Antibody Fragments

169 structures of protein antigens (length >30 amino acids) in complex with antibody fragments have been manually collected from the PDB [41 (link)] of January 2006 at a resolution ≤4Å. Every structure has been manually curated within the IEDB database [1 (link)] and inspected using the EpitopeViewer visualization tool developed by the authors [65 (link)]. Structures in which the antibody binds antigen but involves no CDR residues have been excluded from the analysis; there were four such structures [PDB: 1MHH, 1HEZ, 1DEE, 1IGC]. If a structure contained several complexes in one asymmetric unit (there were 46 such structures in 165) and the authors of the structure observed no structural difference between these complexes, only one complex was selected – those that were specified as a reference complex by the authors of the article describing the structure (primary citation in the PDB); there were 18 such structures out of 46. If the authors didn't provide this information, all complexes in the structure were considered for analysis. The authors of a few structures clearly stated in their papers that antibody-protein contacts in the complexes were different: [PDB: 1MLC, 1NFD, 1OB1, 1P2C, 1QFW]. This initial curation has performed in order to correctly assign the protein-antibody complexes and decrease the number of individual complexes analyzed from 226 to 187 from a total of 169 structures. A total of 24 complexes were formed by one-chain antibody fragments and 163 complexes by two-chain antibody fragments. Alignment of protein chains was performed using the CE algorithm [58 (link)].

In order to test the single step in frame cloning of scFv antibody gene fragments, more than 20 scFv gene fragments obtained by phage display to different antigens from the antibody gene libraries HAL 7/8 [4 (link)] were cloned into the NcoI/NotI cloning site of pCSE2.5-hIgG1Fc-XP to obtain scFv-hIgG1Fc antibody constructs. High quality plasmid preparations for transfection were done using the NucleoBond Xtra Midi Kit according to the manufacturer’s description (Machery Nagel, Düren, Germany).

Co-immunoprecipitation experiments were performed using the Crosslink Immunoprecipitation Kit (Thermo Fisher Scientific, Waltham MA, USA, 26147). This method involves capturing 20 μg antibody on Protein A/G beaded agarose resin and covalently immobilizing it on the support by crosslinking with 2.5 mM disuccinimidyl suberate (DSS). The antibody resin was then incubated at 4 °C for 12 h with 500 μg precleared guinea pig spermatozoa protein extracts, allowing for the antibody-antigen complex to form. Proteins bound to the respective antibodies were eluted, recovered by low-speed centrifugation (3000× g), and stored at −20 °C. Only the antigen was eluted during the procedure, enabling it to be identified and analyzed with minimal interference from antibody fragments. For protein disulfide reduction, supernatant aliquots were boiled for 5 min in 3X Laemmli buffer (pH 10) containing 2-mercaptoethanol, and the proteins were then separated by SDS-PAGE. Next, proteins were transferred to nitrocellulose membranes for immunodetection. The recovered proteins were analyzed by WB using the adequate antibodies.
Negative control was performed by associating 20 µg of an IgG unrelated to NOX2 or NOX4 with the agarose-protein A/G beads. The beads were incubated with 500 µg of sperm protein extract in the same way as before. The WB carried out with the anti-NOX2 antibody did not show the presence of this protein, nor of another (Supplemental Figure S1).

The local flexibility of the single residues during the cMD simulations was determined by calculating the root mean square fluctuations (RMSF). This was achieved by the use of AMBER’s CPPTRAJ implementation [47 (link)]. The structures of the antibody variable fragments without considering the antigens, were therefore aligned on all Cα-atoms of the crystal structure and the fluctuations calculated on the Cα-atoms in a mass-weighted manner [47 (link)].
The obtained simulation trajectories were analyzed with a principal component analysis (PCA) on the Cα-atoms of the CDR 2 loop, and of the binding residues of the CDR 3 loop, as those are the regions directly in contact with the antigen. For this analyses, the PyEMMA 2 python library was used. Additionally, the PCA spaces, using as input features the backbone torsions, and the Cα-atoms of all hypervariable loops individually, were investigated (Supplementary Figures S3–S5).
For the reduction in the dimensionality and the subsequent construction of a Markov State Model (MSM), a time-lagged independent component analysis (tICA) was performed again using the PyEMMA 2 python library. tICA was applied to identify the slowest degrees of freedom [49 (link),50 (link),51 (link)].
The obtained tICA space was therefore clustered geometrically by a k-means clustering algorithm in order to define a set of microstates [52 (link)]. For each simulation, a total number of 120 k-means clusters was defined. These microstates were then coarse-grained into macrostates by use of a fuzzy PCCA+ clustering algorithm, implemented in the used PyEMMA 2 library [49 (link),53 (link)]. This way, kinetically relevant states were defined and transition probabilities between them could be calculated. To construct the Markov-state models we applied a lag-time of 15 ns and evaluated the reliability of the constructed MSM with the so-called Chapman–Kolmogorov test [54 (link),55 (link)].
For the calculation of the contacts between antibody and antigen, as well as for the intermolecular contacts, the GetContacts tool provided by the University of Stanford was used (https://getcontacts.github.io/, accessed on 21 October 2022) [56 ]. This software is able to compute interactions based on pre-defined criteria. For the purpose of this study, the hydrogen bonds beneath a distance cut-off of 3.5 Å between all atoms were computed. Therefore the evolution of contacts for different stages of maturation could be quantified, and the contacts could be directly compared using a flare plot visualization.