Modeling Construction And Process Optimization Of Biological Systems With Enzymatic Regulation

The biological fementation process has been broadly applied in industry to producedesired medical products such as penicillian, bio-energy material such as bioethanol, andmonoclonal antibody, viral vaccine and so on. Research on the modeling of biologicalfermentation process could facilitate learning the internal cell mechanism, and thenimprove the productivity of some desired product and further optimize the wholefermentation process. However, even for the simplest cell structure, it is quite challengingto construct suitable modeling for them. because the cell metabolic networks are composedof thousands upon thousands chemical reactions involving a vast number of cellularmetabolites including extracellular nutrients, internal metabolites and a host of metabolicproducts released into the surrounding medium. Furthermore, all of these reactions arecatalyzed by a vast collection of enzymes with high specific catalytic functions. Eachreaction is catalyzed with a specific enzyme, whose concentration and activity arecontrolled by metabolic in-built regulation mechanism According to different environment,the regulatory mechanism changes enzyme concentration and activity, and thus regulatesthe flux distribution inside the biological system, which could assure the stability andcompatability of the biological systems. Such kind of regulatory mechanism is acquiredthrough thousands years of evolution process. Therefore, in order to fully understand andanalyze the biological fementation process and subsequently provide directional strategyiesfor industry production, it becomes deriable and significant to construct a reasonable andpowerful dynamic model with enzymatic regulation for the biological systems. Generally,there are two critical factors which need to be considered for constructing a perfectdynamic model with enzymatic regulation, i. e. the reaction kinetics of the internalchemical reactions and intermal enzyme regulation mechanism. Until now, there is fewresearch work on considering the enzyme regulation mechanisms in biological fementationsystems based on modeling approach, although there have been some reports onsimultaneously considering the metabolic network and chemical reaction kinetics.Therefore, it will be highly meaningful to conduct this type of research by constructing advanced dynamic model with enzymatic regulation. In our work, we try to fullyunderstand the biological systems with enzyme regulations through constructing effectivemodeling approach, and therefore optimize the biological fementation process. In order toconstruct a perfect dynamic model with enzymatic regulation for the biologicalfementation process, the elemental model analysis is applied to decompose the giantmetabolic networks to the combination of different elemental models which could beanalyzed. These elemental models are related with input variables of the biological systems,i. e. nutrient substrates of the biological systems, and output variables of the biologicalsystems, i. e. production or byproduction of the biological systems. In addition, there aresome inter-mediates and involved enzymes which participate regulation actions betweenthe input and output variables. On the other hand, about the enzyme regulation action inthe biological systems, there have been a set of optimal strategy aiming to opmimal growthability thanks to the long term evolution of the biological systems. Based on this optimalstrategy, the cybernetic variables could be obtained to describe the enzyme actions. Withthe cybernetic variables, the input variable, the output variable and the interaction betweenthe input and output variables could be well described. Based on the above-mentionedmodeling approach, we systematically studied three different kinds of biologicalfementation process by constructing corresponding dynamic model with enzymaticregulation. Furthermore, the prediction ability and process modification are studied forthese dynamic models with enzymatic regulation of these different biological systems. Atlast, we provide a outlook for the dynamic modeling approach with enzymatic regulationand propose the possible extending applications of these modeling approach to otherbiological systems with higher complexity.The research content and main conclusions of this study are as following:1. In order to study the penicillin fementation process, we constructed the dynamic modelwith enzymatic regulation for the Penicillium Chrysogenum which involves theglucose as the dominant nutruitational substrate. Through the simplification of themetabolic network by only considering the important nutritational substrates andreactants while ignoring those intermediates, considering the variation of chemicalreactions under the enzyme regulation actions among these materials, the dynamicmodeling with enzymatic regulation could be accomplished for the penicillin fementation process. The model is further validated through the comparison betweenthe different batches of experimental data and simulation results.2. For mammalian cell culture, we construct dynamic modeling with enzymaticregulation for the mammalian cell culture which involvstwo different substrateshavingthe features of partially complementary and partially substitutable consumptionpatterns, followed by examining the efficiency and prediction ability of the dynamicmodel with enzymatic regulation. Mammalian cell culture involves two nutritionalsubstrates, i. e. glucose and glutamine. Three side products, ammonia, alanine as wellas lactate are released to the medium. Through dynamic model construction withenzymatic regulation for three different batches of fementation process andparameteridentification, three groups of kinetic parameters are derived which have similarvalues. Applying these kinetic parameters to three different batch cultures, thesimulation results of substrates (glucose and glutamine), biomass concentration, aswell as lactate concentration are well consistent with experimental data, which validatethe efficiency of the model. The good prediction performence of the model is reflectedon the ammonina and alanine concentration simulation results, which agree well withexperimental data. In order to further validate the robustness of the model, one groupof kinetic parameter is applied to predict the other two batchs. The simulation resultsare reasonably consistent with experimental data. At last, two involved substrates(glucose and glutamine) in the mammalian cell culture can be qualitatively analyzed tobe partially complementary and partially substitutable through the simulation curveson the time dependent in-built regulation variables, which is consistent with previousliterature.3. In order to study the case of three yeast species fermentation process, we construct thedynamic models with enzymatic regulation appropriate for the fementation processinvolving four different kinds of nutritational substrates. The same metabolic networkis applied for three yeast species. EMA technique is applied to decompose themetabolic network. Three EM groups are then derived. As the number of EMs is toolargh to analyze, MYA is applied to reduce the EM number.According to the specificyield values of four substrates provided in the literature, different EM reduced groupsare obatained for three yeast species. Three dynamic models with enzymatic regulation are then constructed for three yeast species singly based on the reduced EM. Parameteridentification is then proceeded for each species. Three groups of kinetic parameter areobtained for three species. Experimental data from literature are used to validate thesimulation results. The chosen six data sets including four substrate, bioethanol, andbiomass concentration are well validated by comparing simulation results andexperimental data for all three single species fermentaion.4. Based on the above dynamic model with enzymatic regulation for three single species,dynamic model with enzymatic regulation is expended to model mixed cultures. Threeset of mixed cultures composing two species and one set of mixed culture composingthree species are used to study dynamic model with enzymatic regulation in mixedcultures. The kinetic parameter for each species identified in the above single speciesfermentation is adopted for the same species in mixed cultures to predict the substrateand product concentrations in mixed cultures. It is concluded that the prediction resultsby kinetic parameters of single species are quite consistent with experimental datawhen different single species are combined together.5. Though comparing the bioethanol productivity of single species fermentation andmultiple yeast species fermentation, it is found at the initial stage, the bioethanolproductivity of K.maxianus outperforms than others. At the middle stage, mixedcombination of K.maxianus and P.stipitis performs best on the bioethanol productivity.At the last period, single fermentation of P.stipitis is the best. Based on thisphenomenon,’two-step’ fermentation process is designed in order to maximize thebioethanol productiviey during the whole fermentation process. This optimal controlstrategy involves an optimal switching time point which is determined through optimalcaculation. After applying the ’two-step’ fermentation process, the bioethanolproductivity is improved largely.