Open Access

The thioredoxin (Trx) family of proteins comprises many key enzymes in redox signaling, that catalyzes specific reversible redox reactions, e.g. dithiol-disulfide exchange reactions, (de-)glutathionylation, trans-nitrosylation, or peroxide reduction. With the analysis of a large number of proteins, as well as a certain redox couple in [article 1] and [article 4], we demonstrated that electrostatic complementarity is the major distinguishing feature that controls the specific interactions of Trxs with their target proteins. The primary aim of this work was to determine the importance of this specific interaction and the prediction, modulation, and engineering of functional redox interactions of Trx family proteins. To understand the role of electrostatic complementarity for the mammalian Trx1-TrxR complex, we generated more than 20 hTrx1 mutants and systematically engineered the electrostatic potential within and outside the contact area with TrxR [article 1]. The effects of these specific alterations distributed all over the protein surface were analyzed by enzyme kinetics, differential scanning fluorimetry (DSF), circular dichroism (CD) spectroscopy, and MD simulations. Trx family proteins have a broad and very distinct substrate specificity, which is a prerequisite for redox switching. In [article 4], we comprehensively compared the classification of various redoxins from all kingdoms of life based on their similarity in amino acid sequence, tertiary structure, and electrostatic properties. These similarities were then correlated to the existence of common interaction partners. Our analyses confirmed that the primary and tertiary structure similarities do not correlate to the target specificity of the proteins as thiol-disulfide oxidoreductases. However, we demonstrated that the electrostatic properties of the protein from both Trx or Grx subfamilies is the major determinant for their target specificity. Although structurally very similar, CxxC/S-type or class I Grxs act as oxidoreductases and CGFS-type or class II Grxs act as FeS cluster transferases. In [article 3], we re-investigated the structural differences between the two main classes of Grxs to solve the mystery of the missing FeS transferase activity of the CxxC/S-type and the lack of oxidoreductase activity of the CGFS-type Grxs. The presence of a distinct loop structure adjacent to the active site is the major determinant of the Grx function. We confirmed that the function of Grxs can be switched from oxidoreductase to FeS cluster transferase by construction of a CxxC/S-type Grx with a CGFS-type Grx loop and vice versa. Results of several in vitro and in vivo assays together with the detailed structural analyses indicate that not a radically different substrate specificity accounts for the lack of activity, but rather slightly different modes of GSH binding, which is an essential nucleophile required in redox and iron homeostasis. Various processes within the cell depend on GSH, including redox reactions, reversible posttranslational modifications, and iron metabolim. GSH is not only important in the export of FeS precursors from mitochondria, but it is also an essential cofactor for cluster binding in iron sulfur Grxs. In [article 2], we discussed the role of GSH and iron sulfur Grxs in iron metabolism, the physiological role of CGFS-type Grx interactions with BolA- like proteins, and the cluster transfer between Grxs and recipient proteins. The first well characterized physiological function of a Grx-BolA hetero complex is presented with the Grx3/4-Fra2-mediated regulation of iron homeostasis in yeast. In synopsis, the results presented and discussed in these articles and the manuscript support the concept of electrostatic properties as the main determinant in substrate specificity towards functional predictions in Trx family proteins. The mathematical model presented here showed significantly accuracy and precision in function prediction. We are aware that our findings are focused on Trx family proteins as a particular family of proteins, but by using a machine learning strategy this mathematical model is being trained with numerous different protein models for better efficacy and accuracy, that may lead to new insights also in the specific interactions of other protein families. The new concept for the substrate specificity determinant doesn’t eliminate previously described aspects for molecular recognition, instead it reveals a deeper understanding of the protein-protein interaction. The 3D structural elements of a protein play a significant role in the specificity and function. We have been able to activate an inactive protein by replacing defined structural elements. Elimination of the loop structure from CGFS-type Grx5 transformed it from an FeS transferase into an oxidoreductase and the activity was further increased by modification of the active site. We believe that the present findings may be useful to investigate proteins in great detail regarding their function based on structure and electrostatic properties. Understanding the nature of the specific interactions may enable us to specifically modify the signal transduction pathways.