Though defining what makes something “alive” has proved challenging, one crucial feature of living things is the ability to copy genetic information and pass it on to the next generation. Nowadays, enzymes called polymerases copy genetic information encoded within the DNA of living cells. However, when life on Earth began approximately four billion years ago, polymerases had not evolved yet. This means that the genetic information of the first cells had to be copied some other way.

The earliest life on Earth is unlikely to have used DNA to store its genetic information, and probably used a closely related molecule called RNA instead. Like DNA, RNA is made up of four smaller building blocks joined together to form long chains. The building blocks of RNA are commonly referred to using single letters: A, C, G and U. Previous studies have shown that it is possible to copy RNA without enzymes, but for only two of the four RNA letters, namely C and G.

Prywes et al. wanted to know if it was possible to create a chemical system, without polymerases, in which all four RNA letters could be copied. The experiments showed that strings of RNA that were three letters long could catalyze RNA copying, just as long as they were chemically activated. That is to say, these short RNA strings allowed RNA to be copied without enzymes if they had a chemical group at one end that made them more reactive.Each short catalyst helped copy one of the four RNA letters, and adding several into one reaction meant that longer sequences containing all four RNA letters could be copied.

Prywes et al. then used these short catalysts to copy an RNA molecule that itself acts a bit like an enzyme, and confirmed that a significant portion of this molecule could be copied without any polymerases. Further work is now needed to see if it is possible to copy other RNA sequences, and especially longer ones, without enzymes. Another challenge for the future would be to attempt to copy an RNA sequence multiple times without enzymes; a challenge that the earliest ancestors of cells on Earth must have overcome to pass their genetic information down through the generations.

In order to serve as a bridge from prebiotic chemistry to the RNA world (Patel et al., 2015; Powner et al., 2009), the nonenzymatic copying of RNA templates by suitably activated nucleotides (Figure 1a) must occur quickly enough to replicate functional RNA sequences faster than they degrade. However, the nonenzymatic copying of sequences containing all four nucleotides has not yet been possible for a number of reasons (Szostak, 2012), most notably the slow rate of primer extension with adenosine and uridine monomers (Joyce et al., 1987; Wu and Orgel, 1992b; Deck et al., 2011). Indeed, two adjacent A or U residues in the template will stop polymerization entirely (Wu and Orgel, 1992b) under normal conditions; even partial copying requires extreme conditions, such as the eutectic phase of frozen samples (Vogel and Richert, 2007), which are not compatible with replication within protocells. Extensive optimization of reaction conditions with the aim of improving the rate of primer extension with all four monomers, including varying the choice of divalent cation and leaving group, improved the rate and regiospecificity of polymerization for G and C but not A and U (Wu and Orgel, 1992b, c; Deck et al., 2011; Inoue and Orgel, 1981; Hagenbuch et al., 2005; Lohrmann et al., 1980). Recent biophysical studies suggest that low A and U monomer affinity for the template is not the only reason for this problem (Szostak, 2012; Izgu et al., 2015) since A binds only three-fold more weakly than G, but primer extension with A is at least 100-fold slower (Joyce et al., 1987; Wu and Orgel, 1992b; Deck et al., 2011; Heuberger et al., 2015) than with G. Moreover, Deck et al. have shown that downstream helper oligonucleotides (Deck et al., 2011) that provide incoming monomers with an additional binding surface improve the binding of A and U monomers, but the underlying difference in rates remains. The use of helper oligonucleotides allowed primer extension to proceed on a template containing all four bases over the course of several weeks, where previously such templates could not be copied at all.

Figure 1 with 2 supplements see all

Download asset Open asset

Here we show that the presence of a leaving group at the 5′ end of a helper oligonucleotide can be a source of significant catalysis. Activated helper oligonucleotides facilitate the addition of all four RNA monomers individually and in sequence in a single reaction. Finally, when a template is immobilized by attachment to a bead, sequential addition of activated monomers and helper oligonucleotides successfully promoted the synthesis of a significant part of an active hammerhead ribozyme.

Given the historical and continuing problems in demonstrating efficient nonenzymatic template directed primer extension, one might wonder whether this biologically inspired model is appropriate for replication within primordial cells. An alternative scenario that might seem more reasonable in the context of prebiotic chemistry involves template copying by the initial formation of short oligonucleotides, followed by ligation events that generate longer oligonucleotide intermediates and eventually a full length copy (Szostak, 2011; James and Ellington, 1999). To test the viability of this hierarchical assembly model, we compared the rate of primer extension with activated monomers to the rate of ligation of two template bound oligonucleotides (Figure 1b,c reactions 1 and 3). Nucleotides activated with 2-methylimidzole on the 5′-phosphate have been used extensively to model nonenzymatic template copying (Deck et al., 2011; Wu and Orgel, 1992c), irrespective of their prebiotic plausibility, and were used in all experiments reported here. In both the nucleotide polymerization and the oligonucleotide ligation reactions, the chemical reaction is the same, i.e. attack of the 3′-hydroxyl of an oligonucleotide on the 5′-phosphate of a downstream nucleotide, with displacement of the 2-methylimidazole leaving group (Figure 1b). Moreover, the reacting nucleotides are the same in both cases: a G at the 3′-end of the primer and a G monomer or a G at the 5′-end of an oligonucleotide. We expected the ligation reaction to be faster, because the oligonucleotides to be ligated are both stably bound to the template, and because the fully base-paired nicked duplex should be largely in the optimal A-type conformation (Kozlov and Orgel, 2000), whereas the primer/template complex with template bound monomers would be significantly more disordered. To our surprise, the primer extension reaction proceeded almost 100 times more rapidly than the ligation reaction (Figure 1d). This was true for a variety of trimer sequences (Figure 1—figure supplement 2). In our search for an explanation for this unexpected observation, we reasoned that one of the major differences between the two scenarios was the presence of multiple adjacent 5′-activated nucleotides downstream of the primer in the case of monomer addition, but not ligation, where there is only a single activated nucleotide. Indeed, the case for a role of a second downstream activated nucleotide is consistent with the well known difficulty of extending a primer to the last nucleotide of a template (Wu and Orgel, 1992c). Furthermore, Orgel et al. showed over 20 years ago that efficient primer extension requires the presence of two activated monomers adjacent to the primer. Remarkably, the reaction is fastest when both monomers are activated with 2-methylimidazole instead of either or both being activated with imidazole (Wu and Orgel, 1992a), suggesting a possible catalytic role for a physical interaction of the leaving groups of adjacent monomers.

In order to directly test the possibility that a 2-methylimidazole leaving group on the 5′-phosphate of a downstream monomer or oligonucleotide (Figure 1b) would increase the rate of reaction between a primer and an adjacent monomer, we designed a series of templates for nonenzymatic RNA polymerization and ligation (Figure 1c). We then measured the rate of addition of a monomer to a primer in the presence or absence of downstream nucleotides, with or without 5′-activation of the downstream nucleotides (Figure 1c–f). Rates were calculated assuming a pseudo-first order rate equation (Figure 1—figure supplement 1), because the concentrations of primer and template were far below that of the monomer, which would not change significantly during the reaction. Primer extension by addition of a single monomer, in the absence of any additional downstream mono- or oligo-nucleotides, was very slow and comparable in rate to the ligation of the primer to a 5′-activated GAG trinucleotide (Figure 1c–f, reactions 2 and 3). The binding of an unactivated AGC trimer, with either a 5′ hydroxyl or a 5′ phosphate, downstream of an activated G monomer conferred only a modest increase in the rate of addition of the G monomer to the primer (Figure 1c–f, reactions 4 and 5), consistent with previous reports of unactivated downstream ‘helper’ oligonucleotides (Deck et al., 2011; Jauker et al., 2015; Kervio et al., 2010). Such helper oligonucleotides are thought to act largely by increasing monomer affinity to the primer/template complex by providing an additional base-stacking surface. In contrast, when the same AGC trimer was activated as a 5′-phosphoro-2-methylimidazolide and used as a downstream helper oligonucleotide, the rate of primer extension (by addition of a G monomer) was increased by over two orders of magnitude (Figure 1c–f, reaction 6) compared to reactions with unactivated trimers. The rate of primer extension with a single G in the presence of an activated downstream trinucleotide was about twice as fast as the rate when the primer was followed by up to four sequential activated G monomers (Figure 1c–f, reaction 1 and 6).

To systematically study the rate of nonenzymatic primer extension as a function of the length of the downstream activated ‘helper’ oligonucleotide, we synthesized a series of activated oligonucleotides of different lengths (Figure 2a). At saturating concentrations (Figure 2—figure supplement 1) of each of the oligonucleotides, the rate of primer extension improved as the length of the helper increased from mononucleotide to dinucleotide to trinucleotide, and then stayed approximately constant with the tetranucleotide (Figure 2a). At subsaturating concentrations of helper oligonucleotide, the rate of the primer extension reaction increased with increasing concentrations of di- or tri-nucleotide helper (Figure 2—figure supplement 1). Because the maximum rate of the trimer-assisted reaction was significantly faster than that of the dimer-assisted reaction, we proceeded with trimers as the downstream helper oligonucleotides for the remainder of this study.

Figure 2 with 1 supplement see all

Download asset Open asset

We then tested the ability of a downstream 2-methylimidazole-activated AGC trimer to catalyze template-directed RNA primer extension with all four individual 2-methylimidazole activated monomers – A, G, C and U. We also tested the 2-thiouridine monomer, which we have previously shown to be superior to uridine in nonenzymatic primer extension in terms of both increased rate and improved fidelity (Heuberger et al., 2015). In each case, the rate of primer extension increased by at least two orders of magnitude compared to the corresponding reactions without any helper oligonucleotides (Figure 2b). Remarkably, whereas the rate of primer extension with A and U monomers was previously too slow to measure (Wu and Orgel, 1992b; Heuberger et al., 2015) – conservatively estimated to be at least three orders of magnitude slower than primer extension with G or C monomers – the difference in polymerization rates was reduced to roughly one order of magnitude when assisted by downstream activated trimers (Figure 2b).

In order to assess the fidelity of nonenzymatic trimer-assisted primer extension, we measured the rates of monomer addition for all monomer-template combinations, including both matches and mismatches (Figure 3a). We calculated the fidelity for each template as the rate of matched monomer addition divided by the sum of matched and mismatched monomer addition rates. By averaging the fidelities of the four templates we found the overall fidelity of trimer-assisted polymerization to be 98%. Assuming an error threshold of one mutation per genome (Eigen, 1971), this value of fidelity allows for effective genome sizes of 50 nucleotides, long enough to produce functional ribozymes (Ferré-D'Amaré and Scott, 2010). The fastest mismatch reaction was primer extension by G during copying of a template U, which was approximately 5% as fast as primer extension with A on the same template (Figure 3a, Figure 3—figure supplement 1). This is likely due to the formation of a G:U wobble base pair (Heuberger et al., 2015), and is consistent with previous studies of nonenzymatic primer extension reactions (Rajamani et al., 2010). All other mismatches were at least two orders of magnitude slower than their matched counterparts. We also tested the effect of replacing the canonical U monomer with 2-thiouridine and 2-thioribothymine on the polymerization rate and found, in line with previous reports (Heuberger et al., 2015), that this sulfur substitution improved the rates by approximately one order of magnitude (Figure 3b), bringing the rate of primer extension closer to the rates of G and C addition. More rigorous tests of fidelity, in which monomers can compete for binding on a template in all possible sequence contexts, are ongoing.

Figure 3 with 1 supplement see all

Download asset Open asset

Encouraged by the relatively fast and accurate addition of all four monomers to a primer, we then attempted to copy short RNA templates containing all four nucleotides. In order to iterate the process of trimer assisted monomer addition, after a downstream activated trimer catalyzes the addition of the first monomer to the primer, the helper trimer must dissociate to allow for the binding of the next monomer-trimer pair. To test the feasibility of this mode of primer extension, we designed a template that contained binding sites for all four monomers and synthesized the appropriate activated trimers (Figure 4a). We used 2-thiouridine monomer in place of uridine (see above), and all four monomers (A, G, C and 2-thiouridine) were present in every reaction. In the absence of activated trimers, we observed almost no primer extension (lane 1 Figure 4b); even though all four monomers were present, their combined rate of addition was not measurable. Adding just the first helper trimer resulted in rapid addition of the first monomer (C) to the primer, followed by slow addition of the first trimer (AGC) to the growing primer. Similarly, adding two, three or four trimers together resulted in the generation of primers extended by two, three or four monomers respectively (Figure 4b). These reactions converted primer into a product with four sequentially added monomers with an ~80% yield after 16 hr (Figure 4c,d); single monomer addition reactions reached >95% yield after only 10 min (Figure 2b). The difference between the fast rates observed in the trimer-assisted addition of a single base (Figure 2b) and the slower rate of trimer-assisted polymerization of all four bases simultaneously (Figure 4b) may be due to the competition of trimers for overlapping binding sites on the template. This hypothesis is supported by the decreased rate of trimer assisted primer extension with a single monomer in the presence of additional overlapping downstream trimers (Figure 4—figure supplement 1). The copying of templates containing two consecutive A or U monomers is also made possible by the use of activated helper oligonucleotides (Figure 5).

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 5

Download asset Open asset

If nonenzymatic RNA polymerization did indeed precede the RNA world, then it would have been necessary to synthesize RNA sequences long enough to function as ribozymes (Deck et al., 2011). While our method can only produce sequences >7 nucleotides in one reaction in low yields (Figure 6—figure supplement 1) due to the cross-inhibitory effect of trimers with overlapping binding sites (Figure 4—figure supplement 1), we reasoned that a ribozyme could be synthesized by this approach if monomers were added, along with their attendant trimers, one at a time (Figure 6). To that end, we immobilized a template that codes for the sequence of one half of the hammerhead ribozyme (HH2) – on a bead and added matched pairs of monomers and trimers sequentially (Figure 6a). Primer extension was monitored by gel electrophoresis. Despite the fact that no 2-thiouridine monomers were used (in order to protect the catalytic capacity of the ribozyme) a high conversion rate was observed at each step (75–95%) (Figure 6b). The reaction durations required to achieve these conversion rates varied widely, for example, the seventh monomer addition required 21 hr to reach 95% while the eighth addition reached the same level of conversion after just 30 min. Generally, G and C additions were much faster than A or U additions; the reason for the discrepancy between the rates of these reactions on beads and those in solution is unknown. The use of s²U or s²T might offset some of the discrepancy. The final product was released from the bead by hybridization of an excess of a DNA oligonucleotide of identical sequence to the immobilized template. After DNase treatment, the hammerhead substrate, HH1, was added to the crude synthesized pool of oligonucleotides in the presence of 200 mM MgCl₂ for 16 hr (Figure 6c). Despite the fact that this was a crude reaction product, with full-length HH2 comprising only 6% of the reaction products, the product mixture was able to function as a catalyst, with 50% ribozyme cleavage yield, thus demonstrating that a significant portion of an active ribozyme was synthesized nonenzymatically.

Figure 6 with 1 supplement see all

Download asset Open asset

The experiments described above demonstrate for the first time a simple and robust means of nonenzymatically copying mixed sequence RNA templates. Short, activated oligonucleotides – themselves plausibly generated by either templated or untemplated monomer polymerization – are efficient catalysts of high fidelity primer extension with all four RNA monomers. Thus short activated oligonucleotides can play dual roles, acting either as catalysts of chain growth by monomer addition, or directly as replication intermediates, since they can also be incorporated into a growing chain by ligation. Template copying in a complex but realistic milieu containing both activated monomers and oligomers could therefore occur via a hybrid process combining primer extension with monomers and oligonucleotide ligation.

The disparity in rates between the reactions catalyzed by activated vs. unactivated trimers indicates a key role for the leaving group downstream of the polymerizing monomer. The slow rate of activated oligonucleotide ligation is also explained by the lack of a leaving group one base downstream of the site of the primer extension reaction: in this case the downstream position corresponds to the second monomer in the oligonucleotide which necessarily has no leaving group as it is linked to the upstream monomer by a phosphodiester linkage. The nature of the interaction between the downstream leaving group and the reacting monomer remains unclear. It is possible that the downstream leaving group interacts attractively with the reacting leaving group, positioning it optimally for attack by the 3′ hydroxyl of the primer. As observed by Wu et al. (Wu and Orgel, 1992a) this interaction could be very sensitive to steric effects, such that 2-methylimidazole at the downstream position would facilitate polymerization much more effectively than imidazole, though the authors ultimately concluded that such subtle effects are not likely to adequately explain the strength of the interaction. Wu et al. (Wu and Orgel, 1992a) also suggested that the catalytic effect could be due to acid-base catalysis. Another possibility is that the upstream leaving group forms a covalent bond with the phosphate of the downstream monomer, a reaction that is only possible if there is a leaving group on the downstream monomer (Kervio et al., 2016 and T. Walton and J.W. Szostak, unpublished data). This covalent bond could form before the attack of the 3′ hydroxyl of the primer, creating an imidazolium intermediate, or in a concerted reaction where the leaving group of the upstream monomer replaces the leaving group of the downstream monomer. Further understanding of the reaction mechanism could lead to concomitant improvements in the rate and extent of template copying.

Further advances are required to enable the copying of longer and potentially functional RNA sequences in one pot, protocell-compatible conditions. However, the copying of long mixed template sequences under our conditions compares favorably, both in terms of rate and efficiency with reactions in which arbitrary RNA templates are copied by highly evolved ribozyme polymerases (Wochner et al., 2011) (Figure 6—figure supplement 1). In fact, if the rates and fidelities of nonenzymatic RNA polymerization were sufficiently high, the need for a ribozyme polymerase in the RNA world may be circumvented: RNA could be copied by nonenzymatic chemistry alone until the advent of protein polymerases.

We suggest that the identification of a plausible source of chemical energy that could drive the re-activation of monomers and oligomers following hydrolytic loss of the leaving group is a key missing component in efforts to reconstitute nonenzymatic RNA replication. A chemical environment that could maintain a fully activated pool of substrates, and avoid accumulation of the inhibitory by-products resulting from hydrolysis, might be sufficient to drive high yielding and complete copying of longer RNA templates in one pot, thus setting the stage for the emergence of Darwinian evolution.

1. 1. Beaucage SL
   2. Caruthers MH

   (1981) Deoxynucleoside phosphoramidites—A new class of key intermediates for deoxypolynucleotide synthesis

   Tetrahedron Letters 22:1859–1862.

   https://doi.org/10.1016/S0040-4039(01)90461-7

   • Google Scholar
2. 1. Cavaluzzi MJ
   2. Borer PN

   (2004) Revised UV extinction coefficients for nucleoside-5'-monophosphates and unpaired DNA and RNA

   Nucleic Acids Research 32:e13.

   https://doi.org/10.1093/nar/gnh015

   • Google Scholar
3. 1. Deck C
   2. Jauker M
   3. Richert C

   (2011) Efficient enzyme-free copying of all four nucleobases templated by immobilized RNA

   Nature Chemistry 3:603–608.

   https://doi.org/10.1038/nchem.1086

   • Google Scholar
4. 1. Eigen M

   (1971) Selforganization of matter and the evolution of biological macromolecules

   Die Naturwissenschaften 58:465–523.

   https://doi.org/10.1007/BF00623322

   • Google Scholar
5. 1. Ferré-D'Amaré AR
   2. Scott WG

   (2010) Small self-cleaving ribozymes

   Cold Spring Harbor Perspectives in Biology 2:a003574.

   https://doi.org/10.1101/cshperspect.a003574

   • Google Scholar
6. 1. Hagenbuch P
   2. Kervio E
   3. Hochgesand A
   4. Plutowski U
   5. Richert C

   (2005) Chemical Primer Extension: Efficiently Determining Single Nucleotides in DNA

   Angewandte Chemie International Edition 44:6588–6592.

   https://doi.org/10.1002/anie.200501794

   • Google Scholar
7. 1. Heuberger BD
   2. Pal A
   3. Del Frate F
   4. Topkar VV
   5. Szostak JW

   (2015) Replacing uridine with 2-thiouridine enhances the rate and fidelity of nonenzymatic RNA primer extension

   Journal of the American Chemical Society 137:2769–2775.

   https://doi.org/10.1021/jacs.5b00445

   • Google Scholar
8. 1. Inoue T
   2. Orgel LE

   (1981) Substituent control of the poly(C)-directed oligomerization of guanosine 5'-phosphoroimidazolide

   Journal of the American Chemical Society 103:7666–7667.

   https://doi.org/10.1021/ja00415a051

   • Google Scholar
9. 1. Izgu EC
   2. Fahrenbach AC
   3. Zhang N
   4. Li L
   5. Zhang W
   6. Larsen AT
   7. Blain JC
   8. Szostak JW

   (2015) Uncovering the thermodynamics of monomer binding for RNA replication

   Journal of the American Chemical Society 137:6373–6382.

   https://doi.org/10.1021/jacs.5b02707

   • Google Scholar
10. 1. James KD
    2. Ellington AD

    (1999) The fidelity of template-directed oligonucleotide ligation and the inevitability of polymerase function

    Origins of Life and Evolution of Biospheres 29:375–390.

    https://doi.org/10.1023/A:1006544611320

    • Google Scholar
11. 1. Jauker M
    2. Griesser H
    3. Richert C

    (2015) Copying of RNA Sequences without Pre-Activation

    Angewandte Chemie International Edition 54:14559–14563.

    https://doi.org/10.1002/anie.201506592

    • Google Scholar
12. 1. Joyce GF
    2. Inoue T
    3. Orgel LE

    (1984) Non-enzymatic template-directed synthesis on RNA random copolymers. Poly(C, U) templates

    Journal of Molecular Biology 176:279–306.

    https://doi.org/10.1016/0022-2836(84)90425-X

    • Google Scholar
13. 1. Joyce GF
    2. Schwartz AW
    3. Miller SL
    4. Orgel LE

    (1987) The case for an ancestral genetic system involving simple analogues of the nucleotides

    PNAS 84:4398–4402.

    https://doi.org/10.1073/pnas.84.13.4398

    • Google Scholar
14. 1. Kervio E
    2. Hochgesand A
    3. Steiner UE
    4. Richert C

    (2010) Templating efficiency of naked DNA

    PNAS 107:12074–12079.

    https://doi.org/10.1073/pnas.0914872107

    • Google Scholar
15. 1. Kervio E
    2. Sosson M
    3. Richert C

    (2016) The effect of leaving groups on binding and reactivity in enzyme-free copying of DNA and RNA

    Nucleic Acids Research.

    https://doi.org/10.1093/nar/gkw476

    • Google Scholar
16. 1. Kozlov IA
    2. Orgel LE

    (2000) Nonenzymatic template-directed synthesis of RNA from monomers

    Molecular Biology  34:781–789.

    https://doi.org/10.1023/A:1026663422976

    • Google Scholar
17. 1. Lohrmann R
    2. Bridson PK
    3. Orgel LE

    (1980) Efficient metal-ion catalyzed template-directed oligonucleotide synthesis

    Science 208:1464–1465.

    https://doi.org/10.1126/science.6247762

    • Google Scholar
18. 1. Patel BH
    2. Percivalle C
    3. Ritson DJ
    4. Duffy CD
    5. Sutherland JD

    (2015) Common origins of RNA, protein and lipid precursors in a cyanosulfidic protometabolism

    Nature Chemistry 7:301–307.

    https://doi.org/10.1038/nchem.2202

    • Google Scholar
19. 1. Powner MW
    2. Gerland B
    3. Sutherland JD

    (2009) Synthesis of activated pyrimidine ribonucleotides in prebiotically plausible conditions

    Nature 459:239–242.

    https://doi.org/10.1038/nature08013

    • Google Scholar
20. 1. Rajamani S
    2. Ichida JK
    3. Antal T
    4. Treco DA
    5. Leu K
    6. Nowak MA
    7. Szostak JW
    8. Chen IA

    (2010) Effect of stalling after mismatches on the error catastrophe in nonenzymatic nucleic acid replication

    Journal of the American Chemical Society 132:5880–5885.

    https://doi.org/10.1021/ja100780p

    • Google Scholar
21. 1. Szostak JW

    (2011) An optimal degree of physical and chemical heterogeneity for the origin of life?

    Philosophical Transactions of the Royal Society B: Biological Sciences 366:2894–2901.

    https://doi.org/10.1098/rstb.2011.0140

    • Google Scholar
22. 1. Szostak JW

    (2012) The eightfold path to non-enzymatic RNA replication

    Journal of Systems Chemistry 3:2.

    https://doi.org/10.1186/1759-2208-3-2

    • Google Scholar
23. 1. Vogel SR
    2. Richert C

    (2007) Adenosine residues in the template do not block spontaneous replication steps of RNA

    Chemical Communications pp. 1896–1898.

    https://doi.org/10.1039/b702768k

    • Google Scholar
24. 1. Wochner A
    2. Attwater J
    3. Coulson A
    4. Holliger P

    (2011) Ribozyme-catalyzed transcription of an active ribozyme

    Science 332:209–212.

    https://doi.org/10.1126/science.1200752

    • Google Scholar
25. 1. Wu T
    2. Orgel LE

    (1992a) Nonenzymatic template-directed synthesis on hairpin oligonucleotides. 2. Templates containing cytidine and guanosine residues

    Journal of the American Chemical Society 114:5496–5501.

    https://doi.org/10.1021/ja00040a002

    • Google Scholar
26. 1. Wu T
    2. Orgel LE

    (1992b) Nonenzymatic template-directed synthesis on hairpin oligonucleotides. 3. Incorporation of adenosine and uridine residues

    Journal of the American Chemical Society 114:7963–7969.

    https://doi.org/10.1021/ja00047a001

    • Google Scholar
27. 1. Wu T
    2. Orgel LE

    (1992c) Nonenzymatic template-directed synthesis on oligodeoxycytidylate sequences in hairpin oligonucleotides

    Journal of the American Chemical Society 114:317–322.

    https://doi.org/10.1021/ja00027a040

    • Google Scholar

Author details

1. Noam Prywes

   1. Department of Chemistry and Chemical Biology, Harvard University, Cambridge, United States
   2. Department of Molecular Biology and Center for Computational and Integrative Biology, Howard Hughes Medical Institute, Massachusetts General Hospital, Boston, United States

   Contribution

   NP, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0003-2526-2392

2. J Craig Blain

   Department of Molecular Biology and Center for Computational and Integrative Biology, Howard Hughes Medical Institute, Massachusetts General Hospital, Boston, United States

   Present address

   Ra Pharmaceuticals, Cambridge, United States

   Contribution

   JCB, Conception and design, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0001-5664-4065

3. Francesca Del Frate

   Department of Molecular Biology and Center for Computational and Integrative Biology, Howard Hughes Medical Institute, Massachusetts General Hospital, Boston, United States

   Contribution

   FDF, Conception and design, Acquisition of data, Analysis and interpretation of data

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0003-2782-1290

4. Jack W Szostak

   1. Department of Chemistry and Chemical Biology, Harvard University, Cambridge, United States
   2. Department of Molecular Biology and Center for Computational and Integrative Biology, Howard Hughes Medical Institute, Massachusetts General Hospital, Boston, United States

   Contribution

   JWS, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   szostak@molbio.mgh.harvard.edu

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0003-4131-1203

• 3,533

  views

• 815

  downloads

• 122

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.