In this poster I will present my graduate and postdoctoral research on flux quantification in metabolic pathways by isotopomer/bondomer balancing, and the application of flux and systems approaches to investigate plant metabolism and human genetic disorders.
GRADUATE RESEARCH
In my graduate research in Jacqueline Shanks' lab at Iowa State University, I developed a comprehensive flux analysis tool for plant metabolism and applied it to two plant systems: soybean (Glycine max) embryos and Catharanthus roseus hairy roots. Although the systemwide measurement of fluxes can serve as a powerful investigative tool in systems biology, it had received very little attention compared to other 'omic profiling technologies [1, 2]. This is principally because of the complexity and subcellular compartmentation inherent in plant biochemistry, which renders the mathematical models relating labeling data to fluxes highly nonlinear and nontrivial to solve.
Flux quantification in plant systems. Metabolic fluxes were evaluated in soybean embryos and C. roseus hairy roots by using 2-dimensional [13C, 1H] nuclear magnetic resonance (NMR), isotopomer/bondomer balancing, and global χ2 minimization. Comprehensive isotopomer/bondomer balance models were developed to convert the labeling data to fluxes. These were solved by employing recent computational developments in flux analysis and implemented through a computer program, NMR2Flux, developed in this work. Fluxes of parallel pathways in the cytosolic and plastidic compartments were identified in the soybean embryos, in addition to bidirectional fluxes of several pathways in central carbon metabolism. To the extent of our knowledge, this work [3] is the most comprehensive flux analysis of a plant system thus far.
Bondomer and Boolean function mapping. My research introduced a concept called bondomer and a computational technique called Boolean function mapping to simplify flux evaluation from labeling experiments that employ uniformly 13C-labeled (U-13C) carbon sources [4]. These were shown to result in more efficient processing of labeling data, and their efficiency is expected to be valuable while analyzing complex metabolic networks such as those in plant metabolism.
Flux identifiability and optimal experiment design. I also investigated flux identifiability and designed optimal labeling experiments that utilize judicious combinations of labeled varieties of two carbon sources (sucrose and glutamine), to maximize the statistical quality of the evaluated fluxes.
POSTDOC RESEARCH
In my postdoctoral research in Katrina Dipple and James Liao's labs at UCLA, I am applying skills developed in my Ph.D. toward elucidating the role of flux and systems dynamics in glycerol kinase deficiency (GKD), a complex, X-linked, single-gene inborn error of metabolism.
Glycerol kinase (GK) is an important lipogenic enzyme in mammalian liver, adipose tissue, and other organs. It also performs several 'moonlighting' activities unrelated to its biochemical function of phosphorylating glycerol [5]. The inherited disorder GKD exhibits complexities that are not trivially explained by lack of the biochemical activity of GK. Thus far, there has been no correlation between genotype and phenotype in patients this disorder. It has been previously hypothesized that systems dynamics, including flux through metabolic pathways, can play a significant role in imparting a phenotype that is not easily deduced from the genotype [6].
Role of metabolic flux and systems dynamics in GKD. I performed flux analysis of wild type and GK-overexpressing H4IIE rat hepatoma cells by using 13C labeling, gas chromatography-mass spectrometry and isotopomer balancing; which revealed that the GK-overexpressing cell lines displayed a substantially higher flux through the pentose phosphate pathway (PPP). This strengthens our hypothesis of the involvement of flux in the complexity of GKD. The higher PPP flux is likely due to increased NADPH requirement in the cytosol, and may be mediated by glycerol kinase or its network partners. I am currently working on microarray analysis of the wild type and GK-overexpressing cell lines, which will be followed by a network component analysis (NCA) of the microarray data. This is expected to reveal the transcription factor activities altered due to GK overexpression, and together with the flux results, will shed light on network interactions involving GK and the role of systems dynamics in GKD.
Mathematical model of insulin signaling pathway. To elucidate a reported link [7, 8] between GKD and insulin resistance, I developed an extended mathematical model of the insulin signal transduction pathway and am employing it to predict insulin sensitivities and phenotypes from experimental gene expression data including microarray analysis of glycerol kinase-knockout (GK k/o) mice. Analysis using the model predicted that certain genes with altered expression in the GK k/o mice confer decreased insulin sensitivity. Ongoing experimental studies based on the model simulations will shed light on why patients with GKD develop insulin resistance.
Acknowledgments
Katrina M. Dipple, James C. Liao, UCLA (postdoc advisors), and Jacqueline V. Shanks, Iowa State University (Ph. D. advisor).
References
1. Kruger, N.J. and A. von Schaewen, Curr Opin Plant Biol, 2003. 6: 236-46.
2. Sweetlove, L.J., R.L. Last, and A.D. Fernie, Plant Physiol, 2003. 132: 420-425.
3. Sriram, G., et al., Plant Physiol, 2004. 136: 3043-3057.
4. Sriram, G. and J.V. Shanks, Metab Eng, 2004. 6: 116-132.
5. Sriram, G., et al., Am J Hum Genet, 2005. 76: 911-924.
6. Dipple, K.M., J.K. Phelan, and E.R. McCabe, Mol Genet Metab, 2001. 74: 45-50.
7. Guan, H.P., et al., Nat Med, 2002. 8: 1122-8.
8. Gaudet, D., et al., Am J Hum Genet, 2000. 66: 1558-68.