Intellectual Merit. Most pro- and eukaryote genomes encode hundreds of enzymes of unknown function; finding what they do is a critical task for post-genomic biology. Mounting evidence implicates many of these enzymes of unknown function in "metabolite repair", i.e. in reversing damage done to metabolites by unwanted enzymatic side-reactions or chemical degradation. Because metabolites are under constant chemical attack (e.g. by oxidation or hydrolysis) and enzymes make wasteful and toxic catalytic errors, it follows that efficient functioning of meta¬bolic networks demands a support system dedicated to meta¬bolite repair. This system has been glimpsed by classical biochemistry, genetics, and metabolomics but most of it remains hidden. This project will therefore dissect the metabolite repair system by combining chemical biology, comparative genomics, and metabolomics using bacterial models and plants. Specific aims are to: (a) identify 30-50 target metabolites that are highly prone to chemical or enzymatic damage (i.e. that need repair) by cheminformatics, genome-scale metabolic reconstruction, and data mining; (b) predict genes encoding conserved repair enzymes for target metabolites using comparative genomics, and predict chemistries for the repair reactions; (c) test 20 repair predictions by knocking out the repair gene in a model organism, analyzing metabolomic profiles in normal and stress conditions, and identifying structures by cheminformatics; (d) validate repair reactions by mass spectrometric authentication of structures, by biochemical assays of recombinant proteins, and by functional complementation of bacterial or plant mutants; and (e) incorporate validated repair functions in next-generation genome-scale metabolic models. This project integrates modeling in two ways. First, it makes innovative use of modeling to predict a priori the metabolites most likely to need repair. Second, adding validated repair functions to genome-scale bacterial models sets up a virtuous cycle of prediction 'experiment' further prediction to drive discovery in metabolite repair. It also pioneers an essential modeling development: Models that capture the cost of uncontrolled formation and degradation of unwanted metabolites. Research in the emerging field of metabolite repair has the potential to displace a current paradigm of metabolic routes operating with perfect precision by one where the illusion of a flawless system is maintained by a battery of unobtrusive but critical repair functions. Moreover, metabolite repair is almost surely crucial to stress adaptation, to aging, and to metabolic engineering.
Broader Impacts. This project will provide interdisciplinary training in comparative genomics, metabolomics, chemical biology, and integrative modeling to two PhD students and three postdoctorals who will spend time away from their own institution each year at another collaborating institution. Un-dergraduates will participate. In addition, there will be a training outreach component with three facets: (a) Eight two-day hands-on workshops (2 per year) at different universities to train PhD students, post¬doctorals, and faculty in comparative genomics using SEED databases and tools. At least three work¬shops will be at minority-serving institutions. Each workshop will train 10-12 people. (b) Development of a web page in which the instructional content of the workshop will be available for distance learning. (c) Instruction of project postdoctorals and students in how to organize and present workshops, culminating first in their acting as teaching assistants, and ultimately in them teaching themselves.
