Structural,energetic, and functional analysis of a protein-protein interface at distinct stages of affinity maturation

Due to a paucity of studies that synthesize structural, energetic, and functional analyses of a series of protein complexes representing distinct stages in an affinity maturation pathway, the biophysical basis for the molecular evolution of protein-protein interactions is poorly understood. Here, we combine crystal structures and binding-free energies of a series of variant superantigen (SAG)-major histocompatibility complex (MHC) class II complexes exhibiting increasingly higher affinity to reveal that this affinity maturation pathway is controlled largely by two biophysical factors: shape complementarity and buried hydrophobic surface. These factors, however, do not contribute equivalently to the affinity maturation of the interface, as the former dominates the early steps of the maturation process while the latter is responsible for improved binding in later steps. Functional assays reveal how affinity maturation of the SAG-MHC interface corresponds to T cell activation by SAGs.     

Structural basis of affinity maturation and intramolecular cooperativity in a protein-protein interaction

Although protein-protein interactions are involved in nearly all cellular processes, general rules for describing affinity and selectivity in protein-protein complexes are lacking, primarily because correlations between changes in protein structure and binding energetics have not been well determined. Here, we establish the structural basis of affinity maturation for a protein-protein interaction system that we had previously characterized energetically. This model system exhibits a 1500-fold affinity increase. Also, its affinity maturation is restricted by negative intramolecular cooperativity. With three complex and six unliganded variant X-ray crystal structures, we provide molecular snapshots of protein interface remodeling events that span the breadth of the affinity maturation process and present a comprehensive structural view of affinity maturation. Correlating crystallographically observed structural changes with measured energetic changes reveals molecular bases for affinity maturation, intramolecular cooperativity, and context-dependent binding.     

Molecular basis of a million-fold affinity maturation process in a protein-protein interaction

Protein engineering is becoming increasingly important for pharmaceutical applications where controlling the specificity and affinity of engineered proteins is required to create targeted protein therapeutics. Affinity increases of several thousand-fold are now routine for a variety of protein engineering approaches, and the structural and energetic bases of affinity maturation have been investigated in a number of such cases. Previously, a 3-million-fold affinity maturation process was achieved in a protein-protein interaction composed of a variant T-cell receptor fragment and a bacterial superantigen. Here, we present the molecular basis of this affinity increase. Using X-ray crystallography, shotgun reversion/replacement scanning mutagenesis, and computational analysis, we describe, in molecular detail, a process by which extrainterfacial regions of a protein complex can be rationally manipulated to significantly improve protein engineering outcomes.     

Protein energetics in maturation of the early secretory pathway

The early secretory pathway (ESP) consisting of the endoplasmic reticulum (ER), pre-Golgi intermediates and the Golgi stack links protein synthesis to folding and vesicle trafficking to generate the membrane architecture of the eukaryotic cell. The fundamental principles that contribute to organization of the ESP remain largely unknown. We raise the possibility that assembly of the ESP is largely built on a foundation that is influenced by the kinetic and thermodynamic properties of the protein fold. Folding energetics may provide an adjustable platform for adaptor-dependent interactions with the transport machinery, suggesting the possibility that protein cargo energetics plays a central role in directing both trafficking patterns and global compartmental organization of the ESP. In this view, cargo energetics likely coordinates the composition and maturation of ER and Golgi compartments with the physiological state of the cell in different tissue and environmental settings.     

Dissecting the protein-protein interface between beta-lactamase inhibitory protein and class A beta-lactamases

beta-Lactamase inhibitory protein (BLIP) binds and inhibits a diverse collection of class A beta-lactamases at a wide range of affinities. Alanine-scanning mutagenesis was previously performed to identify the amino acid sequence requirements of BLIP for inhibiting TEM-1 beta-lactamase and SME-1 beta-lactamase. Two hotspots of binding energy, one from each domain of BLIP, were identified (Zhang, Z., and Palzkill, T. (2003) J. Biol. Chem. 278, 45706-45712). This study has been extended to examine the amino acid sequence requirements of BLIP for binding to the SHV-1 beta-lactamase, which is a poor binding substrate (Ki= 1.1 microm), and the Bacillus anthracis Bla1 enzyme (Ki= 2.5 nm). The two hotspots previously identified as important for binding TEM-1 and SME-1 beta-lactamase were also found to be important for binding Bla1. The hotspot from the second domain of BLIP, however, does not make substantial contributions to SHV-1 binding. This may explain why BLIP binds to SHV-1 beta-lactamase with much weaker affinity than to the other three enzymes. Three regions, including two loops that insert into the active pocket of TEM-1 beta-lactamase and the Glu-73-Lys-74 buried charge motif, exhibit strikingly different effects on the binding affinity of BLIP toward the various enzymes when mutated and, therefore, act as specificity determinants. Analysis of double mutants of BLIP that combine specificity-determining residues suggests that these residues contribute to the poor affinity between the second domain of BLIP and SHV-1 beta-lactamase.     

Exploring the capacity of minimalist protein interfaces: interface energetics and affinity maturation to picomolar KD of a single-domain antibody with a flat paratope

A major architectural class in engineered binding proteins ("antibody mimics") involves the presentation of recognition loops off a single-domain scaffold. This class of binding proteins, both natural and synthetic, has a strong tendency to bind a preformed cleft using a convex binding interface (paratope). To explore their capacity to produce high-affinity interfaces with diverse shape and topography, we examined the interface energetics and explored the affinity limit achievable with a flat paratope. We chose a minimalist paratope limited to two loops found in a natural camelid heavy-chain antibody (VHH) that binds to ribonuclease A. Ala scanning of the VHH revealed only three "hot spot" side chains and additional four residues important for supporting backbone-mediated interactions. The small number of critical residues suggested that this is not an optimized paratope. Using selection from synthetic combinatorial libraries, we enhanced its affinity by >100-fold, resulting in variants with Kd as low as 180 pM with no detectable loss of binding specificity. High-resolution crystal structures revealed that the mutations induced only subtle structural changes but extended the network of interactions. This resulted in an expanded hot spot region including four additional residues located at the periphery of the paratope with a concomitant loss of the so-called "O-ring" arrangement of energetically inert residues. These results suggest that this class of simple, single-domain scaffolds is capable of generating high-performance binding interfaces with diverse shape. More generally, they suggest that highly functional interfaces can be designed without closely mimicking natural interfaces.     

Dissecting the energetics of an antibody-antigen interface by alanine shaving and molecular grafting.

Alanine-scanning mutagenesis on human growth hormone (hGH) identified 5 primary determinants (Arg 8, Asn 12, Arg 16, Asp 112, and Asp 116) for binding to a monoclonal antibody (MAb 3) (Jin L, Fendly BM, Wells JA, 1992, J Mol Biol 226:851-865). To further analyze the energetic importance of residues surrounding these five, we mutated all neighboring residues to alanine in groups of 7-16 (a procedure we call alanine shaving). Even the most extremely mutated variant, with 16 alanine substitutions, caused less than a 10-fold reduction in binding affinity to MAb3. By comparison, mutating any 1 of the 5 primary determinants to alanine caused a 6- to > 500-fold reduction in affinity. Replacing any of the 4 charged residues (Arg 8, Arg 16, Asp 112, and Asp 116) with a homologous residue (i.e., Arg to Lys or Asp to Glu) caused nearly as large a reduction in affinity as the corresponding alanine replacement. It was possible to graft the 5 primary binding determinants onto a nonbinding homologue of hGH, human placental lactogen (hPL), which has 86% sequence identity to hGH. The grafted hPL mutant bound 10-fold less tightly than hGH to MAb3 but bound as well as hGH when 2 additional framework mutations were introduced. Attempts to recover binding affinity by grafting the MAb3 epitope onto more distantly related scaffolds having a similar 4-helix bundle motif, such as human prolactin (23% sequence identity) or granulocyte colony-stimulating factor, were unsuccessful.(ABSTRACT TRUNCATED AT 250 WORDS)     

Dissecting the energetics of the apoflavodoxin-FMN complex

Many flavoproteins are non-covalent complexes between FMN and an apoprotein. To understand better the stability of flavoproteins, we have studied the energetics of the complex between FMN and the apoflavodoxin from Anabaena PCC 7119 by a combination of site-directed mutagenesis, titration calorimetry, equilibrium binding constant determinations, and x-ray crystallography. Comparison of the strength of the wild type and mutant apoflavodoxin-FMN complexes and that of the complexes between wild type apoflavodoxin and shortened FMN analogues (riboflavin and lumiflavin) allows the dissection of the binding energy into contributions associated with the different parts of the FMN molecule. The estimated contribution of the phosphate is greatest, at 7 kcal mol(-1); that of the isoalloxazine is of around 5-6 kcal mol(-1) (mainly due to interaction with Trp-57 and Tyr-94 in the apoprotein) and the ribityl contributes least: around 1 kcal mol(-1). The stabilization of the complex is both enthalpic and entropic although the enthalpy contribution is dominant. Both the phosphate and the isoalloxazine significantly contribute to the enthalpy of binding. The ionic strength does not have a large effect on the stability of the FMN complex because, although it weakens the phosphate interactions, it strengthens the isoalloxazine-protein hydrophobic interactions. Phosphate up to 100 mM does not affect the strength of the riboflavin complex, which suggests the isoalloxazine and phosphate binding sites may be independent in terms of binding energy. Interestingly, we find crystallographic evidence of flexibility in one of the loops (57-62) involved in isoalloxazine binding.     

Dissecting cooperative calmodulin binding to CaM kinase II: a detailed stochastic model

Calmodulin (CaM) is a major Ca²⁺ binding protein involved in two opposing processes of synaptic plasticity of CA1 pyramidal neurons: long-term potentiation (LTP) and depression (LTD). The N- and C-terminal lobes of CaM bind to its target separately but cooperatively and introduce complex dynamics that cannot be well understood by experimental measurement. Using a detailed stochastic model constructed upon experimental data, we have studied the interaction between CaM and Ca²⁺-CaM-dependent protein kinase II (CaMKII), a key enzyme underlying LTP. The model suggests that the accelerated binding of one lobe of CaM to CaMKII, when the opposing lobe is already bound to CaMKII, is a critical determinant of the cooperative interaction between Ca²⁺, CaM, and CaMKII. The model indicates that the target-bound Ca²⁺ free N-lobe has an extended lifetime and may regulate the Ca²⁺ response of CaMKII during LTP induction. The model also reveals multiple kinetic pathways which have not been previously predicted for CaM-dissociation from CaMKII.     

Nucleic acid binding affinity of fd gene 5 protein in the cooperative binding mode

A sensitive ESR method which allows a direct quantitative determination of nucleic acid binding affinities of proteins under physiologically relevant conditions has been applied to the gene 5 protein of bacteriophage fd. This was achieved with two spin-labeled nucleic acids, (ldT, dT)n and (lA,A)n, which served as macro-molecular spin probes in ESR competition experiments. With the two different macromolecular spin probes, it was possible to determine the relative apparent affinity constants, Kapp, over a large affinity domain. In 20 mM Tris X HCl (pH 8.1), 1 mM sodium EDTA, 0.1 mM dithiothreitol, 10% (w/v) glycerol, 0.05% Triton, and 125 mM NaCl, the following affinity relationship was observed: K(dT)napp = 10(3) KfdDNAapp = 2 X 10(4) K(A)napp = 6.6 X 10(4) KrRNAapp = 1.5 X 10(5) KR17RNAapp. Increasing the [NaCl] from 125 to 200 mM caused considerably less tight binding of gene 5 protein to (lA,A)n, and a typical cooperative binding isotherm was observed, whereas at the lower [NaCl] used for the competition experiments, the binding was essentially stoichiometric. A computer fit of the experimental titration data at 200 mM NaCl gave an intrinsic binding constant, Kint, of 1300 M-1 and a cooperativity factor, omega, of 60 (Kint omega = Kapp) for (lA,A)n.     

Deep learning-based method for predicting and classifying the binding affinity of protein-protein complexes

Protein-protein interactions (PPIs) play a critical role in various biological processes. Accurately estimating the binding affinity of PPIs is essential for understanding the underlying molecular recognition mechanisms. In this study, we employed a deep learning approach to predict the binding affinity (ΔG) of protein-protein complexes. To this end, we compiled a dataset of 903 protein-protein complexes, each with its corresponding experimental binding affinity, which belong to six functional classes. We extracted 8 to 20 non-redundant features from the sequence information as well as the predicted three-dimensional structures using feature selection methods for each protein functional class. Our method showed an overall mean absolute error of 1.05 kcal/mol and a correlation of 0.79 between experimental and predicted ΔG values. Additionally, we evaluated our model for discriminating high and low affinity protein-protein complexes and it achieved an accuracy of 87% with an F1 score of 0.86 using 10-fold cross-validation on the selected features. Our approach presents an efficient tool for studying PPIs and provides crucial insights into the underlying mechanisms of the molecular recognition process. The web server can be freely accessed at https://web.iitm.ac.in/bioinfo2/DeepPPAPred/index.html     

Calorimetric determination of cooperative interactions in high affinity binding processes

It is demonstrated that isothermal titration calorimetry can be used to determine cooperative interaction energetics even for extremely tight binding processes in which the binding affinity constants are beyond the limits of experimental determination. The approach is based on the capability of calorimetry to measure the apparent binding enthalpy at any degree of ligand saturation. When calorimetric measurements are performed under conditions of total association at partial saturation, the dependence of the apparent binding enthalpy on the degree of saturation is a function only of the cooperative binding interactions. The method developed in this paper allows an independent estimation of cooperative energetic parameters without the need to simultaneously estimate or precisely know the value of the association constants. Since total ligand association at partial saturation is achieved only at macromolecular concentrations much larger than the dissociation constants, the method is especially suited for high and very high affinity processes. Biological associations in this category include fundamental cellular processes like cell surface receptor binding or protein-DNA interactions.     

Selection of phage antibodies by binding affinity. Mimicking affinity maturation.

We describe a process, based on display of antibodies on the surface of filamentous bacteriophage, for selecting antibodies either by their affinity for antigen or by their kinetics of dissociation (off-rate) from antigen. For affinity selection, phage are mixed with small amounts of soluble biotinylated antigen (less than 1 microgram) such that the antigen is in excess over phage but with the concentration of antigen lower than the dissociation constant (Kd) of the antibody. Those phage bound to antigen are then selected using streptavidin-coated paramagnetic beads. The process can distinguish between antibodies with closely related affinities. For off-rate selection, antibodies are preloaded with biotinylated antigen and diluted into excess unlabelled antigen for variable times prior to capture on streptavidin-coated paramagnetic beads. To mimic the affinity maturation process of the immune system, we introduced random mutations into the antibody genes in vitro using an error-prone polymerase, and used affinity selection to isolate mutants with improved affinity. Starting with a small library (40,000 clones) of mutants (average 1.7 base changes per VH gene) of the mouse antibody B1.8, and using several rounds of affinity selection, we isolated a mutant with a fourfold improved affinity to the hapten 4-hydroxy-5-iodo-3-nitrophenacetyl-(NIP)-caproic acid (mutant Kd = 9.4(+/- 0.3) nM compared with B1.8 Kd = 41.9(+/- 1.6) nm). The relative increase in affinity of the mutant is comparable to the increase seen in the anti-4-hydroxy-3-nitrophenylacetyl/NIP-caproic acid murine secondary immune response.     

Dissecting the binding energy epitope of a high-affinity variant of human growth hormone: cooperative and additive effects from combining mutations from independently selected phage display mutagenesis libraries

Phage display mutagenesis is a widely used approach to engineering novel protein properties and is especially powerful in probing structure-function relationships in molecular recognition processes. The relative contributions of additive and cooperative binding forces and the influence of conformational diversity in producing a novel protein-protein interface is investigated using as a model an ultra-high-affinity receptor binding variant of human growth hormone (hGHv) that has been previously affinity matured. The modular aspect of how the mutations were grouped in the phage display libraries and combined allowed for a systematic probing of the inherent functional cross-talk between the different secondary structure elements that make up the remodeled hGHv binding surface. We performed an alanine scanning analyses of 35 hGHv residues and determined the kinetics of each variant by surface plasmon resonance (SPR). This analysis showed that there is a significant difference between the additive and cooperative binding forces existing among the selected residues in each library module, and the binding advantage of these residues is maximized over the original wild-type residue when in the context of the other mutations in the library. The degree to which residues in a particular mutagenesis library display binding cooperativity characteristics is generally correlated with the conformational plasticity of the polypeptide chain. Additionally, these cooperativity effects change when the mutations from one library are combined with the mutations from one or several of the other separate libraries. This supports the idea that significant functional cross-talk exists between the combined library modules that can affect the binding energetics of individual residues over a large distance.     

Nucleic binding affinity of bacteriophage T4 gene 32 protein in the cooperative binding mode

This study reports on various parameters which affect the binding stoichiometry for complexes of bacteriophage T4 gene 32 protein (P32) and single stranded polynucleotides (determined by UV absorbance and fluorescence quenching) and presents results of a quantitative electron spin resonance assay to determine physiologically effective binding affinity differences of nucleic acid binding proteins. The assay employs macromolecular spin probes (spin-labeled nucleic acids) which are used to determine the fraction of saturation in competition experiments with unlabeled nucleic acids. It was found that the fraction of complexed spin-labeled polynucleotides can be directly monitored by ESR with a two-component analysis approach when ligands such as poly(L-lysine), gene 5 protein (P5) of filamentous bacteriophage fd, and gene 32 protein (P32) of bacteriophage T4 are used. The ESR data unequivocally show that: 1) the binding stoichiometry for poly(L-lysine), P5 and P32 is nucleotide/lysine, 4 nucleotides/P5 monomer, and 10 nucleotides/P32 monomer, respectively; and 2) under physiologically relevant buffer conditions the relative affinity of P32 in the cooperative binding mode for polythymidylic acid is about 4 times greater than for polydeoxyinosinic acid and about 12 times greater than for polyinosinic acid, and the relative affinity of P32 for polydeoxyinosinic acid is about 3 times greater than for polyinosinic acid.     

Deep geometric representations for modeling effects of mutations on protein-protein binding affinity

Modeling the impact of amino acid mutations on protein-protein interaction plays a crucial role in protein engineering and drug design. In this study, we develop GeoPPI, a novel structure-based deep-learning framework to predict the change of binding affinity upon mutations. Based on the three-dimensional structure of a protein, GeoPPI first learns a geometric representation that encodes topology features of the protein structure via a self-supervised learning scheme. These representations are then used as features for training gradient-boosting trees to predict the changes of protein-protein binding affinity upon mutations. We find that GeoPPI is able to learn meaningful features that characterize interactions between atoms in protein structures. In addition, through extensive experiments, we show that GeoPPI achieves new state-of-the-art performance in predicting the binding affinity changes upon both single- and multi-point mutations on six benchmark datasets. Moreover, we show that GeoPPI can accurately estimate the difference of binding affinities between a few recently identified SARS-CoV-2 antibodies and the receptor-binding domain (RBD) of the S protein. These results demonstrate the potential of GeoPPI as a powerful and useful computational tool in protein design and engineering. Our code and datasets are available at: https://github.com/Liuxg16/GeoPPI.     

Structure and energetics of protein-protein interactions: the role of conformational heterogeneity in OMTKY3 binding to serine proteases

Proteins with flexible binding surfaces can interact with numerous binding partners. However, this promiscuity is more difficult to understand in "rigid-body" proteins, whose binding results in little, or no, change in the position of backbone atoms. The binding of Kazal inhibitors to serine proteases is considered a classic case of rigid-body binding, although they bind to a wide range of proteases. We have studied the thermodynamics of binding of the Kazal serine protease inhibitor, turkey ovomucoid third domain (OMTKY3), to the serine protease subtilisin Carlsberg using isothermal titration calorimetry and have determined the crystal structure of the complex at very high resolution (1.1A). Comparison of the binding energetics and structure to other OMTKY3 interactions demonstrates that small changes in the position of side-chains can make significant contributions to the binding thermodynamics, including the enthalpy of binding. These effects emphasize that small, "rigid-body" proteins are still dynamic structures, and these dynamics make contributions to both the enthalpy and entropy of binding interactions.     

In vitro affinity maturation of a natural human antibody overcomes a barrier to in vivo affinity maturation

Antibodies isolated from human donors are increasingly being developed for anti-infective therapeutics. These antibodies undergo affinity maturation in vivo, minimizing the need for engineering of therapeutic leads for affinity. However, the affinities required for some therapeutic applications may be higher than the affinities of the leads obtained, requiring further affinity maturation in vitro. To improve the neutralization potency of natural human antibody MSL-109 targeting human cytomegalovirus (CMV), we affinity matured the antibody against the gH/gL glycoprotein complex. A phage display library where most of the six complementary-determining regions (CDRs) were allowed to vary in only one amino acid residue at a time was used to scan for mutations that improve binding affinity. A T55R mutation and multiple mutations in position 53 of the heavy chain were identified that, when present individually or in combination, resulted in higher apparent affinities to gH/gL and improved CMV neutralization potency of Fab fragments expressed in bacterial cells. Three of these mutations in position 53 introduced glycosylation sites in heavy chain CDR 2 (CDR H2) that impaired binding of antibodies expressed in mammalian cells. One high affinity (K_D < 10 pM) variant was identified that combined the D53N and T55R mutations while avoiding glycosylation of CDR H2. However, all the amino acid substitutions identified by phage display that improved binding affinity without introducing glycosylation sites required between two and four simultaneous nucleotide mutations to avoid glycosylation. These results indicate that the natural human antibody MSL-109 is close to a local affinity optimum. We show that affinity maturation by phage display can be used to identify and bypass barriers to in vivo affinity maturation of antibodies imposed by glycosylation and codon usage. These constraints may be relatively prevalent in human antibodies due to the codon usage and the amino acid sequence encoded by the natural human repertoire.     

In vitro affinity maturation of a natural human antibody overcomes a barrier to in vivo affinity maturation

Antibodies isolated from human donors are increasingly being developed for anti-infective therapeutics. These antibodies undergo affinity maturation in vivo, minimizing the need for engineering of therapeutic leads for affinity. However, the affinities required for some therapeutic applications may be higher than the affinities of the leads obtained, requiring further affinity maturation in vitro. To improve the neutralization potency of natural human antibody MSL-109 targeting human cytomegalovirus (CMV), we affinity matured the antibody against the gH/gL glycoprotein complex. A phage display library where most of the six complementary-determining regions (CDRs) were allowed to vary in only one amino acid residue at a time was used to scan for mutations that improve binding affinity. A T55R mutation and multiple mutations in position 53 of the heavy chain were identified that, when present individually or in combination, resulted in higher apparent affinities to gH/gL and improved CMV neutralization potency of Fab fragments expressed in bacterial cells. Three of these mutations in position 53 introduced glycosylation sites in heavy chain CDR 2 (CDR H2) that impaired binding of antibodies expressed in mammalian cells. One high affinity (K_D < 10 pM) variant was identified that combined the D53N and T55R mutations while avoiding glycosylation of CDR H2. However, all the amino acid substitutions identified by phage display that improved binding affinity without introducing glycosylation sites required between two and four simultaneous nucleotide mutations to avoid glycosylation. These results indicate that the natural human antibody MSL-109 is close to a local affinity optimum. We show that affinity maturation by phage display can be used to identify and bypass barriers to in vivo affinity maturation of antibodies imposed by glycosylation and codon usage. These constraints may be relatively prevalent in human antibodies due to the codon usage and the amino acid sequence encoded by the natural human repertoire.