Contribution

P450 Enzymes Engineering

BBa_K5327031

P450 enzymes have extensive applications in biomanufacturing, but their engineering poses significant challenges in protein modification. Our research aims to provide new insights and strategies for future research teams in this field.

P450 enzymes are part of a widely distributed heme-thiolate enzyme superfamily that catalyzes mixed-function monooxygenase reactions. However, heterologous expression in other organisms often shows low efficiency. To enhance the catalytic efficiency of the P450 enzymes involved in our pathway, we performed specific modifications to facilitate their expression in S.cerevisiae.

We initially employed computational methods to improve the stability and expression efficiency of two P450 enzymes. Subsequently, we altered the hydrophobicity of the substrate channel by modifying the enzyme structure and adjusting subcellular localization. A linker peptide was inserted between the CYP79F1 and CYP83A1 genes, and a mitochondrial location sequences(MLS) was introduced at the N-terminus of CYP79F1. This allowed the two genes to co-localize and be expressed on the mitochondrial inner membrane, forming a fusion protein.

This fusion protein can catalyze the conversion of short-chain thiol esters into corresponding aldoximes (e.g., 5-methylthio-pentanal oxime) and participate in the biosynthesis of both short-chain and long-chain aliphatic glucosinolates. Further oxidation of the aldoximes leads to the formation of nitrated and nitro-acid compounds.

Modification Plan and Results

1. Computer-aided rational design and evaluation

We utilized computer-aided methods to structurally optimize the CYP79F1 and CYP83A1 genes to enhance the catalytic efficiency of the fusion protein. Initially, we assessed the structure of the original binding proteins using molecular docking and molecular dynamics simulations. The computational analysis revealed that the enzyme's binding sites and amino acid residues significantly affect its catalytic efficiency.

For the modification of CYP79F1, we introduced amino acids with high affinity for di-high methionine but low affinity for aldoxime (e.g., isoleucine and leucine) to improve substrate binding specificity and efficiency. Conversely, for CYP83A1, we optimized polar amino acids with high affinity for aldoxime (e.g., glutamine and asparagine) to enhance its interaction with the substrate.

The optimization steps were conducted using GROMACS software and included:

  1. Preparation of Ligands and Receptors: Ligand molecules were processed using the GAFF force field, and the receptor protein topology was constructed with the AMBER99SB force field.
  2. Simulation Environment Setup: A cubic simulation box with a minimum distance of 1.0 nm was created, incorporating the SPC water model and neutralizing the system with Na+ and Cl- ions.
  3. Energy Minimization and Pre-Equilibration: Structural optimization was performed using the steepest descent and conjugate gradient methods, followed by NVT and NPT pre-equilibration.
  4. Molecular Dynamics Simulation: The docking sites of the enzyme and substrate were identified, and the roles of various amino acid residues were assessed.

After a series of predictive data analyses and docking scores using AutoDock software, we ultimately selected the most suitable modification plan for the fusion protein. The specific semi-rational sequence modifications are as follows:

Fig1. Computer representation of modifications to the CYP79F1 protein sequence.
Fig2. Computer representation of modifications to the CYP83A1 protein sequence.

Results of computer simulation modifications:

Fig3. Molecular dynamics simulation results
Affinity improvement rate Stability improvement rate
79short-swiss-op2 0 31.22%
83short-swiss-op1 7.69% 29.46%

According to the results of the molecular dynamics simulations, the binding stability of the two optimized P450 enzymes to the substrate was increased by 31.22% and 29.46%, respectively.

2.Construction of the fusion protein and subcellular localization modification

Next, we constructed the fusion protein. To enhance reaction efficiency, we introduced a flexible linker peptide between the CYP79F1 and CYP83A1 genes, shortening their spatial distance to create a more functional fusion protein aimed at increasing catalytic efficiency. The introduction of the linker peptide optimized the interactions between the proteins, thereby improving enzyme reaction efficiency.

Fig4. Comparison of computer-simulated homology docking results.

To further enhance the catalytic efficiency of the P450 enzymes, we undertook an ambitious approach: we aimed to modify their electron transfer environment. Specifically, we sought to localize the two P450 enzymes involved in the pathway to the mitochondrial inner membrane, anticipating improved performance. To achieve this, we added a mitochondrial location sequences(MLS) at the N-terminus.

Fig5. The express diagram of CDC19p-MLS-CYP79F1(truncated,Δter)-Linker-CYP83A1(truncated)-HXT7t.

We characterized the size of the fusion protein using yeast microsomes.

Fig6.Fusion protein SDS-PAGE images.

We characterized the subcellular localization of the fusion protein:

Fig7. LSCM image

It can be observed that under the guidance of the mitochondrial location sequence, the fusion protein is successfully directed to the mitochondria.

However, since the substrate for P450 enzymes is NADPH, and the mitochondria lack the necessary co-factors and corresponding electron transfer proteins, we plan to further modify the enzyme system. We aim to introduce cyb5 and pntAB proteins into the mitochondrial inner membrane using the same system, allowing the fusion protein to be localized and expressed within the mitochondrial inner membrane.

Fig8. Expression results of MLS-CYP79F1(truncated,Δter)-CYP83A1(truncated)-pntAB-CYB5. The PCR result; b) The plasmid expression c) The express diagram

We constructed the aforementioned P450s modification system into the yeast plasmid vector pRS425 and transformed it into the engineered strains producing GRA and SFN for fermentation experiments. The results were remarkable: after P450s optimization, the GRA yield increased by approximately 49%, while the SFN yield rose by about 31% compared to the unoptimized results.

Fig8. 96-Hour Fermentation Yield Detection Results

We want to emphasize that this approach not only optimizes our project but also presents our imaginative and groundbreaking solution to the common challenge of P450s modification. This can provide future researchers with additional directions, as well as valuable data and experience.

Back to Top