Exchange, catalysis and amplification of encapsulated RNA driven by periodic temperature changes

doi:10.21203/rs.3.rs-2014540/v1

Download PDF

Article

Exchange, catalysis and amplification of encapsulated RNA driven by periodic temperature changes

https://doi.org/10.21203/rs.3.rs-2014540/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 03 Mar, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Growth and division of biological cells is based on the complex orchestration of spatiotemporally controlled reactions driven by highly evolved proteins. In contrast, it remains unknown how their primordial predecessors could achieve a stable inheritance of cytosolic components before the advent of translation. An attractive scenario assumes that periodic changes of environmental conditions acted as pacemakers for the proliferation of early protocells. Using catalytic RNA (ribozymes) as models for primitive biocatalytic molecules, we demonstrate that the repeated freezing and thawing of aqueous solutions enables the assembly of active ribozymes from inactive precursors encapsulated in separate lipid vesicle populations. Furthermore, we show that encapsulated ligase ribozymes can overcome freezing-induced content loss and successive dilution by freeze-thaw driven propagation in feedstock vesicles. Thus, cyclic freezing and melting of aqueous solvents – a plausible physicochemical driver likely present on early Earth – provides a simple scenario that uncouples compartment growth and division from nucleic acid self-replication, while maintaining the propagation of these replicators inside new vesicle populations.

Early cellular life on Earth is thought to have started with the encapsulation of self-replicating informational subsystems in primitive compartments¹. Encapsulation of biomolecules can occur in a variety of microenvironments such as adhesive mineral surfaces, fatty acid vesicles, liposomes, coacervates, and even peptide-based membranes². The transfer of genetic information is primarily associated with template-dependent replication of nucleic acids,^3,4 but can also be performed to a limited extent with peptides⁵ and small organic molecules⁶. Compartmentalization is required to isolate and protect the autocatalytically replicating information system from environmental perturbations and parasitic replicators and to establish phenotype-genotype coupling^7,8. A popular model for protocells involves a “replicase”, a catalytic nucleic acid that aids templated replication of nucleic acid polymers, which is encapsulated in a membrane vesicle². While initial concepts of such vesicular protocells evolved around fatty acids, which are readily available by prebiotic synthesis⁹, their low stability and lack of tolerance towards physicochemical parameters such as pH and ionic strength made reconciling them with nucleic acid catalysis challenging^10,11. Recent studies demonstrating the non-enzymatic synthesis of diacyl phospholipids either by the phosphorylation of diacyl glycerol¹², the selective alkylation of phosphoglycerol¹³, or the alkylation of an acyl phospholipid precursor¹⁴, strengthen the hypothesis that early protocells may have also had membranes containing simple phospholipids¹⁵.

Vertical transfer of genetic information is a prerequisite for a species’ ability to proliferate. Horizontal gene transfer, on the other hand, is known to be a powerful mechanism for intra- and interspecies dissemination of genetic innovation^16,17. Whilst modern cells rely on protein-based machineries for replication of both compartments and genetic polymers, such systems were not available during early stages of molecular evolution. A possible plausible alternative to the complex extant biological processes that mediate the exchange and spread of genetic molecules is achieved by repetitive physical and / or chemical processes altering compartment permeability, such as periodic pH changes, temperature oscillations or dehydration cycles^18–21. In particular, freeze-thaw (FT) cycles enable content exchange between membrane vesicles without requiring complex protein machineries. For example, repeated liquid nitrogen-based freeze-thaw cycles of pelleted liposomes have been shown to allow the combination of fusion and fission of liposomes with RNA replication by a modern viral RNA replicase, and the distribution of RNA to daughter liposomes²². Previously, our groups investigated the effects of temperature cycling on giant unilamellar vesicles (GUVs), showing that under less extreme freezing conditions and lower centrifugal forces, content exchange can occur independently from fusion and fission²³.

Here, we showcase that when a mixture of GUVs, encapsulating either an RNA substrate or a ribozyme, is subjected to a freeze-thaw cycle, the contents of both GUV populations mix during thawing enabling encapsulated RNA catalysis under unfrozen conditions. The system is independent from the type of ribozyme used and promotes both RNA cleavage and ligation activity. We further demonstrate that the use of an autocatalytic ligase ribozyme replicator enables concurrent ribozyme amplification and vesicle content exchange, thus leading to the proliferation of ribozyme-containing vesicles during serial transfer experiments, when combined with the feeding of substrate-encapsulating GUVs. Together, these experiments present a model for how early RNA-protocells may have been able to grow and proliferate in a plausible geochemical environment exhibiting cyclical freezing and thawing of aqueous solvents.

Freeze-thawing allows content exchange and Hammerhead cleavage.

Initially, we sought to explore whether it is possible to exchange RNA strands encapsulated in different GUV populations by FT cycles, thereby triggering a ribozyme reaction (Fig. 1a). In these experiments, we opted for a well characterized hammerhead ribozyme (HH-min) that cleaves a substrate strand (HH-sub), which contains a fluorescent dye (either FAM or Cy5) at the 5’-end and an optional black hole quencher (BHQ1) at the 3’-end²⁴ (Supplementary Fig. 1a). This system can therefore be characterised both microscopically by the increase in fluorescence upon HH sub-cleavage and by classical denaturing polyacrylamide gel electrophoresis (PAGE). All GUVs were composed of 1-palmitoyl-2-oleoyl-phosphatidylcholine (POPC) as the phase transition temperature of this lipid (-2°C) enables reliable transient membrane permeabilization and content exchange during thawing episodes²³. We encapsulated both HH-min and quenched HH-FQ-sub in separate GUV populations via an inverted emulsion transfer-based method before they were combined in a 1:1 ratio for each experiment. The emulsion transfer method relies on a density gradient between inner and outer phase achieved by equimolar (900 mM) solutions of internal sucrose and external glucose^23,25. The high sucrose concentration was shown to have only a negligible effect on substrate cleavage by HH-min (Supplementary Fig. 2). To distinguish between both GUV populations, labelled lipids (Atto647N-DOPE) were added to the bilayers of the GUVs containing HH-min. We performed a direct quantification of the intravesicular fluorescence of the different vesicle populations using confocal fluorescence microscopy. After a single freeze-thaw step of the samples, which involved freezing for 30 mins in a pre-cooled metal block (-80°C) followed by incubation at room temperature, we observed a strong increase in fluorescence intensity in both GUV populations. This observation suggested that efficient RNA exchange had occurred between both vesicle populations during FT-cycling resulting in HH-min cleavage in both vesicle types (Fig. 1b). We could confirm that RNA cleavage was both HH-min and FT-dependent since samples that were either not subjected to FT-cycling or contained either no HH-min or a catalytically inactive HH variant (HH-mut²⁴, Supplementary Table 1) showed a ~ 7-fold lower fluorescence post-incubation compared to HH-min samples that were exposed to FT-cycles (Fig. 1c). To finally verify that FT-induced content exchange leads to specific HH-sub cleavage, we collected the pre- and post-cycling GUV solutions and analysed the resulting RNA species by PAGE. As expected, almost 80% of the collected HH-sub RNA had undergone site-specific cleavage by HH-min, corroborating the microscopy data that fluorescence increase is indeed a result of RNA catalysis and FT-cycling (Fig. 1d; Supplementary Fig. 3c). In contrast, omitting FT-cycling resulted only in 20% cleavage, which is most likely the result of extravesicular cleavage by leaked HH-min and HH-sub that occurred during sample preparation, e.g. through shear forces-induced rupturing of GUVs. Taken together, these data demonstrate that FT cycles do provide a means for the exchange of RNA between neighbouring protocells and can trigger catalysis of otherwise inactive ribozymes.

FT-driven assembly of active RNA ligase ribozymes in GUVs.

Next, we sought to model more prebiotically relevant ribozyme activities, namely, RNA ligation. We chose an already established modified R3C ligase system²⁶, which catalyses regiospecific 3’-5’ RNA ligation using 5’ triphosphate activation chemistry and whose requirements for Mg²⁺ ions to promote ligation are well compatible with GUV formation (Supplementary Fig. 4). This system employs the ribozyme evolved by Olea and Joyce²⁶, which allows a fluorometric readout of RNA-ligation activity (Fig. 2a). It includes a variant of the original R3C ligase (rt-F) that can use a modified 3’ RNA substrate carrying a TAMRA fluorophore reporter attached to the pyrophosphate leaving group via a 6-carbon linker at the 5’-end (FQ-B, Supplementary Fig. 1b). An additional black hole quencher (BHQ2) at the 3’-end of FQ-B suppresses fluorescence of the substrate prior to ligation. The pyrophosphate attached fluorophore is released from the substrate upon ligation of FQ-B to a short 5’ RNA substrate (rt-A-short) leading to an increase in fluorescence signal (Fig. 2a, Supplementary Fig. 5).

Similar as for the HH-min system, we tested the feasibility of the ligase ribozyme system to assemble into the active ribozyme-substrate complex by FT-driven content exchange between GUVs (Fig. 2b). Equal volumes of GUV populations were mixed, one containing the substrates to be ligated and the other containing the ligase ribozyme. After a single FT cycle of the GUVs, we observed a 1.8- and 4.4-fold increase in TAMRA fluorescence, depending on initial substrate concentration (2 µM and 10 µM, respectively). This observation indicated that RNA ligation had occurred after FT-cycling (Fig. 2c, d). Unlike the assays with the hammerhead system, we observed a substantial heterogeneity in fluorescence intensity per GUV after thawing, indicating non-homogeneous content exchange and / or post-mixing ligation. Variations in ligation efficiency at different substrate concentrations might partly result from inhibitory interactions between the positive choline headgroup of POPC and the negatively charged RNA molecules (Supplementary Fig. 6), as well as content loss that occurs during FT-cycling (~ 20%, Supplementary Fig. 7). The inhibitory effects of high lipid concentrations on catalytic RNA activity has been recently reported, concluding that the ratio of RNA to lipids as well as RNA sequence and length determine RNA-lipid interactions and, consequently, how ribozyme activity is modulated (enhancement or inhibition)²⁷. The increase in overall ligation yields at higher RNA concentrations can therefore be explained by a mitigation of both of these effects. Surprisingly, no further increase in fluorescence was detected when post-thawing GUV samples were further incubated for one hour at 37°C. This observation suggests that catalytic activity had already come to completion during sample thawing to room temperature. In summary, these proof-of-concept experiments revealed the compatibility of our model protocellular system with the catalytic activity of a ligase ribozyme.

FT cycling enables assembly and self-replication of encapsulated ligase ribozymes.

Having shown that the modified R3C system is active under FT-cycling conditions, we further set out to explore if the same system can be leveraged to establish autocatalytic RNA self-replication inside vesicles. We aimed to model a potential prebiotic scenario addressing a fundamental feature of early protocells, namely the autocatalytic replication of genetic material and its horizontal transfer via a simple, environment-driven inter-protocellular exchange mechanism. To this end, we selected an optimized self-replicating R3C enzyme, which was previously obtained by directed evolution by Robertson and Joyce⁴. In this system, the ligase ribozyme (F1) replicates itself by joining the two oligonucleotide substrates (Supplementary Fig. 1c). During an initial characterization of the F1 self-replicator under bulk conditions, we observed substantial background formation of F1 from A and B even in absence of the full-length ribozyme suggesting a residual activity of both fragments, likely resulting from their association even in the absence of F1 ribozyme (Supplementary Fig. 8, 9). We observed no background ligation of the full-length ribozyme in assays containing a variant of substrate A that was derived from the original “wildtype” ribozyme E1⁴ (before in vitro evolution of F1) and which differs only by the single transition G38A from the F1-substrate A (Supplementary Fig. 10). From here onwards, we will refer to the catalytically active F1 substrate A as Hyper-A and the inactive, E1-derived and Cy5-labelled substrate as Cy5-A. Substantial Hyper-A background ligation activity was undesirable for our model system as it might obscure F1-dependent self-replication during serial dilution experiments, especially if substrates are encapsulated in the same GUV population. However, using Cy5-A as a reporter substrate, we found that, at Hyper-A concentrations below or equal to 1 µM and in the absence of any initial full-length F1, only ~ 12% of ligation product was formed after 60 min of incubation at 37°C. In contrast, the addition of 1 µM “seed” F1 ribozyme led to an almost complete ligation (~ 90%) of Cy5-A and B already after 30 min (Supplementary Fig. 11–12).

After establishing suitable self-replication activities of F1 under bulk conditions, we probed whether F1 self-replication could be triggered by FT-driven exchange of substrate containing GUVs with vesicles encapsulating the full-length ribozyme. Specifically, we aimed to explore if R3C-based replicators can “propagate” in vesicle populations through repeated “infection” of substrate filled GUVs driven by FT-cycling. To this end, we mixed equal amounts of vesicles containing 1 µM F1 with vesicles containing the two substrates. To exclude Hyper-A based background ligation before content mixing, we separated the substrates Hyper-A (10 µM with 10% Cy5-A as reporter strand) and B (15 µM) in different GUV populations. As expected, no ligation occurred during co-incubation of the three populations at 42°C. However, when the system was exposed to repeated FT-cycling, we observed increasing levels of F1 self-ligation from Hyper-A and B (Fig. 3b; Supplementary Fig. 13): After one FT cycle, ~ 2.6 µM of F1 was present inside the GUV population, and further cycling of the same samples increased concentrations to ~ 3.2 µM and ~ 3.7 µM after 2 and 3 cycles, respectively. We could show that the majority of the activity resulted from content mixing of F1 with both of its substrates, as GUV mixtures in which F1-containing GUVs had been replaced with empty vesicles showed only 0.8–1.0 µM of F1 following the FT cycles (Fig. 3b). The observed correlation between the number of FT cycles and ribozyme concentration in systems with F1 containing GUVs suggested that repeated cycles enhanced content exchange between the different populations and increased the total fraction of ligated product despite the concurrent content loss that occured during transient membrane disruption and vesicle fragmentation (Supplementary Figs. 7 and 14).

FT-cycling enables sustained self-replication of GUV-encapsulated ligase ribozymes.

After identifying suitable RNA concentrations and the optimal FT-cycling protocol (10-fold excess substrates; three FT cycles followed by an hour at 42°C), we aimed to explore whether R3C-based replicators can persist in vesicle populations during serial transfer experiments by escaping selective pressure in an open system through repeated “infection” of substrate filled GUVs driven by FT-cycling. Specifically, we set out to leverage the combination of FT cycle-induced content mixing and autocatalytic replication to simulate survival of a ribozyme under serial transfer conditions, i.e. when vesicles containing self-replicators are challenged with ongoing dilution. Such a setting implements a selection pressure for an autocatalytic system at the most basic level, as “death” by dilution can only be counteracted by a sufficient level of autocatalysis⁸.

In these experiments, GUVs containing an initial full-length F1 starting concentration (1 or 2 µM) were mixed in equal amounts with separate vesicles containing either 10 µM Hyper-A (10% Cy5-A) or 15 µM B. To maximize content mixing, the tripartite vesicle population was then subjected to three consecutive FT cycles followed by an incubation period of 1 hour at 42°C. Subsequently, half of the resulting vesicles were combined with an equal volume of fresh substrate vesicles and the entire process was repeated for several generations (Fig. 4a). When the system was seeded with GUVs containing 1 µM F1, we observed a strong increase in encapsulated F1 levels reaching almost ~ 3.8 µM after the first FT cycle and almost complete regeneration of this concentration (~ 3.4 µM) after the first transfer followed by three FT cycles and 1 hour at 42°C (Fig. 4b). Further generations showed a stepwise decrease of F1-levels down to about 1.5 µM after the sixth transfer. To investigate if higher seed-concentrations might stabilize F1-levels at higher steady state concentrations, we increased the initial seed concentrations of F1 to 2 µM and repeated the experiment. While the initial boost in F1 yields after the first cycle were similar to 1 µM starting concentration, the system behaved quite differently in further generations. Already after four generations, F1 levels reached a stable steady-state concentration of ~ 2.8 µM (post incubation) suggesting that, under these conditions, self-sustained intravesicular self-replication was indeed successful (Fig. 4b; Supplementary Fig. 15).

The approach of a stable equilibrium between ribozyme self-replication and F1 loss due to the serial transfers as well as vesicle leakage was even more evident when the post-cycling concentration was plotted against the dilution of the initial F1 seed concentration (Fig. 4c). Whilst increasing amounts of dilution challenged complete regeneration of ribozyme throughout the experiment for 1 µM F1 starting concentration, the initiation with 2 µM changed the picture completely: Here, F1 levels post-cycling plateaued at a steady state concentration of ~ 3 µM after as few as 3 cycles and the system enabled ongoing regeneration of encapsulated F1 in all further dilutions. This striking difference between 1 µM and 2 µM initial F1 concentration in later generations is somewhat surprising given the similar concentrations of F1 after the three initial FT cycles. One explanation for this observation is that a higher initial F1 concentration allows a more even distribution of F1 into the substrate vesicles. Such a more homogeneous spread of F1 into substrate vesicles might then facilitate survival under conditions where 50% of the vesicles get removed after every iteration. We could generally exclude that the F1 steady state concentrations are simply the result of Hyper-A background activity followed by ongoing F1 accumulation: In control samples with empty (0 µM) F1 GUVs, F1-formation became nearly FT-cycle independent and never exceeded ~ 0.5 µM (Fig. 4b-c). Thus, although Hyper-A-dependent F1 formation could be observed, it was apparently not sufficient to achieve a threshold level of F1 where self-replication can counteract content loss and serial dilution. In summary, it was possible in our model system to generate an encapsulated auto-replicator capable of surviving multiple generations of serial dilution by propagating amongst vesicles via environment-induced horizontal RNA transfer.

In this work, we sought to investigate how primordial protocells, void of proteins, may have survived in a plausible geochemical environment on early Earth. Inspired by previous work combining nucleic acid catalysis and membrane vesicles, we introduced an encapsulated autocatalytic enhancing ribozyme system into a geological scenario exhibiting cyclic freezing and melting with transient eutectic ice formation. The freezing of salted water leads to the formation of interstitial brines amongst the growing ice crystals, which concentrates dilute solutes and ions in solution²⁸. For this reason, liquid brine phases in water-ice matrices are attractive candidates for various prebiotically relevant environmental niches as they favour reactions that are otherwise incompatible with dilute aqueous conditions. For example, UV-irradiation of urea solutions followed by freeze-thaw cycles enables prebiotic synthesis of purines and pyrimidines²⁹. Frozen conditions have also been shown to promote both nucleotide activation and non-enzymatic templated copying of RNA³⁰, as well as host more complex ribozyme reactions that increase genotype / phenotype complexity of the starting RNA pools^31–33. Even a plausible pathway for the synthesis of peptide-RNA is promoted by water-ice formation and the pH change that accompanies it³⁴. Milder temperature fluctuations that did not necessarily involve phase-transitions in the bulk solvents might have also been of importance for early prebiotic processes. For example, temperature-induced membrane phase transitions in fatty acid vesicles enabled the generation of “daughter protocells” with reshuffled content²⁰. Whilst more complex than fatty acid membranes, the superior stability of phospholipid membranes across pH, salt and temperature gradients makes them readily compatible with encapsulated nucleic acid catalysis^{18,25, 35–37} (Supplementary Fig. 4). As shown in this study, transient temperature changes can enable the exchange of intravesicular components amongst other vesicles in the vicinity. Since content exchange relies on the membrane instability during the phase transition temperature of lipids²³, phospholipids with higher melting temperatures may also enable content exchange at temperatures above the freezing point of water. Our findings support the argument that temperature fluctuations, including those involving eutectic ice formation, provide a suitable environment to drive a variety of reactions relevant to the emergence and evolution of life, potentially supporting a continuous path from the prebiotic chemical synthesis of monomers, their subsequent activation and non-enzymatic polymerization to the point where they allow tertiary folds and associated catalysis including – but not limited to – autocatalytic replication. We succeeded in developing a system exhibiting life-like behaviours with the aid of environmental fluctuations: the oscillation of ribozyme concentration in the vesicle population in response to dilution and freezing / thawing. The simultaneous synthesis of nucleic acid precursors and phospholipids in an environment supporting the cyclical freezing and thawing of water could have resulted in compartment formation and encapsulation of molecules³⁸, while exchanging resources with the outer environment and other protocells by means of the temperature-induced membrane permeabilization.

In vitro transcription. ssDNA oligonucleotides were ordered from Integrated DNA Technologies (IDT). RNA was either ordered from IDT or transcribed in vitro (Supplementary Table 1). Briefly, a partially dsDNA template with double-stranded T7 promoter region was created by incubating equimolar amounts of the DNA oligomers at 85°C and then placing them on ice. In vitro transcription (IVT) reactions (100 µL final volume) were carried out as follows: 1 µM DNA template, 50 mM TRIS pH 7.8, 30 mM MgCl₂, 5 mM of each NTP (Jena Bioscience), 10 mM DTT, 2 mM Spermidine (Sigma-Aldrich), 0.1 U E. coli Inorganic Pyrophosphatase (NEB), and 0.5 µM of homemade T7 polymerase (purified from a recombinant source in our laboratory). The reaction was incubated at 37°C for 4–6 hours. After column purification (Monarch RNA Cleanup Kit, NEB), the transcribed RNAs were gel purified after 12–20% PAGE. The transcript band was visualized by UV shadowing on a TLC plate, excised, crushed in a 2 mL tube with a syringe plunger and soaked overnight in 2 v/w of 0.3M NaOAc pH 5.2 at 4°C on a rotator. Gel debris was removed by spin filtration in 0.45 µm Spin-X cellulose acetate filters (Costar) for 2 minutes at 17,000 g, 4°C. The eluted RNA was precipitated in 1.2 volumes of cold isopropanol, cooled 45–60 minutes at -20°C and centrifuged for 90 minutes at 21,000 g, 4°C. The supernatant was discarded and the pellet was washed with 1 mL of 80% ethanol and centrifuged another 20 minutes at 21,000 g, 4°C. The pellet was suspended in ultrapure water and quantified on a Nanodrop measuring absorbance at 260 nm using their specific extinction coefficient calculated using OligoCalc (http://biotools.nubic.northwestern.edu/OligoCalc.html).

Synthesis of FQ-B Substrate. TAMRA-NHS-Ester was bought from Sigma. gamma-aminohexyl-GTP was bought from Jena Bioscience. A 5’-phosphorylated pentamer with a 3’-end quencher (BHQ2) was ordered from IDT. The substrate was first in vitro transcribed with the modified initiator nucleotide. The reaction differed from the IVT described above by having 6 mM ATP, UTP and CTP, 2 mM GTP, and 4 mM gamma-aminohexyl-GTP. The resulting RNA was column purified (Monarch RNA Cleanup Kit) and reacted with the TAMRA-NHS-ester according to manufacturer’s instructions. Briefly, around 60 ODs of the 5’ aminohexyl RNA was dried under vacuum and resuspended in 600 µL sodium tetraborate pH 8.4 then combined with 100 µL of TAMRA-NHS-ester in DMSO at 10 µg/µL of fluorophore and incubated for 4 hours in a dry block at 25°C, 600 rpm. The RNA was precipitated in 1.2 volumes of isopropanol, washed once with 80% ethanol, resuspended in ultrapure water and quantified on a Nanodrop as described above. Subsequently, the fluorophore-tagged RNA was ligated to the 3’ quencher-tagged pentamer with T4 RNA ligase 2 (NEB) and a DNA splint complementary to the 15 nucleotides from the 3’-end. The reaction setup (200 µL – 1 mL) was as follows: 3 µM of TAMRA-B, 3.5 µM Pentamer, 3.5 µM DNA splint, 8% DMSO, 1X T4 RNA ligase 2 buffer, 9 mM MgCl₂, 100 µM ATP, 50 U T4 RNA ligase 2, 15% PEG8000. The nucleic acids were mixed with water and DMSO, heated to 85°C and placed on ice for annealing. The remaining reagents were added in the order listed. The reaction was incubated at 25°C for 1 hour, then at 16°C for 16 hours. After isopropanol precipitation, the pellet was suspended in Formamide and gel purified on a 20% PAGE. Recovery and quantification of RNA was as described above.

Preparation of GUVs for microscopy. POPC lipids dissolved in chloroform were added to a glass vial containing chloroform to obtain 250 µL of a 6 mM solution. Fluorescent GUV membranes were obtained by adding 0.5 µL of a 1 mg/mL solution of Atto 647N-DOPE lipids in chloroform (~ 0.03 mol%). The chloroform was evaporated under nitrogen flow (10 min) and the lipids were dried under reduced pressure for 1 h to remove residual traces of chloroform. A lipid-mineral oil solution (400 µM lipids) was prepared by adding 3 mL of mineral oil to the dried lipids followed by sonication for 1 h at elevated temperature (40–60°C). 750 µL of lipid-mineral oil solution was layered on top of 2 mL of outer phase solution in a 5 mL Eppendorf tube. Incubation for 15 min at RT allows lipid monolayer formation at the water-oil interface. Meanwhile, 750 µL of lipid-mineral oil solution and 15 µL of inner phase solution were combined in a 1.5 mL Eppendorf tube. A water-in-oil emulsion was generated by rubbing the tube over an Eppendorf rack. The water-in-oil emulsion was immediately and carefully pipetted to the biphasic mixture in the 5 mL Eppendorf tube. To generate lipid GUVs, gradual centrifugation (10 min at 300 g followed by 2.5 min at 1500 g) was performed to push the denser inverse micelles (surrounded by a lipid monolayer) through the interfacial lipid monolayer. Vesicle pellets at the tube bottom were visible by eye and could thus be easily taken up with a pipette. A 384-glass bottom microtiter plate was first passivated with 50 µL of a Pluronic F-127 solution (10 mg/mL) and then washed once with 50 µL of outer phase solution. Next, the pellet and outer phase solution (~ 50 µL) were transferred to the plate. Another GUV population was added and mixing of both populations was achieved by gently pipetting up and down (10 times). Before freeze-thawing, the plate was centrifuged for 5 min at 1500 g for GUV accumulation and high GUV densities. All steps including mineral oil were performed in a custom-made glove box under reduced humidity (< 10% relative humidity) except for the centrifugation step for GUV formation.

Freezing and thawing of well plate for microscopy. Freezing and thawing was performed at high cooling and low heating rates for all experiments. Before sample freezing, glass bottom microtiter plates were sealed with the aluminium-coated sealing film ROTALIBO®. A block made of stainless steel pre-cooled in a -80°C freezer was pressed against the glass bottom leading to sample freezing within a few seconds. Plates were then stored at -80°C for up to 30 min, taken out and thawed at RT while holding them at an angle to avoid GUV dispersal, thus maintaining high GUV densities required for an efficient content exchange. In time-lapse experiments (Supplementary Fig. 14b-d), plates remained attached to the microscope stage during freeze-thawing. Here, fast-freezing was achieved by liquid nitrogen poured into the wells surrounding the sample well. Condensation on the glass bottom was avoided by blowing in N₂ gas from below to enable sample access at sub-zero temperatures with a dry objective. Sample thawing at RT started once the liquid nitrogen was fully evaporated.

Preparation of GUVs for PAGE assays with freezing and thawing. POPC lipids dissolved in chloroform were dried under nitrogen flow, and placed under vacuum for 1 hour to remove traces of chloroform. The dried lipids were subsequently dissolved in light mineral oil (Carl Roth) to obtain a final concentration of 400 µM, incubated at 60°C for 5 minutes then vigorously vortexed for 1 minutes, and finally sonicated for 1 hour at 40–60°C in a water bath to disperse aggregates and completely resuspend dried lipids. Next, 1.5 mL tubes were prepared with 600 µL of outer phase containing ribozyme buffer (as specified per experiment) and 900 mM glucose, layered with 300 µL of lipids in oil suspension and was allowed to rest for 30 minutes to properly form the lipid monolayer at the oil-water interface. The emulsion was prepared by adding 24 µL inner phase solution (containing 900 mM sucrose) to 400 µL of lipids in oil solution and mechanically agitating the tube over a tube rack until the suspension turned milky white. The emulsion was carefully transferred to the oil layered on the outer phase, and spun at 300 g for 5 minutes. The vesicle pellet was recovered by piercing the bottom of the open tube with a needle and ejecting it into a clean tube by closing the cap. GUV populations were united, 700 µL fresh glucose outer phase was added and the vesicles were spun at 2000 g for 5 minutes. Most of the supernatant was discarded. The pellet was resuspended in ~ 25 µL buffered glucose outer phase, and 10 µL samples were placed in PCR tubes and spun again at 2000 g for 5 minutes to form a vesicle pellet. Control samples were held at room temperature (RT) and the rest were subjected to cycles of freezing and thawing as follows: 5 minutes in aluminum block at -80°C, 5 minutes at room temperature, and incubation at the desired reaction temperature. When mentioned, the freeze thaw cycle was applied multiple times.

RNA recovery from GUVs and PAGE analysis. To recover the encapsulated RNA from GUVs and analyze activity via polyacrylamide gel electrophoresis, the following protocol was employed: samples of 5–10 µL were added to 500 µL of unbuffered glucose outer phase, mixed by inverting the tube and spun at 2000 g for 5 minutes. 450 µL of supernatant was then replaced with an equal volume of fresh glucose outer phase. This wash step serves to exchange the outer phase and remove any biomolecules or salts that leaked from GUVs into the outer phase. After exchanging the outer phase, the GUVs were pelleted by spinning at 2000 g for 5 minutes, the supernatant was discarded leaving only around 10 µL of pellet, and 10–20 µL of GUV loading buffer was added (containing 25 mM EDTA, 0.3% Triton X-100, 0.01% bromophenol blue, 95% formamide) and vortexed vigorously for 5–10 seconds. Finally, the samples were denatured by incubating for 5 minutes at 85°C and placed on ice. A few microliters were then loaded on a 12–20% denaturing urea-PAGE and run for 60 minutes at a constant 25 W. Gels were imaged immediately after the run on the Azure RGB Sapphire and analyzed using the Azure Spot software.

Hammerhead assay in solution. Reactions were carried out in volumes of 50 µL. Substrate cleavage was monitored in a reaction containing 5 µM hammerhead ribozyme, 2.5 µM Cy5-tagged substrate in hammerhead buffer that consisted of 20 mM Tris-HCl pH 8.3, 900 mM sucrose, and 4 mM MgCl₂. The reaction was set up on ice and started by the addition of MgCl₂ and incubating at 25°C. Samples were taken at specific time points by quenching 5 µL of reaction with 20µL of hammerhead loading buffer containing 10 mM EDTA, 0.01% bromophenol blue, 98% formamide and placed on ice. After all samples were collected, they were denatured at 85°C for 5 minutes and cooled quickly on ice before loading onto a 20% denaturing urea-PAGE. Gel imaging and analysis were performed as described above. Data was fitted to an exponential plateau function on GraphPad Prism.

Hammerhead assay in GUVs for microscopy. Two GUV populations encapsulating the hammerhead buffer (900 mM sucrose, 20 mM Tris-HCl pH 8.3, 4 mM MgCl₂) were prepared and combined in a 1:1 ratio as described above. One population contained the substrate (HH-FQ-sub − 2.5 µM) and the other membrane-labelled population contained either minimal hammerhead ribozyme (HH-min − 5 µM), the mutated ribozyme (HH-mut − 5 µM) or no ribozyme (negative control - NC). The outer phase consisted of hammerhead buffer and 900 mM of glucose to ensure both a density gradient and isosmotic conditions between inner and outer phase. Sample freezing and thawing was applied as described above. Images were taken at RT before and after freeze-thawing.

Hammerhead assay in GUVs with PAGE analysis. GUVs encapsulating the hammerhead ribozyme were prepared as described above with the hammerhead buffer (900 mM sucrose, 20 mM Tris-HCl pH 8.3, 4 mM MgCl₂) and 5 µM hammerhead ribozyme. Substrate GUVs were also prepared the same way except that it included 2.5 µM Cy5-tagged substrate RNA instead of the ribozyme. The outer phase was the hammerhead buffer but with 900 mM of glucose rather than sucrose, which is necessary for the density gradient. The two GUV populations were united and washed once with 700 µL of buffered outer phase followed by spinning at 2000 g for 5 minutes. The volume was reduced to ~ 25 µL and split into two samples of 10 µL each. One sample was a control held at room temperature while the other was subjected to a cycle of freezing and thawing. Both samples were subsequently incubated at 37°C for 1 hour before the GUVs were washed, the RNA was recovered from the GUVs and analyzed by 20% PAGE, as described above.

R3C ligase assay in GUVs for microscopy. GUVs encapsulating the R3C buffer (900 mM sucrose, 50 mM EPPS pH 8.5, 20 mM MgCl₂) were prepared as described above. One population was membrane-labelled and contained the ribozyme rt-F, the other contained the substrates rt-A-short and FQ-B. Ribozyme and substrates were encapsulated at varying concentrations in two experimental approaches. Concentrations were either 20/15/10 µM or 5/5/2 µM (rt-F/rt-A-short/FQ-B, respectively). The outer phase consisted of 900 mM of glucose and R3C buffer. Freezing and thawing was implemented as described above. Images were taken at RT before and after freeze-thawing. Further images were taken after a subsequent 1 h incubation at 37°C.

R3C autocatalytic ligase assay in solution. To determine the activity of the R3C ligase ribozyme in vitro, only the Cy5-tagged inactive A substrate was used as a reporter for ligation yields. 10 µL reactions were prepared on ice, with 5 µM Cy5-A.,7.5 µM B, and varying concentrations of either ribozyme F1 or the unmodified substrate Hyper-A. The R3C buffer consisted of 50 mM EPPS pH 8.5, 900 mM sucrose, and 20 mM MgCl₂. The reaction was started by the addition of MgCl₂ and incubating it at 42°C. Samples were consistently taken at specific time points by quenching 1 µL of the reaction with 9 µL of R3C RNA loading buffer (containing 25 mM EDTA, 95% formamide and 0.01% bromophenol blue) and placed on ice. After all samples were collected, they were denatured at 85°C for 5 minutes and immediately placed on ice, before loading onto a 12% denaturing urea-PAGE. Gel imaging and analysis were performed as described above. The data points of the ribozyme F were fitted to an exponential plateau function while the data points of the Hyper-A substrate were fitted to an exponential growth function on GraphPad Prism.

FT-dependent activity of R3C ligase. The different GUV populations were prepared independently as described above. Briefly, ribozyme GUVs encapsulated 1 µM of ribozyme F1 in R3C buffer (50 mM EPPS, 20 mM MgCl₂, 900 mM sucrose), substrate A GUVs contained a mixture of 9 µM Hyper-A and 1 µM Cy5-A in R3C buffer, substrate B GUVs encapsulated 15 µM of B in R3C buffer, empty GUVs were used as a negative control and contained only R3C buffer with no RNA. The outer phase also contained the R3C buffer but with 900 mM glucose instead of sucrose. The GUV populations were combined in a 1:1:1 ratio to obtain a starting sample with both substrate vesicles and either ribozyme GUVs or empty GUVs as the control. The GUVs were washed once with 700 µL buffered glucose outer phase and centrifuged at 2000 g for 5 minutes. The supernatant was discarded, and the pellet was resuspended in ~ 25 µL buffered glucose outer phase. 10 µL samples were aliquoted in PCR tubes and subjected to different numbers of FT cycles (none, one, two or three successive cycles) followed by a 1-hour incubation at 42°C. Samples were subsequently washed twice in 500 µL unbuffered glucose and prepared for 12% denaturing urea PAGE as described above.

Serial dilution of autocatalytic R3C ligase GUVs. The GUV populations were prepared as above, but a population of only substrate vesicles (both A and B) was also prepared, to be used as a fresh feedstock. After GUV populations were combined (1:1:1), washed and centrifuged, they were subjected to 3 FT cycles, followed by a 1-hour incubation period at 42°C. Then, settled GUVs were resuspended, a 5 µL sample was taken and washed in 500 µL unbuffered glucose outer phase twice and subsequently resuspended in 10 µL of the GUV loading buffer. To the remaining 5 µL of GUVs, 5 µL of fresh substrate GUV feedstock was added. The new generation was centrifuged at 2000 g for 5 minutes to pellet the GUVs and subjected again to 3 FT cycles followed by incubation. This process was repeated for several generations. After all generations were sampled, they were heat denatured and loaded on a 12% denaturing urea-PAGE and analyzed as above. Data points were fitted to a single exponential decay function on GraphPad Prism for descriptive purposes.

Laser scanning confocal fluorescence microscopy. All images were taken on a laser scanning microscope LSM 780 (Carl Zeiss, Germany) equipped with a water immersion objective (C-Apochromat 40 x / 1.20 W) for acquisition at room temperature before and after FT, or a dry objective (Plan-Apochromat 40 x / 0.95) for time-lapse imaging during FT-cycling. Samples were excited at 488 nm for FAM / Atto-488, 561 nm for TAMRA/Alexa Fluor-568 excitation and 633 nm for Atto-647N. At room temperature, colours were excited separately to avoid cross-talk whilst time-lapse experiments required simultaneous illumination. Images were typically recorded using a 1 airy unit pinhole, a resolution of 512 x 512 pixels and pixel dwell times ranging from 12–25 µs. Time-lapse imaging was typically performed with 3 µs pixel dwell times and 1 s intervals. For tile-scan imaging, a group of individual adjacent image fields (tiles) was recorded and stitched together by the Zeiss software (Zen Black, v13.0.2.518).

Image analysis. All confocal images shown were analysed using Fiji (v1.53c). For visual presentation, images were merged from separately recorded channels and adjusted for brightness and contrast. Brightness and contrast adjustments were applied homogenously for all acquired images. Image stacks derived from time-lapse experiments were analysed by defining regions of interest using the built-in ROI manager. Data processing and plotting was performed using OriginPro (2020b, v9.7.5.184).

Acknowledgments: Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 364653263 – TRR 235, Project P14 (H.M and P.S.). H.M. is grateful for support from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 802000, RiboLife).

Author Contributions: PS and HM conceived the project. ES and BP designed experiments, collected the data, performed the analysis. All authors wrote the paper. All authors read and approved the final manuscript for publication.

Competing Interest Statement: The authors declare no competing interests.

Materials and Correspondence:

[email protected], [email protected]

Szostak, J. W., Bartel, D. P. & Luisi, P. L. Synthesizing life. Nature vol. 409 387–390 (2001).
Joyce, G. F. & Szostak, J. W. Protocells and RNA Self-Replication. Cold Spring Harb. Perspect. Biol. 10, a034801 (2018).
von Kiedrowski, G. A Self-Replicating Hexadeoxynucleotide. 25, 932–935 (1986).
Robertson, M. P. & Joyce, G. F. Highly efficient self-replicating RNA enzymes. Chem. Biol. 21, 238–245 (2014).
Lee, D. H., Granja, J. R., Martinez, J. A., Severin, K. & Ghadiri, M. R. A self-replicating peptide. Nature 382, 525–528 (1996).
Breslow, R. On the mechanism of the formose reaction. Tetrahedron Lett. 1, 22–26 (1959).
Tawfik, D. S. & Griffiths, A. D. Man-made cell-like compartments for molecular evolution. Nat. Biotechnol. 16, 652–656 (1998).
Eigen, M. Selforganization of matter and the evolution of biological macromolecules. Naturwissenschaften 58, 465–523 (1971).
Nooner, D. W. & Oro, J. Synthesis of Fatty Acids by a Closed System Fischer-Tropsch Process. in Hydrocarbon Synthesis from Carbon Monoxide and Hydrogen. 178, 159–171 (American Chemical Society, 1979).
Anella, F. & Danelon, C. Reconciling Ligase Ribozyme Activity with Fatty Acid Vesicle Stability. Life 4, 929–943 (2014).
Adamala, K. P., Engelhart, A. E. & Szostak, J. W. Collaboration between primitive cell membranes and soluble catalysts. Nat. Commun. 7, 1–7 (2016).
Fayolle, D. et al. Crude phosphorylation mixtures containing racemic lipid amphiphiles self-assemble to give stable primitive compartments. Sci. Rep. 7, (2017).
Bonfio, C. et al. Length-Selective Synthesis of Acylglycerol-Phosphates through Energy-Dissipative Cycling. J. Am. Chem. Soc. 141, 3934–3939 (2019).
Liu, L. et al. Enzyme-free synthesis of natural phospholipids in water. Nat. Chem. 12, 1029–1034 (2020).
Fiore, M. et al. Synthesis of Phospholipids Under Plausible Prebiotic Conditions and Analogies with Phospholipid Biochemistry for Origin of Life Studies†. Astrobiology 22, 598–627 (2022) doi:10.1089/AST.2021.0059.
Ochman, H., Lawrence, J. G. & Grolsman, E. A. Lateral gene transfer and the nature of bacterial innovation. Nature 405, 299–304 (2000).
Woese, C. R. On the evolution of cells. Proc. Natl. Acad. Sci. 99, 8742–8747 (2002).
Sarkar, S., Dagar, S., Verma, A. & Rajamani, S. Compositional heterogeneity confers selective advantage to model protocellular membranes during the origins of cellular life. Sci. Rep. 10, 1–11 (2020).
Morasch, M. et al. Heated gas bubbles enrich, crystallize, dry, phosphorylate and encapsulate prebiotic molecules. Nat. Chem. 11, 779–788 (2019).
Rubio-Sánchez, R. et al. Thermally Driven Membrane Phase Transitions Enable Content Reshuffling in Primitive Cells. J. Am. Chem. Soc. 143, 16589–16598 (2021).
Joshi, M. P., Sawant, A. A. & Rajamani, S. Spontaneous emergence of membrane-forming protoamphiphiles from a lipid–amino acid mixture under wet–dry cycles. Chem. Sci. 12, 2970–2978 (2021).
Tsuji, G., Fujii, S., Sunami, T. & Yomo, T. Sustainable proliferation of liposomes compatible with inner RNA replication. Proc. Natl. Acad. Sci. U.S.A: 113, 590–595 (2016).
Litschel, T. et al. Freeze-thaw cycles induce content exchange between cell-sized lipid vesicles. New J. Phys. 20, (2018).
Drobot, B. et al. Compartmentalised RNA catalysis in membrane-free coacervate protocells. Nat. Commun. 9, 1–9 (2018).
Moga, A., Yandrapalli, N., Dimova, R. & Robinson, T. Optimization of the Inverted Emulsion Method for High-Yield Production of Biomimetic Giant Unilamellar Vesicles. Chembiochem 20, 2674 (2019).
Olea, C. & Joyce, G. F. Real-Time detection of a self-replicating RNA Enzyme. Molecules 21, (2016).
Czerniak, T. & Saenz, J. P. Lipid membranes modulate the activity of RNA through sequence-dependent interactions. Proc. Natl. Acad. Sci. U. S. A. 119, (2022).
Menor-Salván, C. & Marín-Yaseli, M. Prebiotic chemistry in eutectic solutions at the water–ice matrix. Chem. Soc. Rev. 41, 5404–5415 (2012).
Menor-Salván, C. & Marín-Yaseli, M. R. A New Route for the Prebiotic Synthesis of Nucleobases and Hydantoins in Water/Ice Solutions Involving the Photochemistry of Acetylene. Chem. – A Eur. J. 19, 6488–6497 (2013).
Zhang, S. J., Duzdevich, D., Ding, D. & Szostak, J. W. Freeze-thaw cycles enable a prebiotically plausible and continuous pathway from nucleotide activation to nonenzymatic RNA copying. Proc. Natl. Acad. Sci. 119, (2022).
Attwater, J., Wochner, A., Pinheiro, V. B., Coulson, A. & Holliger, P. Ice as a protocellular medium for RNA replication. Nat. Commun. 1, 1–9 (2010).
Vlassov, A. V., Johnston, B. H., Landweber, L. F. & Kazakov, S. A. Ligation activity of fragmented ribozymes in frozen solution: implications for the RNA world. Nucleic Acids Res. 32, 2966–2974 (2004).
Mutschler, H., Wochner, A. & Holliger, P. Freeze–thaw cycles as drivers of complex ribozyme assembly. Nat. Chem. 7, 502–508 (2015).
Roberts, S. J., Liu, Z. & Sutherland, J. D. Potentially Prebiotic Synthesis of Aminoacyl-RNA via a Bridging Phosphoramidate-Ester Intermediate. J. Am. Chem. Soc. 144, 4254–4259 (2022).
Monnard, P. A., Apel, C. L., Kanavarioti, A. & Deamer, D. W. Influence of Ionic Inorganic Solutes on Self-Assembly and Polymerization Processes Related to Early Forms of Life: Implications for a Prebiotic Aqueous Medium. Astrobiology 2, 139–152 (2004).
Budin, I. & Szostak, J. W. Physical effects underlying the transition from primitive to modern cell membranes. Proc. Natl. Acad. Sci. U. S. A. 108, 5249–5254 (2011).
Jin, L., Kamat, N. P., Jena, S. & Szostak, J. W. Fatty Acid/Phospholipid Blended Membranes: A Potential Intermediate State in Protocellular Evolution. Small 14, e1704077 (2018).
Costa, A. P., Xu, X. & Burgess, D. J. Freeze-Anneal-Thaw Cycling of Unilamellar Liposomes: Effect on Encapsulation Efficiency. Pharm. Res. 31, 97–103 (2014).

There is NO Competing Interest.

SalibiSI.docx
Supplementary Information

Download PDF

Journal Publication

published 03 Mar, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Exchange, catalysis and amplification of encapsulated RNA driven by periodic temperature changes

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1