A paper published in Nature Chemistry in May 2025 describes what its authors call the first demonstration of exponential RNA replication by a polymerase ribozyme under conditions that could plausibly have occurred on the early Earth. The work comes from Dr James Attwater and Dr Philipp Holliger at the MRC Laboratory of Molecular Biology in Cambridge, with co-authors at UCL Chemistry.
This is one paper addressing one specific obstacle in a long-running debate about how life began. The finding is worth taking seriously, but it should not be read as the final word on the origin of life, nor as confirmation that the RNA world hypothesis is settled.
The problem the paper set out to solve
The RNA world hypothesis holds that the earliest self-replicating chemistry on Earth was based on RNA molecules, which are capable of both storing genetic information and acting as catalysts. Under this model, RNA came before DNA and before protein-based enzymes, and the first crucial step was an RNA molecule that could copy itself.
Getting that to work in the laboratory has run into a persistent obstacle the field calls the strand separation problem. When an RNA strand copies itself, it produces a complementary partner strand, and the two bind together into a stable double helix. The stability is the problem: RNA duplexes form quickly and hold tightly, like velcro that zips shut faster than it can be pulled apart. Before the copies can serve as templates for the next generation of replication, the strands need to be separated. In modern cells, protein-based enzymes handle this. In the world before proteins, there was nothing to do it. Prior work in this area had demonstrated template copying and pieces of the replication process, but not a complete, repeatable replication cycle that could plausibly have operated before biology existed to assist it.
This is the gap the Attwater-Holliger paper addresses. The abstract of the paper states the matter plainly: “both enzymatic and non-enzymatic RNA replication cycles are impeded by the ‘strand separation problem’, a form of product inhibition arising from the extraordinary stability of RNA duplexes and their rapid reannealing kinetics.”
What the researchers did
The key innovation was the use of trinucleotides: building blocks composed of three RNA letters rather than the single-letter nucleotides used in standard replication. Trinucleotides do not appear in biology today. The team used trinucleotide triphosphates as the substrate for a polymerase ribozyme, an RNA molecule capable of catalysing the copying of other RNA strands.
The chemistry worked as follows. RNA strands in solution were first subjected to acid and heat, which separated the double helix. The solution was then neutralised and frozen. In the thin liquid channels that form between ice crystals during freezing, the trinucleotides concentrated and coated the separated RNA strands, holding them in a single-stranded state and preventing them from zipping back together. Replication of those strands then proceeded in those liquid veins. Thawing and repeating the cycle, driven by alternating pH and temperature, allowed exponential replication to continue over multiple rounds.
Crucially, replication was exponential and open-ended. The paper reports that both the positive and negative strands of the RNA duplex were replicated, and that the system was applied to random RNA sequence pools, yielding either defined replicating sequences or gradually diversifying pools. The authors also observed that sequence composition drifted toward what they describe as hypothesised primordial codons, the earliest precursors to the genetic code.
The proposed environmental analogue is a geothermal freshwater setting: a warm spring or pool where heat from underground rock meets a cold surface atmosphere, producing repeated freeze-thaw cycles across the day. Saltwater does not work; the presence of salt disrupts the freezing process and prevents the required concentration from building up. Evaporation is also ruled out as a concentrating mechanism, because RNA degrades at the elevated temperatures involved.
What makes this result distinct from prior work
It is worth being precise here, because the RNA world literature has produced several significant results in recent years, and this one is easy to conflate with adjacent findings.
A 2024 paper from Gerald Joyce’s group at the Salk Institute demonstrated an RNA polymerase ribozyme capable of copying strands with substantially higher accuracy than earlier versions, including a small self-cleaving RNA. That work addressed the fidelity problem in RNA replication. A separate line of work from Holliger’s own laboratory, published in early 2026, described a small ribozyme called QT45 capable of synthesising both itself and its complementary strand, though not yet simultaneously in one pot.
The May 2025 paper addresses a different problem entirely: not fidelity, and not the capacity to copy RNA at all, but the strand separation bottleneck that was blocking any replication cycle from turning over. Without strand separation, you can copy an RNA strand once but cannot propagate the copies. The trinucleotide-freeze-thaw mechanism provides a physical, chemistry-based solution to that bottleneck that requires no protein machinery and no biological infrastructure. That is what the “first time” claim in the paper’s framing refers to.
The limits of the result
The paper does not provide an end-to-end account of how life began. The authors do not claim otherwise. Attwater is quoted in the UCL press release as noting that LUCA, the Last Universal Common Ancestor of all known life, is “a pretty complex entity” with a great deal of evolutionary history hidden behind it. The RNA world hypothesis itself describes a period that predates any known fossil or molecular trace.
The trinucleotide building blocks used in the experiment do not occur in biology today, which the team acknowledges. Their logic is that the earliest life forms were likely quite different from anything we know; the chemistry of the very first replicators was probably simpler and messier than what has been preserved in any organism. This is a reasonable inference but not a confirmed historical fact.
The origin of life is not, in the authors’ or anyone else’s view, an RNA-only story. The current consensus in the field is that RNA, peptides, lipids, and simple metabolic chemistry were probably all involved, and that these components emerged and began interacting in a prebiotic environment we cannot directly observe. Several other UCL and MRC groups, including those led by Dr John Sutherland and Professor Matthew Powner, have been working in parallel on how nucleotides, amino acids, lipids, and vitamins could have assembled from simpler precursors. This paper addresses one step in what remains an incompletely solved problem.
What to watch next
The most immediate question the paper opens is whether the trinucleotide-freeze-thaw mechanism can be extended to longer RNA sequences and, eventually, to the self-replication of the ribozyme itself under the same prebiotically plausible conditions. As Holliger put it in the press release: “Life is separated from pure chemistry by information, a molecular memory encoded in the genetic material that is transmitted from one generation to the next.” The gap between a replication cycle that works on short defined sequences in a controlled laboratory and a self-sustaining system capable of evolution remains real, and the field will take some time to cross it.
The paper’s observation that replicated random RNA sequences drifted toward hypothesised primordial codons is, if it holds up to scrutiny and replication by other groups, a separately interesting result: it suggests that the replication chemistry itself may have imposed structural biases on the early genetic code, rather than the code emerging purely through selection. That is a speculative interpretation at this stage, and the authors frame it carefully.
A result of this kind is reproduced and stress-tested by other laboratories over months and years. What the paper does now is give the field a mechanism to work with, and a set of conditions specific enough to test further.
