Spring 2021 challenge: phase 2 contenstant

Analysis of Heme Binding to Procoagulant Factor VIIIa

Diana Imhof, University of Bonn, Bonn, Germany
Screening of the primary sequence of activated FVIII (FVIIIa) with HeMoQuest resulted in 13 potential heme-binding motifs (HBMs) after manual refinement. The synthesized motifs (in form of nonapeptides) were analyzed for their heme-binding affinity by UV/Vis spectroscopy, which allowed for a classification in HBMs with high (4) and moderate (6) affinity. In order to support these results on the protein level, we selected a subset of high-affinity HBMs for molecular docking, which confirmed the experimental results. Next steps are: 1) Independent MD simulation of FVIIIa for insights into the conformational dynamics of the HBMs upon structure equilibration and analysis of better druggable sites (DoGSiteScorer). 2) Redocking of heme to the new MD snapshots of FVIIIa and longer MD simulations of the most promising heme-FVIII complexes to determine binding properties (Hyde, MMPBSA). 3) In vitro binding studies on the protein level. 4) Functional assays in the presence of heme.

After 3 months, Diana has achieved the following goals:

  1. The DoGSiteScorer algorithm in SeeSAR was run on the initial crystal structure of the B-domain-deficient FVIII (pdb: 2r7e) in order to find druggable binding pockets. A total of 29 unoccupied binding sites has been identified. Seven of our 13 predicted heme-binding sites are located in very close proximity to these sites. Most interestingly, the HBM with one of the highest experimentally determined heme-binding affinity was part of one of the best predicted druggable binding pockets comprising the following properties: Number of residues: 37, DoGSiteScore: 0.52, number of donors: 28, number of acceptors: 29, hydrophobicity: 0.67, surface: 647.28 Å2, and volume: 864.00.
  2. Docking heme to FVIIIa was not directly possible in SeeSAR due to the nature of heme as a complex containing the protoporphyrin IX ring system coordinating one Fe(III) ion. Hence, a first round of docking has been performed in YASARA. A cubic docking simulation cell (15 Å) was created around either the deprotonated imidazole nitrogen of a His or the phenol oxygen of a Tyr residue. Subsequently, the ensemble docking methodology was applied. Docked poses with a distance of <4 Å between the Fe(III) ion (heme) and the nitrogen or oxygen atoms (His and Tyr side chains of protein) were considered as heme-binding. Selected poses were energy minimized to refine the bound conformation. The energy-minimized complexes were provided to SeeSAR and heme was redocked using the standard docking method (FlexX). So far, we identified 3 motifs with good heme-binding potential. Redocking of these complexes resulted in HYDE scores that are in qualitative agreement with experimentally determined affinities.
  3. MD simulations of these first 3 complexes were started, uncovering very interesting results: In a MD simulation (~5 ns) of heme bound to a single His residue (as part of a predicted HBM) in FVIIIa a conformational change was observed that brought another His residue in close proximity to the distal side of heme. The distance sharply reduced from over 6 Å to less than 3 Å resulting in a hexacoordinated binding of heme. This binding has the potential to cause conformational changes to the protein that might trigger the exposition of other motifs provoking further heme binding. This might significantly impact the functionality of FVIIIa, a phenomenon observed in other heme-regulated proteins as well (e.g., Hopp et al. 2021). A fully-solvated complex system of FVIII-heme is comprised of ~190000 atoms and, hence, MD simulations with the standard hardware (8 cores/16 threads) is prohibitively slow. However, since the simulations yielded extremely useful results, these will be continued.
Scroll Up