Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Mar 19;146(13):8887–8894. doi: 10.1021/jacs.3c10813

High-Fidelity RNA Copying via 2′,3′-Cyclic Phosphate Ligation

Adriana Calaça Serrão †, Sreekar Wunnava †, Avinash V Dass †,‡, Lennard Ufer †, Philipp Schwintek †, Christof B Mast †, Dieter Braun †,*
PMCID: PMC10995993  PMID: 38503430

Abstract

graphic file with name ja3c10813_0006.jpg

Templated ligation offers an efficient approach to replicate long strands in an RNA world. The 2′,3′-cyclic phosphate (>P) is a prebiotically available activation that also forms during RNA hydrolysis. Using gel electrophoresis and high-performance liquid chromatography, we found that the templated ligation of RNA with >P proceeds in simple low-salt aqueous solutions with 1 mM MgCl2 under alkaline pH ranging from 9 to 11 and temperatures from −20 to 25 °C. No additional catalysts were required. In contrast to previous reports, we found an increase in the number of canonical linkages to 50%. The reaction proceeds in a sequence-specific manner, with an experimentally determined ligation fidelity of 82% at the 3′ end and 91% at the 5′ end of the ligation site. With splinted oligomers, five ligations created a 96-mer strand, demonstrating a pathway for the ribozyme assembly. Due to the low salt requirements, the ligation conditions will be compatible with strand separation. Templated ligation mediated by 2′,3′-cyclic phosphate in alkaline conditions therefore offers a performant replication and elongation reaction for RNA on early Earth.

Introduction

Nucleic acid replication is essential for the propagation of genetic information and therefore is a central step for the origin of life.1 An early form of molecular evolution that preceded catalytic polymers (i.e., enzymes and ribozymes) required a nonenzymatic copying mechanism. While primer extension by the addition of mono-, di-, or tri-nucleotides has been demonstrated to copy shorter sequences,2−5 its processivity is limiting, and the combination with strand separation is a considerable hurdle. A simple way to overcome these issues could be templated ligation, which reduces the number of steps-per-length required to generate an oligonucleotide, and if possible, operating at low magnesium concentration would offer an easier strand separation, particularly when coupled to nonequilibrium microenvironments with salt and pH cycling.6 Additionally, ligation chain reactions are known to offer exponential replication.7

The state-of-the-art method for templated ligation makes use of phosphoramidates, such as phosphorimidazolides.8,9 The need for a separate (ex situ) presynthesis step with condensing agents, coupled with their short half-life, reduces their prebiotic likelihood. Furthermore, in situ activation with prebiotically plausible organocatalysts in ligation-compatible scenarios has not been shown. Disregarding the need for multiple synthesis steps required to make the phosphorimidazolides,10,11 imidazole-activated oligonucleotides are less reactive than their mononucleotide counterparts, lowering the yield of templated ligation compared to that of polymerization.8 Moreover, studies demonstrating the assembly of a long catalytic RNA by templated ligation of imidazole-activated oligonucleotides required a high concentration of Mg2+, which leads to product inhibition.8,9,12 This strongly supports the need for an efficient ligation system compatible with strand separation.

The quest for such a system led us to cyclic phosphates since they generate short oligomers in the dry state,13−15 which retain the active >P ends and could act as raw material for ligation. More importantly, they represent a simple and endogenous activated group, minimizing the need for complex multistep synthesis and ex situ activation. 2′,3′-Cyclic phosphate (>P) endings are likely readily available in the prebiotic pool since they are the primary product of prebiotic nucleotide synthesis16 and phosphorylation reactions.17,18 Moreover, ribozymatic19 and alkaline hydrolysis of RNA strands via transesterification20−22 produce >P ends, which are substrates for ribozymes catalyzing phosphodiester bond formation.23,24

Hydrolysis of >P results in 2′- or 3′-monophosphate, the recycling of which, i.e., the recyclization to activated oligonucleotides with >P, under prebiotic conditions was also demonstrated,17,18,23 suggesting a way for the in situ recycling of hydrolyzed substrates. Biochemical protocols using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) at low temperatures are also common.24−26 The widespread presence of these activated phosphate groups in established prebiotic pathways and their relatively longer half-life motivated us to investigate their role in ligating RNA oligomers in a templated setting and their relevance to early RNA replication.

Previous work with >P containing oligonucleotides has demonstrated template copying only through ligation with DNA/RNA chimeras, resulting in low yield and a predominance of 2′–5′ linkages.26,27 Nontemplated ligation of random sequences with >P has also been shown to proceed in eutectic phase at low rates.28 However, the potential of ligation of RNA sequences with >P ends for quantitative genetic copying remained largely unexplored.

Our study explored the ligation of a completely RNA-based system in simple conditions devoid of additional organocatalysts. We investigated the impact of reaction conditions on yield, kinetics, and sequence fidelity, achieving a maximum yield at 5 °C and pH 10. This was achieved under reduced salt conditions, enabling compatibility of the settings with strand separation. Moreover, we observed an increased ratio of 50% of the canonical 3′–5′ linkages at the ligation site, an improvement over previously reported aqueous condensations involving >P.26,29,30 Additionally, we demonstrated that the reaction is highly sequence-selective, even for a single nucleotide at the ligation site. Finally, we could demonstrate that multistep ligations within a splinted RNA system using >P can generate long RNA molecules on the length scale of 100 nucleotides through a cross-templating reaction.

This work serves as a proof of principle for nonenzymatically replicating and generating long RNA, in a sequence-specific manner using a simple and ubiquitous phosphate chemistry, albeit at conditions of elevated pH. Such alkaline conditions are however common in fresh-water volcanic lakes,31,32 indicating it is a possible scenario on early Earth.

Results and Discussion

Learning from our previous work15 and other studies on ligation with >P,26,27,29,33 we decided to investigate the ligation reaction with >P under alkaline conditions. This was also corroborated by a broader pH screen (Supporting Information S4, Figure S4.1). All the reactions were performed in aqueous alkaline conditions with 50 mM 2-(cyclohexylamino)ethane-1-sulfonic acid (CHES) buffer and an equimolar concentration of the participating strands. A possible influence of the buffer itself was not supported in a buffer screen at pH 9 (Supporting Information S4, Figure S4.2). The two ligating strands are labeled as primer a and b and the template, BA. All the reactions were performed in the presence of 1 mM Mg2+ as it was found to be optimal (Supporting Information S4, Figure S4.3). The analysis of the ligation reaction was done either by polyacrylamide gel electrophoresis (PAGE, Figure 1c; see Supporting Information S1.3 and S2.2) or high-performance liquid chromatography (HPLC, Figures 1d, Supporting Information S1.4 and S2.1), or a combination of both.

Figure 1.

Figure 1

Open in a new tab

Nonenzymatic template-directed ligation of short RNA strands. (a) Schematics of reaction design. Both primers a andb bind on the complementary template, BA. The primer a has a 2′,3′-cyclic phosphate, while b contains a 5′-OH group. (b) 5′-OH performs a nucleophilic attack on the cyclic phosphate group and forms a phosphodiester bond between the two primers, leading to the ligation product strand ab. As a side reaction, the cyclic phosphate in a can also hydrolyze, rendering a inactive. (c) Denaturing polyacrylamide gel electrophoresis (PAGE) analysis of ligation reaction over time. Reaction contained 1 μM primers, 1 μM template, 50 mM CHES, pH 10, and 1 mM MgCl2 at 5 °C. (d) Stacked HPLC chromatograms (absorbance at 260 nm) of the same reaction mixtures as in (c). The product peak increases over time as the primers get depleted.

To study the impact of temperature and pH on the ligation reaction, three temperatures of 5, 10, and 25 °C and pH 9, 10, and 11 were tested, as shown in Figure 2. The reaction yielded both the ligation product (ab) and the inactivated primer side product (a–P) while consuming the two primers a > P and b. The template (BA) was not consumed by the reaction. Thus, its concentration was used to correct for small pipetting errors through normalization.

Figure 2.

Figure 2

Open in a new tab

Kinetic study of temperature and pH influence in ligation reaction. (a) Schematics of the ligation reaction. The hydrolysis of the cyclic phosphate competes with the ligation reaction. (b) Screening the concentration of primers and product over a 7 day period for varied pH (9–11) and temperature (5, 10, 25 °C). The highest yield at 7 days is highlighted in red, for pH 10 and 5 °C. At both high temperature and pH (25 °C and pH 11), there is additional hydrolysis of the backbone, particularly after 7 days (bottom, right–most graph). The full lines correspond to the exponential fit of the data as a guide to the eye. (c) Product yield obtained (%) at 7 days for all of the conditions tested. Maximum obtained yield was for pH 10 at 5 °C (red square). This reported yield was calculated for the limiting concentration of a > P. (d) Observed initial pseudo-first order rate of product and inactive primer formation for all the conditions (Supporting Information S8). Reactions contained 1 μM primers, 1 μM template, 50 mM CHES, with varied pH, and 1 mM MgCl2. Concentrations were measured with HPLC UV detection at 260 nm. Data are represented as mean ± standard deviation of three independent replicates.

The concentration over time of the strands a > P, b, ab, a–P, for different temperature–pH combinations is plotted in Figure 2b (see Figure 2a for color-coded schematics). It is important to note that in all the experiments, the initial concentration of a–P is on average 26% of total a, suggesting that a part of a > P was already hydrolyzed in the stock solutions. This capped the maximum concentration of ab at 0.74 μM, the concentration of initial a > P. The yield of ab depends on the temperature and pH combination of the reaction. The highest concentration of ab at 7 days (0.28 μM) was obtained at 5 °C, pH 10 (Figure 2b,c, red rectangle). Calculating for the real initial amount of a > P, the yield was 38%. The obtained yield was measured under a limiting concentration of the primer with a cyclic phosphate end (a > P), meaning that the formation of hydrolyzed a–P contributed to the observed partial conversion. However, the addition of excess primer a did not improve the yield (Supporting Information S10), indicating product inhibition, confirmed by our estimate that the product strands have an off-rate of about 40 days at 5 °C. Comparable yields have been reported for ligation with phosphorimidazolide activation under similarly low Mg2+ concentrations.8

Since the formation of the primer–primer–template complex through base-pairing proceeds at much faster time scales (kon ≈ 1 μM–1 s–1, literature values for similar-sized oligonucleotides34−36) than that of the nucleophilic attack on the cyclic phosphate, it can be assumed that the reaction is of pseudo-first order (Supporting Information S8.1). Thus, first-order kinetic rate constants (kobs) for product formation and hydrolysis of cyclic phosphate were fitted to the data, see Figure 2d (Supporting Information S8, Figure S8.1). The obtained rates are between 0.001 and 0.03 h–1, which are in the same order of magnitude for the ligation of native RNA at 20 mM MgCl2 with 2-Me imidazolide chemistry.9

A few salient features of the plots in Figure 2b–d are the rates and yields of ligation (ab) and hydrolysis (a–P). Both the rates of ligation and hydrolysis increase with an increase in either pH or temperature, keeping the other parameter constant. Figure 2d shows that while the observed ligation rate mostly increases from pH 9–10, the inactivation rate increases from pH 10–11. At higher pH, the ligation kinetics is slightly faster; however, the final yield drops due to the competing inactivation rate. The rate of ligation is in general higher than the rate of the hydrolysis at 5 and 10 °C, while 25 °C favors the inactivation. It has been reported that low temperatures reduce the entropic cost of the ligation reaction and shift the reaction equilibrium from hydrolysis to ligation.23 The maximum 7 day yield obtained is at pH 10 and 5 °C. Significant RNA backbone hydrolysis is observed when both temperature and pH are maximal (25 °C and pH 11).

The attack of the 5′-OH on the cyclic phosphate (Figure 1b) can form either a 2′–5′ or a 3′–5′ phosphodiester bond (Figure 3a,b, respectively). Previous studies on the polymerization and ligation with cyclic phosphates have reported varying ratios of 3′–5′ and 2′–5′ linkages, depending largely on the experimental conditions. For example, dry-state polymerization resulted in a natural linkage enrichment ratio of 2:1,13,37 while an aqueous state (with 0.5 M diamine, pH 8, and 0 °C) was reported to lead to at least 97% of 2′–5′.30 Templated cleavage and ligation at 25 °C, pH 9, and 5 mM MgCl2 in an aqueous solution was reported to also show a predominance of 2′–5′ linkages (about 95%).29 Conversely, templated ligation in the eutectic phase resulted in an excess of 3′–5′ linkages.38 For the templated ligation reaction described here, we found no significant regioselectivity under the tested conditions (Figure 3c). This difference in comparison to previous studies is potentially due to the different systems and conditions tested. To investigate this, the reaction was quenched by ethanol precipitation, and the samples were digested with nuclease P1 following the manufacturer’s protocol (Supporting Information S1.5). Nuclease P1 specifically lyses the 3′–5′ linkages, which in this case would digest all the strands a, b, and BA completely but digest ab either completely or result in a UC dimer.

Figure 3.

Figure 3

Open in a new tab

Linkage analysis of the reaction product ab through digestion with nuclease P1. Scheme of the ligation site with a 2′–5′ linkage (a, green) and a 3′–5′ linkage (b, pink). (c) Total product yield obtained for three condition sets, with the corresponding relative amounts of 2′–5′ and 3′–5′ linkage. The ligation with 2′,3′-cyclic phosphates does not exhibit regioselectivity as both linkages are equally represented for the studied conditions. After the reactions with 10 μM primers, 10 μM template, 50 mM CHES, pH 10, and 1 mM MgCl2, they were digested with nuclease P1 (Supporting Information S1.5). The concentrations of the samples before and after digestion were measured with HPLC UV detection at 260 nm (Supporting Information S2.1 and S5). Data represent the mean of independent duplicates.

The concentration of total product predigestion and UC dimer postdigestion was determined using HPLC UV absorbance (Supporting Information S2.1 and S5). We observed that both types of linkages were formed equally (Figure 3c), indicating that the reaction is not regioselective. However, a slight enrichment of the canonical linkage over time for the 5 °C conditions can be seen, possibly due to the favored hydrolysis of 2′–5′ linkages, particularly in double-stranded RNA in alkaline solutions.39 The presence of noncanonical linkages, however, does not render the product strands obsolete. Such mixed backbone RNA has been demonstrated to still fold into functional structures.40 The stability of the RNA duplex has been documented to be reduced for strands fully composed by 2′–5′ linkages in comparison with that of RNA with canonical linkages.41,42 For this system, however, one single 2′–5′ linkage at the ligation site would likely not have a considerable destabilizing effect as the duplex could accommodate the structural disruption.40 Furthermore, Supporting Information S7, Figure S7.3a shows that ab with a 2′–5′ linkage can still template the formation of BA through >P mediated ligation, highlighting the possibility of a replication cycle.

To answer the question if the ligation reaction is sequence-specific, we conducted the reaction with each of the 16 different nucleotide combinations (four each on the 3′ end of a and the 5′ end of b) at the ligation site (Figure 4a). Additionally, we tested two different templates, one with GA and the other with UA at the position complementary to the ligation site. Except for the ligation site, the remainder of the sequence was fully complementary to the template. This is similar to the approaches undertaken for the fidelity calculation for ligases.43−45 These reactions were carried out at pH 10 and 5 °C for 168 h, i.e., conditions with the highest yield of ligation among the ones tested (Figure 2c). Figure 4b,c shows that of all the combinations of the primers tested, the highest yield of ab is obtained for the sequence with the correct nucleotides at the ligation site (marked in red, CU for the template GA and AU for the template UA). However, mismatched ligations did occur, albeit with much lower relative yields, which were especially reduced for the template UA.

Figure 4.

Figure 4

Open in a new tab

Ligation site specificity for two different template sequences. After 7 days, the yields were measured for reactions using primers that had each of the four possible nucleotides at the 3′ end of a (Na) and the 5′ end of b (Nb). This resulted in a total of 16 different primer combinations being evaluated. (a) Schematics of the ligation sites and the two templates tested. The templates differed only at the dimer complementary to the ligation site with either 5′ GA 3′ (b) or 5′ UA 3′ (c). The maximum yield obtained in both cases was for the correct combination of complementary primers (nucleotides highlighted in red). Guanine/uracil (G/U) wobble pairing is represented as a tilted nucleotide in the cartoon representation. For most combinations, one single mutation at the ligation site reduced its relative yield considerably or prevented ligation, even though the other primer is fully bound to the template. The snippet below the heat map in (b,c) corresponds to the PAGE of the bottom row, showing the ligated product ab. Reactions were performed with 10 μM primers, 10 μM template, 50 mM CHES, pH 10, and 1 mM MgCl2 for 7 days at 5 °C. Data are represented as mean of independent triplicates. (d) The fidelity of ligation was extrapolated to a per-nucleotide replication fidelity using primers of varying length using single-mutation-sensitive binding calculation of the primers with NUPACK (Supporting Information S9.1). The fidelity decreased for longer primers. Also, a G at the ligation site leads to lower fidelity due to the G–U wobble pairs (Supporting Information S9.3).

Interestingly, G/U wobble pairing of the 5′ U of primer b at the ligation site led to a high relative yield (78%) compared to that of the complementary primers for the template GA. When considering one single mutation at the ligation site either on a or b, the reaction yield drops on average by 91 or 82%, respectively, relative to the nonmutated complex. We term this the ligation fidelity for one mutation. If we consider two mutations at the ligation site, the average experimental yield drops to 12% (template AG) and 5% (template AU), and thus, the ligation fidelity for the respective template is 88 and 95% for two mutations.

A prebiotic replication through ligation would take place from a diverse pool of oligonucleotides consisting of different sequences of varied lengths. In such a scenario, the likelihood of unstable primer–template complexes is high. Two contributions of the nucleotides surrounding the ligation site can be identified as having a significant effect on the stability of the complex, the amount of mismatches, and the length of the binding region. To compare the performance of the ligation with a base-by-base replication, we calculated the per-nucleotide replication fidelity as detailed in Supporting Information S9. This corresponds to the minimum fidelity that a base-by-base replicator would require, for each incorporation, to create the same number of errors within the ligated strand. We combined thermodynamic analysis, using NUPACK, for one or two mutations in the ligating strands outside of the ligation site with the experimental ligation errors obtained by mutating both nucleotides at the ligation site (Figure 4b,c). These extrapolated experimental–theoretical per-nucleotide fidelities reached 89–92% for the tested system and 95–98% for primer length of 6-mer. Unlike shorter strands where a single point mutation would destabilize the complex, longer primers are more tolerant to single point mutations at cold temperatures, setting a fidelity-based length limit to the ligating strands. This is the reason for the increase in fidelity for shorter strands. Interestingly, the optimal lengths for ligation are reached by dry oligomerization from 2′,3′-cyclic nucleotides at the same alkaline pH.15

After understanding that the nonenzymatic ligation with 2′,3′-cyclic phosphates is a reliable copying mechanism, we aimed to investigate the potential for elongation. This would have been an important characteristic of a potential prebiotic replication mechanism as it would establish a link between nontemplated nucleotide condensation and the faster replication by long ribozymes, with tens or hundreds of nucleotides.46−48

We explored the possibility of bridging these two oligonucleotide length regimes by designing splint strands with short (4- or 8-mer) binding regions that can both cross-template and ligate (Figure 5a). Each system has two strands (labeled c and d) with >P. The strands are designed such that the 5′ half of the strand d is a reverse complement to the 5′ half of c, and the same goes for the 3′ halves such that they bind and form a long network of ccc... bound to ddd... (Figure 5a, see Supporting Information S3 for sequence information). The formed secondary structure allows for multiple ligations, resulting in homopolymers of c and d. Figure 5b shows that up to 5 concatenations (n = 6) could be detected for both the 16-mer system (Figure 5b) and 8-mer system (Figure 5d) resulting in 96- and 48-mer RNA, respectively. Figure 5c,e shows the gel quantification of the respective gels in Figure 5b,d. For the shorter system, the yield was reduced by about an order of magnitude for each additional concatenation and was generally lower than that for the 16-mer system, which was likely due to the slower kinetics in a frozen state at −20 °C. Specifically for the case of the 16-mer at 5 °C, 5.1 μM of the strands was incorporated into the concatemers, and 4.9 μM remained. The remaining strands c and d contain either the active or inactive phosphate group as the two species do not resolve through PAGE. We propose that for these conditions, as for the system in Figure 2, the main limitation to the yield is the hydrolysis of the cyclic phosphate.

Figure 5.

Figure 5

Open in a new tab

Assembly of long RNA via splinted ligation of 2′,3′-cyclic phosphate containing oligonucleotides. (a) Schematics of sequence design. Two strands labeled c and d with complementary subregions and not corresponding to the complete reverse complement, bind to form a long chain with repeating units of each. Both c and d contain 2′,3′-cyclic phosphate, and the ligation can yield all possible length multiples of the initial strands. For the case where c and d are 16-mer long, denaturing PAGE of ligation reaction at −80, −20 °C (frozen), and 5 °C (b) revealed products of up to five concatenations. The concentration of each ligation product (up to n = 5), obtained from the SYBR Gold fluorescence analysis, is plotted in (c). Similarly, for an initial strand size of 8-mer, denaturing PAGE (d) and product concentration (e) are shown. The optimal temperature for splinted ligation depends on oligomer length as 5 °C yields a higher concentration for the 16-mer, while −20 °C is better for 8-mer. Reactions were performed with 10 μM primers, 10 μM template, 50 mM CHES, pH 10, and 1 mM MgCl2 for 7 days. Data are represented as the mean of three independent replicates.

It is interesting to note that the maximum yield for the 16-mer system was obtained at 5 °C, whereas for the 8-mer system, it was at −20 °C. We believe this to result from the low duplex stability of the 4-mer duplex region at 5 °C. We tested this by studying the ligation of shorter systems (Supporting Information S6). Systems with 4- and 6-mer binding regions were designed to have the same sequence at the ligation site and the same GC content as the strands used in Figures 1 and 2. While there was a significant reduction in yield for both systems, we found that for the 6-mer system, the yields were higher at 5 °C, whereas for the 4-mer system, −20 °C was more favorable (Supporting Information, Figure S6.1). Additionally, the obtained yields at 7 days were very low (about 3%) for the latter. This suggests that to stabilize the duplex for short oligonucleotides, a compromise between the slower rate of ligation and the low probability of duplex formation must be made. We demonstrate that the ligation reaction is robust since the only requirement for the templated ligation via 2′,3′-cyclic phosphate, besides the formation of a duplex, seems to be an alkaline pH, needing no additional molecules or external activations. Also, the shown formation of long RNA, bridges the length gap toward a regime where long, functional ribozymes can evolve.

Conclusions

A nonenzymatic replicator chemistry on the early Earth should have the capacity, under plausible conditions, to elongate strands and undergo further replication steps all while being highly accurate and processive. We demonstrate with this work that ligation with >P RNA fulfills these criteria.

First, we show that the template-directed replication mechanism only requires salts for stabilizing the duplex and alkaline pH, making the ligation with >P RNA robust and reproducible in both aqueous and frozen solutions. For the aqueous case, a 38% yield was observed in contrast to the previously reported yield of 16% for similar reactions. It was found that a combination of high pH and low temperature promotes ligation over the hydrolysis of the cyclic phosphate moiety. Such conditions are thought to be plausible on early Earth, where the fainter sun contributed to a cold surface temperature that would still allow liquid water.49 Additionally, Hadean oceans were potentially alkaline due to the sequestering of CO2 in carbonate minerals,50 and alkaline conditions present in freshwater volcanic lakes31,32 have been proposed to foster early metabolism.

However, a fraction of the >P still hydrolyzes, which contributes to its incomplete conversion. While the yield could be further improved by adding reagents that aid in the recyclization of the monophosphate moiety, such as diamidophosphate in combination with imidazole,23 this would increase the complexity of the system. Moreover, this reaction has low salt requirements (1 mM MgCl2, Supporting Information, Figure S4.3), ensuring RNA backbone integrity and compatibility with strand separation. It can also proceed in a wide pH range, even unbuffered (Supporting Information S4).

The elongation of short RNA was demonstrated with splinted systems that yielded up to six-copy concatemers of short RNA strands of either 16- or 8-mer, resulting in long RNA on the scale of 100-mer (Figure 5). This length range approaches the average length of replicating ribozymes,46−48 representing a significant step toward assembling functional RNAs by plausible means. Even very short >P RNA fragments (with 4-mer base-pairing regions) ligate under frozen conditions (Supporting Information S6), establishing a bridge from the single nucleotide condensation reactions, yielding very short RNA strands, to a regime where templated ligation reactions could dominate.

Furthermore, we evaluated the copying accuracy with varying nucleotides at the ligation site. An experimentally measured fidelity of at least 82% was obtained upon screening all possible single base mutations at either the 3′ or 5′ end of the ligation site (Figure 4d), with the remaining primer being entirely complementary. To compare the ligation fidelity with a base-by-base replication, we extrapolated a per-nucleotide fidelity of 89–92% for adding eight nucleotides or a fidelity of 95–98% when adding six nucleotides.

Contrary to previous studies on >P, we found that under the tested conditions, the reaction was not regioselective, producing equal amounts of 2′–5′ and 3′–5′ linkages at the ligation site (Figure 3). While approximately half of the linkages were noncanonical, we argue this does not diminish the applicability of the reaction in a prebiotic context. Strands with 2′–5′ linkages have been shown to fold into functional structures,40 and these noncanonical linkages have also been demonstrated to be more labile than 3′–5′ and have potential for interconversion.39,51 Furthermore, Supporting Information, Figure S6.3 shows that product ab with a 2′–5′ linkage at the ligation site could still template the reverse ligation reaction. The noncanonical linkage in the template at the ligation site did not impede ligation, paving the way for exponential replication cycles.

Strand separation, driven by nonequilibrium environments with thermal, salt, or pH oscillations would allow for the implementation of a ligation chain reaction, similar to that reported by Edeleva et al., with the added benefit of not generating deleterious side-products by a prebiotically implausible EDC.7 An air–water interface with continuous feeding would allow for both the denaturation and replenishment of activated primers, without necessarily high temperature which could degrade the cyclic phosphate and the RNA backbone.6 This suggests that such scenarios could provide a niche, where the ligation reactions by 2′,3′-cyclic phosphate could evolve toward a ribozymatic replicator.

Considering these results, ligation with >P is an interesting framework to produce diverse pools of long RNA that could undergo molecular evolution. We show that the system described in the current study enables the generation of long RNA, with high fidelity. This was demonstrated for a range of lengths, sequence combinations, reaction conditions, and temperatures, suggesting that ligation of RNA with >P holds a central position in the general conception of the RNA world.

Acknowledgments

The authors acknowledge the support from the European Research Council (ERC Evotrap, grant number 787356), the CRC 235 Emergence of Life (project-ID 364653263), the CRC 392 Molecular Evolution in Prebiotic (project-ID 521256690) and the Excellence Cluster ORIGINS (under Germany’s Excellence Strategy—EXC-2094–390783311) funded by the Deutsche Forschungsgemeinschaft (DFG), the Simons Foundation (327125), and the Center for NanoScience (CeNS).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c10813.

  • Experimental procedures and quantification, nucleic acid sequences, buffer and salt conditions screening, control experiments, kinetic constant fitting, and per-nucleotide fidelity extrapolation parameters and expressions (PDF)

Author Contributions

§ A.C.S and S.W. contributed equally to this paper.

The authors declare no competing financial interest.

Supplementary Material

ja3c10813_si_001.pdf (2MB, pdf)

References

  1. Gilbert W. Origin of Life: The RNA World. Nature 1986, 319 (6055), 618. 10.1038/319618a0. [DOI] [Google Scholar]
  2. 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. U.S.A. 2022, 119 (17), e2116429119 10.1073/pnas.2116429119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Li L.; Prywes N.; Tam C. P.; O’Flaherty D. K.; Lelyveld V. S.; Izgu E. C.; Pal A.; Szostak J. W. Enhanced Nonenzymatic RNA Copying with 2-Aminoimidazole Activated Nucleotides. J. Am. Chem. Soc. 2017, 139 (5), 1810–1813. 10.1021/jacs.6b13148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Vogel S. R.; Deck C.; Richert C. Accelerating Chemical Replication Steps of RNA Involving Activated Ribonucleotides and Downstream-Binding Elements. Chem. Commun. 2005, 39, 4922–4924. 10.1039/b510775j. [DOI] [PubMed] [Google Scholar]
  5. Sosson M.; Pfeffer D.; Richert C. Enzyme-Free Ligation of Dimers and Trimers to RNA Primers. Nucleic Acids Res. 2019, 47 (8), 3836–3845. 10.1093/nar/gkz160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Salditt A.; Karr L.; Salibi E.; Le Vay K.; Braun D.; Mutschler H. Ribozyme-Mediated RNA Synthesis and Replication in a Model Hadean Microenvironment. Nat. Commun. 2023, 14 (1), 1495. 10.1038/s41467-023-37206-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Edeleva E. V.; Salditt A.; Stamp J.; Schwintek P.; Boekhoven J.; Braun D. Continuous Nonenzymatic Cross-Replication of DNA Strands with in Situ Activated DNA Oligonucleotides. Chem. Sci. 2019, 10 (22), 5807–5814. 10.1039/C9SC00770A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Wachowius F.; Holliger P. Non-Enzymatic Assembly of a Minimized RNA Polymerase Ribozyme. ChemSystemsChem 2019, 1, 1–4. 10.1002/syst.201900004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Zhou L.; O’Flaherty D. K.; Szostak J. W. Template-Directed Copying of RNA by Non-enzymatic Ligation. Angew. Chem. 2020, 132, 15812–15817. 10.1002/ange.202004934. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Joyce G. F. Non-enzymatic template-directed synthesis of RNA copolymers. Orig. Life Evol. Biosph. 1984, 14 (1–4), 613–620. 10.1007/BF00933711. [DOI] [PubMed] [Google Scholar]
  11. Prywes N.; Blain J. C.; Del Frate F.; Szostak J. W. Nonenzymatic Copying of RNA Templates Containing All Four Letters Is Catalyzed by Activated Oligonucleotides. Elife 2016, 5, e17756 10.7554/eLife.17756. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Zhou L.; O’Flaherty D. K.; Szostak J. W. Assembly of a Ribozyme Ligase from Short Oligomers by Nonenzymatic Ligation. J. Am. Chem. Soc. 2020, 142 (37), 15961–15965. 10.1021/jacs.0c06722. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Verlander M. S.; Lohrmann R.; Orgel L. E. Catalysts for the Self-Polymerization of Adenosine Cyclic 2′,3′-Phosphate. J. Mol. Evol. 1973, 2 (4), 303–316. 10.1007/BF01654098. [DOI] [PubMed] [Google Scholar]
  14. Verlander M. S.; Orgel L. E. Analysis of High Molecular Weight Material from the Polymerization of Adenosine Cyclic 2′, 3′-Phosphate. J. Mol. Evol. 1974, 3 (2), 115–120. 10.1007/BF01796557. [DOI] [PubMed] [Google Scholar]
  15. Dass A. V.; Wunnava S.; Langlais J.; von der Esch B.; Krusche M.; Ufer L.; Chrisam N.; Dubini R. C. A.; Gartner F.; Angerpointner S.; Dirscherl C. F.; Rovó P.; Mast C. B.; Šponer J. E.; Ochsenfeld C.; Frey E.; Braun D. RNA Oligomerisation without Added Catalyst from 2′,3′-Cyclic Nucleotides by Drying at Air-Water Interfaces**. ChemSystemsChem 2023, 5 (1), e202200026 10.1002/syst.202200026. [DOI] [Google Scholar]
  16. Powner M. W.; Gerland B.; Sutherland J. D. Synthesis of Activated Pyrimidine Ribonucleotides in Prebiotically Plausible Conditions. Nature 2009, 459 (7244), 239–242. 10.1038/nature08013. [DOI] [PubMed] [Google Scholar]
  17. Lohrmann R.; Orgel L. E. Prebiotic Synthesis: Phosphorylation in Aqueous Solution. Science 1968, 161 (3836), 64–66. 10.1126/science.161.3836.64. [DOI] [PubMed] [Google Scholar]
  18. Gibard C.; Bhowmik S.; Karki M.; Kim E. K.; Krishnamurthy R. Phosphorylation, Oligomerization and Self-Assembly in Water under Potential Prebiotic Conditions. Nat. Chem. 2018, 10 (2), 212–217. 10.1038/nchem.2878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lafontaine D. A.; Beaudry D.; Marquis P.; Perreault J.-P. Intra- and Intermolecular Nonenzymatic Ligations Occur within Transcripts Derived from the Peach Latent Mosaic Viroid. Virology 1995, 212 (2), 705–709. 10.1006/viro.1995.1528. [DOI] [PubMed] [Google Scholar]
  20. Buzayan J. M.; Gerlach W. L.; Bruening G. Non-Enzymatic Cleavage and Ligation of RNAs Complementary to a Plant Virus Satellite RNA. Nature 1986, 323 (6086), 349–353. 10.1038/323349a0. [DOI] [Google Scholar]
  21. Li Y.; Breaker R. R. Kinetics of RNA Degradation by Specific Base Catalysis of Transesterification Involving the 2‘-Hydroxyl Group. J. Am. Chem. Soc. 1999, 121 (23), 5364–5372. 10.1021/ja990592p. [DOI] [Google Scholar]
  22. Breslow R..; Labelle M.. Sequential General Base-Acid Catalysis in the Hydrolysis of RNA by Imidazole. J. Am. Chem. Soc. 1986, 108 (10), 2655–2659. 10.1021/ja00270a025. [DOI] [Google Scholar]
  23. Song E. Y.; Jiménez E. I.; Lin H.; Le Vay K.; Krishnamurthy R.; Mutschler H. Prebiotically Plausible RNA Activation Compatible with Ribozyme-Catalyzed Ligation. Angew. Chem. Int. Ed. 2021, 60 (6), 2952–2957. 10.1002/anie.202010918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mutschler H.; Holliger P. Non-Canonical 3′-5′ Extension of RNA with Prebiotically Plausible Ribonucleoside 2′,3′-Cyclic Phosphates. J. Am. Chem. Soc. 2014, 136 (14), 5193–5196. 10.1021/ja4127714. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mutschler H.; Wochner A.; Holliger P. Freeze-Thaw Cycles as Drivers of Complex Ribozyme Assembly. Nat. Chem. 2015, 7 (6), 502–508. 10.1038/nchem.2251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lutay A. V.; Chernolovskaya E. L.; Zenkova M. A.; Vlassov V. V. The Nonenzymatic Template-Directed Ligation of Oligonucleotides. Biogeosci. Discuss. 2006, 3 (3), 243–249. 10.5194/bg-3-243-2006. [DOI] [Google Scholar]
  27. Lutay A. V.; Chernolovskaya E. L.; Zenkova M. A.; Vlasov V. V. Nonenzymatic Template-Dependent Ligation of 2’,3′-Cyclic Phosphate-Containing Oligonucleotides Catalyzed by Metal Ions. Dokl. Biochem. Biophys. 2005, 401 (1–6), 163–166. 10.1007/s10628-005-0060-4. [DOI] [PubMed] [Google Scholar]
  28. Mutschler H.; Taylor A. I.; Porebski B. T.; Lightowlers A.; Houlihan G.; Abramov M.; Herdewijn P.; Holliger P. Random-Sequence Genetic Oligomer Pools Display an Innate Potential for Ligation and Recombination. Elife 2018, 7, e43022 10.7554/eLife.43022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lutay A. V.; Zenkova M. A.; Vlassov V. V. Nonenzymatic Recombination of RNA: Possible Mechanism for the Formation of Novel Sequences. Chem. Biodivers. 2007, 4 (4), 762–767. 10.1002/cbdv.200790062. [DOI] [PubMed] [Google Scholar]
  30. Renz M.; Lohrmann R.; Orgel L. E. Catalysts for the Polymerization of Adenosine Cyclic 2′,3′-Phosphate on a Poly (U) Template. Biochim. Biophys. Acta, Nucleic Acids Protein Synth. 1971, 240 (4), 463–471. 10.1016/0005-2787(71)90703-9. [DOI] [PubMed] [Google Scholar]
  31. Toner J. D.; Catling D. C. A Carbonate-Rich Lake Solution to the Phosphate Problem of the Origin of Life. Proc. Natl. Acad. Sci. U.S.A. 2020, 117 (2), 883–888. 10.1073/pnas.1916109117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Toner J. D.; Catling D. C. Alkaline Lake Settings for Concentrated Prebiotic Cyanide and the Origin of Life. Geochim. Cosmochim. Acta 2019, 260, 124–132. 10.1016/j.gca.2019.06.031. [DOI] [Google Scholar]
  33. Lutay A. V.; Grigoriev I. V.; Zenkova M. A.; Chernolovskaya E. L.; Vlassov V. V. Nonenzymatic Recombination of RNA by Means of Transesterification. Russ. Chem. Bull. 2007, 56 (12), 2499–2505. 10.1007/s11172-007-0398-0. [DOI] [Google Scholar]
  34. Kinjo M.; Rigler R. Ultrasensitive Hybridization Analysis Using Fluorescence Correlation Spectroscopy. Nucleic Acids Res. 1995, 23 (10), 1795–1799. 10.1093/nar/23.10.1795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wetmur J. G.; Davidson N. Kinetics of Renaturation of DNA. J. Mol. Biol. 1968, 31 (3), 349–370. 10.1016/0022-2836(68)90414-2. [DOI] [PubMed] [Google Scholar]
  36. Cisse I. I.; Kim H.; Ha T. A Rule of Seven in Watson-Crick Base-Pairing of Mismatched Sequences. Nat. Struct. Mol. Biol. 2012, 19 (6), 623–627. 10.1038/nsmb.2294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Usher D. A.; Yee D. Geometry of the Dry-State Oligomerization of 2′,3′-Cyclic Phosphates. J. Mol. Evol. 1979, 13 (4), 287–293. 10.1007/BF01731369. [DOI] [PubMed] [Google Scholar]
  38. 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. 2004, 32 (9), 2966–2974. 10.1093/nar/gkh601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rohatgi R.; Bartel D. P.; Szostak J. W. Nonenzymatic, Template-Directed Ligation of Oligoribonucleotides Is Highly Regioselective for the Formation of 3′-5′ Phosphodiester Bonds. J. Am. Chem. Soc. 1996, 118 (14), 3340–3344. 10.1021/ja9537134. [DOI] [PubMed] [Google Scholar]
  40. Sheng J.; Li L.; Engelhart A. E.; Gan J. H.; Wang J. W.; Szostak J. W. Structural Insights into the Effects of 2 ’-5 ’ Linkages on the RNA Duplex. Proc. Natl. Acad. Sci. U.S.A. 2014, 111 (8), 3050–3055. 10.1073/pnas.1317799111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Giannaris P. A.; Damha M. J. Oligoribonucleotides Containing 2′,5′-Phosphodiester Linkages Exhibit Binding Selectivity for 3′,5′-RNA over 3′,5′-ssDNA. Nucleic Acids Res. 1993, 21 (20), 4742–4749. 10.1093/nar/21.20.4742. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Wasner M.; Arion D.; Borkow G.; Noronha A.; Uddin A. H.; Parniak M. A.; Damha M. J. Physicochemical and Biochemical Properties of 2‘,5‘-Linked RNA and 2‘,5‘-RNA:3‘,5‘-RNA “Hybrid” Duplexes. Biochemistry 1998, 37 (20), 7478–7486. 10.1021/bi980160b. [DOI] [PubMed] [Google Scholar]
  43. Lohman G. J. S.; Bauer R. J.; Nichols N. M.; Mazzola L.; Bybee J.; Rivizzigno D.; Cantin E.; Evans T. C. A High-Throughput Assay for the Comprehensive Profiling of DNA Ligase Fidelity. Nucleic Acids Res. 2016, 44 (2), e14 10.1093/nar/gkv898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wu D. Y.; Wallace R. B. Specificity of the Nick-Closing Activity of Bacteriophage T4 DNA Ligase. Gene 1989, 76 (2), 245–254. 10.1016/0378-1119(89)90165-0. [DOI] [PubMed] [Google Scholar]
  45. Luo J. Improving the Fidelity of Thermus Thermophilus DNA Ligase. Nucleic Acids Res. 1996, 24 (15), 3071–3078. 10.1093/nar/24.15.3071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Jaeger L.; Wright M. C.; Joyce G. F. A Complex Ligase Ribozyme Evolved in Vitro from a Group I Ribozyme Domain. Proc. Natl. Acad. Sci. U.S.A. 1999, 96 (26), 14712–14717. 10.1073/pnas.96.26.14712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Horning D. P.; Joyce G. F. Amplification of RNA by an RNA Polymerase Ribozyme. Proc. Natl. Acad. Sci. U.S.A. 2016, 113 (35), 9786–9791. 10.1073/pnas.1610103113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Shechner D. M.; Grant R. A.; Bagby S. C.; Koldobskaya Y.; Piccirilli J. A.; Bartel D. P. Crystal Structure of the Catalytic Core of an RNA-Polymerase Ribozyme. Science 2009, 326 (5957), 1271–1275. 10.1126/science.1174676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rosing M. T.; Bird D. K.; Sleep N. H.; Bjerrum C. J. No Climate Paradox under the Faint Early Sun. Nature 2010, 464 (7289), 744–747. 10.1038/nature08955. [DOI] [PubMed] [Google Scholar]
  50. Kadoya S.; Krissansen-Totton J.; Catling D. C. Probable Cold and Alkaline Surface Environment of the Hadean Earth Caused by Impact Ejecta Weathering. Geochem., Geophys., Geosyst. 2020, 21 (1), e2019GC008734 10.1029/2019GC008734. [DOI] [Google Scholar]
  51. Jarvinen P.; Oivanen M.; Lonnberg H. Interconversion and Phosphoester Hydrolysis of 2’,5′- and 3′,5′-Dinucleoside Monophosphates: Kinetics and Mechanisms. J. Org. Chem. 1991, 56 (18), 5396–5401. 10.1021/jo00018a037. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ja3c10813_si_001.pdf (2MB, pdf)

Articles from Journal of the American Chemical Society are provided here courtesy of American Chemical Society

RESOURCES