Slipping in extra benzene rings creates a broader DNA double helix that is similar to, but different from, natural DNA. Importantly, it can encode more genetic information — and that could have wide implications.
Compared with proteins, which can have any number of jumbled forms, the DNA double helix — that most famous of nature's molecular structures1— is comfortingly regular. Over the past few years, however, Eric Kool and his colleagues have re-engineered this predictable duplex, producing an 'expanded' version that they call xDNA2. Writing in the Journal of the American Chemical Society3, these researchers further elucidate the structure of this architectural delight. Their results indicate that, like a highway with a lane added to convey more traffic, the wider duplex can carry more genetic information than can standard DNA.
Probing and manipulating the dimensions of nucleic acids is nothing new for chemists. In fact, the first synthesis and characterization of size-expanded DNA bases was carried out in the 1970s4, and expanded bases are still being developed and used to study protein specificity, and as a route to drug development5,6. But these studies all concentrate on the single basic unit of DNA — the nucleotide — rather than the oligonucleotide polymer from which DNA strands are made. Full-blown, double-stranded DNA is constructed by the selective pairing of the bases of two oligonucleotides through hydrogen bonding to form the 'rungs' of the duplex. Kool and colleagues' study is unique in that it exploits the expanded bases to createnew criteria for base-pairing among oligonucleotides, and thus produces an entirely different duplex structure.
xDNA is made up of analogues of the nucleotides of natural DNA — guanine (G), cytosine (C), adenine (A) and thymine (T) — that are dubbed xG, xC, xA and xT. These nucleotides carry bases that are identical to their natural counterparts, but they are expanded at right angles to the duplex's long axis by the addition of a benzene ring (Fig. 1). In natural DNA, G always pairs up with C, and A with T. The expanded analogues maintain their natural hydrogen-bonding topology, so they can match up with their complementary base to form pairs that are wider than normal by a single benzene ring. But accommodating a single expanded pair within a natural duplex requires significant distortion of the duplex backbone, and is therefore energetically disfavoured. Likewise, a pair formed between two natural bases is too narrow to fit into a duplex in which all the other pairs are expanded.
Previous studies by Kool and colleagues2,7 have shown that two strands of modified DNA can combine to form a stable xDNA duplex provided that just one member of every base pair is expanded. This is true regardless of whether the expanded bases are all within one strand, or spread between both. Given the hydrogen-bonding requirements, there are thus four viable partnerships in the expanded duplex: xG–C, G–xC, xA–T and A–xT. This is twice as many as in normal DNA, so the expanded duplex can encode more information.
In their latest paper3, Kool et al. present us with the first picture of xDNA containing all four possible base pairs. The studies confirm that such an xDNA duplex bears many of the same features as B-DNA (the right-handed-twisting variant that is the most common natural form of DNA). But it has several interesting differences, too. For example, the new structure reveals extensive inter- and intra-strand stacking between expanded bases. This stacking probably explains the remarkable stability of the xDNA duplex. And, to accommodate the increased diameter of the bases, twelve base pairs are required to complete one turn of the expanded duplex, rather than the ten of the natural duplex. The so-called major and minor grooves of the xDNA duplex are therefore also markedly wider and shallower. As these grooves mediate many of the interactions that underlie protein and small-molecule binding to DNA, such molecules might interact with xDNA in different, and perhaps useful, ways.
The spontaneous assembly of xDNA into a unique and regular duplex containing four unique base pairs could make the material useful in various bio- and nanotechnological applications. Unlike natural DNA, xDNA bases are fluorescent, making spectroscopy easier to perform. And as xDNA can form twice as many base pairs as natural DNA, it could also provide an unexplored avenue towards expanding the genetic alphabet8.
Daunting challenges remain. Most notably, taking advantage of the increased information potential of xDNA will require, at a minimum, a polymerase capable of replicating the duplex. The polymerases that replicate natural DNA are very sensitive to the geometry of a base pair9, so it seems unlikely that any natural polymerase will accommodate the altered dimensions of the xDNA. Although the directed evolution of polymerases — creating variants of the enzyme with specific functions through in vitro mutagenesis and selection — has had some success10, something as radical as producing an xDNA polymerase has never been attempted. This would in itself be a milestone in directed evolution.
The structure of xDNA raises as many fascinating questions as it answers. It will be particularly interesting to see what kinds of duplex structures DNA forms when its bases are expanded in different directions. Kool and colleagues have already reported several other types of expanded base11, so answers could be swift in coming. Further biochemical and biophysical study of these duplexes should uncover more surprises and also provide general information about interactions between nucleic acids and proteins. It is no small accomplishment to re-engineer one of nature's oldest and most universal molecules. Like its slimmer progenitor, xDNA should be around for years to explore, study and enjoy.