Predicting the Structure of Enzymes with Metal Cofactors: The Example of [FeFe] Hydrogenases.

The advent of deep learning algorithms for protein folding opened a new era in the ability of predicting and optimizing the function of proteins once the sequence is known. The task is more intricate when cofactors like metal ions or small ligands are essential to functioning. In this case, the combined use of traditional simulation methods based on interatomic force fields and deep learning predictions is mandatory. We use the example of [FeFe] hydrogenases, enzymes of unicellular algae promising for biotechnology applications to illustrate this situation. [FeFe] hydrogenase is an iron-sulfur protein that catalyzes the chemical reduction of protons dissolved in liquid water into molecular hydrogen as a gas. Hydrogen production efficiency and cell sensitivity to dioxygen are important parameters to optimize the industrial applications of biological hydrogen production. Both parameters are related to the organization of iron-sulfur clusters within protein domains. In this work, we propose possible three-dimensional structures of Chlorella vulgaris 211/11P [FeFe] hydrogenase, the sequence of which was extracted from the recently published genome of the given strain. Initial structural models are built using: (i) the deep learning algorithm AlphaFold; (ii) the homology modeling server SwissModel; (iii) a manual construction based on the best known bacterial crystal structure. Missing iron-sulfur clusters are included and microsecond-long molecular dynamics of initial structures embedded into the water solution environment were performed. Multiple-walkers metadynamics was also used to enhance the sampling of structures encompassing both functional and non-functional organizations of iron-sulfur clusters. The resulting structural model provided by deep learning is consistent with functional [FeFe] hydrogenase characterized by peculiar interactions between cofactors and the protein matrix.

Probing protein stability: towards a computational atomistic, reliable, affordable, and improvable model.

We present an improved application of a recently proposed computational method designed to evaluate the change of free energy as a function of the average value of a suitably chosen collective variable in proteins. The method is based on a full atomistic description of the protein and its environment. The goal is to understand how the protein melting temperature changes upon single-point mutations, because the sign of the temperature variation will allow us to discriminate stabilizing vs. destabilizing mutations in protein sequences. In this refined application the method is based on altruistic well-tempered metadynamics, a variant of multiple-walkers metadynamics. The resulting metastatistics is then modulated by the maximal constrained entropy principle. The latter turns out to be especially helpful in free-energy calculations as it is able to alleviate the severe limitations of metadynamics in properly sampling folded and unfolded configurations. In this work we apply the computational strategy outlined above in the case of the bovine pancreatic trypsin inhibitor, a well-studied small protein, which is a reference for computer simulations since decades. We compute the variation of the melting temperature characterizing the folding-unfolding process between the wild-type protein and two of its single-point mutations that are seen to have opposite effect on the free energy changes. The same approach is used for free energy difference calculations between a truncated form of frataxin and a set of five of its variants. Simulation data are compared to in vitro experiments. In all cases the sign of the change of melting temperature is reproduced, under the further approximation of using an empirical effective mean-field to average out protein-solvent interactions.

Modelling Protein Plasticity: The Example of Frataxin and Its Variants.

Frataxin (FXN) is a protein involved in storage and delivery of iron in the mitochondria. Single-point mutations in the FXN gene lead to reduced production of functional frataxin, with the consequent dyshomeostasis of iron. FXN variants are at the basis of neurological impairment (the Friedreich's ataxia) and several types of cancer. By using altruistic metadynamics in conjunction with the maximal constrained entropy principle, we estimate the change of free energy in the protein unfolding of frataxin and of some of its pathological mutants. The sampled configurations highlight differences between the wild-type and mutated sequences in the stability of the folded state. In partial agreement with thermodynamic experiments, where most of the analyzed variants are characterized by lower thermal stability compared to wild type, the D104G variant is found with a stability comparable to the wild-type sequence and a lower water-accessible surface area. These observations, obtained with the new approach we propose in our work, point to a functional switch, affected by single-point mutations, of frataxin from iron storage to iron release. The method is suitable to investigate wide structural changes in proteins in general, after a proper tuning of the chosen collective variable used to perform the transition.