8.5 Tinkering with Nature
129
Benner’s group (Hoshika et al, 2019) synthesized eight-letter DNA. What next?
Romesberg (2019) declares: Our goal is to further blur the boundary between the
inanimate and the living, to create manmade parts that work within living systems,
and thereby seamlessly alter or increase what they are capable of, ultimately creating new organisms with new attributes. Nobody knows, of course, what the consequences will be. But once something becomes possible, it also becomes unstoppable.
A less dangerous direction is making use of genetic machinery for non-genetic
aims. I mentioned in the preceding section how attached DNA strings can facilitate
the formation of colloidal crystals. In this case, nanoparticles play a passive role: a
variety of structures can be built from pure DNA, exploiting the ability of complementary nucleotide bases to selectively link together. Joining DNA strands by their
sticky ends has a long history (Cohen et al, 1973); the conjunction was even inserted
into E. coli and proved to be functional. In this way, it was also possible to assemble
a cubic DNA molecule (Chen and Seeman, 1991) and “DNA origami” with fancy
shapes (Rothemund, 2006), as shown in the bottom panels of Fig. 8.11. Far from being just toys, DNA origami were later proposed as vehicles for drug delivery (Jiang
et al, 2012).
Other polyhedra were constructed (He et al, 2008) by preparing junctions built
from three single strands, as shown in the upper left-hand panel of Fig. 8.11, to be
placed at vertices of a three-dimensional supramolecule, like those shown on the
same figure. Niles Pierce and coworkers (Yin et al, 2008) assembled the various
structures shown in the upper right-hand panels of Fig. 8.11 through catalytic formation of branched junctions that spontaneously combine in irregular forms. They
also synthesized an autonomous bipedal DNA walker capable of moving along a
DNA track, just as natural protein molecular motors walk along protein filaments
(Sect. 5.2). RNA structures can be designed and manipulated just as simply and
flexibly, while retaining their catalytic functions (Guo, 2010).
Early diffraction images of microcrystalline DNA (Franklin and Gosling, 1953)
were instrumental in the discovery of its double helix structure by James Watson and
Francis Crick (1953). The design techniques developed in later years enabled selfassembly of varied crystalline structures (Paukstelis and Seeman, 2016). DNA crystals were proposed as scaffolds for orienting and positioning guest molecules, such
as protein enzymes, and smart porous materials for molecular separation. These
tasks are made precise thanks to the rigidity of DNA molecules and the specificity
of interactions between base pairs. But the ultimate unnatural application of specific
interactions between nucleotide bases is molecular computing.
Already at the dawn of computer technology, Richard Feynman (1961) envisaged
miniaturization going right down to the molecular level. DNA is, in its essence, a
digital coding element, and can be viewed therefore as a natural tool for molecular computing. Duplication or protein coding proceeds, like a Turing machine, by
scanning DNA as a data tape. The first attempt to use DNA for computation was
undertaken by Leonard Adleman (1994). He set out to solve the classical traveling
salesman problem, that is, to find a route on a graph that visits each node exactly
once, which is called a Hamiltonian path. The way to do this was first to synthesize
129
Benner’s group (Hoshika et al, 2019) synthesized eight-letter DNA. What next?
Romesberg (2019) declares: Our goal is to further blur the boundary between the
inanimate and the living, to create manmade parts that work within living systems,
and thereby seamlessly alter or increase what they are capable of, ultimately creating new organisms with new attributes. Nobody knows, of course, what the consequences will be. But once something becomes possible, it also becomes unstoppable.
A less dangerous direction is making use of genetic machinery for non-genetic
aims. I mentioned in the preceding section how attached DNA strings can facilitate
the formation of colloidal crystals. In this case, nanoparticles play a passive role: a
variety of structures can be built from pure DNA, exploiting the ability of complementary nucleotide bases to selectively link together. Joining DNA strands by their
sticky ends has a long history (Cohen et al, 1973); the conjunction was even inserted
into E. coli and proved to be functional. In this way, it was also possible to assemble
a cubic DNA molecule (Chen and Seeman, 1991) and “DNA origami” with fancy
shapes (Rothemund, 2006), as shown in the bottom panels of Fig. 8.11. Far from being just toys, DNA origami were later proposed as vehicles for drug delivery (Jiang
et al, 2012).
Other polyhedra were constructed (He et al, 2008) by preparing junctions built
from three single strands, as shown in the upper left-hand panel of Fig. 8.11, to be
placed at vertices of a three-dimensional supramolecule, like those shown on the
same figure. Niles Pierce and coworkers (Yin et al, 2008) assembled the various
structures shown in the upper right-hand panels of Fig. 8.11 through catalytic formation of branched junctions that spontaneously combine in irregular forms. They
also synthesized an autonomous bipedal DNA walker capable of moving along a
DNA track, just as natural protein molecular motors walk along protein filaments
(Sect. 5.2). RNA structures can be designed and manipulated just as simply and
flexibly, while retaining their catalytic functions (Guo, 2010).
Early diffraction images of microcrystalline DNA (Franklin and Gosling, 1953)
were instrumental in the discovery of its double helix structure by James Watson and
Francis Crick (1953). The design techniques developed in later years enabled selfassembly of varied crystalline structures (Paukstelis and Seeman, 2016). DNA crystals were proposed as scaffolds for orienting and positioning guest molecules, such
as protein enzymes, and smart porous materials for molecular separation. These
tasks are made precise thanks to the rigidity of DNA molecules and the specificity
of interactions between base pairs. But the ultimate unnatural application of specific
interactions between nucleotide bases is molecular computing.
Already at the dawn of computer technology, Richard Feynman (1961) envisaged
miniaturization going right down to the molecular level. DNA is, in its essence, a
digital coding element, and can be viewed therefore as a natural tool for molecular computing. Duplication or protein coding proceeds, like a Turing machine, by
scanning DNA as a data tape. The first attempt to use DNA for computation was
undertaken by Leonard Adleman (1994). He set out to solve the classical traveling
salesman problem, that is, to find a route on a graph that visits each node exactly
once, which is called a Hamiltonian path. The way to do this was first to synthesize