A Modular 3D-Printed Design to Investigate Prebiotic Chemical Systems in Hot Spring Pools

2.1. Materials and reagents

DA, DO, and both phospholipids (POPC 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine and POPG 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol) were purchased from Sigma-Aldrich and used as received. All water used was purified to 18 MΩ (Milli-Q). A stock solution (0.1 M) of phosphate buffer (PB) was prepared by dissolving appropriate amounts of NaH₂PO₄ and Na₂HPO₄ in water, followed by adjustment of pH (7.4) using either 5 M HCl or 5 M NaOH (Lowy solutions). Similarly, 50 mM HEPES [4-(2-hydroxyethyl)−1-piperazineethanesulfonic acid] buffer, which was used for samples that contained phospholipids, was prepared by dissolving HEPES in water and adjusting the pH to 7.4. The basalt sample was collected from a recent, fresh lava flow just outside Reykjavik, Iceland (exact coordinates not recorded). The kaolinite sample was collected from altered basaltic rocks of the 3.49 Ga North Star Basalt that underlies the stromatolite-bearing Dresser Formation at the Dresser Mine site in the North Pole Dome, Pilbara Craton (UTM Zone 50 K, E752698, N7659202). The pyranine/HPTS (8-hydroxypyrene-1,3,6-trisulfonic acid trisodium salt, Sigma Aldrich Australia) stock solution was prepared in water and used as an encapsulation dye for both fatty acid and phospholipid-based vesicles. Either 0.1 M PB or water was used to rehydrate the lipids during the WD cycles. For experiments requiring the use of RNA, 1 mg mL⁻¹ of RNA (ribonucleic acid from torula yeast, Type VI, Sigma-Aldrich) solution was prepared using 0.1 M PB (pH 7.4) containing 10 mM EDTA (ChemSupply Australia). The RNA solution was adjusted to pH 7.4 with 5 M NaOH. The fluorescent dye in a QuantiFluor^® RNA System kit (Promega) was used for imaging RNA. Adjustment of pH for all samples was carried out using an Orion Star A121 pH meter with an Orion 8103BN mini-probe (Thermo Fisher).

2.2. Designing, 3D printing, and Pool configuration

To mimic hot spring pools on early Earth, a modular and scalable hot-spring pool simulator was developed (Fig. 1, Supplementary Figs. S1 and S2). The design aimed to produce small volume (2–3 mL), thermally responsive, and chemically robust units capable of integration into interconnected arrays for studies on prebiotic chemistry and protocell dynamics.

OPEN IN VIEWER

FIG. 1.

Photograph of the model hot spring pools system showing the assembled pools with temperature probes placed in the cups and heating pads inserted underneath for controlling temperature. Power supply provides electricity for heating pads, and an Arduino Uno is connected via USB-type-C cable to a laptop for PID (Proportional-Integral-Derivative) control. Blue-colored LEDs in the Arduino Uno turn off when the setpoint is reached and back on to maintain the temperature in the rehydration phase of the WD cycles. Water flow control is managed by the mini-peristaltic pumps connected to the water control unit.

2.3. Programmed water and temperature control for multipool systems

To simulate the dynamic physicochemical environment of ancient volcanic hot spring terraces, a programmable control system was developed to regulate water levels and temperatures independently across multiple interconnected cup units. This system enables precise spatial and temporal control over key experimental parameters critical to prebiotic chemistry research, which include WD cycling, thermal gradients, and fluid exchange.

2.4. Rock coating of the model pools to mimic surficial early Earth pools

To simulate the mineral surfaces of early Earth hot springs, a robust and repeatable rock coating methodology was developed for application on 3D-printed cup interiors. The aim was to reproduce chemically and structurally relevant mineral surfaces (e.g., basalt and kaolinite) that could endure WD cycling without degradation, thereby enabling prebiotic chemistry experiments on realistic geological substrates. Individual cup geometry facilitated balanced rotational motion during coating, improving coating uniformity and minimizing structural distortion during oven curing.

The rock coating process was executed in three distinct stages: (1) preparation of a PDMS intermediate layer, (2) formulation of the rock slurry, and (3) application and curing of the coating. Particular attention was paid to coating uniformity, adhesion, and resistance to aqueous dissolution.

Polydimethylsiloxane (PDMS) served as a hydrophobic adhesive substrate for the rock particles. A 10:1 ratio of PDMS base to curing agent was mixed and diluted with hexane in a 1:1 solvent-to-PDMS ratio to reduce viscosity and promote even spread on curved cup surfaces. Approximately 0.5 g of this solution was applied per cup, manually rotated to ensure even coating, and cured at 60°C for 2 h. Following curing, plasma treatment was performed using a Harrick plasma cleaner for 5 min to temporarily convert the PDMS surface from hydrophobic to hydrophilic to enhance adhesion of the aqueous rock slurry.

Comparative testing of the coating behavior on PLA compared with printed Durable Resin cups indicated superior adhesion and uniformity on PLA (Supplementary Fig. S3). The resolution of FDM printing was demonstrated to be more advantageous for our application, as the coarser surface finish increased the surface area of the cups and enabled greater adhesion of the rock coating to the cup surfaces, compared with using a durable resin material.

The geologically relevant materials kaolinite and basalt were selected for testing. Submillimeter-sized grains were prepared by a Fritsch Jaw Crusher and a Beige Ring mill, then ground with an agate mortar and pestle for 5 mins. The powders were then passed through a 106 µm sieve to remove the largest particles. Slurries of each powder were mixed with water and lime [Ca(OH)₂] as an additive to improve adhesion and reduce cracking during drying. The slurry was applied and manually spread across the PDMS-treated surface. Drying was carried out at 40°C for 3–5 min to avoid rapid solvent evaporation and associated cracking. Microscopic examination at 10× and 50× magnification confirmed that a 5% lime slurry (rock:water: lime at 1:2.5:0.18) yielded the most homogeneous coatings, with minimal cracking and high surface coverage (Supplementary Fig. S4). Qualitative water droplet and submersion tests demonstrated coating resilience for the 5% lime slurry samples under fluid delivery, with minimal delamination observed after multiple wetting cycles (Supplementary Fig. S5).

2.5. Vesicle preparation

Vesicles were made via the thin-film hydration method. The appropriate amounts of lipid [DA, DA:DO (4:1), and phospholipids] were dissolved in chloroform (Sigma-Aldrich), and then the chloroform was evaporated under a gentle stream of nitrogen followed by the addition of a hydration buffer (0.1 M PB). Samples of the resulting solutions were then imaged to confirm vesicle self-assembly. Samples undergoing WD cycles were heated to a temperature of ∼60°C–70°C in the 3D-printed cups until all the solvent evaporated, after which the materials were then rehydrated with the respective buffer or water. As a control to compare morphological differences between experiments, one WD cycle was also performed in a glass vial that contained lipids and was heated at the same temperatures used for the model pool system experiments.

To obtain contrast for imaging, fatty acid vesicle samples were diluted 3× in water and phospholipid vesicles (2.5 mM) diluted 6× in 50 mM HEPES. Stock solutions of DA vesicles (150–200 mM) were diluted such that the final concentration of the vesicles remained above the critical vesicle concentration (CVC ∼40 mM). For RNA-containing DA samples, the pH was adjusted to 7.35–7.40, and the vesicles were either washed twice using 100 kDa molecular weight cut off filters (5000 g for 10 mins) or diluted 3× in 0.1 M PB to obtain improved contrast for imaging. Experiments were repeated at least twice; for each repeat, all vesicle samples were freshly prepared on the same day as the subsequent WD cycling and imaging.

2.6. Microscope imaging

Images were captured using a Nikon Eclipse TE-2000 inverted microscope with a 100× Ph3 objective [Plan Fluor, numerical aperture (NA) = 1.3] equipped with a camera (pco.edge 4.2 sCMOS). For all images captured, the focus was set to the middle planes (solution phase) of the sample to image the products representative of the experimental solutions and avoids imaging vesicles that had been immobilized by the glass cover slips or glass slides, or those that had grown on the glass surfaces. Image analysis such as adjustment of brightness or contrast as well as insertion of scale bars was done using ImageJ/FIJI.

3.1. Modular design considerations

The design of our initial hot-spring simulator prototype was based on an interconnected modular cup-based design, with each cup representing one hot-spring pool, that is, one set of conditions. This prototype consists of a three-pool system in which each cup is connected and fluid flow is driven by mini-peristaltic pumps and gravity (Fig. 1). Heating pads, which can be inserted under each of the cups, are covered with a layer of sand to minimize any direct burning of the cup resin and allow for an efficient exchange of heat between the cup and the heating pad. Two adjacent cups are connected via a closed channel and kept at the same surface level, while a third cup is at a lower height and connected only to the second cup via an open terraced channel to simulate conditions observed in natural hot spring systems. This unique design can simulate dynamic water flow that, along with temperature control within individual cups, can promote prebiotic chemical reactions in the different cups. The design enables different prebiotically relevant chemical reactions to occur in different pools due to WD cycling of fluids in the different cups (e.g., condensation of monomers and encapsulation of polymers such as RNA in lipid vesicles), and the collection of samples for analyses after the different fluids and reaction products are mixed in the third cup. Because various minerals in the weathering zones of acidic hot springs, particularly montmorillonite, have also been shown to be important in prebiotic chemistry (Ferris, 2006; Markússon and Stefánsson, 2011), the cups were designed so that they could be coated with rock and/or mineral particles for a more faithful simulation of the conditions of actual hot spring pools, as discussed in Section 2.4 and Supplementary Figures S3–S5.

3.2. Vesicle formation was not affected by the 3D-printed material (PLA vs. durable resin) of the cups

We initially investigated whether the 3D-printed resin of the cups affected the formation of vesicles (Fig. 1). PLA and durable resin were used to print the cups as they are among the most commonly used polymeric materials for 3D printing. These polymers differ in their material properties that include biodegradability, flexibility, strength, and durability, resistance to heat, and surface finish. It was found that the surface finish, which can significantly affect how lipids will assemble, was visibly smooth for durable resin and rough for PLA. Nonetheless, surficial features smaller than the wavelength of visible light, such as roughness at the nanometer scale, could be similar.

The 3D-printed cups were tested at elevated temperatures to mimic the geothermal environment of hot spring pools and geysers (∼60°C–70°C). Evaporation rates and any deformations of the cups were monitored during the experiments. Placing a thin layer of sand between the cups and the heat source (a heating pad) provided a more even distribution of heat in the cups as monitored by temperature probes positioned in the cups. An even heat distribution makes the evaporation of water more efficient. Complete evaporation of the fluid resulted in adsorption of lipids to the cups’ surfaces. Although PLA and durable resin are known to be inert, the cups were immediately rewetted after complete dehydration to avoid potential degradation and any chemical reactions with and between the adsorbed lipids.

It was found that vesicles formed after multiple WD cycles (Fig. 2) independent of the composition of the cup material. The PLA resin, possessing a larger degree of surface roughness, was found to be more conducive for coating with mineral particles, and is thus recommended for ongoing work. OPEN IN VIEWER

We performed WD cycles both in a glass vial and in the cups in order to test whether the WD cycles in the cups yielded vesicles with different morphologies. DA was used instead of long-chain oleic acid or other fatty acids to mimic the prebiotically plausible short-chain fatty acids. The morphological variation of the vesicles was compared with that of phospholipid vesicles also subjected to WD cycles.

It was observed that lipids subjected to WD cycles produced heterogeneous populations of vesicles such as thin-walled unilamellar or oligolamellar vesicles and thick-walled multilamellar vesicles as well as lipid aggregates (Fig. 2). WD cycling in the cups yielded fewer vesicles compared with the glass vial. Glass vials are known to have smooth and negatively charged surfaces, which could trigger the release of lipids upon rehydration as a result of electrostatic repulsion between the glass surface and negatively charged headgroups of the phospholipids and fatty acids. In contrast to glass vials, the rougher surfaces of the cups may have a greater tendency to keep the lipids adsorbed and/or vesicles trapped in surficial microstructures.

3.4. Encapsulation is favored in thin-walled vesicles

We then investigated whether different vesicle morphologies influenced the encapsulation of organic material such as RNA. For ease of investigation, we first used the fluorescent reporter dye HPTS as a model prebiotic material and evaluated its encapsulation in both fatty acid and phospholipid-based vesicles after multiple WD cycles. As observed in the initial experiments, the cups triggered the formation of multiple types of vesicle populations, especially thin-walled oligolamellar or unilamellar vesicles (Fig. 3). We were particularly interested in the thin-walled unilamellar vesicles as they best resemble extant biological cells. We previously demonstrated that one WD cycle in an Eppendorf tube leads to the formation of some unilamellar vesicles and enhanced encapsulation of both HPTS and fluorescently labeled RNA (Steller et al., 2022). Unilamellar vesicles have various advantages over their multilamellar counterparts, such as ease of permeation of nutrients and waste across the bilayer, a potentially necessary feature for early protocells to sustain internal metabolic reactions. OPEN IN VIEWER

Here, we observed that multiple WD cycles yielded many oligolamellar or unilamellar vesicles, and fluorescence microscopy revealed enhanced encapsulation of HPTS in these vesicle morphologies compared with thick-walled multilamellar vesicles. As explained by Steller and coworkers, this result is potentially a consequence of the apparent pK_a of the bilayers changing during WD cycling and the thin-walled structures being more osmotically accessible than thicker lipid stacks (Steller et al., 2022). During the dehydration phase of the WD cycles, the ionic strength of the solution increases, leading to a decrease in apparent pK_a that would favor the formation of a bilayer phase even at lower pHs. When the salt concentration is high, recruitment of protons to the bilayer surface is expected to become less efficient, leading to a decrease in apparent pK_a and thereby enabling lipids to stack in a layer-by-layer fashion while incorporating cations (e.g., Na⁺) along with other encapsulated solutes. Such physical preorganization of lipids in layers is considered crucial in the formation of giant unilamellar vesicles, as well as in the encapsulation of solutes. The thinner bilayer stacks containing the most solutes should swell fastest upon rehydration, consistent with the observation of thinner-walled vesicles encapsulating the most dye (Steller et al., 2022).

The morphology of the vesicles can vary depending on their lipid composition and external environment (e.g., pH, temperature, and salt concentration). For instance, it is known that the addition of DO to DA vesicles stabilizes the membranes by facilitating hydrogen bonding between hydroxyl and carboxylate/carboxylic acid head groups, respectively, even at higher pHs (Apel et al., 2002).

To investigate whether vesicles that contained mixtures of fatty acid and fatty alcohol affected the encapsulation of solutes, vesicles were prepared using a 4:1 mixture of DA and DO and subjected to WD cycles in the model pools (Fig. 3). We observed that the addition of 20 mol% DO promoted the formation of oligolamellar or unilamellar vesicles and enhanced the encapsulation of HPTS, while also lengthening the shelf-life of the high concentration samples (∼400 mM). In contrast, DA alone was prone to an abrupt decline in pH at high concentrations due to protonation and formation of neat oil at the air-water interface. Some of these differences can potentially be attributed to the different critical packing parameters of the membrane constituents (Israelachvili, 2011).

3.6. Encapsulation of RNA in vesicles is similar to dye encapsulation

After optimizing the experimental setup with HPTS and developing an analysis protocol, we next evaluated the encapsulation of RNA in the vesicles under similar conditions.

In this experiment, DA was used to form vesicles in 0.1 M PB, and the encapsulation of RNA from torula yeast (0.05–0.1 mg mL⁻¹) was investigated. The RNA was dissolved in a 0.1 M PB that contained 10 mM EDTA, which serves as a chelating agent for trace divalent cations such as Ca²⁺ and Mg²⁺ and thereby prevents RNA degradation by nucleases. The DA along with RNA, Promega RNA dye (dilution factor indicated in figure captions, e.g., 800×), and buffer (0.1 M PB + 10 mM EDTA) were combined, and the pH was adjusted to ∼7.4 using 5 M NaOH. The mixture was then subjected to one WD cycle in a glass vial, rehydrated with water, readjusted to pH 7.4, and analyzed with fluorescence microscopy. Similar to the HPTS dye encapsulation that we observed in the DA vesicles, encapsulation of RNA in the inner core of the vesicles was observed after one WD cycle (Fig. 4). However, higher concentrations of RNA and prolonged storage after sample preparation both resulted in a significant drop in pH and the formation of clumped vesicles (Supplementary Figs. S6 and S7). These results confirmed that HPTS is a good proxy for encapsulation of low concentrations of larger anionic solutes, such as RNA, and can provide an inexpensive means for further experimentation. OPEN IN VIEWER

4.1. Programmed water flow control

Preliminary work for automated fluid control included design simulations of water transfer using peristaltic pumps and solenoid valves governed by PID control, as implemented in LabView in previous iterations. Future integration will involve pump control via Arduino-driven relays, which will allow researchers to program inflow/outflow cycles, emulate WD cycles, and coordinate flow across pools at predetermined intervals or thresholds (e.g., temperature or time-based triggers).

The system design anticipates the need for directional flow between pools facilitated by top-mounted channels and programmable volume displacement. Such features are critical to experiments involving protocell transfer, dilution, and mineral diffusion across a synthetic hydrothermal network.

4.2. Rock coatings to simulate hot spring pool surfaces

In the future, we are focused on performing vesicle WD experiments that utilize the mineral-coated surfaces to closely simulate the rocky surface of geothermal hot spring pools. The use of different surface topographies, chemistries, porosities, and surface roughness could influence the versatility of the vesicles’ population and encapsulation of RNA, peptides or other polymers. This added feature of rock-coating in the pools will aid our understanding of how different early-Earth pools harbored vesicles and facilitated the evolution of protocells during WD cycles.