Conventional methods for enzyme modification depend on re-sampling and in vitro confirmation studies, which makes them time-consuming and costly exercises. Two of such most widely accepted and followed methods for engineering enzymes are Rational Design and Directed Evolution. However, they fall short of providing biologically relevant information on stability and activity, thus the number of success stories using these approaches have been far less than the failures encountered, mainly due to:
- Silent substitutions - dissimilar sequences that yield similar structural characteristics in comparison to wild type without any changes in the conformation
- Even, major changes lead to futile changes in the functional properties of an enzyme
- Predicting accurate enzyme modifications for better physical and catalytic properties like thermal/ pH stability, activity, substrate specificity, allosteric regulation, controlled enzyme activity (individually or in combination)
- Generating quantifiable (‘n’ log or fold) increase in activity of modified enzyme models
- Eliminating the need for iterative in vitro validation experiments
There is no such integrated platform that can produce thousands of modifications and predict activation energy and rate limiting steps at different process conditions to achieve the required biocatalysis process.
Connecting the dots: Polyclone’s in silico enzyme engineering framework (eEF)
Polyclone’ in silico Enzyme Engineering Framework (eEF) is a smart integration of proprietary algorithms and processes for high throughput and high confidence predictions of enzyme modifications for desired properties such as better activity, thermal/pH stability and substrate specificity. eEF platform integrates computational library design, directed evolution and statistical modeling of sequence–function relationships, while considering the intermediary stages of enzyme-substrate interaction to achieve its goals.
- Tailor-made solution for enhancing the physical and catalytic properties of enzymes
- The process takes into account the intermediary transition states to provide biologically relevant predictions
- Fast turn-around time and huge costs cut down in comparison to contemporary enzyme engineering methods
eEF platform explores the conformational spaces of biological interactions using computational methods by incorporating statistical checkpoints at all possible modes during the enzyme-substrate interaction. The framework helps in reducing the number of validation experiments in vitro by providing crucial information about the enzymes; hence eEF platform is used as a high throughput screening system for enzyme engineering.
The components of eEF platform:
- hot spots identification: eEF platform identifies hot spots that are potential sites for mutations/ modifications by simulating enzymes using different sampling techniques and the resultant trajectories to calculate interatomic weights for the residues in and around the active site. These weights are based on the physical contact (favourable and unfavourable) and contact energies (favourable and unfavourable) like Van der Waals forces, electrostatic energies, hydrogen bond energies, pi-pi and cat-pi energies etc. Also, during the simulation process eEF platform explores the path taken by the active substrates (high and low activity), inactive substrates (no activity) and substrates of interest (may be inactive or less active) to enter the active site of the enzyme; which helps in identifying modification sites even away from the binding site. Also, this information helps in modifying the enzyme such that it will facilitate in substrate entry to the active site and product dissociation from the active site, thereby optimizing the whole enzymatic reaction.
- In silico mutagenesis: eEF models thousands of mutations all around the enzyme in the identified hotpots as detailed above.
- Filtering mutations / modification: Multiple levels of filters are used to predict the potential mutations; during this process eEF platform uses intramolecular and intermolecular energies of the enzyme and substrates involved in the reaction. In fact, the biasing of scoring functions and transformations of scoring functions vary from enzyme to enzyme. The scoring functions contain combination of many standard empirical and statistical scoring terms. Certain scoring terms are transformed (changing the weights) based on the type of interaction, for example, if it is a charge based interaction then we change the weight by giving more importance to the electrostatic interactions. These transformations are calibrated over a set of experimental enzyme/s activity data (which belongs to the same class of the enzyme of interest) i.e. the weights and scoring functions are biased to reflect the experimental activity values of the enzyme against different substrates or different variants of the enzyme against a substrate. In addition, eEF platform uses transformations of desolvation parameters, intra hydrogen bond energies etc., to compute the stability of enzymes.
- In silico E-S reaction of top modifications: The filtered modifications enter the “Transition State Analysis” of eEF platform, which generates reaction coordinates of different stages of transition states that are mentioned below. This is carried out by QM/MM simulations, steered molecular dynamics and transition path sampling. Activation energy and rate limiting steps of the reaction are predicted using these reaction coordinates.
Different transition states:
1. E-S transition state: The conformational changes and the energy raise due to substrate binding will be computed (induced fit modes)
2. E-SE-P transition state: During this transition state, the enzyme alters the electronic structure by protonation, proton abstraction, electron transfer, geometric distortion etc., by either bond breaking or bond making. 3. E+P transition state: Conformational changes and the energy required for dissociation of product from the enzyme and the energy of enzyme required to initiate the next catalytic cycle
Why Polyclone's eEF?
- eEF can be used to engineer enzymes – Against thermal & pH denaturation
- Predict hotspots and enzyme modifications throughout the protein for better activity – The hotspots are predicted in the catalytic site as well sites aiding in catalysis, which can be away from the active site
- In silico experiments conducted to predict hotspots are carried out in different process conditions like desired solvent, pH, temperature etc.
- Never misses a potential modification! – Produces and screens about 100,000 enzyme modifications in one iteration
– Increase activity on novel substrates
– Attain activity against a synthetic substrate etc.
- Molecular dynamics and Integrated 4D-QSAR formalism, which uses more than 8000 descriptors to identify Hotspots
- Homology and ab initio modeling methods are used to model the 100,000 3D modifications of an enzyme
- The modification are screened using – Ensemble docking analysis
- Transitions states are calculated using – Transition path sampling – QM/MM studies – Steered MD
– Using transformed scoring functions based on the contact energies like Van-der Waals forces, electrostatic potential, hydrogen bond, pi-pi energies & cat-pi interactions and GCOD descriptors