Model¶
Our team used modeling to gain insight into how our project works or should be implemented.
Purpose of modeling:
To simulate the processes of sesquiterpenoids production by MVA pathway in eukaryotic system and MEP pathway in prokaryotic system. The synthesis pathway was modeled by ordinary differential equation (ODE) to predict the concentration of sesquiterpenes, which proved the feasibility of sesquiterpenes production in eukaryotic and prokaryotic systems. We can make a conclution that the production of patchoulol in yeast and other organisms by targeted modification of synthetic biology was feasible.
1. ODE modeling of MVA pathway in eukaryotic systems¶
In the first step, Aceto Acetyl-CoA was generated from the substrate acetyl-CoA under the catalysis of enzyme ERG10:
Among them, K_ERG10 represents the forward equilibrium constant of ERG10 enzymatic reaction, and [Ace-CoA] represents the reaction concentration of Acetyl-CoA. K_deERG10 represents the reverse equilibrium constant of ERG10 enzymatic reaction, and [Di-Ace-CoA] represents the reaction concentration of Aceto acetyl-CoA.
Acetyl-CoA and Aceto acetyl-CoA generate HMG-CoA under the catalysis of enzyme ERG13:
In this formula, K_ERG13denotes the forward equilibrium constant of the enzymatic reaction of ERG13, [Ace-CoA] represents the reaction concentration of Acetyl-CoA, and [Di-Ace-CoA] denotes the reaction concentration of Aceto acetyl-CoA. K_deERG13denotes the reverse equilibrium constant of ERG13 enzymatic reaction, and [HMG-CoA] is the reaction concentration of HMG-CoA.
Mevalonate (MVA) is formed from HMG-CoA under the catalysis of enzyme HMGR:
Here, K_HMGR represents the forward equilibrium constant of the enzymatic reaction of HMGR, and [HMG-CoA] represents the reaction concentration of HMG-CoA. K_deHMGR represents the reverse equilibrium constant of enzymatic reaction of HMGR, and [MVA] represents the reaction concentration of MVA.
The reaction of IPP generation from MVA could be divided into three steps. Here, we integrated these three steps into one-step reaction, thus, the reaction of IPP generation from MVA can be expressed by the following formula:
Among them, K_(S_3 ) represents the forward equilibrium constant of the three-step combined reaction, and [MVA] represents the reaction concentration of MVA. K_(deS_3 ) represents the reverse equilibrium constant of the combined enzyme reaction, and [IPP] represents the reaction concentration of IPP.
IPP generates FPP under the catalysis of enzyme FPS:
In this formula, K_FPS represents the forward equilibrium constant of FPS enzymatic reaction, and [IPP] represents the reaction concentration of IPP. K_deFPS represents the reverse equilibrium constant of FPS enzymatic reaction, and [FPP] represents the reaction concentration of FPP.
FPP is catalyzed by the enzyme PTS to generate Sesquiterpene (here, patchoulol as the final product) :
K_PTS represents the forward equilibrium constant of the enzymatic reaction of PTS, and [FPP] represents the reaction concentration of FPP. K_dePTSrepresents the reverse equilibrium constant of the enzymatic reaction of PTS, and [Patchoulol] represents the reaction concentration of Patchoulol.
Eukaryotic¶
As shown in the figure, after the beginning of the reaction, the intracellular concentration of ACE-CoA began to slowly decrease to generate HMG-CoA, and the concentration of HMG-CoA increased rapidly, followed by the concentration of intermediates IPP, FPP and Patchoulol. Due to the fast generation of FPP but insufficient reaction rate of HMG-CoA in the early stage, the concentration of IPP decreased first and then increased in the early stage. At the same time, the saturation concentration of FPP was lower than that of HMG-CoA and Patchoulol due to the high equilibrium constant in the process of IPP generation of FPP. The concentration of Patchoulol also reached the peak after various intermediates reached the plateau stage.
2. ODE modeling of MVA pathway in Prokaryotic systems¶
Pyruvate and PGAL generate DXP under the catalysis of enzyme DXS:
Among them, K_DXS represents the forward equilibrium constant of DXS enzymatic reaction, [Pyruvate] represents the reaction concentration of Pyruvate, and [PGAL] represents the reaction concentration of PGAL. K_deDXS represents the reverse equilibrium constant of DXS enzymatic reaction, and [DXP] represents the reaction concentration of DXP.
DXP generates MEP under the catalysis of enzyme DXR:
Among them, K_DXR represents the forward equilibrium constant of DXR enzymatic reaction, and [DXP] represents the reaction concentration of DXP. K_deDXR represents the reverse equilibrium constant of DXR enzymatic reaction, and [MEP] represents the reaction concentration of MEP.
The reaction from MEP to IPP was divided into five steps here, and these five steps are integrated into one-step reaction. Therefore, the reaction from MEP to IPP can be expressed by the following formula:
K_(S_5 ) represents the forward equilibrium constant of five-step combined reaction, and [MEP] represents the reaction concentration of MEP. K_(deS_5 ) represents the reverse equilibrium constant of five-step combined reaction, and [IPP] represents the reaction concentration of IPP.
IPP generates FPP under the catalysis of enzyme FPS:
Among them, K_FPS represents the forward equilibrium constant of FPS enzymatic reaction, and [IPP] represents the reaction concentration of IPP. K_deFPS represents the reverse equilibrium constant of FPS enzymatic reaction, and [FPP] represents the reaction concentration of FPP.
FPP is catalyzed by the enzyme PTS to generate Sesquiterpene (here, patchoulol as the final product) :
K_PTSrepresents the forward equilibrium constant of the enzymatic reaction of PTS, and [FPP] represents the reaction concentration of FPP. K_dePTS represents the reverse equilibrium constant of the enzymatic reaction of PTS, and [Patchoulol] represents the reaction concentration of Patchoulol.
Prokaryotic¶
As shown in the figure, after the beginning of the simulation, the concentration of Pyruvate in the cells began to decrease slowly, and then the concentration of DXP,FPP and Patchoulol began to rise, among which DXP first reached a relatively stable concentration. Due to the rapid conversion rate of IPP to FPP and the large equilibrium constant, the concentration of IPP was maintained at a relatively low concentration after the rise. The concentration of Patchoulol also reached the peak after various intermediates reached the plateau stage.
3.ERG¶
> ERG20 amino acid sequence MASEKEIRRERFLNVFPKLVEELNASLLAYGMPKEACDWYAHSLNYNTPGGKLNRGLSVVDTYAILSNKTVEQLGQEEYEKVAILGWCIELLQAYFLVADDMMDKSITRRGQPCWYKVPEVGEIAINDAFMLEAAIYKLLKSHFRNEKYYIDITELFHEVTFQTE LGQLMDLITAPEDKVDLSKFSLKKHSFIVTFKTAYYSFYLPVALAMYVAGITDEKDLKQARDVLIPLGEYFQIQDDYLDCFGTPEQIGKIGTDIQDNKCSWVINKALELASAEQRKTLDENYGKKDSVAEAKCKKIFNDLKIEQLYHEYEESIAKDLKAKISQVDE SRGFKADVLTAFLNKVYKRSK
3D structure of ERG20 Original Protein
General Ramachandran Plot of ERG20 Original Protein Structure
We used swiss-model to simulate the 3D protein structure of the original ERG20, the structure plot shows a high fraction of α-helix in the protein structure. Also, the Ramachandran plot indicates the rationality of the origin protein structure.
> ERG20 (FPS) K197H amino acid sequence (CAC) MASEKEIRRERFLNVFPKLVEELNASLLAYGMPKEACDWYAHSLNYNTPGGKLNRGLSVVDTYAILSNKTVEQLGQEEYEKVAILGWCIELLQAYFLVADDMMDKSITRRGQPCWYKVPEVGEIAINDAFMLEAAIYKLLKSHFRNEKYYIDITELFHEVTFQTELGQLMDLITAPED KVDLSKFSLKKHSFIVTFHTAYYSFYLPVALAMYVAGITDEKDLKQARDVLIPLGEYFQIQDDYLDCFGTPEQIGKIGTDIQDNKCSWVINKALELASAEQRKTLDENYGKKDSVAEAKCKKIFNDLKIEQLYHEYEESIAKDLKAKISQVDESRGFKADVLTAFLNKVYKRSK
3D structure of ERG20 Protein (CAC)
General Ramachandran Plot of ERG20 Protein(CAC) Structure
> ERG20 (FPS) K197P amino acid sequence (CCT) MASEKEIRRERFLNVFPKLVEELNASLLAYGMPKEACDWYAHSLNYNTPGGKLNRGLSVVDTYAILSNKTVEQLGQEEYEKVAILGWCIELLQAYFLVADDMMDKSITRRGQPCWYKVPEVGEIAINDAFMLEAAIYKLLKSHFRNEKYYIDITELFHEVTFQTELGQLMDLITAPE DKVDLSKFSLKKHSFIVTFPTAYYSFYLPVALAMYVAGITDEKDLKQARDVLIPLGEYFQIQDDYLDCFGTPEQIGKIGTDIQDNKCSWVINKALELASAEQRKTLDENYGKKDSVAEAKCKKIFNDLKIEQLYHEYEESIAKDLKAKISQVDESRGFKADVLTAFLNKVYKRSK.
3D structure of ERG20 Protein (CCT)
General Ramachandran Plot of ERG20 Protein(CCT) Structure
Further, We simulated the 3D protein structure of the ERG20(CAC) and ERG20(CCT), the structure plot and Ramachandran plot above all indicates the rationality of the optimized protein structure, it also proves the rationality of our improvement.
4.Multiplex-prediction model by DeepLearning¶
On the basis of the experiment, in order to provide a theoretical basis for the industrialization of the project and predict the optimal conditions required for industrial production, we used the neural network algorithm in Deep Learning to predict the concentration of Patchoulol and the parameters of the growth curve fitting equation. We add all the conditions that may theoretically affect the yield to the Input layer (pH, temperature, concentration of culture medium, composition and concentration of additives).
For Output layer, we can directly use the model to predict the concentration of Patchoulol, or we can combine the model with the logistic equation, and calculate the parameters k, a and b of the fitting equation of Logestic growth curve in the model, so as to predict the growth status of yeast. Reducing the cost of mass production of yeast cultures while maintaining high Patchoulol concentrations.
