Mathematical modeling of phenylalanine and lignin biosynthetic networks in plants
2019-05-14T17:39:04Z (GMT) by
L-phenylalanine (Phe) is an important amino acid which is the precursor of various plant secondary metabolisms. Its biosynthesis and consumption are governed by different levels of regulatory mechanisms, yet our understanding to them are still far from complete. The plant has evolved a complex regulation over Phe, likely due to the fact that a significant portion of carbon assimilated by photosynthesis is diverted to its downstream products. In particular, lignin as one of them, is among the most abundant polymers in plant secondary cell wall. Studies have unraveled the interconnected metabolism involved in lignin biosynthesis, and a hierarchical gene regulatory network on top of it is also being uncovered by different research groups. These biological processes function together for sufficient lignification to ensure cell wall hydrophobicity and rigidity for plant normal growth. Yet on the other hand, the presence of lignin hinders the efficient saccharification process for biofuel production. Therefore, it is fundamental to understand lignin biosynthesis and its upstream Phe biosynthesis in a systematic way, to guide rational metabolic engineering to either reduce lignin content or manipulate its composition in planta.
Phe biosynthesis was predominantly existed in plastids according to previous studies, and there exists a cytosolic synthetic route as well. Yet how two pathways are metabolically coordinated are largely under-explored. Here I describe a flux analysis using time course datasets from 15N L-tyrosine (Tyr) isotopic labeling studies to show the contributions from two alternative Phe biosynthetic routes in Petunia flower. The flux split between cytosolic and plastidial routes were sensitive to genetic perturbations to either upstream chorismate mutase within shikimate pathway, or downstream plastidial cationic amino-acid transporter. These results indicate the biological significance of having an alternative biosynthetic route to this important amino acid, so that defects of the plastidial route can be partially compensated to maintain Phe homeostasis.
To understand the metabolic dynamics of the upstream part of lignin biosynthesis, we developed a multicompartmental kinetic model of the general phenylpropanoid metabolism in Arabidopsis basal lignifying stems. The model was parameterized by Markov Chain Monte Carlo sampling, with data from feeding plants with ring labeled [13C6]-Phe. The existence of vacuole storage for both Phe and p-coumarate was supported by an information theoretic approach. Metabolic control analysis with the model suggested the plastidial cationic amino-acid transporter to be the step with the highest flux controlling coefficient for lignin deposition rate. This model provides a deeper understanding of the metabolic connections between Phe biosynthesis and phenylpropanoid metabolism, suggesting the transporter step to be the promising target if one aims to manipulate lignin pathway flux.
Hundreds of gene regulatory interactions between transcription factors and structural genes involved in lignin biosynthesis has been reported with different experimental evidence in model plant Arabidopsis, however, a public database is missing to summarize and present all these findings. In this work, we documented all reported gene regulatory interactions in Arabidopsis lignin biosynthesis, and ended up with a gene regulatory network consisting of 438 interactions between 72 genes. A network is then constructed with linear differential equations, and its parameters were estimated and evaluated with RNA-seq datasets from 13 genetic backgrounds in Arabidopsis basal stems. We combined this network with a kinetic model of lignin biosynthesis starting from Phe and ending with all monolignols participated in lignin polymerization. This hierarchical kinetic model is the first model integrating dynamic information between transcriptional machinery and metabolic network for lignin biosynthesis. We showed that it is able to provide mechanistic explanations for most of experimental findings from different genotypes. It also provides the opportunity to systematically test all possible genetic manipulation strategies targeting to lignification relevant genes to predict the lignin phenotypes in silico.