The variety of functions performed by proteins of the thioredoxin superfamily, including glutaredoxins, involves the wide range of redox potential associated with the -Cys-X-X-Cys- motif found in their active sites. The determinants of these differences in redox potential are still obscure. A better understanding requires a detailed characterization of the reduced state of these enzymes, especially because the lowered p K a of the reduced N-terminal active-site cysteine is a key feature of these enzymes' chemistry, including their redox potential. Analysis of the factors controlling this p K a is complicated by the apparent structural heterogeneity of the reduced active sites across glutaredoxins. In this family, pig glutaredoxin (pGrx) was one of the first to be functionally characterized, including some intriguing mutagenesis data, but a structure of its reduced state has been lacking. We used long molecular dynamics simulations and electrostatic calculations to analyze the structure, dynamics and electrostatics of reduced pGrx and some of its mutants. Comparison with experimental data is drawn whenever possible. It is shown that a dynamic model is essential to capture the structural properties of the cationic side-chains around the -Cys22-Pro23-Phe24-Cys25- sequence in the pGrx active site. Examples include Arg26, which can swing to stack on this sequence, and Lys19 which can contact the thiolate. However, contrary to a commonly held hypothesis, these cationic side-chains provide little stabilization for the thiolate, implying that they affect the enzymatic activity via other mechanisms. The p K a value of nucleophilic cysteine 22 (p K a 22) is dominated by local hydrogen-bonds, formed only in a well-defined active-site conformation, supported by a comparison between the calculated and experimental values of p K a 22. The edge of the aromatic ring of Phe24 is polar enough to contribute to stabilize the thiolate, consistent with the conserved aromatic side-chain at this position in the glutaredoxin motif. The locality and directionality of the hydrogen bonds in the active site suffice to explain the vast difference between the p K a values of its two cysteine residues. A control of the cysteine p K a values by local hydrogen bonds implies that the peripheral ionized side-chains can evolve independently of the maintenance of these p K a values, maybe guided instead by substrate recognition. Comparison with other glutaredoxins indicates that the calculated p K a values of the N-terminal cysteine are better conserved than those of the C-terminal cysteine. Overall, a methodological strategy to systematically compare all reduced enzymes of this family emerges.
Read full abstract