2003 Progress Report: Analysis and Management of Fluxes in Bacillus Pathways for Pesticide and Protein ProductionEPA Grant Number: R829589
Title: Analysis and Management of Fluxes in Bacillus Pathways for Pesticide and Protein Production
Investigators: Grossmann, Ignacio E. , Domach, Michael M.
Institution: Carnegie Mellon University
EPA Project Officer: Richards, April
Project Period: January 1, 2002 through December 31, 2004
Project Period Covered by this Report: January 1, 2003 through December 31, 2004
Project Amount: $180,000
RFA: Technology for a Sustainable Environment (2001) RFA Text | Recipients Lists
Research Category: Sustainability , Pollution Prevention/Sustainable Development
The overall objective of this research project is to modify the metabolic network of Bacillus species to attain near-normal growth without acid byproduct production. Bacillus species is a target because it generally is regarded as safe, and can serve a means for producing natural pesticides. Achieving this research objective will improve the prospects for Bacillis-based bioprocesses because waste products are reduced. Specifically, we are putting pyruvate kinase under inducible control in Bacillus species. Additionally, cloning the Escherichia coli enzyme that converts phosphoenolpyruvate to oxaloacetate (phosphoenolpyruvate carboxylase) into Bacillus is underway. A second supportive effort involves augmenting a general computerized metabolic engineering design platform that will enable the characterization of the metabolically engineered cells as well as provide a general tool for other workers. There are three objectives for the platform: (1) identifying flux alternatives based on a network picture entered by the user; (2) automating the design of contrast 13C nuclear magnetic resonance (NMR) experiments so that alternative flux distributions can be better resolved; and (3) improving the capability of going from spectra to fluxes, which would enable troubleshooting the biological "designs."
Last year, we found vectors that appear to enable the integration of a desired gene into the B. subtilis genome when green fluorescent protein (GFP) is used as a tracer. We aim for integration into a gene that B. subtilis can "live without" under standard laboratory cultivation conditions (e.g., amylase). We were encouraged with regard to integration by the following positive results after attempting transformation: (1) the correct alteration of antibiotic resistance was obtained; (2) the chromosomal gene into which the new gene was inserted was not expressed; and (3) some cells acquired GFP expression capability.
This year, we confirmed successful transformation and have created constructs with inducible gene expression that allow one to regulate gene activity as opposed to using pure knockouts or inserts. We have found that one strain makes nil acid when grown in a rich medium supplemented with glucose. Wild-type cells, in contrast, use glucose in a sloppy fashion under such circumstances, resulting in the formation of copious amounts of acid.
Last year, we reported that a series of chemostat experiments were completed with B. thuringiensis (Bt). These experiments suggest that the acid reduction strategy that works for acid reduction in B. subtilis (add a small amount of citrate) also may also work for Bt. This year, we found that supplanting the growth medium with some amino acids for Bt improves the process. Acid analysis via high performance liquid chromatography is nearly complete, and a publication has been drafted based on the results gathered so far and to guide "polishing" work. Interestingly, there are some small, but significant, differences between B. subtilis and Bt that have been worked out.
A version with a graphical user interface has been completed. This version predicts all flux scenarios possible for mutation and/or objective. The scenarios are, in turn, used to drive the design of 13 NMR experiments so that the best labeled precursor is chosen (in that one can discriminate maximally between flux distribution alternatives). Based on this work, we were invited to contribute an article to a special edition of the journal Computers in Chemical Engineering in 2004 that was accepted "as is without revision," which is rare.
The computation has been expanded significantly in 2004. We have completed a signal metabolite ranking system and successfully applied it to a scaled-down test problem. First, how the spectrum changes as fluxes change is computed to determine an analyte's potential "sensitivity" to flux changes. More importantly, how "uniquely" the information in the metabolite's spectrum inverts to a flux distribution now can be determined using Branch and Reduce Optimization Navigator software and some algorithms that we developed. In practice, the "contour" for optimizing the fit of fluxes to the information present in a spectrum can be "flat" or have local "close optima." We now can enumerate all such solutions, and the more such "close solutions" that exist for a candidate tracer metabolite, the less unique its information is and thus the more likely the analyte's spectrum is to yield an erroneous flux distribution result despite the signal intensity being well above the noise. Overall, we can determine "sensitivity" and "uniqueness" profiles for all measurable NMR analytes. With such information in hand, a subset can be picked that together provides "sensitivity" and minimized probability of providing an incorrect flux solution when inverted. In 2004, a Web site was created (see Relevant Web Sites below).
We have found that under some circumstances, Bacillus can form helical or other types of filaments that are comprised on 5-100 cells. There are sketchy reports of this in the literature. We are exploring whether this phenomena may affect transformation and if it can be harnessed for cell culture broth separations.
We will submit the Bt results for publication. A contribution to a special edition of Computers and Chemical Engineering will be in galley form hopefully by the end of summer 2004. Abstracts have been accepted for the American Institute of Chemical Engineers 2004 Annual Meeting to be held in November, and two presentations will be made.
We will apply the "sensitivity" and "uniqueness" computation to larger problems, which represent computational efficiency challenges. By the end of summer 2004, we plan to have a mathematical proof that the computational means for testing analyte uniqueness is rigorous (using Hessian gradient analysis). Also, we are opening up the double substrate labeling realm to expand the scope of the in silico experimental design of new molecular entity experiments.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
|Other project views:||All 17 publications||4 publications in selected types||All 4 journal articles|
||Ghosh S, Zhu T, Grossmann IE, Ataai MM, Domach MM. Closing the loop between feasible flux scenario identification for construct evaluation and resolution of realized fluxes via NMR. Computers & Chemical Engineering 2005;29(3):459-466.||
||Zhu T, Phalakornkule C, Ghosh S, Grossmann IE, Koepsel RR, Ataai MM, Domach MM. A metabolic network analysis & NMR experiment design tool with user interface-driven model construction for depth-first search analysis. Metabolic Engineering 2003;5(2):74-85.||