Molecular Modeling: Engineering at the Atomic Level

March 2010
Topics: Biotechnology, Computational Biology
MITRE researcher Steven Fairchild explains how computational molecular modeling allows researchers to predict how different atomic configurations will affect a molecule's behavior, enabling the engineering of molecules with unique characteristics.

Computational molecular modeling allows researchers to predict how different atomic configurations will affect a molecules behavior, thereby enabling the engineering of molecules with unique structural and functional characteristics.

Very Small Building Blocks

Atoms are the basic building blocks used to construct all molecules on earth. They can be combined in different ways to create molecular systems with diverse functions. For example, sulfur, phosphorus, oxygen, nitrogen, carbon, and hydrogen can be combined to make biomolecules like DNA, proteins, and lipids. However, those same six elements can also be combined in different ways to create liquid fuels for automobiles and rockets or to construct plastic polymers for storage or construction.

Ultimately, each molecule's properties depend on the configuration of its atoms. Engineering molecules with specific functions therefore requires a method for understanding how different atomic configurations alter a molecule's properties.

Molecular systems have traditionally been designed through intuition and random chance. Researchers know the properties of many molecules and can use this information to predict the behavior of new molecules. However, this is a slow process that often requires expensive and time-consuming tests for each new design. Additionally, it becomes increasingly difficult to predict a molecule's behavior as the number of atoms increases. Thus, molecules with a desired property are often found by screening large libraries of pre-synthesized compounds.

Similarly, biomolecules are often designed by making millions of random changes to the molecule and seeing if any changes cause the desired effect. While such screens can return positive hits, they require extensive experimental procedures and often lead to sub-optimal results.

Modeling Molecular Systems

Computational molecular modeling (CMM) is a method for rapidly understanding and engineering molecular systems. CMM involves the use of computational models based on classical and quantum physics to predict the behavior of atomic systems. A major benefit of CMM is that it enables rapid testing of molecular function prior to experimental testing. Given modern computing capabilities, this often means that millions of molecular designs can be tested computationally in the time required for one experimental laboratory test.

These models can have varying levels of detail. More coarse-grained approaches approximate each atom as a single sphere and fix certain degrees of freedom within the system (e.g., atomic bond lengths). More detailed quantum mechanical models account for electrons surrounding each nucleus.

Both types of models can predict various properties about molecules, including favorable conformations, electronic characteristics, and intermolecular interactions. This latter area is especially useful when understanding complex macromolecular systems such as biological systems.

Characterizing and Creating Compounds

One main area where CMM is being applied is in characterizing molecules. For example, CMM can be used to predict molecules' structures and electronic characteristics and to model chemical reactions. This kind of information can help determine such properties as the expected behavior of electrons in photovoltaic systems or the energetic characteristics of fuels and explosives.

CMM can also be used for molecular engineering. For example, it has been used to enable construction of nanotubes with specific dimensions and strengths. It has also been used to create better dyes for visualizing biological cells and for optimizing industrial polymer production.

MITRE is employing CMM in a variety of projects. In one, MITRE is using CMM to predict the signature of explosive compounds, providing a method for rapidly and inexpensively constructing sensor libraries. MITRE's nanotechnology group is also using quantum mechanical modeling to develop equations that describe a molecule's electronic behavior from the nano- to macro-scale levels. These equations could provide a useful mechanism for designing bulk materials with desired electronic properties.

Building Better Biomolecules

Another area where CMM is being applied is in modeling and engineering biological systems. CMM enables researchers to understand molecular interactions in living organisms and then apply this knowledge to build biological systems with desired characteristics. For example, CMM can be used to predict the functions of unknown proteins, to create proteins with enhanced stability, to model networks of molecular interactions, and to design highly specific therapeutics.

CMM can also be used to design single-celled organisms capable of creating useful compounds. In one application, algae are being developed that can naturally combine the sun's energy and atmospheric carbon dioxide to create biofuels. Such algae could potentially provide a low-cost, renewable energy resource. In another application, scientists are designing organisms that can synthesize expensive pharmaceutical compounds at a fraction of the cost for traditional chemical techniques.

MITRE is using CMM to design biomolecular systems with enhanced capabilities. In one project, MITRE is helping the U.S. Army engineer enzyme-based countermeasures that can effectively neutralize chemical warfare nerve agents. Such countermeasures could offer a dramatic improvement over existing countermeasures since a single enzyme can destroy multiple nerve agent molecules.

In the Pipeline

This research is benefiting from MITRE's own Protein Design Pipeline (PDP) software. The PDP combines state-of-the-art molecular modeling packages with a multi-dimensional genetic algorithm procedure to enable rapid engineering of proteins with specific binding and enzymatic properties. The software accomplishes this by determining how changes to the protein's structure will affect its activity, and then repeatedly building on these changes until they result in the desired molecular state.

With the software, MITRE researchers can examine tens of thousands of protein variants a day. This significantly exceeds what can be tested experimentally, enabling rapid prototyping of the molecular system. The computational models also return additional information—atomic structure, protein stability, and substrate interaction mechanism—that is difficult to obtain experimentally.

Bridging the Gap

Information about molecular systems is expanding rapidly, especially in the areas of nanotechnology and biochemistry. However, the processing of this raw data has not kept pace. CMM provides an effective method for bridging this gap since it enables researchers to rapidly characterize molecules' properties and engineer molecules with desired properties. Given the growth of raw data and the benefits of molecular engineering, CMM is expected to be an area where MITRE can assist government agencies for the foreseeable future.

—by Steven Fairchild


Publication Search