1. Development of the single molecule assay.
The first objective was to set up a reliable Single Molecule assay to study the kinetics of a proteolytic reaction. we had to develop different strategies based on the engineering of protein substrates, that can be stretched with an Atomic Force Microscope. Four different approaches have been developed (Figure1).
Finally, with the combination of an engineered octamer substrate and the design of an adequate force protocol, we successfully obtained experimental recordings that could be used to measure the reaction rates.
2. Extensive measurement of kinetics with different conditions
With this optimized assay (Figure 2), I was able to repeat the experiments in different conditions to measure the effect of Tev concentration and force. Both conditions greatly affect the kinetics of cleavage. A total of 3500 traces could be used across more than 100 experiments. The increase of Tev concentration shows a hyperbolic acceleration of the reaction with a plateau at 0.8 s-1, defining a limiting reaction step that corresponds to the catalytic rate, kcat, measured in bulk.
Alongside, the force dependency displays an unexpected two-regime behavior that could not be explained with polymer physics models.
3. Molecular Modelling of protein stretched by force
Single molecule experiments show that the reaction rate is modulated by force and enzyme concentration. In a simple kinetics model, we propose that the observed rate of cleavage is due to two consecutive reaction steps, the complex formation followed by the catalysis.
We elucidated how forces impact the conformation of the pulled chain and the association rate via molecular simulations.
First, we used steered molecular dynamics (SMD) on two different simplistic chains (10-alanines and 100-alanines). However, these 20 ns long SMD simulations were not sufficient to describe the complete conformational space of the chain. In our experiments, this kinetics occurred in a millisecond-second time scale, a time significantly slower than the ns simulations.
For this reason, we have developed a new method based on free energy calculations and that overcomes the limitations of SMD simulations.
This approach combined simulations with Adaptive Biased Forces and statistical mechanics to describe the elasticity of proteins with a new molecular description. The resulting model successfully provided the new possibility to estimate how the stretching of proteins can modulate the substrate conformation and the association rate. Remarkably, we observed a close correlation between the modeled and experimental force dependencies.
4. Generalization of the approach
We were able to predict the force dependency for different proteases from a simple analysis of the backbone conformation of the substrate bound to its cognate enzyme.
For instance, the holo xray structure of thrombin suggests an optimal cleavage force in the range of the unfolding force (~170pN). A critical strength of the “proof of concept” assay with the Tev was the ability to mechanically unfold the substrate in a range of force that prevents the reaction. Because of this feature, we cannot expand our assay to broadly study other proteases, however, a collaboration in the lab led to a new study that expands the general concept of our approach.
