Construction And Application Of The New Generation Industrial Microorganisms Biological Network Models

As a system biology tool,biological network model is widely used in the study of microbial phenotype.In this dissertation,centring on the development of biological models,by collecting these published genome-scale metabolic models?GSMMs?,a consensus database?IMGMD?was built in the LAMP?Linux+Apache+MySQL+PHP?system,which provided a platform for the integration and standardization of the first generation of biological models.Then,of these models,colleceted by IMGMD database,typical industrial microorganisms were selected as research objects.An enzyme-constrained model of Escherichia coli K12?ec_iML1515?,and a whole-cell model of Saccharomyces cerevisiae S288C?WM_S288C?were constructed,respectively.With these developed second generation of biologic network,phenotype of microorganisms can be predicted more accurately.Based on ec_iML1515 model,a series of strategies were developed to optimize the lysine production performance of E.coli.The main results were described as follows:1.According to the existed biologic models,and mining from literatures,the existed GSMMs of industrial microorganisms was collected and processed.For these GSMMs,which were consisted of two sheets,including reactions list,and metabolites list.Each model was unified as the same format,such as EXCEL,so that can be read and analysis by the COBRA toolbox.Eventually,328 GSMMs constructed for 139 microorganisms were integrated and standardized,which including 7948metabolites,and 20849 reactions.Using LAMP framework,the In silico Microbial Genome-scale Metabolic Models Database?IMGMD?was built,which can be accessed at http://imgmd.jiangnan.edu.cn/database.The IMGMD database can help microbial researchers download manually curated GSMMs,rapidly reconstruct standard GSMMs,design pathways,and identify metabolic targets for strategies on strain improvement.Since launched in 2016,the database has been accessed more than 2300 times by users from 16countries and regions.2.Based on the newest published E.coli GSMM iML1515,an enzyme-constrained model,named ec_iML1515,was constructed by incorporating enzymatic properties,with GECKO method.This model covers the enzymatic properties of 1286 proteins(molecular weights,k_cat values,etc).Compared with iML1515 model,the reaction and metabolite numbers were increased 125.0%,and 92.3%,respectively.The simulation results were well agreed with the chemostat data,under the dilution rate of 0.15 h^-1,0.2 h^-1,0.25 h^-1,0.3 h^-1,0.35 h^-1,and 0.4h^-1,respectively.The simulated growth rate under 24 different culture conditions achieved a0.57 PCC value?p-value=0.03?.Furthermore,we verified the substrate utilization capacity of92 carbon sources,51 nitrogen sources,16 phosphorus sources,and 12 sulfur sources.And the accuracy was 95.7%,82.4%,100%,and 66.7%,respectively.Using ec_i ML1515 model,we identified 415 genes that were essential in at least 1 of 16 conditions.183 genes were essential in all conditions,and 190 were essential in specific conditions.Eventually,the ec_iML1515predicted gene essentiality in 16 conditions with an accuracy of 92.7%.Lastly,using ec_iML1515,we explored the ability of wide type strain to produce 3 amino acids?L-lysine,L-valine,and L-threonine?,and 3 organic acids?fumarate,L-malate,and succinate?.3.Using ec_iML1515 model,20 targets were identified.Based on the lysine producing strain E.coli CCTCC M2019435,which was acquired by our lab through NTG mutagenesis,these targets were further verified by metabolic engineering.With the over-expressing of dihydrolipoamide dehydrogenase?lpdA?,aquacobalamin reductase/NAD?P?H-flavin reductase?fre?,acetyl-CoA synthetase?acs?,diaminopimelate decarboxylase?lysA?,and aspartate kinase?lysC?,lysine production was increased 63.8%,108.7%,55.6%,50.0%,and123.6%,respectively.These genes can be classified into three modules,such as precursors accumulation,production synthesis,and energy supplyment.Further regulating gene expression strength with different ribosome binding sites?RBS?,and the results showed that when kept fre at high level,and lysC at medium level,the lysine production was 95.7±0.7 g·L-1,increased 169.1%.The simulationed result and 7.5-L fed-batch fermentation culture condition proved that NH₄⁺was suitable for lysine production.Compared with urea,uracil,nitrite,and nitrate,lysine production was increased 69.4%,19.8%,15.9%,and 28.2%,respectively.When the amino nitrogen level was maintained at 0.15%,lysine production was up to 123.7±1.5 g·L^-1.And the dynamic FBA results showed that with the increase of oxygen uptake rate,the fermentation cycle was shortened by 47.7%,and the lysine synthesis rate was increased by 107.7%.The 7.5-L fed-batch fermentation proved that when the DO level was high?50%?,lysine production was 193.6±1.8 g·L^-1,which was 55.8%higher than low?10%?DO level.4.The complex intracellular activities of saccharomyces cerevisiae were modularized into 26 cellular processes.And these processes were represented by different mathematical models,such as flux-balance analysis?FBA,using for metabolism process?,Boolean logic?for chromosome segregation?and Poisson processes?for RNA degradation?.These 26 sub-models were integrated into a unified model using 15 cellular states.The accuracy and repeatability of the unified model was further verified based on experimental results.Using WM_S288C to predict phenotypes of S.cerevisiae,the functions of 1140 essential genes were characterized and linked to phenotypes at five levels?DNA,RNA,protein,metabolite,and geometry?.During the cell cycle,dynamic allocation of intracellular molecules could be tracked in real time,and the demamd of cofactors in different cell processes could be identified.Additionally,one-third of non-essential genes were identified to affect cell growth via regulating nucleotide concentrations.And the cAMP signaling pathway is vital for intracellular nucleotide regulation.