Team:HK SSC/Model

Home Description
Team Attributions Collaborations
Contribution Engineering Experiments Implementation Notebook Results Model
Parts Improve
SAFETY
IHP Communication

Project Modeling

In this year, our modeling work focused on the protein redesign of the 3-hydroxybutyrate dehydrogenase (CF3HBD) to favor the presence of 5AVA in the active site, to increase the catalytic efficiency from 5AVA to valerolactam. We used protein modeling software Rosetta to perform point mutations and analysis. Further analysis was also planned, using the molecular dynamics software, GROMACS.

Rationale of choosing CF3BD

Several enzymes possess the ability to cyclize 5AVA into valerolactam, for instance CALB, MBP_ORF26, CaiC, Act and CF3BD. All the enzymes were compared against another, and it was discovered that some enzymes need an energy source, such as MBP_ORF26 requires ATP and Mg2+ and Act requires Acetyl-CoA. Also, some enzymes like CALB have an optimum temperature that does not match our strain’s temperature range. Thus after careful judgment and consideration, the cyclase CF3BD was chosen due to its optimum reaction temperature of 40ºC, which fits the optimum 38ºC to 41ºC growth temperature range of S.elongatus UTEX 2973. Its reaction also doesn’t require any energy source or activators, minimizing the energy used for the cyclization of 5AVA to valerolactam1 When compared to other enzymes, CF3BD best fits into the growth conditions of the S.elongatus UTEX 2973 strain.

Protein design

Introduction to Rosetta

We aim to perform point mutations on the selected protein and analyze the proposed design changes. Rosetta software suite was chosen as the main tool to evaluate our design, as it provides multiple pre-combined algorithms as protocols to perform our intended design procedures. Protocols used in this modelling were Relax, a protocol for simple all-atom refinement of input structure2, Backrub, a protocol for refining the sidechain prediction3 and Point Mutant Scan Application, a protocol for mutating the protein structure at specific residues and analyzing its stability, are used in the rational design process4.
Aim:
1. Perform point mutation on the residues of reported catalytic sites
2. Analysis of stability of mutated candidates

Preparation of initial protein structure

Structure of CF3HBD (PDB ID: 5YSS) was obtained from the RCSB protein data bank (RCSB PDB), downloaded as a protein data bank (PDB) file. Then, the structure was cleaned while keeping the ligand5. Relaxation was performed to idealize the structure, ensuring less steric clashes and inappropriate geometry were present in the structure.

Analysis of proteins

After relaxing the structure, the PDB candidates were ranked using the scoring function. The scoring function evaluates candidates' different properties, such as Lennard-Jones potential6, Coulombic electrostatic potential6 and Disulfide geometry potential6 etc. For a complete list of energy terms and weights used in the scoring function, it can be found on the Rosetta Scoring Tutorial Documentation. Because of the involvement of statistical terms, the scores were presented in custom Rosetta Energy Units (REU). The top-scored structure was selected to be used in the Backrub application.

Backrub application

After the structure was relaxed, the PDB is backrubbed. Backrub is an application in Rosetta that involves internal backbone rotations about axes between C-alpha atoms3 respective to its neighboring sidechain’s atoms, lowering of the Gibbs-free energy, hence, more realistic structures is produced before performing point mutation. This method is chosen as the backrubbing is shown to improve the accuracy of sidechain conformational variability and sidechain predictions3, which will be important for later molecular modeling of the calculating interaction energies between 5AVA and the sidechain of the amino acid in the catalytic sites. The candidates produced during this application were also analyzed through the scoring function in Rosetta, with the best model used in the point mutation application.

Point mutation application

According to the documentation of Rosetta7 and Rosetta Guide for the iGEM beginner made by Team iGEM2016 Technion and iGEM2016 TU Eindhoven, point mutations for CF3HBD were performed in Rosetta 3.13. A mutation list concerning the proposed active site was constructed based on previous literature studying CF3HBD7. Candidates were scored using the scoring function in Rosetta.


Results

Description Average ΔΔG Average total energy
5yss.V140K,Q193E 131.191 50.30
5yss.S139M 113.465 76.07
5yss.Y152I,W184R 100.833 92.55
5yss.K149P,Q193V 99.089 82.52
5yss.Y152K,Q193C 96.689 84.80
5yss.Y152T,V190E 96.443 94.29
5yss.Q91M,V190I 96.369 96.34
5yss.K149T,Q193C 96.109 85.50
5yss.V140K,V190W 95.319 97.18
5yss.V140K,H141A 94.806 85.18
Table 1. Table showing the top 10 stabilizing mutated candidates

From the above table , we can see the ΔΔG values, generally, ΔΔG > 0.5 was considered a destabilizing mutation8, but since we ran the “Altered Specificity” mode in point mutation protocol, potential candidates should be having high ΔΔG and low average total energy, as Rosetta’s point mutant scan application tends to be a bit off in its predictions and thus high average energy could be even more unfavourable than predicted. 10 structures with ΔΔG values and lowest total average energy were presented in the above table and these are the mutants passed onto molecular dynamics analysis:

V140K,Q193E
S139M
Y152I,W184R
K149P,Q193V
Y152K,Q193C
Y152T,V190E
>Q91M,V190I
K149T,Q193C
V140K,V190W
V140K,H141A

Y152A mutant was added to the list as this is the mutant that has increased catalytic activity in previous literature7.

Molecular Dynamics

Due to difficulties in preparing the structure in GROMACS, our team is unable to run molecular dynamics analysis of the proposed candidates; we planned to continue the analysis in the next phase of our project.


Aims of the molecular dynamics model:

  • Determine the interaction energy of the mutated 5YSS with 5AVA

Methods

Software Package and Force Field

Molecular Dynamics simulation was carried out in GROMACS9 (GROningen MAchine for Chemical Simulations). We have also parameterized CHARMM3610 (Chemistry at HARvard Macromolecular Mechanics) force field to suit our simulation.


Reparameterization of Force Field and Creating Topology for the NAD+

Although NAD+ was defined in GROMACS, the naming of the atoms was not standardized with Rosetta, we are currently working on reparameterization of the force field to continue creating the topology of mutated 5YSS.

Solvation

The simulation box will be 1.0 nm of distance from the solute to the box borders. spc216 was used for solvation.

Energy minimization

Energy minimization will be carried out to ensure that the system has no steric clashes or inappropriate geometry. Steepest gradient algorithm will be used.

NVT

After energy minimization, we will obtain a reasonable starting structure. To begin molecular dynamics, we must equilibrate the solvent, NVT ensemble will be carried out to bring the system to the right temperature and establish the proper orientation about the solute (protein and 5AVA). After arriving at the correct temperature (based on kinetic energies), we will apply pressure to the system until it reached the proper density. One modification we plan to make to the parameter file is that we will increase the equilibration time to 1 ns according to literature11.

NPT

Then, NPT ensemble was carried out to equilibrate the system. One modification we will be making to the parameter file is increasing the equilibration time to 1 ns.

Produce MD

Finally, we can run the MD simulation. This concludes the molecular dynamic modelling of our mutated CF3BD with 5AVA.

By ranking the total interaction energy of our mutated proteins with the 5AVA, we can determine the best candidate for favoring the 5AVA in its active site (i.e. the one with the highest total interaction energy). Since molecular dynamics simulates the whole dynamic environment instead of static binding energy, we hope our simulation can serve as a better method to further evaluate the mutated candidates. Top candidates will be further tested in vivo to determine yield from complete pathways compared to wildtype.

References:

  1. Gordillo Sierra, A. R., & Alper, H. S. (2020). Progress in the metabolic engineering of bio-based lactams and their ω-amino acids precursors. Biotechnology Advances, 43, 107587. https://doi.org/10.1016/j.biotechadv.2020.107587
  2. Relax application. (n.d.). Retrieved October 18, 2021, from https://www.rosettacommons.org/docs/latest/application_documentation/structure_prediction/relax
  3. Smith, C. A., & Kortemme, T. (2008). Backrub-like backbone simulation recapitulates natural protein conformational variability and improves mutant side-chain prediction. Journal of Molecular Biology, 380(4), 742–756. https://doi.org/10.1016/j.jmb.2008.05.023
  4. Point mutant ("pmut") scan application, pmut_scan_parallel. (n.d.). Retrieved October 18, 2021, from https://www.rosettacommons.org/docs/latest/application_documentation/design/pmut-scan-parallel
  5. How to prepare structures for use in Rosetta. (n.d.). Retrieved October 18, 2021, from https://www.rosettacommons.org/docs/latest/rosetta_basics/preparation/preparing-structures
  6. Scoring tutorial. (n.d.). Retrieved October 18, 2021, from https://new.rosettacommons.org/demos/latest/tutorials/scoring/scoring
  7. Yeom, S. J., Kim, M., Kwon, K. K., Fu, Y., Rha, E., Park, S. H., Lee, H., Kim, H., Lee, D. H., Kim, D. M., & Lee, S. G. (2018). A synthetic microbial biosensor for high-throughput screening of lactam biocatalysts. Nature Communications, 9(1). https://doi.org/10.1038/s41467-018-07488-0
  8. Free energy of mutation calculations in Cyrus Bench. (n.d.). Retrieved September 20, 2021, from https://cyrusbio.com/wp-content/uploads/Rosetta-cartesian-DDG-2019.pdf.
  9. Lindahl, Abraham, Hess, & van der Spoel. (2020, July 9). GROMACS 2020.3 Manual (Version 2020.3). Zenodo.http://doi.org/10.5281/zenodo.3923644
  10. Vanommeslaeghe K., et. al. CHARMM general force field: A force field for drug‐like molecules compatible with the CHARMM all‐atom additive biological force fields. J. Comput. Chem., 2010. 31: 671-90.
  11. Mahnam, K et al. “Design of a novel metal binding peptide by molecular dynamics simulation to sequester Cu and Zn ions.” Research in pharmaceutical sciences vol. 9,1 (2014): 69-82.