pKa of lipid anions in the micelles, which affects their stability. For example, when saturated fatty acids were added to 1.0 M NaCl that is made acidic with 1.0 mM HCl (pH 3) and melted by increasing the temperature, the protonated fatty acids did not assemble into organized structures (micelles or bilayers) even though salt was present. Instead, they formed fluid droplets that floated on the surface. When the pH was then increased by titrating with NaOH to the pH that is also the pK of the carboxyl groups, the droplets dissolved into the aqueous phase as micelles, and subsequently formed vesicles if the fatty acid was sufficiently concentrated. In this case, the sodium ions brought along hydroxyl ions that adjusted the pH to the pK of the fatty acid, at which point half of the carboxyl groups became carboxylate anions (Maurer and Nguyem, Reference Maurer and Nguyem2015).
Salt is the biggest driver of membrane stability
Our findings suggest that at neutral pH, salt plays an important role in membrane formation and stability. For our simulations of deprotonated C12 (lauric acid), C14 (myristic acid) and C16 (palmitic acid), we did not observe stable bilayer membranes in pure water at pH 7, regardless of the fatty acids' chain length, or temperature. As the salinity increases, fatty acids aggregate into larger spherical micelles, cylindrical micelles and eventually stable bilayer membranes. These phases align with the existing literature on the self-assembly of amphiphiles, which is driven by the ‘packing problem’ (Schäfer et al., Reference Schäfer, Kolli, Christensen, Bore, Diezemann, Gauss, Milano, Lund and Cascella2020). Previous studies have shown that the binding of counter ions can lead to changes in micellar morphology, allowing stable cylindrical micelles to form (Schäfer et al., Reference Schäfer, Kolli, Christensen, Bore, Diezemann, Gauss, Milano, Lund and Cascella2020). The long-range interactions resulting from the presence of counter ions are responsible for the observed morphology transitions from spherical to cylindrical micelles before the fatty acid membranes stabilize. The packing of fatty acids in a manner similar to sodium dodecyl sulphate, as studied by Schäfer et al. (Reference Schäfer, Kolli, Christensen, Bore, Diezemann, Gauss, Milano, Lund and Cascella2020), further supports the role of forces in aggregation and determining the final shape. Additionally, Sangwai and Sureshkumar (Reference Sangwai and Sureshkumar2011) have shown that salt induces a phase transition from spheres to rods at high concentrations. The observed phase transitions from spherical to cylindrical micelles align with previous findings, reinforcing the significance of salt in shielding electrostatic interactions.
Once bilayers had formed, increasing salt concentration was found to lead to more densely packed and stable bilayers. Figure 6 illustrates MD snapshots of stable C16 membranes formed at 0.15 M NaCl and 5 M NaCl, which demonstrate the high salt affinity of the fatty acid bilayer surface. The positively charged Na+ ions are attracted to and accumulate on the bilayer surface, particularly around the hydroxyl group and between the carboxylic acid head groups of the fatty acids.
MD snapshots of (a) a stable C16 membrane below T m at 0.15 M NaCl. The salt ions (yellow is sodium, teal is chloride) accumulate on the bilayer surface, between the fatty acid head groups. (b) Stable C16 membrane below T m at 5.0 M NaCl. In both cases, the sodium ions were found to strongly interact with the bilayer surface, locating between the fatty acid head groups. Water molecules are not shown for clarity.
The NaCl ions likely shield the electrostatic repulsion between fatty acid head groups, allowing for closer grouping of fatty acids. Ions can weakly bond to the charged head groups, creating an electric bilayer between the micelle–water surface, which results in a lower surface area per surfactant (Sangwai and Sureshkumar, Reference Sangwai and Sureshkumar2011). The formation of a charged layer was also observed with the addition of antagonistic salts, which formed an electric double layer at the solvent interface and stabilized ordered structures such as membranes (Sadakane and Seto, Reference Sadakane and Seto2018). The specific ions present play a role in how different salts impact charge shielding and the grouping of fatty acids (Hess and van der Vegt, Reference Hess and van der Vegt2009).
Longer chains and low temperature support formation
The only observed stable bilayer membrane at a moderate salt concentration of 0.15 M NaCl was formed by C16 fatty acids at a temperature below T m, indicating that long chains and low temperatures support the formation of stable fatty acid membranes. This is further supported by the high S CD values in this simulation, which indicate a higher degree of order in the fatty acid tails. However, the C16 0.15 M NaCl simulation above the melting temperature did not maintain a stable bilayer membrane; instead, it transitioned into a cylindrical micelle. The gel phase below T m appears to promote membrane formation.
Salt increases fatty acid packing and tail ordering
Stable membranes formed at high salinity exhibited high deuterium order parameters, indicating a high degree of ordering in the fatty acids. In contrast, cylindrical and spherical micelle phases showed lower deuterium order parameters, corresponding to lower fatty acid ordering. The increased ordering of fatty acids supports the formation and retention of stable bilayer membranes, which is also reflected in the lower area per lipid and higher density of the stable membranes.
The role of salt for membrane formation and stability
While the presence of salt has been found to disrupt fatty acid membranes in some studies (for instance in Monnard et al., Reference Monnard, Apel, Kanavarioti and Deamer2002), our findings align with recent literature that supports the role of salts in shielding electrostatic repulsion and facilitating the formation of stable membranes. It is important to note that the impact of salts on membrane formation can be complex and dependent on various factors. In our work, we specifically investigated the role of ion interactions resulting from environmental salt concentrations and their effect on the formation of stable fatty acid membranes.
Salt seems to be a significant factor in membrane stability. In highly saline environments, such as seawater tide pools or the Dead Sea, the presence of salt facilitates the formation of bilayer membranes from fatty acids. This suggests that a WLP located near a high salinity body of water, or a seawater tide pool, is a promising candidate for the emergence of protocells. Over time, salt concentrations in WLPs near less salty water sources, or tide pools, could increase through flooding by salty water and evaporation during wet–dry cycles. The salt ions play a crucial role in shielding the electrostatic repulsion between the hydrophilic heads of the fatty acids, allowing for closer contact and arranging themselves in a tightly packed and well-ordered membrane configuration, resulting in the formation of stable bilayer membranes. These findings are summarized in Fig. 7.
Summary of all results. Depending on temperature and salinity, fatty acids are expected to form micelles, cylindrical micelles or membranes. Seawater tide pools and hydrothermal springs should be most beneficial for the formation of membranous structures.
Although the high salinity values may appear to be unlikely in an early Earth environment, new estimates of halogen abundance in Earth's mantle suggest that seawater may have contained higher salt concentrations than modern Earth. The Halogen budget model presented by Guo and Korenaga proposes that halogens underwent early degassing and subsequent regassing, suggesting that halogen abundance may have been earlier than previously suggested for early Earth (Guo and Korenaga, Reference Guo and Korenaga2021). Given that the current majority of halogens remain concentrated in surface reservoirs, including oceans (Kendrick et al., Reference Kendrick, Hémond, Kamenetsky, Danyushevsky, Devey, Rodemann, Jackson and Perfit2017), an extremely salty body of water is a plausible environment for the emergence of prebiotic life.
Conclusion
The environmental conditions for the origin of life are not well constrained, but as membranes were key to the origin of life and the first ones were likely composed primarily of fatty acids, we conducted all-atom MD computer simulations to investigate the formation of bilayer membranes composed of C12, C14 and C16 fatty acids below and above their respective melting temperatures, and different salt concentrations to mimic freshwater ponds, saline solution, seawater, seawater tide pools, salty tide pools and dead sea salt concentrations. No bilayers formed in the absence of salt. The presence of NaCl facilitated fatty acid membrane formation by mitigating electrostatic repulsion, enabling the formation of tightly packed and highly ordered bilayer membranes. Longer fatty acid chains and lower temperatures also supported membrane formation. All fatty acids were found to undergo a phase sequence from spherical micelles to cylindrical micelles, and eventually bilayers. Phase coexistence was observed in the simulations. While the long-chained C16 can form bilayers already at physiological salt concentrations, C14 membranes would form in seawater tide pools only, and C12 bilayers require elevated salt concentrations of at least salty tide pools. Our findings suggest that fatty acids could likely serve as the fundamental building blocks for simple membranous vesicles and protocells in WLPs connected to salty water bodies, or seawater tide pools. Elevated salt levels in these ponds and pools can be maintained through the influx of salty water and the evaporation of water during wet–dry cycles. While the formation of membranes is a fundamental aspect of cellular organization, it is not an evolutionary process in the traditional sense but rather a result of the physicochemical properties of molecules and their interactions. Evolutionary processes would have subsequently shaped the diversity and complexity of life, including the structures and functions of cellular membranes.
Materials and methods
Fatty acids
All simulations were set up using the CHARMM-GUI (Jo et al., Reference Jo, Kim, Iyer and Im2008) bilayer builder (Wu et al., Reference Wu, Cheng, Jo, Rui, Song, Dávila-Contreras, Qi, Lee, Monje-Galvan, Venable, Klauda and IM2014) using the built-in structures for lauric acid, myristic acid and palmitic acid. Table 3 lists these fatty acid properties.
List of all fatty acids used in this study, their chemical formula, the corresponding chain melting temperatures and the temperatures of the MD simulations
The simulation temperatures above and below the melting point were calculated as being 15°C below (referred to as ‘Below T m’) and 15°C above (‘Above T m’) to simulate a realistic environmental temperature fluctuation.
Molecular dynamics (MD) simulations
Each fatty acid (C12, C14, C16) was simulated below and above the corresponding T m with (1) no added ions, (2) 0.15 M NaCl, (3) 0.5 M NaCl, (4) 1.0 M NaCl, (4) 3.0 M NaCl and (5) 5.0 M NaCl. All systems contained 200 fatty acid molecules and 10 000 TIP3 water molecules (Mark and Nilsson, Reference Mark and Nilsson2001). Salt molecules were added in accordance with the concentration and varied from no added sodium or chloride ions (0 M) to a maximum of 1176 sodium ion molecules and 976 chloride ion molecules, for a maximum system size of 12132 molecules in the systems containing 5.0 M NaCl. Transition temperatures are listed in Table 3. The plain membrane systems (systems which were only examining fatty acids on their own for the formation of bilayer membranes) all contained 100 deprotonated fatty acids on both the upper and lower leaflets. The membranes composed of equal parts C12, C14 and C16 deprotonated fatty acids contained 99 lipids on the upper and lower leaflets. The size of the simulation box was 419.4 nm3 (5.477 nm along the x- and y-axes, and 13.98 nm along the z-axis) for the plain membrane systems, and 415.0 nm3 (5.450 nm along the x- and y-axes, and 13.98 nm along the z-axis) for the mixed membranes. All systems contained 50 water molecules per one lipid molecule. The selected temperatures were about 15°C below and above T m of each fatty acid. The bilayer builder in CHARMM-GUI was selected from the membrane builder input generator to set up the simulations. Each membrane was prepared with a rectangular box. The length of z was determined by the hydration number of 50 water molecules per 1 lipid molecule, and the length of xy was determined by the number of lipid components. The number and type of lipids used for each respective simulation were selected from the fatty acid drop-down menu to finalize the system information. NaCl ions were added in the appropriate concentration using the replacement method for all systems that included ions. Once the ions and water box were generated, the CHARMM36m force field (Huang et al., Reference Huang, Rauscher, Nawrocki, Ran, Feig, de Groot, Grubmüller and MacKerell2017) was chosen and the GROMACS input was selected. Finally, the temperature was adjusted to the corresponding simulation.
All-atom simulations were run using GROMACS 5.1.2 (Berendsen et al., Reference Berendsen, van der Spoel and van Drunen1995; Lindahl et al., Reference Lindahl, Hess and van der Spoel2001; Van Der Spoel et al., Reference Van Der Spoel, Lindahl, Hess, Groenhof, Mark and Berendsen2005; Hess et al., Reference Hess, Kutzner, van der Spoel and Lindahl2008; Pronk et al., Reference Pronk, Páll, Schulz, Larsson, Bjelkmar, Apostolov, Shirts, Smith, Kasson, van der Spoel, Hess and Lindahl2013; Abraham et al., Reference Abraham, Murtola, Schulz, Páll, Smith, Hess and Lindahl2015; Páll et al., Reference Páll, Abraham, Kutzner, Hess, Lindahl, Markidis and Laure2015) on two GPU-accelerated computer workstations. The systems were first energy minimized using the steepest-descent algorithm. Minimization was followed by 250 ps of stepwise equilibration using an NVT ensemble with the reference temperature (corresponding to the specific simulation) maintained with the Berendsen thermostat. NVT equilibration was followed by 1625 ps of stepwise NPT equilibration at constant temperature (corresponding to the specific simulation) using the Berendsen thermostat. The pressure of 1.0 bar was maintained using the Berendsen barostat and semi isotropic weak coupling (τ = 5 ps). Equilibration was confirmed with gmx energy to check for plateaus in the area per lipid and total energy. Following equilibration, simulations were performed for 50 ns using a 2 fs time step over 25 000 000 steps, a periodic boundary cut-off of 1.2 nm, the particle-mesh Ewald method using a real-space cut-off of 1.2 nm, fourth-order interpolation and 0.16 nm grid spacing to solve for long-range electrostatics. The parallel LINCS (P-LINCS) algorithm was used to determine bond constraints (Hess, Reference Hess2008). Temperature and pressure were maintained at respective simulation temperatures (below Tm and above Tm)and 1.0 bar using a Nosé–Hoover thermostat (Nosé, Reference Nosé1984a, Reference Nosé1984b; Hoover, Reference Hoover1985) at 30°C (τ = 1.0 ps) and Parrinello–Rahman semi isotropic weak coupling (τ = 5 ps) (Parrinello and Rahman, Reference Parrinello and Rahman1981). All analyses were performed on the entire simulations and systems were run for 50 ns.
All MD simulations in this study were run at neutral pH 7 with respect to the CHARMM-GUI input generation. The challenge in explicitly defining pH directly in MD simulations arises from the fact that pH is a property that emerges from the behaviour of a large number of water molecules, ionizable solutes and their interactions (Hess and van der Vegt, Reference Hess and van der Vegt2009; Sadakane and Seto, Reference Sadakane and Seto2018). At a pH of 7, a concentration of 10−7 M of H3O+ with water at 55 M means that effectively no water molecules are dissociated in a typical sized simulation box with less than 105 water molecules. We note that in fatty acid membranes, the protonation state of fatty acids can vary depending on the pH of the surrounding environment. Fatty acids contain a carboxylic acid group (–COOH), which can become protonated or deprotonated depending on the pH of the solution. At low pH (acidic conditions), the carboxylic acid group tends to be protonated, meaning it has picked up an extra hydrogen ion (H+) to form the –COOH group. In this protonated state, fatty acids have a net positive charge. At high pH (alkaline conditions), the carboxylic acid group tends to be deprotonated, meaning it has lost a hydrogen ion (H+) to form the carboxylate anion (–COO–). In this deprotonated state, fatty acids have a net negative charge.
C12–C16 fatty acids have pKa values in the range 4.95–5.3 (Human Metabolome Database; ICSC 0530; PubChem. Hazardous Substances Data Bank (HSDB); PubChem. Lauric Acid; PubChem. Myristic Acid; PubChem. Palmitic Acid). Using the Henderson–Hasselbalch equation, the ratio of protonated to deprotonated fatty acids was calculated for C12, C14 and C16 fatty acids with their respective pKa values. All three fatty acids are almost entirely deprotonated with pH above 7.0, so the acid is effectively dissociated. Using fully deprotonated fatty acids for all simulations essentially mimics a basic environment with pH >7. Our simulations are thus in agreement with reports that fatty acid membranes form in weakly alkaline conditions (pH 7–9) (Maurer and Nguyem, Reference Maurer and Nguyem2015).
Data analysis
GROMACS scripts were used to obtain the area per lipid, bilayer thickness and order parameter for the last 20 ns of each respective simulation. To calculate the area per lipid, gmx energy was used to extract the x− and y− box vectors from the portable energy file of the final output stage. The two vectors were then multiplied and subsequently divided by the number of lipids per leaflet (100 for the pure membranes analysed). The estimated error per x− and y− box vector was provided in the GROMACS output and propagated to the area per lipid.
gmx density was employed to compute the partial density of fatty acids across the simulation box. Only fatty acids were selected as the input to exclude the ions and water. The resulting density.xvg files were uploaded to Python and the peak values of the high salt membranes were extracted using find peaks from scipy.signal. The change in thickness for stable membranes was calculated by taking the absolute difference of the x values (which correspond to the positions of fatty acids within the simulation box) associated with the density peaks of high salt membranes.
The order parameter was calculated using gmx order, which computes the order parameter for the atoms of carbon tails. An index file was first created which only contained the carbon atoms associated with the fatty acids. The index file was then used with gmx order to generate the order and the deuterium order parameter. Standard deviation was calculated from the average S CD. The deuter.xvg file was imported into Python to generate plots using matplotlib (Hunter, Reference Hunter2007), and simulation output images were created with visual molecular dynamics (Humphrey et al., Reference Humphrey, Dalke and Schulten1996).
Acknowledgements
We would like to thank Dr David Deamer, UC Santa Cruz, for fruitful discussions and advice on this project.
Financial support
This research was supported by the National Science and Engineering Council Canada (NSERC), the Canadian Foundation for Innovation (CFI) and McMaster University. M. C. R. is the recipient of an Early Researcher Award of the Province of Ontario and a McMaster University Scholar.
Competing interest
None.
Open access
