Programming molecular structures and systems

Background - nanofabrication by nucleic acid self-assembly

The interactions between molecules of the nucleic acids DNA and RNA are controlled by the sequences in which the component bases A, C, G, T/U are concatenated. The information stored in their base sequences can therefore be used to program their assembly and behaviour. We use this control to build molecular-scale structures by self-assembly, to program cascaded reactions that compute, and to create synthetic molecular machinery.

Programming assembly

Four strands of DNA are designed to assemble a DNA tetrahedron1,2

Strands of DNA (and RNA) interact through predictable Watson-Crick pairing between the bases A, C, T, G: two strands bind (hybridize) to form an antiparallel duplex in which C binds G, A binds T/U. The figure above shows a simple example: four strands of DNA are designed to bind to each other to form a DNA tetrahedron with 7 nm edges. Nucleic acid nanostructures can often be made by simply annealing the component strands. This is massively parallel nanofabrication with near-atomic precision – with a component cost of the order of £10-15 per copy.

The assembly of more complex nanostructure raises interesting physical questions - such as how to engineer a structure to minimize assembly defects3.

1. R.P. Goodman et al., Science 310, 1661-1665 (2005)
2. T. Kato et al., Nano Lett. 9, 2747-50 (2009)
3. K. Dunn et al., Nature 525, 82-86 (2015)

Programming dynamics

A simple DNA machine10

It is possible to make nucleic acid nanostructures structures that can respond to changes in their environment or to specific molecular cues. In the simple demonstration above, molecular tweezers, comprising double-helix arms connected by a flexible hinge of single-stranded DNA, close and open in response to sequential addition of two signalling strands of DNA. (We observe these transitions by measuring changes in fluorescence from attached dye molecules). The use of auxiliary strands of DNA to trigger state changes suggests that responsive nanostructures could be designed to respond to natural RNA signalling molecules for applications in biology and medicine. More sophisticated cascades of DNA binding and competitive displacement can be used to compute. Other structural motifs can respond to recognition of specific molecular triggers (such as proteins) or temperature or pH change.

10. B. Yurke et al., Nature 406, 605-8 (2000)

Given that we can build atomically precise molecular structures that compute, sense and respond to stimuli, what should we do?

Current projects include ...

Programming chemical synthesis (a synthetic ribosome)

Life depends on precise, sequence-controlled polymer synthesis by the ribosome, a natural molecular machine that reads genes and writes the corresponding proteins by concatenating amino-acid building blocks. Our ambition is to create a synthetic ribosome, a molecular machine capable of translating a synthetic genetic code into a completely synthetic, sequence-defined polymer.


Our research is based on the principle of DNA-templated synthesis20,21. Building blocks for polymer synthesis are attached to DNA adapters (analogous to tRNAs) that serve to identify them. These adapters are used by our molecular machinery to select building blocks and bring them into proximity, enabling their concatenation in a programmed sequence.

Autonomous, programmed synthesis22

With Rachel O'Reilly's group (Birmingham Chemistry) we have demonstrated programmed bond formation using both natural and unnatural backbone chemistries (peptide and Wittig, respectively), including sequence-programmed synthesis of 10-unit oligomers23 and the autonomous creation of DNA-tagged combinatorial libraries22. We are developing better ways to integrate the chemistry of bond formation with the operation of the underlying DNA machinery to improve yield and reduce synthesis errors.

Genetically programmed synthesis by molecular machinery will allow exploration of vast new regions of chemical space by selection and evolution using libraries of products (1012 ~ 1015) created from combinatorial libraries of genes. This will open up many exciting applications including the directed evolution of non-natural polymers to match and extend the functionalities of peptides and proteins. Molecular machinery for programmed synthesis will also provide a platform to recreate complex biological behaviours such as gene regulation, bringing the creation of artificial living systems a step closer.

20. M.L. McKee et al., Angew. Chem. Int. Ed. 49, 7948 (2010)
21. C.T. Calderone, et al., Angew. Chem. Int. Ed. 41, 4104-8 (2002)
22. W. Meng et al., Nature Chem. 8, 542-548 (2016)
23. P.J. Milnes et al., Chem. Commun. 48, 5614-6 (2012)

DNA robotics

Bipedal molecular motor30

Nanostructures that sense, compute and actuate have all the properties necessary to implement molecular robotics. Responsive nanostructures underpin our work on synthetic molecular motors (above), programmed chemical synthesis, and molecular computation.

Actuation of a reconfigurable nanostructure31

Transitions between states can be driven by strand-displacement reactions in which one DNA or RNA molecule replaces another in a duplex (cf. the opening and closing transitions of the device shown above). Strand displacement, like nanostructure assembly, is controlled through design of the base sequences of the reacting strands: the dynamic behaviour of nucleic acid structures and systems is thus programmable. Strand displacement creates new species which can initiate further hybridization reactions: cascaded strand displacement is the basis of many schemes for molecular computation32. Transitions can also be triggered by integrated sensors, including motifs that bind to natural RNA molecules that transmit information within the cell. Use of aptamers, single-stranded molecules of DNA or RNA that are selected from random sequence pools for their ability to bind specifically to selected molecules or structures33,34, greatly increases the range of potential triggers to include proteins and small molecules. Engineered competition for binding to a sensing motif, between its ligand and a base-pairing interaction elsewhere in the nanostructure, provides a flexible interface to integrate sensing with molecular computation or actuation of a mechanism.

Current projects include new ways to provide the energy required to drive synthetic molecular machinery, the creation of a molecular 3-D printer, and control hardware for natural gene-editing apparatus (see next section).

30. S.J. Green et al., Phys. Rev. Lett. 101, 238101 (2008)
31. R.P. Goodman et al., Nat. Nanotechnol. 3, 93-6 (2008)
32. L. Qian & E Winfree, Science 332, 1196-201 (2011)
33. A.D. Ellington & J.W. Szostak, Nature 346, 818-22 (1990)
34. C. Tuerk & L. Gold, Science 249, 505-10(1990)

Nanostructures into cells

in cell.png
DNA nanostructures can cage proteins40,41 and enter cells42

The figures above represent experiments that motivate our research on applications of DNA nanostructures within cells. We have shown that DNA cages can enter cells and survive substantially intact for at least 48 hours42. Cages carrying biochemically active cargoes30,31 and incorporating sensing motifs that open in response to specific stimuli could be transformed into 'smart' drug delivery vehicles.

Current research projects include:

  • targeted intracellular delivery of DNA nanostructures;
  • DNA-based control mechanisms for CRISPR/Cas9 gene editing43 - a potential technology for the treatment of heritable diseases and a model application for self-assembled molecular robots operating within the cell;
  • nanostructure tools for the imaging of cellular structure and function by electron microscopy.

40. C. Erben et al., Angew. Chem. Int. Ed. 45, 74414-17 (2006)
41. R. Crawford et al., Angew. Chem. Int. Ed. 52, 2284-88 (2013)
42. A.S. Walsh et al., ACS Nano 5, 5427-32 (2011)
43. M. Jinek et al., Science 337, 816-21 (2012)

DNA templates for biology and physics

Peptide assembly templated by DNA nanostructures50

Although DNA assembly is precise, predictable and robust, DNA in itself has limited functionality beyond its natural role as an information store. In contrast, the chemical diversity of the natural amino acids underpins the folding of a wide variety of protein structures with highly evolved functions including architectural control, signalling, transport, and catalysis. However, the relationship between the amino acid sequence of a polypeptide and its structure and function is complex, which makes de novo protein design extremely challenging. By creating hybrid nanostructures, by attaching biomolecular components to self-assembled DNA templates, we can combine the architectural control provided by nucleic acid assembly with the functionalities of peptides and proteins. We are collaborating with Dek Woolfson's group (Bristol Chemistry) to study peptide biophysics50 and Richard Berry's group (Oxford Physics) to investigate the assembly of large protein complexes51, using the bacterial flagellar motor as a model.

The near-atomic resolution of DNA assembly also has applications in the physical sciences and engineering. With the Ardavan group (Oxford Physics) and collaborators we are developing methods for the assembly and measurement of circuits built from individual molecules. Our goal is to create the first scalable technology for molecular electronics.

50. J. Jin et al., ACS Nano 13, 9927-35 (2019)
51. M.A.B. Baker (2016) et al., Nat Struct Mol Biol 23 197-203 (2016)