Impact of protein and ligand impurities on ITC-derived protein–ligand thermodynamics

Background

The thermodynamic characterization of protein–ligand interactions by isothermal titration calorimetry (ITC) is a powerful tool in drug design, giving valuable insight into the interaction driving forces. ITC is thought to require protein and ligand solutions of high quality, meaning both the absence of contaminants as well as accurately determined concentrations.

Methods

Ligands synthesized to deviating purity and protein of different pureness were titrated by ITC. Data curation was attempted also considering information from analytical techniques to correct stoichiometry.

Results and conclusions

We used trypsin and tRNA-guanine transglycosylase (TGT), together with high affinity ligands to investigate the effect of errors in protein concentration as well as the impact of ligand impurities on the apparent thermodynamics. We found that errors in protein concentration did not change the thermodynamic properties obtained significantly. However, most ligand impurities led to pronounced changes in binding enthalpy. If protein binding of the respective impurity is not expected, the actual ligand concentration was corrected for and the thus revised data compared to thermodynamic properties obtained with the respective pure ligand. Even in these cases, we observed differences in binding enthalpy of about 4 kJ ⋅ mol^− 1, which is considered significant.

General significance

Our results indicate that ligand purity is the critical parameter to monitor if accurate thermodynamic data of a protein–ligand complex are to be recorded. Furthermore, artificially changing fitting parameters to obtain a sound interaction stoichiometry in the presence of uncharacterized ligand impurities may lead to thermodynamic parameters significantly deviating from the accurate thermodynamic signature.     

Variation of protein binding cavity volume and ligand volume in protein–ligand complexes

We have systematically analyzed the variation of protein binding cavity volume of 200 protein–ligand complexes belonging to eight protein families. Wide variation in protein binding cavity volume for the same protein is observed on binding different ligands. Analysis of individual protein families shows high correlation between atom–atom interactions in binding site and ligand volume. This study implies the significance of protein flexibility in docking small molecule inhibitors on the basis of protein binding cavity volume with respect to ligand volume.     

Evaluation of small ligand–protein interaction by ligation reaction with DNA-modified ligand

A method for the evaluation of interactions between protein and ligand using DNA-modified ligands, including signal enhancement of the DNA ligation reactions, is described. For proof of principle, a DNA probe modified by biotin was used. Two DNA probes were prepared with complementary sticky-ends. While one DNA probe was modified at the 5′-end of the sticky-end, the other was not modified. The probes could be ligated together by T4 DNA ligase along the strand without biotin modification. However, in the presence of streptavidin or anti-biotin Fab, the ligation reaction joining the two probes could not occur on either strand.     

Mass spectrometry-based approaches to protein–ligand interactions

One of the greatest current challenges in proteomics is to develop an understanding of cellular communication and regulation processes, most of which involve noncovalent interactions of proteins with various binding partners. Mass spectrometry plays an important role in all aspects of these research efforts. This article provides a survey of mass spectrometry-based approaches for exploring protein–ligand interactions. A wide array of techniques is available, and the choice of method depends on the specific problem at hand. For example, the high-throughput screening of compound libraries for binding to a specific receptor requires different approaches than structural studies on multiprotein complexes. This review is directed to readers wishing to obtain a concise yet comprehensive overview of existing experimental techniques. Specific emphasis is placed on emerging methods that have been developed within the last few years.     

Force spectroscopy studies on protein–ligand interactions: A single protein mechanics perspective

Protein–ligand interactions are ubiquitous and play important roles in almost every biological process. The direct elucidation of the thermodynamic, structural and functional consequences of protein–ligand interactions is thus of critical importance to decipher the mechanism underlying these biological processes. A toolbox containing a variety of powerful techniques has been developed to quantitatively study protein–ligand interactions in vitro as well as in living systems. The development of atomic force microscopy-based single molecule force spectroscopy techniques has expanded this toolbox and made it possible to directly probe the mechanical consequence of ligand binding on proteins. Many recent experiments have revealed how ligand binding affects the mechanical stability and mechanical unfolding dynamics of proteins, and provided mechanistic understanding on these effects. The enhancement effect of mechanical stability by ligand binding has been used to help tune the mechanical stability of proteins in a rational manner and develop novel functional binding assays for protein–ligand interactions. Single molecule force spectroscopy studies have started to shed new lights on the structural and functional consequence of ligand binding on proteins that bear force under their biological settings.     

Trimethylsilyl tag for probing protein–ligand interactions by NMR

Protein–ligand titrations can readily be monitored with a trimethylsilyl (TMS) tag. Owing to the intensity, narrow line shape and unique chemical shift of a TMS group, dissociation constants can be determined from straightforward 1D ¹H-NMR spectra not only in the fast but also in the slow exchange limit. The tag is easily attached to cysteine residues and a sensitive reporter of ligand binding also at sites where it does not interfere with ligand binding or catalytic efficiency of the target protein. Its utility is demonstrated for the Zika virus NS2B–NS3 protease and the human prolyl isomerase FK506 binding protein.     

Computation of temperature-dependent dissociation rates of metastable protein–ligand complexes

Molecular simulations are often used to analyse the stability of protein–ligand complexes. The stability can be characterised by exit rates or using the exit time approach, i.e. by computing the expected holding time of the complex before its dissociation. However determining exit rates by straightforward molecular dynamics methods can be challenging for stochastic processes in which the exit event occurs very rarely. Finding a low variance procedure for collecting rare event statistics is still an open problem. In this work we discuss a novel method for computing exit rates which uses results of Robust Perron Cluster Analysis (PCCA+). This clustering method gives the possibility to define a fuzzy set by a membership function, which provides additional information of the kind ‘the process is being about to leave the set’. Thus, the derived approach is not based on the exit event occurrence and, therefore, is also applicable in case of rare events. The novel method can be used to analyse the temperature effect of protein–ligand systems through the differences in exit rates, and, thus, open up new drug design strategies and therapeutic applications.     

Intrinsic affinity of protein – ligand binding by differential scanning calorimetry

Differential scanning calorimetry (DSC) determines the enthalpy change upon protein unfolding and the melting temperature of the protein. Performing DSC of a protein in the presence of increasing concentrations of specifically-binding ligand yields a series of curves that can be fit to obtain the protein–ligand dissociation constant as done in the fluorescence-based thermal shift assay (FTSA, ThermoFluor, DSF). The enthalpy of unfolding, as directly determined by DSC, helps improving the precision of the fit. If the ligand binding is linked to protonation reactions, the intrinsic binding constant can be determined by performing the affinity determination at a series of pH values. Here, the intrinsic, pH-independent, affinity of acetazolamide binding to carbonic anhydrase (CA) II was determined. A series of high-affinity ligands binding to CAIX, an anticancer drug target, and CAII showed recognition and selectivity for the anticancer isozyme. Performing the DSC experiment in buffers of highly different enthalpies of protonation enabled to observe the ligand unbinding-linked protonation reactions and estimate the intrinsic enthalpy of binding. The heat capacity of combined unfolding and unbinding was determined by varying the ligand concentrations. Taken together, these parameters provided a detailed thermodynamic picture of the linked ligand binding and protein unfolding process.     

An ant colony optimization approach to flexible protein–ligand docking

The prediction of the complex structure of a small ligand with a protein, the so-called protein–ligand docking problem, is a central part of the rational drug design process. For this purpose, we introduce the docking algorithm PLANTS (Protein–Ligand ANT System), which is based on ant colony optimization, one of the most successful swarm intelligence techniques. We study the effectiveness of PLANTS for several parameter settings and present a direct comparison of PLANTS’s performance to a state-of-the-art program called GOLD, which is based on a genetic algorithm and frequently used in the pharmaceutical industry for this task. Last but not least, we also show that PLANTS can make effective use of protein flexibility giving example results on cross-docking and virtual screening experiments for protein kinase A. This article is based on a paper that won the best paper award at ANTS 2006, the 5th International Workshop on Ant Colony Optimization and Swarm Intelligence held in Brussels, Belgium, 2006. This article includes new types of experiments and also the possibility of considering flexibility of protein side-chains.     

Mapping protein–ligand interactions using whole genome phage display libraries

The function of many genes cannot be deduced from sequence similarity, and biochemical methods are usually required. Whole genome sequences can be thought of as not only a set of genes but also collections of functional domains. These domains can be studied by affinity methods whereby identification of the ligand can provide information on biochemical function. To take advantage of this method, one must express all functional domains in a form suitable for affinity studies. Phage display technology provides a means for accomplishing this. The pJuFo phage display system, based on the interaction between the leucine zippers Jun and Fos, has been modified and used to create a genomic phage display library from Escherichia coli MG1655. The system has been tested by using the library to map the dominant binding epitopes for an anti-RecA protein polyclonal antibody sera. This methodology provides a general biochemical approach to functional analysis of protein–ligand interactions on a genome-wide basis.     

Development of a novel PPARγ ligand screening system using pinpoint fluorescence-probed protein

The activation of peroxisome-proliferator-activated receptor-γ (PPARγ), which plays a central role in adipocyte differentiation, depends on ligand-dependent co-activator recruitment. In this study, we developed a novel method of PPARγ ligand screening by measuring the increase in fluorescent polarization accompanied by the interaction of a fluorescent co-activator and PPARγ. Sterol receptor co-activator-1 (SRC-1), a major PPARγ co-activator, was probed by fluorescent TAMRA by the Amber codon fluorescence probe method. Polarization was increased by adding PPARγ ligands to a solution containing labeled SRC-1 (designated TAMRA-SRC-S) and PPARγ. The disassociation constants (Kd) of the PPARγ synthesized ligands, pioglitazone (221 nM), troglitazone (83.0 nM), and 15-deoxy-Δ12,14-prostaglandin J(2) (15d-ΔPGJ(2)) (156 nM), were determined by this method. Farnesol (2.89 μM) and bixin (21.1 μM), which we have reported to be PPARγ ligands, increased the fluorescent polarization. Their Kd values were in agreement with the ED(50) values obtained in the luciferase assay. The results indicate that the method is valuable for screening natural PPARγ ligands.     

Identification of individual protein–ligand NOEs in the limit of intermediate exchange

Interactions of proteins with small molecules or other macromolecules play key roles in many biological processes and in drug action, and NMR is an excellent tool for their structural characterization. Frequently, however, line broadening due to intermediate exchange completely eliminates the signals needed for measuring specific intermolecular NOEs. This limits the use of NMR for detailed structural studies in such kinetic situations. Here we show that an optimally chosen excess of ligand over protein can reduce the extent of line broadening for both the ligand and the protein. This makes observation of ligand resonances possible but reduces the size of the measurable NOEs due to the residual line broadening and the non-stoichiometric concentrations. Because the solubility of small molecule drug leads are often limited to high micromolar concentrations, protein concentrations are restricted to even lower values in the low micromolar range. At these non-stoichiometric concentrations and in the presence of significant residual line broadening, conventional NOESY experiments very often are not sensitive enough to observe intermolecular NOEs since the signals inverted by the NOESY preparation pulse sequence relax prior to significant NOE build up. Thus, we employ methods related to driven NOE spectroscopy to investigate protein–ligand interactions in the intermediate exchange regime. In this approach, individual protein resonances are selectively irradiated for up to five seconds to build up measurable NOEs at the ligand resonances. To enable saturation of individual protein resonances we prepare deuterated protein samples selectively protonated at a few sites so that the 1D ¹H spectrum of the protein is resolved well enough to permit irradiation of individual protein signals, which do not overlap with the ligand spectrum. This approach is suitable for measuring a sufficiently large number of protein–ligand NOEs that allow calculation of initial complex structures, suitable for structure-based optimization of primary drug leads obtained from high-throughput screening. The method was applied to measure individual intermolecular NOEs between the anti-apoptotic protein Bcl-xL at 25 μM and a “first generation” small-molecule ligand, for which the spectrum is entirely broadened at stoichiometric concentrations. This approach is general and can also be used to characterize protein–protein or protein–nucleic-acid complexes.     

Accuracy and precision of protein–ligand interaction kinetics determined from chemical shift titrations

NMR-monitored chemical shift titrations for the study of weak protein?Cligand interactions represent a rich source of information regarding thermodynamic parameters such as dissociation constants (K _D ) in the micro- to millimolar range, populations for the free and ligand-bound states, and the kinetics of interconversion between states, which are typically within the fast exchange regime on the NMR timescale. We recently developed two chemical shift titration methods wherein co-variation of the total protein and ligand concentrations gives increased precision for the K _D value of a 1:1 protein?Cligand interaction (Markin and Spyracopoulos in J Biomol NMR 53: 125?C138, 2012). In this study, we demonstrate that classical line shape analysis applied to a single set of ¹H?C¹⁵N 2D HSQC NMR spectra acquired using precise protein?Cligand chemical shift titration methods we developed, produces accurate and precise kinetic parameters such as the off-rate (k _off ). For experimentally determined kinetics in the fast exchange regime on the NMR timescale, k _off ?~?3,000?s^?1 in this work, the accuracy of classical line shape analysis was determined to be better than 5?% by conducting quantum mechanical NMR simulations of the chemical shift titration methods with the magnetic resonance toolkit GAMMA. Using Monte Carlo simulations, the experimental precision for k _off from line shape analysis of NMR spectra was determined to be 13?%, in agreement with the theoretical precision of 12?% from line shape analysis of the GAMMA simulations in the presence of noise and protein concentration errors. In addition, GAMMA simulations were employed to demonstrate that line shape analysis has the potential to provide reasonably accurate and precise k _off values over a wide range, from 100 to 15,000?s^?1. The validity of line shape analysis for k _off values approaching intermediate exchange (~100?s^?1), may be facilitated by more accurate K _D measurements from NMR-monitored chemical shift titrations, for which the dependence of K _D on the chemical shift difference (????) between free and bound states is extrapolated to ?????=?0. The demonstrated accuracy and precision for k _off will be valuable for the interpretation of biological kinetics in weakly interacting protein?Cprotein networks, where a small change in the magnitude of the underlying kinetics of a given pathway may lead to large changes in the associated downstream signaling cascade.     

The environment of amide groups in protein–ligand complexes: H-bonds and beyond

A comprehensive structural analysis of interactions involving amide NH and C=O groups in protein-ligand complexes has been performed based on 3,275 published crystal structures (resolution < or =2.5 A). Most of the amide C=O and NH groups at the protein-ligand interface are highly buried within the binding site and involved in H-bonds with corresponding counter-groups. Small percentages of C=O and NH groups are solvated or embedded in hydrophobic environments. In particular, C=O groups show a higher propensity to be solvated or embedded in a hydrophobic environment than NH groups do. A small percentage of carbonyl groups is involved in weak hydrogen bonds with CH. Cases of dipolar interactions, involving carbonyl oxygen and electrophilic carbon atoms, such as amide, amidinium, guanidium groups, are also identified. A higher percentage of NH are in contact with aromatic carbons, interacting either through hydrogen bonds (preferably with the NH group pointing towards a ring carbon atom) or through stacking between amide plane and ring plane. Comprehensive studies such as the present one are thought to be important for future improvements in the molecular design area, in particular for the development of new scoring functions. [Figure: see text].     

Detection of protein–ligand interactions by NMR using reductive methylation of lysine residues

We show that reductive methylation of proteins can be used for highly sensitive NMR identification of conformational changes induced by metal- and small molecule binding, as well as protein-protein interactions. Reductive methylation of proteins introduces two (13)C-methyl groups on each lysine in the protein of interest. This method works well even when the lysines are not actively involved in the interaction, due to changes in the microenvironments of lysine residues. Most lysine residues are located on the protein exterior, and the exposed (13)C-methyl groups may exhibit rapid localized motions. These motions could be faster than the tumbling rate of the molecule as a whole. Thus, this technique has great potential in the study of large molecular weight systems which are currently beyond the scope of conventional NMR methods.     

NLDB: a database for 3D protein–ligand interactions in enzymatic reactions

NLDB (Natural Ligand DataBase; URL: http://nldb.hgc.jp) is a database of automatically collected and predicted 3D protein–ligand interactions for the enzymatic reactions of metabolic pathways registered in KEGG. Structural information about these reactions is important for studying the molecular functions of enzymes, however a large number of the 3D interactions are still unknown. Therefore, in order to complement such missing information, we predicted protein–ligand complex structures, and constructed a database of the 3D interactions in reactions. NLDB provides three different types of data resources; the natural complexes are experimentally determined protein–ligand complex structures in PDB, the analog complexes are predicted based on known protein structures in a complex with a similar ligand, and the ab initio complexes are predicted by docking simulations. In addition, NLDB shows the known polymorphisms found in human genome on protein structures. The database has a flexible search function based on various types of keywords, and an enrichment analysis function based on a set of KEGG compound IDs. NLDB will be a valuable resource for experimental biologists studying protein–ligand interactions in specific reactions, and for theoretical researchers wishing to undertake more precise simulations of interactions.