People: Faculty
David W. Wood
Assistant Professor B.S., Biology and Chemical Engineering, California Institute of Technology, 1990
M.S., Chemical Engineering, Rensselaer Polytechnic Institute, 1997
Ph.D., Rensselaer Polytechnic Institute, 2000
Room: A-417 Engineering Quadrangle
Phone: (609) 258-5721
E-mail: dwood@princeton.edu
Publications
Research Interests
Despite the fact that biological processes have been commonly used in the production of foods and chemicals for millennia, recent advances in genetic manipulation and protein engineering have only now begun to unlock their potential. Greater benefits can be realized, however, if these systems can be adapted to chemical engineering approaches and goals. This can be difficult because these systems are often impossible to characterize with a high degree of predictive accuracy. Even a process as fundamental as protein folding requires years of computation to simulate. At the other end of the spectrum, a precise characterization of the thousands of interdependent chemical reactions taking place in a single growing cell, where critical control mechanisms are still being discovered, is also unrealistic. For these reasons, traditional chemical engineering approaches for the rational design of enzymes and biological pathways are limited.
A recent advance which shows promise in overcoming some of these difficulties is to mimic cycles of variation and selection observed in natural evolution. This effectively forces the evolution of more useful biological processes relative to the proposed applications without requiring detailed knowledge of their mechanisms. A key issue is the development of powerful screening systems which can rapidly differentiate desirable and undesirable variants. The most powerful screens are those wherein desirable variants are allowed to grow, while undesirable variants are killed. The development and generalization of these types of selections form a critical component of my research program.
Combining protein and metabolic engineering with directed evolution. A rough understanding of the structure and mechanism of a parent protein element can allow somewhat optimized variants to be rationally designed. Often, however, the variant’s activity must be fine-tuned before it can be used. One strategy involves joining the variant activity to a critical metabolic process. If designed properly, this combination can provide a conditionally lethal selection for optimized activity. Because undesirable mutants are killed, the isolation of a few useful mutants from a background of tens of thousands is possible during rounds of variation and selection. This approach of combining rational engineering with directed evolution can lead to the construction and optimization of individual enzymes or entire engineered pathways through genetic selection in vivo. Successful application of this technique further promises to provide breakthroughs in the understanding and design of new metabolic pathways in living systems.
Enzyme theory and design. Nature has evolved countless variants of a vast array of enzymes, and genomic databases and the tools to analyze bioinformatic data are also evolving. It is possible that all of the data required to start down the path of rational protein design already exists in nature and will soon be available to research. Evolutionary theory, combined with phylogenetic analysis using a variety of models, can be used to assemble a basic understanding of protein and enzyme design in nature. Artificially evolved enzymes add to this understanding, providing clues as to how enzymes might evolve for use in the highly unnatural environments used by chemical engineers. Our goal is to use these data to develop computer-based strategies which would mimic evolution, allowing rational protein design based on accumulated modifications in the target structure. This approach can also be applied to altering substrate specificity, and in combination with in vitro evolutionary data, may eventually be extrapolated into non-native environments where no naturally occurring enzyme examples exist.
Bioseparations using self-cleaving binding domains. The development of a self-cleaving protein element during prior work has resulted in a new and useful purification system. The on-column self-cleavage used by this system represents a significant departure from conventional bioseparation approaches, opening new avenues for the development and optimization of this technique. One is to develop new binding domains with extremely high affinities for inexpensive resins, thus reducing the cost while increasing the efficiency of these separations. Another is to develop new self-cleaving elements using in vitro selection systems, such as phage or yeast-surface display, with different modes of control for use in systems ranging from simple bacterial expression hosts to transgenic vertebrates. A third area is to develop splicing elements for use in highly specific covalent immobilization of proteins and enzymes, where the target peptide is immobilized directly from the cell extract without purification or complex chemistry.
Despite the fact that biological processes have been commonly used in the production of foods and chemicals for millennia, recent advances in genetic manipulation and protein engineering have only now begun to unlock their potential. Greater benefits can be realized, however, if these systems can be adapted to chemical engineering approaches and goals. This can be difficult because these systems are often impossible to characterize with a high degree of predictive accuracy. Even a process as fundamental as protein folding requires years of computation to simulate. At the other end of the spectrum, a precise characterization of the thousands of interdependent chemical reactions taking place in a single growing cell, where critical control mechanisms are still being discovered, is also unrealistic. For these reasons, traditional chemical engineering approaches for the rational design of enzymes and biological pathways are limited.
A recent advance which shows promise in overcoming some of these difficulties is to mimic cycles of variation and selection observed in natural evolution. This effectively forces the evolution of more useful biological processes relative to the proposed applications without requiring detailed knowledge of their mechanisms. A key issue is the development of powerful screening systems which can rapidly differentiate desirable and undesirable variants. The most powerful screens are those wherein desirable variants are allowed to grow, while undesirable variants are killed. The development and generalization of these types of selections form a critical component of my research program.
Combining protein and metabolic engineering with directed evolution. A rough understanding of the structure and mechanism of a parent protein element can allow somewhat optimized variants to be rationally designed. Often, however, the variant’s activity must be fine-tuned before it can be used. One strategy involves joining the variant activity to a critical metabolic process. If designed properly, this combination can provide a conditionally lethal selection for optimized activity. Because undesirable mutants are killed, the isolation of a few useful mutants from a background of tens of thousands is possible during rounds of variation and selection. This approach of combining rational engineering with directed evolution can lead to the construction and optimization of individual enzymes or entire engineered pathways through genetic selection in vivo. Successful application of this technique further promises to provide breakthroughs in the understanding and design of new metabolic pathways in living systems.
Enzyme theory and design. Nature has evolved countless variants of a vast array of enzymes, and genomic databases and the tools to analyze bioinformatic data are also evolving. It is possible that all of the data required to start down the path of rational protein design already exists in nature and will soon be available to research. Evolutionary theory, combined with phylogenetic analysis using a variety of models, can be used to assemble a basic understanding of protein and enzyme design in nature. Artificially evolved enzymes add to this understanding, providing clues as to how enzymes might evolve for use in the highly unnatural environments used by chemical engineers. Our goal is to use these data to develop computer-based strategies which would mimic evolution, allowing rational protein design based on accumulated modifications in the target structure. This approach can also be applied to altering substrate specificity, and in combination with in vitro evolutionary data, may eventually be extrapolated into non-native environments where no naturally occurring enzyme examples exist.
Bioseparations using self-cleaving binding domains. The development of a self-cleaving protein element during prior work has resulted in a new and useful purification system. The on-column self-cleavage used by this system represents a significant departure from conventional bioseparation approaches, opening new avenues for the development and optimization of this technique. One is to develop new binding domains with extremely high affinities for inexpensive resins, thus reducing the cost while increasing the efficiency of these separations. Another is to develop new self-cleaving elements using in vitro selection systems, such as phage or yeast-surface display, with different modes of control for use in systems ranging from simple bacterial expression hosts to transgenic vertebrates. A third area is to develop splicing elements for use in highly specific covalent immobilization of proteins and enzymes, where the target peptide is immobilized directly from the cell extract without purification or complex chemistry.
