Model

This is a video, made by our mentor Fiona Kearns, using VMD to visualize the denaturing of PETase as temperature increases.

Methodology

---

Our team used a tool called Molecular Dynamics Simulations to construct virtual, computational models of all of our PETase types, and subjected the different types of PETase to a range of temperatures, from 310 K to 510 K, measuring the PETase types’ structure and function as they began falling apart and losing their function.

Molecular dynamics simulations model all atoms and molecules as hard spheres with certain properties, like mass and charge. Those hard spheres interact with one another through a molecular mechanical force field, which approximates chemical interactions and chemical energies:

For example the energy of a chemical bond may look like this mathematical equation in the molecular mechanical force field:

This equation is based on Hooke’s spring force law where Kij is the force constant of a spring connecting two spheres, l0ij is the equilibrium distance, when molecules are either not moving are are moving with constant speed, between those two spheres and lij is the distance between those two spheres at any given step in the simulation.

Additionally, charges between two spheres can interact via Coulomb's Law which takes this form:

Where qi is the charge of sphere i and qj is the charge on sphere j and rij is the distance between them.

The molecular dynamics program we used, NAMD3.0, calculates the derivative of the energy on all atoms in the system (the energy over time) to get the force on each atom in the system because force = negative derivative of the potential energy. Then NAMD3.0 uses the force on each atom to predict, according to force = mass * acceleration, in which direction each atom will move and after a certain amount of time where each atom will be. The molecular dynamics program then does this integration of Newton’s laws of motion over and over again to produce a “movie” of how molecules move over time. Each simulation is conducted at a specific temperature (310K, 360K, 410K, 460K, or 510K), and if you raise the temperature in the simulation you can “see” or predict how the protein or molecule might fall apart as the temperature in the simulation increases. From these simulations you can then track specific molecular properties of interest to pinpoint which properties “fall apart” the fastest with respect to temperature.

To construct our PETase protein models, we used a tool called Visual Molecular Dynamics (VMD), which has protein model building tools. With these tools, for each “mutant” protein, either the “Double Mutant”, “HotPETase”, or our own designed “Westview Petase” we enacted the mutations through the mutate command in VMD tools. To model PETase in a biologically relevant context, we then added water around the protein and 150 mM NaCl, to have a final simulation setup that looks like this:

Figure: Picture of the Westview PETase protein (blue) in a box of water (grey transparent cube) with NaCl ions. The water molecules are represented as a grey transparent cube for clarity, but all water atoms are represented explicitly during these simulations.

For each PETase type, we then conducted a long “restrained” and controlled simulation at 310K for a molecular time of 50 nanoseconds. The restraint in this first 50ns was to keep the catalytic triad residues (Ser160, Asp206, His237) in close proximity to one another. After the first 50ns, we removed this restraint, then simulated for another 10ns at 310K, then simulated at 360K for 10ns, 410K for 10ns, 460K for 10ns, and 510 K for 10ns. For each PETase type, we repeated these simulations 3x for three trials each. We then used collaborative python notebooks through Google Collab to perform analysis on all resultant simulations with MDAnalysis, a python package for collecting data from MD simulations.

For the catalytic triad, we used MD Analysis to calculate the distances between the different catalytic triad residues (160, 237, and 206), and then we plotted the change in distance over the simulation time measured in frames. For the Radius of Gyration, we used a function that took simulation results for the Radius of Gyration of our protein (which used a tool in MDAnalysis to solve the equation Rix=i*mi[r2iy+r2iz]i*mi, where rix represents the radius of gyration around the x-axis, mi is the mass of atom i, and ri is the position of atom i) and graphically represented them in a way that related time and Radius of Gyration, and temperature as well. We then ran the function three trials, used code to calculate the average and standard deviation of all three trials, then combined them and plotted them on a fourth graph. To find the root mean square deviation, we utilized another package in MD Analysis to obtain our results. This package allowed us to compare how the catalytic triad residues (160, 237, and 206) varied from the stable orientation of this triad as temperature increased. In order to find the Root Mean Square Fluctuation (RMSF) of our many proteins, we first used a theoretical PETase model to obtain data of each amino acid’s trajectory as temperature increased over time. By again using MD Analysis, we then used calculations of each amino acid’s RMSF to create three sample graphs that were then averaged together to make a cohesive model of the entire protein’s RMSF.

Results and Analysis

---

The catalytic triad consists of three amino acids, Serine (S160), Histidine (H237), and Dalanine (D206), within the structure of PETase. The changing relationships of the positions of these amino acids is what allows PETase to break down PET plastics. The most important distance is between the serine and histidine.

The distance between D206 and S160, represented by the red line, and the distance between H237 and D206, represented by the blue line, appears to follow a similar pattern throughout all 4 types of PETase. However, the S160 to H237 distance, represented by the black line, changes significantly among PETase types. At 410K, this distance in PETase is shown to increase initially, but then the rate of increase slows down. If we fit these graphs to a function, and then take the derivative at about 410K, the derivative at 410K is lower for the PETase than for the HotPETase, the Westview PETase, and the Double Mutant PETase. This distance increases at a faster rate in the three other types of PETase (HotPETase, Westview PETase, Double Mutant PETase) and then the rate of increase flattens out, similar to the PETase. At 460K, the S160 to H237 distance is much higher for Westview, Double Mutant, and HotPETase than for regular PETase. At 460K, the S160 to H237 distance for Double Mutant PETase decreases while the HotPETase increases, and the Westview PETase and PETase S160 to H237 distances both seem to increase and then decrease. At 510 K, PETase and Double Mutant PETase experience an increase in the distance between S160 and H237, while the HotPETase and Westview PETase experience a decrease.

Overall, the melting point of Westview PETase appears to have extreme fluctuations in catalytic triad distances at around 410-460 K. And this is also true of HotPETase, so we concluded that the melting point of Westview PETase is likely the same as the melting point of HotPETase. Therefore, we may not have made a more thermostable PETase, however, we did not make the enzyme less stable at higher temperatures.

The radius of gyration measures the rotation and movement of the PETase enzyme for increasing temperatures. When comparing all four variations, we originally noted that there were very few discrepancies between them. However, some noticeable differences were between PETase and the other three PETase types at around 410K. PETase has a jump in the radius of gyration, but the other graphs show a flat line, indicating little movement at this temperature. In the Westview PETase model, while the 410K region is comparatively thermostable, we see a loss in stability represented through a spiking and increasingly delinear slope around the 460K temperature region. Both Westview PETase and HotPETase showed a more spiking and varying radius of gyration at around 460 K, whereas the Double Mutant PETase and PETase showed a fairly constant increase. And at 510 K, only the PETase showed a constant increase—all other three types of PETase showed multiple “spikes” in the radius of gyration as they increased and decreased multiple times. Therefore, while the Westview PETase becomes equally as unstable as other variants in the 510 K temperature region, it performs similarly to other variants in lower temperature ranges (310-410) in terms of thermostability. Overall, the thermostability of the Westview PETase performed similar to other variants.

When comparing the PETase and Double Mutant PETase, looking at the 410K region, there is a significant spike in the Root Mean Square Deviation (RMSD) of the PETase, and an even larger spike in the RMSD of the Double Mutant PETase at a slightly higher temperature, closer to 460K. In the 410 K region, the RMSD of the Westview PETase and the HotPETase seem to behave similarly, increasing at an almost constant rate (except for when the HotPETase begins to flatten out). At around 460K, PETase and HotPETase increase in RMSD, whereas the Westview PETase and Double Mutant PETase increase quickly before decreasing. At 510K, the RMSD of Westview PETase and HotPETase form spikes as they increase and then decrease, while PETase and Double Mutant PETase increase. Given this data recorded for our Westview PETase molecule, we can tell that the temperature the protein begins to lose stability is approximately 460 Kelvin. In comparison to the Double Mutant PETase and wild type PETase, the Westview PETase does appear to remain thermostable for a larger temperature range, as the spike in RMSD is not until a higher temperature. By using RMSD as a means of determining if Westview PETase is more thermodynamically stable than other forms of PETase, we can conclude that there is a slight increase in thermostability.

Root Mean Square Fluctuation (RMSF), which helped us identify areas of possible mutations to improve PETase’s thermostability, displayed different results through all four types of PETase. The most important lines to look for are the dark blue and white lines, representative of 310 K (36.85 C) and 410 K (136.85 C) respectively. At these temperatures, PETase (40 C) and HotPETase (82.5 C) will have surpassed their melting points allowing, through analysis of their RMSF fluctuations, to find possible mutations that could strengthen a future protein. From PETase to HotPETase, especially at the amino acids 180-200 and 200-230, mutations of less thermostable amino acids regions reduced the melting point of PETase and improved the weakest points in the protein. However, for Westview PETase, its RMSF appeared very similar to HotPETase. This is most likely due to Westview PETase’s changes being only variations of a HotPETase base. This changes when looking at Double Mutant PETase. The mutations around amino acids 30-50 and 50-100 were able to lower the RMSF in both 310 K and 410K temperatures. With fewer fluctuations at higher temperatures, compared to PETase, mutated types of this enzyme can now degrade PET plastics more efficiently as heat “softens” the material as it reaches its glass transition temperature (which for PET plastics is around 60-70 degrees C).

Sources

---

Our best resource was our mentor, Fiona Kearns! We want to thank her so much for all her help this season!

Aguda, Adeleke H., et al. "Structural Basis of Collagen Fiber Degradation by Cathepsin K." Proceedings of the National Academy of Sciences. PNAS, www.pnas.org/doi/10.1073/pnas.1414126111f. Accessed 9 Oct. 2023.

Austin, Harry P., et al. "Characterization and Engineering of a Plastic-degrading Aromatic Polyesterase." Proceedings of the National Academy of Sciences. PNAS, www.pnas.org/doi/10.1073/pnas.1718804115. Accessed 9 Oct. 2023. Originally published in Proceedings of the National Academy of Sciences, 8 May 2018.

Shosuke Yoshida et al.,A bacterium that degrades and assimilates poly(ethylene terephthalate).Science351,1196-1199(2016).DOI:10.1126/science.aad6359

Bell, E.L., Smithson, R., Kilbride, S. et al. Directed evolution of an efficient and thermostable PET depolymerase. Nat Catal 5, 673–681 (2022). https://doi.org/10.1038/s41929-022-00821-3

Han, X., Liu, W., Huang, JW. et al. Structural insight into catalytic mechanism of PET hydrolase. Nat Commun 8, 2106 (2017). https://doi.org/10.1038/s41467-017-02255-z

Alissa Rauwerdink and Romas J. Kazlauskas ACS Catalysis 2015 5 (10), 6153-6176 DOI: 10.1021/acscatal.5b01539

Roth C, Wei R, Oeser T, Then J, Föllner C, Zimmermann W, Sträter N. Structural and functional studies on a thermostable polyethylene terephthalate degrading hydrolase from Thermobifida fusca. Appl Microbiol Biotechnol. 2014 Sep;98(18):7815-23. doi: 10.1007/s00253-014-5672-0. Epub 2014 Apr 13. PMID: 24728714.

image sources: https://en.wikipedia.org/wiki/Force_field_(chemistry)