Loop Structure of Antibody Refinement

Creative Biolabs possesses unchallenged experience in antibody structure modelling. With our well-developed antibody structure modelling platform, multiple modelling strategies are available here. We offer high-quality loop structure refinement service to assist you to generate an expected antibody structure.

Loop Structure of Antibody

Antibodies have several different forms, varying from large, multi-chain and multi-domain complexes, to small, single domain molecules. Nevertheless, binding always happens in a similar form, via interactions between the antigen and many loops on the antibody complementarity determining regions (CDRs). Generally, there are six CDRs loops in mammalian antibodies, three on the heavy chain and three on the light chain (L1, L2, L3 and H1, H2, H3 respectively). For L1, L2, L3, H1 and H2, only a small number of different forms have been discovered, generating a set of discrete conformational classes called canonical structures. Different from the five CDRs, the H3 loop has not been classified into canonical forms; a great number of structures have been discovered. It is considered that H3 loop contribute the most to an antibody’s antigen-binding properties. Thus, antibody H3 modeling is the most important and difficult step of antibody structure modeling.

Loop Modelling

The first problem is a group of missing residues in an antibody structure, where the sequence of the missing segment is known, whereas the three-dimensional structure of those residues is not. There are three major steps of predicting the structure of the loop: decoy generation, filtering, and ranking.

• The initial step decoy generation refers to form a number of candidate conformations, or decoys, which related to the residues on either side of the gap in the protein structure. As some of the decoys obtained will not be physically possible, for example, atoms may be too close together.
• Therefore, a filtering step is needed to remove these structures, and it may be integrated with the other parts of the loop modelling process.
• After all decoys have been formed, a ranking system is required to choose a final prediction. This is an essential step, although decoys close to the native structure have been produced at a previous stage, a useless ranking system implies that the structure selected as the final prediction will be inaccuracy.

Fig 1. The major steps in an H3 loop modelling algorithm. Part A shows the inputs to the algorithm, which are an antibody structure with a missing loop, and the sequence of that loop. Part B shows the decoy generation, part C shows the filtering of structures, and part D shows the ranking and selection of the final prediction. (Marks, C., 2017)

Loop Structure Refinement

In order to further improve the accuracy of antibodies structure prediction, Creative Biolabs has also develop a two-phase strategy, which combines:

• CReF (Central Residue Fragment-based) expert prediction of an approximate 3-D structure;
• And refinement depended on molecular dynamics (MD) simulations of a physics-based potential energy function.

The refinement method takes advantage of simulations by the MD approach. In this way, the optimization of the loop regions and consequent progress to the overall conformation of the predicted 3-D structure will be well performed.

Fig 2. Examples of successful refinement in CASP11. (Lee, G. R., 2016)

With our advanced loop structure refinement service, designing and engineering novel antibodies with desired properties is available. We customize the service according to the specific requirements from our customers, and our service will contribute greatly to the success of your project. We also provide other antibody structure modelling services. Please contact us for more information and a detailed quote.

References

1. Marks, C., (2017). “Antibody H3 Structure Prediction.” Computational and structural biotechnology journal, 15, 222-231.
2. Dall’Agno, K. C., (2013). “An expert protein loop refinement protocol by molecular dynamics simulations with restraints.” Expert Systems with Applications, 40(7), 2568-2574.
3. Lee, G. R., (2016). “Effective protein model structure refinement by loop modeling and overall relaxation.” Proteins: Structure, Function, and Bioinformatics, 84(S1), 293-301.

All services provided on this site are intended to support preclinical research only. Do not use our services or final products on humans.