NATURE 

Origins of life: The cooperative gene

The origin of life on Earth remains one of the great unsolved mysteries. A new study suggests that cooperation among molecules could have contributed to the transition from inanimate chemistry to biology.

Cooperation operates at all scales of life, from whole organisms, such as wolves hunting in packs, to individual cells acting in a coordinated fashion during development or organ function. In a paper published on Nature's website today, Vaidya et al.¹ describe networks of RNA molecules that assemble one another, suggesting that cooperation may be as old as life itself.

The molecular architecture of modern-day organisms is based around a division of labour: the nucleic acids DNA and RNA are used mainly for the storage and processing of genetic information, with proteins fulfilling metabolic and structural roles. However, there is compelling evidence for a primordial biology that lacked DNA and proteins and instead relied on RNA for both heredity and metabolism². A cornerstone of this 'RNA world' is self-replication by RNA molecules that also mutate and hence evolve towards ever more efficient self-replication.

But how did such a self-replicating RNA — the original 'selfish gene' — arise from the chemical ingredients present on the early Earth? Recent advances in prebiotic chemistry² (the study of the chemical reactions that might have led to the formation of the molecules typical of today's organisms) offer glimpses of how RNA's building blocks could have accumulated and polymerized into short chains². Indeed, even some very short RNAs can perform chemical reactions³ (and are therefore called RNA enzymes, or ribozymes). But it seems likely that the more complex functionalities required for self-replication would necessitate the assembly of longer, structurally more complex ribozymes, which known prebiotic reactions do not produce.

Vaidya and colleagues' remarkable work points to a possible strategy to begin bridging this gap, based on a principle of self-organization first proposed more than 30 years ago⁴. In this scenario, self-replicating RNA entities go beyond simply making copies of themselves and act on other replicators through a cyclic network of reinforcing loops called hypercycles (Fig. 1). The authors' laboratory had previously described a ribozyme — from an Azoarcus bacterium — that had the ability to assemble itself when fragmented⁵. Now Vaidya et al. show that variants of such RNA fragments can assemble and act on one another to form cooperative self-assembly cycles very much like the proposed hypercycles, in which ribozyme 1 aids assembly of ribozyme 2; 2 aids 3; and 3 aids 1 (Fig. 1).

Figure 1: The emergence of hypercycles.

a, A primordial replicator molecule (R) enhances its own assembly from substrate molecules (S) in a simple autocatalytic cycle. b, Imperfect replication generates a set of related replicators, each promoting the synthesis of all the others. c, d, The introduction of biases in replicator specificity gives structure to the network and can lead to selfish subsystems (c) or to a cooperative 'hypercycle' (d), akin to the system described by Vaidya and colleagues¹. Such hypercycles remain globally autocatalytic, but are more resistant to the accumulation of mutations, enabling replicators to specialize and to acquire new functions. Thick and dashed red arrows indicate increased and decreased efficacy, respectively, at enhancing replicator assembly.

• Full size image (189 KB)

The authors' key finding is that, through such cooperative cycles, participating RNAs gain an advantage and can outcompete selfish replication cycles, in which a particular fragment assembles itself. Cooperation also allowed full-length ribozyme assembly from sets of four different RNA fragments. Thus, cooperation between small RNA molecules can aid the emergence of longer, more complex RNAs.

The authors describe a certain three-member cooperative cycle in great detail, but the data in one of their experiments hint at the potential for much larger cycles and networks of cooperating RNAs. This observation in particular suggests many lines of investigation that could advance our understanding of molecular cooperation and its significance to the RNA world. Questions for future enquiry include how such networks develop over time, and whether network complexity scales with efficiency — that is, whether larger or more interconnected networks always replicate more efficiently than simpler alternatives.

How might such networks have arisen (and persisted) in the pools of random RNA chains generated on the early Earth? In the present study, all members of the pool are derived from a set of 'prefabricated' fragments of the Azoarcus ribozyme. It will be important to determine how cooperative RNA networks perform in the presence of many unrelated and potentially interfering RNAs, and how much sequence variation within the Azoarcus fragments can be tolerated before self-assembly is abolished. The present work is encouraging in this respect, as it shows that limited sequence diversity in the three-member system yielded better assembly than defined fragments, demonstrating that some sequence variation can be harnessed for gains in efficiency.

Comparison with an earlier two-component system, in which two ribozymes catalysed each other's synthesis from a mixture of four fragments⁶, is illustrative. This system displayed exponential self-replication and, when seeded with fragments of defined sequence variation, yielded a diverse pool of recombinant molecules, some of which were more efficient replicators than the initial ones. Thus, such molecular systems can harness the powerful evolutionary potential of recombination to reassort themselves into more active replicators. Vaidya and colleagues' use of a ribozyme system that had a larger degree of freedom in the choice of assembly partners has now enabled networks to develop beyond this two-component system.

However, the need for defined RNA components is likely to constrain the evolutionary potential of such systems, because recombinants are unable to break away from the prescribed component structure. A more general capacity for self-replication and evolution would require a different type of system that has the ability to copy genetic information — akin to present-day biology, in which RNA or DNA sequences are replicated by polymerase enzymes 'letter by letter' from monomer units. Although RNA-polymerizing ribozymes have been described⁷, their activity falls short of self-replication, despite recent improvements⁸.

The excursions into 'molecular ecology' described by Vaidya et al. suggest that cooperative networks might be designed to harness the best of both types of system, if synthesis of short RNAs by polymerizing ribozymes could be coupled to a ribozyme system capable of self-assembly⁹. Such networks might outperform replicators that go it alone, and exploit recombination to resist the gradual accumulation of harmful mutations¹⁰ and the concomitant deterioration of the encoded genetic information. Finally, complete covalent assembly might not be essential for higher-order functions such as molecular kin recognition and polymerizing activities. Indeed, non-covalent assembly of multiple RNA chains into functional complexes has precedents in modern-day biology, notably the ribosome, a large complex of multiple RNA and protein chains, which catalyses protein synthesis and may date back to the RNA world.

The precise molecular events that led to the origin of life on Earth are likely to be lost in time, but science can construct molecular 'doppelgängers' of the ancestral molecules and explore the plausibility of different ways in which the transition from prebiotic to biotic matter might have occurred. Vaidya and colleagues make a persuasive case for the benefits of cooperation even at this nascent stage of life. The first genes may not have been so selfish, after all.