Analysis and prediction of VH/VL packing in antibodies

Antibody modeling assessment II. Structures and models

To assess the state-of-the-art in antibody structure modeling, a blinded study was conducted. Eleven unpublished Fab crystal structures were used as a benchmark to compare Fv models generated by seven structure prediction methodologies. In the first round, each participant submitted three non-ranked complete Fv models for each target. In the second round, CDR-H3 modeling was performed in the context of the correct environment provided by the crystal structures with CDR-H3 removed. In this report we describe the reference structures and present our assessment of the models. Some of the essential sources of errors in the predictions were traced to the selection of the structure template, both in terms of the CDR canonical structures and VL/VH packing. On top of this, the errors present in the Protein Data Bank structures were sometimes propagated in the current models, which emphasized the need for the curated structural database devoid of errors. Modeling non-canonical structures, including CDR-H3, remains the biggest challenge for antibody structure prediction.

SAbDab: the structural antibody database

Structural antibody database (SAbDab; http://opig.stats.ox.ac.uk/webapps/sabdab) is an online resource containing all the publicly available antibody structures annotated and presented in a consistent fashion. The data are annotated with several properties including experimental information, gene details, correct heavy and light chain pairings, antigen details and, where available, antibody-antigen binding affinity. The user can select structures, according to these attributes as well as structural properties such as complementarity determining region loop conformation and variable domain orientation. Individual structures, datasets and the complete database can be downloaded.

N332-Directed broadly neutralizing antibodies use diverse modes of HIV-1 recognition: inferences from heavy-light chain complementation of function

Dozens of broadly neutralizing HIV-1 antibodies have been isolated in the last few years from the sera of HIV-1-infected individuals. Only a limited number of regions on the HIV-1 spike, however, are recognized by these antibodies. One of these regions (N332) is characterized by an N-linked glycan at residue 332 on HIV-1 gp120 and is recognized by antibody 2G12 and by the recently reported antibodies PGT121-137, the latter isolated from three donors. To investigate the diversity in mode of antibody recognition at the N332 site, we used functional complementation between antibody heavy and light chains as a means of assessing similarity in mode of recognition. We examined a matrix of 12 PGT-heavy chains with each of 12 PGT-light chains. Expression in 96-well format for the 144 antibodies (132 chimeric and 12 wild-type) was generally consistent (58±10 µg/ml). In contrast, recognition of HIV-1 gp120 was bimodal: when the source of heavy and light chains was from the same donor, recognition was good; when sources of heavy and light chains were from different donors, recognition was poor. Moreover, neutralization of HIV-1 strains SF162.LS and TRO.11 generally followed patterns of gp120 recognition. These results are consistent with published sequence, mutational, and structural findings, all of which indicate that N332-directed neutralizing antibodies from different donors utilize different modes of recognition, and provide support for a correlation between functional complementation of antibody heavy and light chains and similarity in antibody mode of recognition. Overall, our results add to the growing body of evidence that the human immune system is capable of recognizing the N332-region of HIV-1 gp120 in diverse ways.

Antibody affinity maturation using yeast display with detergent-solubilized membrane proteins as antigen sources

Antigen preparations in the form of detergent-solubilized cell lysates could, in principle, render membrane proteins (MPs) compatible with in vitro antibody engineering technologies. To this end, detergent-solubilized cell lysates were coupled with the yeast surface display platform to affinity mature an anti-transferrin receptor (TfR) single-chain antibody (scFv). Lysates were generated from TfR-expressing HEK293 cells by solubilization with detergent-containing buffer after undergoing plasma membrane-restricted biotinylation. Lysate-resident TfR was then combined with a mutagenic anti-TfR scFv library in a competitive, dissociation rate screen, and scFvs were identified with up to 4-fold improved dissociation rates on the surface of yeast. Importantly, although the lysates contained a complex mixture of biotinylated proteins, the engineered scFvs retained their TfR binding specificity. When secreted by yeast as soluble proteins, mutant scFvs bound to cell surface TfR with 3-7-fold improvements in equilibrium binding affinity. Although a known MP antigen was targeted for purposes of this study, employing biotin tagging as a means of antigen detection makes the lysate-based approach particularly flexible. We have previously shown that yeast display can be used to identify lead antibodies using cell lysate-resident MP antigens, and combined with this work showing that antibodies can also be quantitatively engineered using cell lysates, these approaches may provide a high-throughput platform for generation and optimization of antibodies against MPs.

Natural and man-made V-gene repertoires for antibody discovery

Antibodies are the fastest-growing segment of the biologics market. The success of antibody-based drugs resides in their exquisite specificity, high potency, stability, solubility, safety, and relatively inexpensive manufacturing process in comparison with other biologics. We outline here the structural studies and fundamental principles that define how antibodies interact with diverse targets. We also describe the antibody repertoires and affinity maturation mechanisms of humans, mice, and chickens, plus the use of novel single-domain antibodies in camelids and sharks. These species all utilize diverse evolutionary solutions to generate specific and high affinity antibodies and illustrate the plasticity of natural antibody repertoires. In addition, we discuss the multiple variations of man-made antibody repertoires designed and validated in the last two decades, which have served as tools to explore how the size, diversity, and composition of a repertoire impact the antibody discovery process.

A systematic comparison of free and bound antibodies reveals binding-related conformational changes

To study structural changes that occur in Abs upon Ag binding, we systematically compared free and bound structures of all 141 crystal structures of the 49 Abs that were solved in these two forms. We found that many structural changes occur far from the Ag binding site. Some of them may constitute a mechanism for the recently suggested allosteric effects in Abs. Within the binding site itself, CDR-H3 is the only element that shows significant binding-related conformational changes; however, this occurs in only one third of the Abs. Beyond the binding site, Ag binding is associated with changes in the relative orientation of the H and L chains in both the variable and constant domains. An even larger change occurs in the elbow angle between the variable and the constant domains, and it is significantly larger for binding of big Ags than for binding of small ones. The most consistent and substantial conformational changes occur in a loop in the H chain constant domain. This loop is implicated in the interaction between the H and L chains, is often intrinsically disordered, and is involved in complement binding. Hence, we suggest that it may have a role in Ab function. These findings provide structural insight into the recently proposed allosteric effects in Abs.

Computer-aided antibody design

Recent clinical trials using antibodies with low toxicity and high efficiency have raised expectations for the development of next-generation protein therapeutics. However, the process of obtaining therapeutic antibodies remains time consuming and empirical. This review summarizes recent progresses in the field of computer-aided antibody development mainly focusing on antibody modeling, which is divided essentially into two parts: (i) modeling the antigen-binding site, also called the complementarity determining regions (CDRs), and (ii) predicting the relative orientations of the variable heavy (V(H)) and light (V(L)) chains. Among the six CDR loops, the greatest challenge is predicting the conformation of CDR-H3, which is the most important in antigen recognition. Further computational methods could be used in drug development based on crystal structures or homology models, including antibody-antigen dockings and energy calculations with approximate potential functions. These methods should guide experimental studies to improve the affinities and physicochemical properties of antibodies. Finally, several successful examples of in silico structure-based antibody designs are reviewed. We also briefly review structure-based antigen or immunogen design, with application to rational vaccine development.

Cytokinergic IgE Action in Mast Cell Activation

Some 10 years ago it emerged that at sufficiently high concentrations certain monoclonal mouse IgEs exert previously unsuspected effects on mast cells. Thus they can both promote survival and induce activation of mast cells without the requirement for antigens. This was a wake up call that appears to have been missed (or dismissed) by the majority of immunologists. The structural attributes responsible for the potency of the so-called “highly cytokinergic” or HC IgEs have not yet been determined, but the events that ensue when such IgEs bind to the high-affinity receptor, FcεRI, on mast cells have been thoroughly studied, and are strikingly similar to those engendered by antigens when they form cross-linked complexes with the receptors. We review the evidence for the cytokinergic activity of IgE, and the structural features and known properties of immunoglobulins, and of IgE in particular, most likely to be implicated in the phenomenon. We suggest that IgEs with cytokinergic activity may be generated by local germinal center reactions in the target organs of allergy. We consider also the important implications that the existence of cytokinergic IgE may have for a fuller understanding of adaptive immunity and of the action of IgE in asthma and other diseases.

Germline VH/VL pairing in antibodies

Antibodies are key molecules of the adaptive immune response and are now a major class of biopharmaceuticals. Pairing of heavy and light chains is one of the ways of generating antibody diversity and, while little is known about mechanisms governing V(H)/V(L) pairing, previous studies have suggested that the germline source from which chains are paired is random. By selecting paired antibody protein sequences from human and mouse antibodies from the KabatMan database and mapping them onto their corresponding germline sequences, we find that pairing preferences do exist in the germline, but only for a small proportion of germline gene segments; others are much more promiscuous showing no preferences. The closest equivalent human and mouse gene families were identified and pairing preferences compared. This work may impact on the ability to generate more stable antibodies for use as biopharmaceuticals.

Evaluation of three different formats of a neutralizing single chain human antibody against toxin Cn2: neutralization capacity versus thermodynamic stability

The single-chain antibody fragment (scFv) 6009F, obtained by directed evolution, neutralizes the effects of the Cn2 toxin, which is the major toxic component of Centruroides noxius scorpion venom. In this work we compared the neutralization capacity and the thermodynamic stability of scFv 6009F with those of two other derived formats: Fab 6009F and diabody 6009F. Additionally, the affinity constants to Cn2 toxin of the three recombinant antibody fragments were determined by means of BIAcore. We found a correlation between the thermodynamic stability of these antibody fragments with their neutralization capacity. The order of thermodynamic stability determined was Fab≫scFv>diabody. The Fab and scFv were capable of neutralizing the toxic effects of Cn2 and whole venom but the diabody was unable to fully neutralize intoxication. In silico analysis of the diabody format indicates that the reduction of stability and neutralization capacity could be explained by a less cooperative interface between the heavy and the light variable domains.

Computer-assisted modeling of antibody variable domains

Antibody modeling is an interesting option to gain structure-function insights and to allow rational engineering of these molecules in the absence of experimental data. Among a diversity of algorithms, software packages, and specialized Web servers, the protocol described here presents the following main features: (1) nonautomatic modeling process guided by direct application of antibody modeling principles; (2) local generation of molecular models using free software which can be used in most common operational systems; and (3) the resulting model quality is comparable to models generated by Web servers which represent the current standard of antibody modeling. Briefly, hybrid models of heavy- and light-chain variable domains are separately built by grafting segments from homologous templates (framework regions and complementarity-determining regions). Next, hybrid models are mutated to comply with the target's sequence and associated by fitting into a template structure that closely matches the predicted packing angle for the target variable domains. After a few cycles of energy minimization the model can be submitted to CDR-H3 optimization or its quality can be directly assessed.

Predicting antibody complementarity determining region structures without classification

Antibodies are used extensively in medical and biological research. Their complementarity determining regions (CDRs) define the majority of their antigen binding functionality. CDR structures have been intensively studied and classified (canonical structures). Here we show that CDR structure prediction is no different from the standard loop structure prediction problem and predict them without classification. FREAD, a successful database loop prediction technique, is able to produce accurate predictions for all CDR loops (0.81, 0.42, 0.96, 0.98, 0.88 and 2.25 Å RMSD for CDR-L1 to CDR-H3). In order to overcome the relatively poor predictions of CDR-H3, we developed two variants of FREAD, one focused on sequence similarity (FREAD-S) and another which includes contact information (ConFREAD). Both of the methods improve accuracy for CDR-H3 to 1.34 Å and 1.23 Å respectively. The FREAD variants are also tested on homology models and compared to RosettaAntibody (CDR-H3 prediction on models: 1.98 and 2.62 Å for ConFREAD and RosettaAntibody respectively). CDRs are known to change their structural conformations upon binding the antigen. Traditional CDR classifications are based on sequence similarity and do not account for such environment changes. Using a set of antigen-free and antigen-bound structures, we compared our FREAD variants. ConFREAD which includes contact information successfully discriminates the bound and unbound CDR structures and achieves an accuracy of 1.35 Å for bound structures of CDR-H3.

Antigen recognition by antibody C836 through adjustment of V(L)/V(H) packing

C836 is a neutralizing monoclonal antibody to human interleukin IL-13 generated by mouse immunization. The crystal structure of the C836 Fab was determined at 2.5 Å resolution and compared with the IL-13-bound form determined previously. This comparison indicates an induced-fit mechanism of antigen recognition through rigid-body rotation of the V(L) and V(H) domains. The magnitude of this rearrangement is one of the largest observed for antibody-protein interactions.

Probing the stability-limiting regions of an antibody single-chain variable fragment: a molecular dynamics simulation study

Antibody single-chain variable fragments (scFvs) offer particular advantages over the full-size antibodies, including easy expression, efficient local concentration and fast body clearance. However, scFvs typically show low thermal stability that limits their biomedical and biotechnological applications. In this study, we examined the thermal stability of the human and murine vascular endothelial growth factor antibody scFv fragment by molecular dynamics simulations. A consistent observation was the dissociation of the light-chain (VL) and heavy-chain (VH) domains and loss of the native structures of both domains in the simulations at the elevated temperatures. The stability-limiting structural elements in the protein were revealed from the detailed analyses on the native contacts. We found that dissociation of the VL-VH domains was the first event leading to the unfolding of the native structure of the protein and the disruption of the VL-VH interface was largely due to the break of the interfacial hydrophobic and aromatic interactions while the hydrogen-bonding interaction between Gln38 in VL and Gln39 in VH remained. Within the β-barrel structure of the VL and VH domains, β-strands β6, β2 and β11 appeared to be the least stable. In addition, we found that the VH domain was more thermally resistant than the VL domain. Based on these findings, we discussed potential strategies to improve the stability of this therapeutically important scFv fragment.

The association of heavy and light chain variable domains in antibodies: implications for antigen specificity

The antigen-binding site of immunoglobulins is formed by six regions, three from the light and three from the heavy chain variable domains, which, on association of the two chains, form the conventional antigen-binding site of the antibody. The mode of interaction between the heavy and light chain variable domains affects the relative position of the antigen-binding loops and therefore has an effect on the overall conformation of the binding site. In this article, we analyze the structure of the interface between the heavy and light chain variable domains and show that there are essentially two different modes for their interaction that can be identified by the presence of key amino acids in specific positions of the antibody sequences. We also show that the different packing modes are related to the type of recognized antigen.

Crystal structure of human germline antibody 3-23/B3

The human antibody repertoire is dominated by a few combinations of germline sequences, with the genes 3-23 for the heavy chain and B3 for the light chain being among the most frequently used. Despite this fact, there was no experimental structure of the antibody composed of 3-23 and B3. The crystal structure of the Fab fragment of the synthetic antibody composed of the 3-23 and B3 germline sequences was determined to provide a template for antibody modeling. The antigen-binding loops were found in the canonical conformations. Comparison to the other structures where either 3-23 or B3 is paired with a different chain reveals a significant deviation in the orientation of the variable domains.