This 48G7 hybridoma line Fab, binding to the hapten 5-(para-nitrophenylphsponate)-pentanoic acid, was derived from the germline esterolytic antibody fragments with the PDB codes 2RCS and 1AJ7.32Clustering of the 1s metadynamics simulation resulted in 120 cluster representatives that were simulated for each 100 ns. canonical clusters and additional dominant solution structures in the micro-to-millisecond timescale for all CDR loops, independent of length and sequence composition. Besides identifying all relevant conformations in solution, our results revealed that various canonical cluster medians actually belong to the same kinetic minimum. Additionally, we reconstruct the kinetics and probabilities of the conformational transitions between canonical clusters, and thereby extend the model of static canonical structures to reveal a dynamic conformational ensemble in solution as a new paradigm in the field of antibody structure design. Abbreviations:CDR: Complementary-determining region; Fv: Antibody variable fragment; PCCA: Perron cluster analysis; tICA: Time-lagged Methylene Blue independent component analysis; VH: Heavy chain variable region; VL: Light chain variable region KEYWORDS:Antibody design, canonical clusters, molecular dynamics simulations, conformational transitions, ensembles in solution == Introduction == The importance of characterizing and engineering the structure of antibodies to improve specificity, stability and suitability as biotherapeutics has increased substantially in the past decades.14Natural occurring antibodies are symmetric Y-shaped proteins, and each symmetric unit consists of a heavy and light chain. Sequence and structural diversity of antibodies is concentrated on six hypervariable loops, also known as complementarity-determining regions (CDRs), located within each of the two antibody antigen-binding domains. Three hypervariable loops (CDR-H1, CDR-H2, CDR-H3 and CDR-L1, CDR-L2, CDR-L3) are located on the heavy and light chain, respectively.5Various studies have focused on classifying five of the six CDR loops into canonical conformations, except the CDR-H3 loop, assuming that, depending on the length and sequence composition, antibody CDR loops only adopt a limited number of main-chain conformations.6 The highest variability in length, sequence, and structure can be observed for the CDR-H3 loop. It is known that the CDR-H3 loop samples a large number of conformations during V(D)J recombination and somatic hyper-mutations.7,8Together with the CDR-H3 loop, the CDR-L3 loop is situated hRPB14 in the center of the paratope and contributes to antigen recognition. The CDR-L3 loop reveals a comparable diversity to the CDR-H3 loop, but, without the contribution of a D gene, the degree of variability is lower.9Besides the CDR-H3 and CDR-L3 loops, the CDR-H1 and CDR-H2 loops also play a vital role in many antibodyantigen interactions and are therefore targeted for mutagenesis in synthetic libraries. With the substantial rise in the number of antibody crystal structures, the number of antibody databases and sequence-based classification servers has also increased significantly. 1013Numerous studies have tried to classify antibody CDR loops structurally and sequentially, and correlate them with their locus and sequence to improve fast antibody structure prediction and design.1316Additionally, several numbering systems for antibodies have been developed that are similar in the framework region but differ around the CDRs.1720The PyIgClassify database assigns conformational clusters by Methylene Blue determining the CDR sequences and lengths using the international ImMunoGeneTics information system nomenclature18and calculating the dihedral angles and of the residues in each CDR.13Recent studies using molecular dynamics simulations extended the model of static canonical clusters to a characterization of the CDR-L3 and CDR-H3 loop as ensembles in solution by capturing several conformational transitions between different canonical structure medians.21,22 We analyzed the conformational diversity of all CDR loops to identify transition probabilities and timescales between canonical CDR loop conformations of the same length, and to kinetically characterize the CDR loop ensembles in the solution for all CDR loops. Therefore, we chose the median crystal structures of the high-populated canonical clusters with a good resolution and used this median X-ray structure as the starting point for molecular dynamics simulations. == Results == == CDR-L1 loop Methylene Blue ensemble in solution == The most common and highest populated CDR-L1 loop length observed in antibody crystal structures is 11 residues. With this loop length, there exist three canonical clusters L1-11-1, L1-11-2 and L1-11-3. As the starting structure for our simulations, we chose the antigen-binding fragment (Fab) binding to the tumor-associated human leukocyte antigen (HLA) HLA-A1.MAGE-A1 (Protein Data Bank (PDB) accession code: 1W72), which is the median crystal structure of the canonical L1-11-3 cluster and is the only one of the three available canonical clusters composed of lambda light chains.23As described in the methods section, we clustered the obtained 1 s metadynamics simulation and used the resulting 132 cluster representatives as starting structures for each 100 ns simulations. This approach does not only allow the generation of a broad ensemble of CDR-L1 loop conformations but also the characterization Methylene Blue of the ensemble in solution kinetically.Figure 1ashows the obtained free energy surface of 13.2 s trajectories with the projection of the available canonical cluster medians (2D7 T, 1ZAN, 1W72) into the time-lagged independent component analysis (tICA) space. We observed that the 1W72 median crystal structure lies in the global free energy minimum in solution, while the two highest populated cluster medians.