Computational Studies of AEGIS Molecular Components

One of the most important outcomes of the last quarter-century of synthetic biology is the recognition that the biopolymers that have been delivered to us by 4 billion years of biological evolution are not the only molecules that might support genetics, inheritance, evolution, and catalysis. Developing new biopolymers and characterizing their properties in living organisms not only has significance for testing our ideas about the functional optimization of the existing "molecules of life" but also opens new opportunities in biotechnology. Such work also permits insights into questions of the uniqueness of terrean biology and whether life in other parts of the universe could be constructed using alternate chemistries. Considering just DNA and RNA (collectively xNA), it is now clear that the four distinct "standard" building blocks (A, G, C, T and its equivalent U) for xNA do not exhaust the constraints imposed by the two rules guiding Watson-Crick pairing in natural nucleic acids. For example, the number of nucleobase pairs can be increased from two to six by merely rearranging hydrogen bond donor and acceptor groups. Efforts to implement this observation in practice using chemical synthesis, have (so far) led to two "generations" of novel heterocycles that can be incorporated into precursors suitable for use in automated xNA synthesis, thereby yielding artificially expanded genetic information systems (AEGIS).

Although exploiting altered patterns of hydrogen bonding to obtain novel nucleobase pairs that are (in principle) "orthogonal" to A:T and G:C appears straightforward, the practical realization of these ideas has proven surprisingly problematic. For example, some potential heterocycles have highly populated tautomeric forms with altered hydrogen bonding patterns; these can base pair with standard nucleobases in either duplex DNA or within the active sites of polymerases, thereby giving rise to unanticipated mutations or the loss of the AEGIS nucleobases during replication. Being able to predict tautomer populations in solution or within enzyme active sites prior to chemical synthesis would be a significant step in improving the efficiency with which new nucleobase pairs can be discovered. Even were these "design" problems to be resolved, little is known about how the incorporation of these non-natural nucleobases into xNA affects the conformational preferences and dynamical properties of these complex molecules, which are fundamental to the interaction of "standard" xNA with proteins, such as polymerases, and transcription factors. We note that there has been a dearth of studies aimed at understanding how AEGIS nucleobases, which have altered electrostatic properties (dipole moments, charge distribution), might perturb xNA structure in both free solution and when bound within polymerase active sites. Finally, the validation of the force field parameters needed to model AEGIS nucleobases by, for example, comparing calculated free energies of interaction between xNA and proteins with experimental measurements has not yet been reported.
Work in this project will therefore seek to address the problems outlined above by (i) developing and validating new computational methods for determining the populations of tautomeric forms of AEGIS nucleobases in water and in protein environments, and (ii) using advanced MD-based methods to understand how the incorporation of AEGIS nucleobase pairs affects the conformational and dynamical properties of duplex DNA and its interactions with DNA-binding proteins and polymerases. In particular, free energy perturbation methods will be used to study how replacing "standard" Watson-Crick bases by an AEGIS nucleobase pair changes the affinity of the DNA-binding domain of the human SETMAR transcription factor. The successful accomplishment of this aim will lay a foundation for obtaining novel endonucleases capable of cleaving AEGIS-containing duplex DNA.

DNA and RNA (collectively xNA) are fascinatingly diverse molecules that display a multitude of detailed conformations with biological significance. In recent years, it has been recognized that, by merely rearranging hydrogen bond donor and acceptor groups, the number of nucleobase pairs can be increased from two to six, thereby yielding an Artificially Expanded Genetic Information System (AEGIS). Efforts to reduce this observation to practice using chemical synthesis, have (so far) led to two "generations" of novel heterocycles that can be incorporated into precursors suitable for use in automated xNA synthesis. In addition, protein engineering using a variety of strategies has created polymerases that can copy and PCR amplify oligonucleotides containing a variety of these "non-natural" nucleobase bases, opening up a substantial number of technological applications. The long-term development of AEGIS-based applications is constrained, however, by a lack of theoretical "infrastructure" that permits us to understand fundamental chemical aspects of expanded genetic systems. In this project, we will carry out theoretical studies at multiple levels to improve our understanding of xNA molecules containing novel nucleobases and nucleobase pairs. The project has two specific aims, which can be summarized as (1) to develop and validate new computational methods for determining the populations of tautomeric forms of AEGIS nucleobases in water and in protein environments, and (2) to understand how the incorporation of AEGIS nucleobase pairs affects the conformational and dynamical properties of duplex DNA and its interactions with DNA-binding proteins. Not only will these studies facilitate the discovery of novel non-natural nucleobases but they will also drive the development of reagents for the creation of organisms in which AEGIS play a functional role.

The creation of artificial genetic systems capable of Darwinian evolution is a central theme in the emerging field of synthetic biology. Work in this project will drive the development and evaluation of computational strategies that can be used (i) to investigate the molecular properties, including tautomer preferences, of heterocycles that are potential non-natural nucleobases, and (ii) to obtain novel restriction endonucleases and polymerases for manipulating DNA that is built from expanded genetic alphabets. Access to such enzymes will be an essential element in permitting synthetic biologists to construct plasmids containing an expanded number of nucleobases, which will, in turn, accelerate their ability to develop modified bacteria with the capability of producing proteins that contain novel amino acids, with significant implications for the sustainable production of fine chemicals by engineered microorganisms. Indeed, the topic also lends itself to articles in the popular press given the general interest in the nature of life that may exist elsewhere in the Universe. We anticipate that the dynamical simulations of duplex DNA can form the basis of podcasts that can used by school teachers to enhance science teaching.

This project involves an extensive number of collaborations, primarily with USA-based researchers, and so our studies will positively impact the international reputation of UK-based research in chemical biology. We anticipate that the PDRA appointed on this grant will interact closely with our collaborators, both using Skype-based conference calls and by travel to Indiana and Florida. The expertise of our collaborators in the design, chemical synthesis and the characterization of novel heterocycles as novel nucleobases, and in the structure and biophysical properties of nucleic acids is also very strong, and so the project PDRA will gain unique cross-disciplinary training and insight into this emerging sub-discipline of synthetic biology. This will enhance their skill set, and have a positive impact on his/her competitiveness for a career in either academia or the industrial/biotechnology sector.

The Benner laboratory at the Foundation for Applied Molecular Evolution has a strong track record of working with industrial partners to exploit the properties of novel nucleobases in diagnostic reagents. Any intellectual property developed in this study will be identified through periodic discussions with, and disclosures to, Research Innovation Services at Cardiff University and the other institutions involved in this collaborative research project.