RNA made in its own mirror image

An RNA enzyme has been generated that can assemble a mirror-image version of itself. The finding helps to answer a long-standing conundrum about how RNA molecules could have proliferated on prebiotic Earth. See Letter p.440

Many organic and biological molecules come in right-handed and left-handed versions that are mirror-image twins of one another. These variations are referred to as D- and L-enantiomers, respectively. Modern RNA molecules are linear polymers that are synthesized from ribonucleotide monomers, and take the D-form. But on page 440 of this issue, Sczepanski and Joyce¹ suggests that early evolution may have involved an interplay between the D- and L-structures of RNA.

Before DNA and proteins existed, RNA may have evolved as the primordial macromolecule that could both store information like DNA does and catalyse chemical reactions like many proteins do. According to this 'RNA world hypothesis'², one of the functions of these RNA enzymes (called ribozymes) was to replicate other RNA molecules by using their sequences as templates to make complementary strands. This function, called polymerization, involves the chemical joining of ribonucleotide monomers or oligonucleotides (short sequences of monomers).

Some 30 years ago, a conundrum arose concerning how RNA molecules first proliferated through prebiotic chemical reactions. This was because of the demonstration by Joyce et al.³ that the non-enzymatic copying of an RNA template to form a complementary RNA strand could be brought to a screeching halt by the incorporation into the growing polymer of monomers of opposite handedness to the template. This phenomenon was termed 'enantiomeric cross-inhibition'. Given that both D- and L-enantiomers of RNA molecules were probably present as substrates on prebiotic Earth, how could template-directed polymerization have proceeded? Sczepanski and Joyce now revisit this issue by creating a ribozyme that not only catalyses template-directed polymerization in the presence of both D- and L-enantiomers, but actually prefers mononucleotides and oligonucleotides of the opposite handedness to itself as its substrates.

The authors synthesized a pool of right-handed D-RNA polymers of random sequences and linked them covalently to a left-handed L-RNA template in the presence of left-handed oligonucleotide substrates. They then used in vitro selection^4,5,6 to isolate RNA species from the pool that could join (polymerize) the substrates. After ten rounds of selection and amplification of catalytic molecules; pruning of superfluous sequences; insertion of another randomized segment to create a new pool; and then another six rounds of selection and amplification, a D-ribozyme was isolated that could perform template-directed joining of L-substrates about a million times faster than in the uncatalysed reaction¹.

As with ribozymes previously selected and further optimized for polymerization activity^7,8,9, this ribozyme resembles modern-day polymerization enzymes (polymerases) in several ways. First, it can operate on completely separate template–substrate complexes, implying that sequence-independent contacts form between it and the complexes. Second, it can perform limited polymerization by catalysing the sequential joining of several mononucleotides. In addition, as long as the oligomeric substrates are bound to their complementary templates, the ribozyme seems to be indifferent to substrate length. In fact, the authors observed that it can connect 11 L-oligonucleotides to form a mirror copy of itself, a remarkable first demonstration of an enzyme (RNA or protein) being synthesized by its own enantiomer. Importantly, the D-ribozyme and its L-enantiomer efficiently catalyse their respective joining reactions even in a mixture containing both D- and L- versions of the substrates and templates. In other words, the enantiomeric cross-inhibition that thwarted non-enzymatic template-mediated replication does not occur.

This work adds weight to the notion of a primordial RNA world in which cycles of cross-handed replication used mirror-image forms of RNA (Fig. 1). Such mutualistic coupling of D- and L-RNA polymerases might have conferred several advantages on RNA evolution, and may now benefit researchers who aspire to create RNA polymerase systems that can self-replicate. Because of their shapes, D- and L-RNA molecules cannot form consecutive Watson–Crick base pairs with each other¹⁰, just as left and right hands cannot properly handshake one another. Consequently, Sczepanski and Joyce's polymerase is unlikely to exhibit sequence preferences or restrictions owing to duplex formation between complementary sequences in the ribozyme and in templates or reaction products. These problems have confounded the experimental search for a general polymerase that can copy RNA sequences without bias, and may similarly have affected the course of early macromolecular evolution. Furthermore, the dependence on two distinct, coupled polymerases makes the replication cycle less susceptible to invasion by molecular parasites that could usurp the chemically activated substrates needed for the polymerization reaction.

Figure 1: Possible mechanism for RNA replication on prebiotic Earth.

Sczepanski and Joyce¹ have generated an RNA enzyme (a ribozyme) that catalyses the polymerization of oligonucleotides of the opposite handedness to itself: the right-handed D-ribozyme yields the left-handed L-ribozyme, and vice versa. This adds weight to the idea that a cross-handed cycle involving both D- and L-ribozymes may have replicated RNA on prebiotic Earth. In the cycle, the L-ribozyme acts on a complex formed between a D-template RNA strand and D-oligonucleotides, joining the latter together to form a duplex RNA product. Separation of the duplex's strands liberates the D-ribozyme. This then catalyses formation of the L-ribozyme from the left-handed template–oligonucleotide complex.

Viewing the work in the context of evolution begs the question of how an all D- or all L-ribozyme could have arisen to begin with. Sczepanski and Joyce suggest that simple forms of nucleic acids that lacked the mirror twin served as templates for polymerization of RNA mono- or oligonucleotide substrates on prebiotic Earth. It remains unclear, however, whether RNA polymerization from such templates would be immune to deleterious enantiomeric cross-inhibition.

Beyond the implications for the RNA world hypothesis, the new ribozyme may have practical value for the production of spiegelmers¹¹ — L-versions of functional D-RNAs (the name derives from the German word for 'mirror': Spiegel). Spiegelmers resist degradation by nucleases — the enzymes that degrade nucleic acids — and seem to avoid detection by the immune system, making them attractive therapeutic candidates and sensors for biological ligands. However, because natural enzymes do not recognize L-nucleotides, spiegelmers can be made only by chemical synthesis, which limits access to longer spiegelmers. Ribozyme-catalysed cross-handed polymerization might enable convenient enzymatic access to spiegelmers, and eventually render them directly amenable to in vitro selection methods.

Sczepanski and Joyce's twin polymerases will probably require further engineering before they can copy long RNA templates of any sequence efficiently and accurately. Nevertheless, successive improvements have been made for other in vitro-selected ribozymes^8,9, providing reason for optimism in this case.