(245n) A Kinetic Model of Lignin Biosynthesis in Arabidopsis thaliana for Improved Biofuel Production

Rohit Jaini¹*, Longyun Guo², Peng Wang², Natalia Dudareva², Clint Chapple², John A. Morgan^1,2

¹School of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, ²Department of Biochemistry, Purdue University, West Lafayette, Indiana 47907

Plants naturally channel almost 20-30% of the carbon fixed via photosynthesis towards the phenylpropanoid pathway for its conversion to lignin, a hetero-aromatic polymer that imparts mechanical support and hydrophobicity to plant vasculature while impedes cellulosic biofuel production. The past two decades witnessed several genetic engineering efforts targeting the phenylpropanoid pathway to manipulate lignin content and composition to improve forage digestibility and saccharification efficiency. Despite the progress, several questions pertaining to control and regulation of carbon flux in this pathway remain unanswered, revealing a significant gap in the mechanistic understanding of flux control in this pathway. In an attempt to elucidate this underlying regulatory framework, we are developing a kinetic model for the phenylpropanoid network in Arabidosis thaliana. In this study, we present a holistic strategy for kinetic parameter estimation and validation using targeted metabolomics data obtained by stable-isotope labeling experiments conducted on wild-type (WT) and transgenic plants with knockdown of the gene encoding caffeoyl-shikimate esterase (CSE) respectively. Arabidopsis cse lines, compared to WT, exhibit a phenotype with reduced lignin content that is enriched in p-hydroxyphenyl units presenting an interesting system to investigate.

As a first step, we developed a novel and comprehensive analytical method based on liquid chromatography and tandem mass spectrometry (LC-MS/MS) for accurate quantitation of the phenylpropanoid pathway intermediates. Employing two separate extraction strategies allowed successful profiling of all the intermediates of the pathway, including the previously un-detectable and labile hydroxycinnamyl CoA esters. Soluble extracts from basal section of stems were analyzed on a reverse phase LC-MS/MS platform in the negative ion mode harnessing the sensitivity conferred by multiple reaction monitoring (MRM). Chromatography and compound ionization was optimized for achieving low limits of detection (LODs) for the metabolites of interest. WT and transgenic Arabidopsis plants were fed with three different concentrations (100, 300 μM & 1 mM) of [ring-¹³C₆]-phenylalanine as part of our stable isotope labeling experiments. Concentrations and label enrichments of phenylpropanoid metabolites measured at seven different time points after feeding constituted the metabolomics data sets for parameter estimation.

The analytical method allowed the detection and quantification of 15 of the 19 compounds of the pathway, for which standards were available, in WT tissue. This data was used to extend our previously presented kinetic model on the first three reactions of the lignin biosynthesis pathway catalyzed by phenylalanine ammonia-lyase (PAL), cinnamic acid 4-hydroxylase (C4H), 4-(hydroxy) cinnamoyl CoA ligase (4CL) to now include a key section consisting of enzymes hydroxycinnamoyl-CoA: shikimate hydroxycinnamoyl transferase (HCT), p-coumaroylshikimate 3´-hydroxylase (C3´H), and CSE. Kinetic parameters were identified by simultaneously fitting fractional label and pool sizes of metabolites using a covariance matrix adaptation evolution strategy (CMA-ES), a stochastic and derivative-free numerical optimization method widely applied to ill-conditioned systems. Our current model captures pathway dynamics over the range of feeding treatments considered for the study. Parameter uncertainties were evaluated using a Markov Chain Monte Carlo (MCMC) sampling scheme. The model was validated on Arabidopsis cse lines by comparing dynamic metabolite-profiling data to predicted simulations. We believe our study has laid foundation to extend the kinetic model to the remaining metabolic pathway showing promise as a predictive tool to aid in rational metabolic engineering of lignocellulosic feedstock for efficient biofuel production.