RNA copies itself without a catalyst

The RNA-world hypothesis has a famous hole in the middle of it. For RNA to have carried heredity before proteins existed, it must have been able to replicate itself without enzymes, yet decades of experiments found that self-copying RNA was either too slow, too short, or too dependent on carefully engineered conditions. A paper in the 12 February issue of Science reports a step that has been missing for most of that history: short RNA molecules, roughly 45 bases long, that copy themselves into complementary templates and then use those templates to reconstruct the original strands, all without any protein catalyst.

The demonstration works in two stages, and the two-stage structure is precisely what makes it interesting. In the first stage, an RNA strand directs the assembly of its complement, base by base, from activated building blocks. In the second, that freshly made complement does the same job in reverse, regenerating the original sequence. Each strand therefore acts as both template and product, and the cycle can in principle continue as long as the building blocks last. The authors found three such molecules, and the modest number is itself a clue: this is a property that some sequences have and others do not, which is exactly the kind of selectivity a primitive genetic system would need.

Two caveats are stated plainly in the paper, and they matter. First, no single RNA molecule has yet been shown to perform both reactions, copying its own complement and then copying itself from that complement, in one continuous cycle. The two steps have been demonstrated, but stitched together rather than observed as a closed loop. Second, the copying is slow. The experiments ran at freezing temperatures, where ice crystals concentrate the reactants and shield them from water, and synthesising a new RNA strand took on the order of 72 days. That is not a speed that would impress a modern polymerase, but it is a pace at which an ancient, enzyme-free chemistry could plausibly operate over geological time.

The choice of cold conditions is not a weakness but part of the argument. One of the persistent objections to non-enzymatic RNA replication is that the reaction is poisoned by water, which competes with the incoming nucleotides and promotes error and strand degradation. Freezing concentrates the reagents in liquid veins between ice crystals and effectively removes much of the water from the reaction, solving in the laboratory a problem that the early Earth would have solved naturally in seasonal ice. If RNA chemistry could only proceed at useful rates in frozen environments, it would place the origin of replication in a particular kind of landscape, one with repeated freeze-thaw cycles, and that is a falsifiable prediction rather than a hand-wave.

The finding does not close the RNA-world debate, and it is not presented as such. A self-replicating system needs more than template copying: it needs to copy its complement faithfully enough to survive selection, and it needs error rates low enough that information is not erased within a few generations. Those thresholds remain to be measured. What has changed is that the most obvious physical obstacle, the absence of any enzyme at all, has been shown to be surmountable with nothing more exotic than a short sequence and a cold winter.