Analytical methods for obtaining binding parameters of drug-protein interactions: A review.

Analytica Chimica Acta 1219 (2022) 340012 Available online 1 June 2022

0003-2670/© 2022 Elsevier B.V. All rights reserved.

Analytical methods for obtaining binding parameters of drug–protein interactions: A review

Lijuan Wang a, Wenmei Zhang a, Yunlong Shao a, Dongtang Zhang a,*, Guangsheng Guo a,b,

Xiayan Wang a,** a Center of Excellence for Environmental Safety and Biological Effects, Beijing Key Laboratory for Green Catalysis and Separation, Department of Chemistry and Biology,

Beijing University of Technology, Beijing, 100124, PR China b Minzu University of China, Beijing, 100081, PR China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Methods for the investigation of drug- protein interactions are reviewed.

• The process of obtaining binding pa­ rameters for each method is discussed.

• The methods are classified according to binding parameters.

A R T I C L E I N F O Keywords:

Drug–protein interactions Binding parameters Quantitative analysis

Drug–protein binding Method selection A B S T R A C T

The study of drug–protein interactions can reveal the corresponding binding mechanisms, providing valuable information for the early phase drug development and development of new drugs. This article reviews the methods used for obtaining the binding parameters of drug–protein systems. The methods include equilibrium dialysis, high-performance affinity chromatography, capillary electrophoresis, spectroscopy, calorimetry, competition and displacement, mass spectrometry, fluorescence resonance energy transfer, and thermal stability shift analysis. Relevant parameters include the association constant, number of binding sites, thermodynamic properties, binding force types, binding site types, binding distances, changes in protein conformation, and changes in protein stability. In addition, the review also summarizes the principles, advantages, and limitations of each method in detail. The comparison of parameter information can not only guide method selection but also provide valuable reference information for in-depth exploration of drug–protein interaction mechanisms.

1. Introduction As the essential building blocks and most significant components in organisms, proteins play a crucial role in the structural and dynamic functions of the body. Proteins interact with a variety of substances such as other proteins, DNA, RNA, and drugs [1]. The study of these in­ teractions has improved our molecular-level understanding of biological

* Corresponding author.

** Corresponding author.

E-mail addresses: zhangdongtang@bjut.edu.cn (D. Zhang), xiayanwang@bjut.edu.cn (X. Wang).

Contents lists available at ScienceDirect Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca https://doi.org/10.1016/j.aca.2022.340012

Received 20 December 2021; Received in revised form 25 May 2022; Accepted 27 May 2022

Analytica Chimica Acta 1219 (2022) 340012 2 processes. Drugs play an important role in regulating body functions, curing various diseases, and maintaining health. The study of drug–­ protein interactions supports the discovery of protein binding sites for targeted therapies. The information on the binding sites enables the development of new drugs. A drug enters the blood circulatory system through oral administration or injection, and then reversibly binds with plasma proteins [such as human serum albumin (HSA) and alpha-1-acid glycoprotein] to form a bound complex [2,3]. The latter generally serves as a reservoir for the free drug. Part of the free drug in the body is excreted through metabolism, while the other part passes through a variety of biological membranes to reach the target sites and exert its therapeutic effect [3,4]. Therefore, the study of the binding of drugs to target proteins contributes to the assessment of the pharmacodynamics and toxicology of drugs. Similarly, the study of the binding of drugs to plasma proteins plays a key role in their pharmacokinetics (i.e., ab­ sorption, distribution, metabolism, and elimination) and pharmacodynamics.

Previous reviews of drug–protein interactions focused on intro­ ducing the relevant methods and principles [2,4–10]. This article de­ scribes each method from the new perspective of obtaining information on the binding parameters. In particular, we systematically review the process of obtaining parameter information with each method. The parameters considered here include the association constant and num­ ber of binding sites [11–15], thermodynamic properties and binding force types [12,13,16,17], binding site types [11–13,18,19], binding distances [11,12,16], changes in protein conformation [11–13,20], and changes in protein stability [21–25]. We also discuss the principles and inherent advantages and limitations of each method.

2. Association constant and number of binding sites

Drug-protein interactions are mostly reversible and rapid equilib­ rium processes governed by the law of mass action [3]. In the simplest case, assuming that there is only one reversible binding site between the protein and the drug, the binding between the protein and the drug can be described by Equation (1) [26]: [

Df ] + [ Pf ] ⇄ kon koff [DP] (1) where [Df], [Pf], [DP] are the concentrations of free drug, free protein and drug-protein complex, respectively, and kon and koff are the asso­ ciation rate constant and dissociation rate constant. At equilibrium, the association rate constant equals the dissociation rate constant.

Therefore, the association constant (Ka) can be defined as shown in

Equation (2) [26]:

Ka = kon koff = [DP] [ Df ] ⋅ [ Pf ] (2) The association constant is a crucial quantitative parameter in drug–protein interactions. The dissociation constant (Kd) is also a frequently considered quantity, which is the reciprocal of Ka. When the drug–protein association constant Ka > 108 L mol−1, the binding of the two is a strong interaction; when the drug–protein association constant

Ka = 106–108 L mol−1, the binding of the two is moderate interaction; when the drug–protein association constant Ka < 106 L mol−1, the binding of the two is a weak interaction [27]. The number of binding sites can help determine the number of drug binding sites on the protein, which is very important for the accurate measurement of the association constant. Various methods can provide the association constant and number of binding sites for drug–protein systems. Common approaches include equilibrium dialysis (ED), high-performance affinity chroma­ tography (HPAC), capillary electrophoresis (CE), spectroscopy, and calorimetry. Table 1 summarizes the formulas used for calculating the two parameters, along with the advantages and limitations of each method.

2.1. Equilibrium dialysis ED is often used as a reference method in drug–protein interaction studies [5,39–43]. The ED device is composed of two compartments separated by a semipermeable membrane, as shown in Fig. 1. One compartment contains a mixture of drug and protein, and the other contains only the buffer. The semipermeable membrane is equivalent to a molecular sieve, allowing only free drugs to pass through. After reaching equilibrium, the solutions in the two compartments change.

One compartment contains a solution of protein and drug–protein complexes, while the other contains a solution of free drug. The con­ centration of the latter is determined by high-performance liquid chro­ matography (HPLC) or mass spectrometry (MS) [39]. The interactions of plasma protein with pazopanib [39], paclitaxel [40], and 99mTc dieth­ ylenetriaminepentaacetic acid [41] have been investigated by the ED method. The association constant and the number of binding sites of drug–protein systems can be determined by the Scatchard equation [Equation (3)] [28,29]: r =

∑ m i=1 niKai [ Df ] 1 + Kai [ Df ] (3) Abbreviation

AC affinity chromatography ACE affinity capillary electrophoresis

△H binding enthalpy BSA bovine serum albumin CD circular dichroism

CE capillary electrophoresis CS cromolyn sodium CETSA cellular thermal shift assay

CZE capillary zone electrophoresis DSC differential scanning calorimetry

ΔS entropy change ED equilibrium dialysis EPN epinastine hydrochloride

ESI electrospray ionization FA frontal analysis FACCE frontal analysis continuous capillary electrophoresis

FRET fluorescence resonance energy transfer △G Gibbs free energy change

△CP heat capacity change HD Hummel–Dreyer HDX hydrogen–deuterium exchange

HPAC high-performance affinity chromatography HSA human serum albumin

ITC isothermal titration calorimetry MRE mean residual ellipticity

MS mass spectrometry MINS mononaphthalimide spermidine

NMR nuclear magnetic resonance UV–Vis ultraviolet–visible

TPP thermal proteome profiling VP vacancy peak ZE zonal elution

MNPs magnetic iron oxide nanoparticles L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 3 Table 1

Calculation of association constant and number of binding sites, along with advantages and disadvantages of different methods used to assess drug–protein interactions.

ED HPAC CE Spectroscopic Calorimetry ZE FA ACE CZE, FA, HD,

VP UV–Vis Absorption Fluorescence quenching CD ITC

DSC How to obtain Ka or n [Df] tR, [I] [A], tA μeff, [Pt] [Df]

A F, [Q] Δ[θ], [Df] Q, [A], △H, [T] T0, TM, △H0, △Cp, [L]TM

Calculation formula r = [Dt] −[Df] [Pt] = ∑ m i=1 niKai[Df]

1 + Kai[Df] 1 k = tR −tM tM = Ka,ILVM[I] Ka,ALmL +

VM Ka,ALmL mL, app = (VA − VM) × [A] 1 mL,app = 1 KamL[A] +

1 mL μeff = μf + μcKa[Pt] 1 + Ka[Pt] r = [Dt] −[Df] [Pt]

= ∑ m i=1 niKai[Df] 1 + Kai[Df] 1 A −A0 = 1 Amax −A0

+ 1 (Amax −A0)Ka[Dt] log [F0 −F F ] = logKa + nlog[Q]

Δ[θ] [Pt] [ Df ] = Ka ( ε −Δ[θ] [Pt] ) Q = ΔH⋅V⋅[A]tot⋅ n⋅KaT

1 + Ka[T] Ka(TM) = exp {−ΔH0 nR ( 1 TM − 1 T0 ) + ΔCP nR ( ln TM

T0 + T0 TM −1 )} −1 [L]TM Advantages Simple and convenient operation; Low cost; True equilibrium in solution; 96- well plates are available

The affinity column can be reused many times; Kinetic data can be obtained; Pure samples are not required

Fast and efficient separation; Kinetic data can be obtained;

Easy to automate; Pure samples are not required Simple and convenient operation; Protein structure information can be obtained

Thermodynamic information can be directly obtained

Drawbacks Long equilibrium time (>6 h); Nonspecific adsorption;

Donnan effect; Df requires additional quantitative methods

Long preparation time for affinity column (>16 h);

Change the storage solution frequently Protein adsorption capillary wall; Generation of joule heat

Poor reproducibility; Multiple equilibria are not applicable; Low sensitivity; High sample consumption;

Highly pure reagents High sample consumption; Highly pure reagents

Ref [28,29] [30,31] [28,32–34] [13,35] [36–38] ED: equilibrium dialysis; HPAC: high-performance affinity chromatography; CE: capillary electrophoresis; ZE: zonal elution; FA: frontal analysis; ACE: affinity capillary electrophoresis; CZE: capillary zone electrophoresis;

HD: Hummel–Dreyer; VP: vacancy peak; CD: circular dichroism; ITC: isothermal titration calorimetry; DSC: differential scanning calorimetry; r: number of moles of bound drug per mole of total protein; [Dt]: total drug concentration; [Df]: free drug concentration; [Pt]: total protein concentration; n: number of binding sites; Ka: association constant; k: retention factor; tR: retention time; tM: dead time; Ka,AL: association constant of interaction between analyte and affinity ligand; Ka,IL: association constant of interaction between competing agent and affinity ligand; mL: total number of moles of all binding sites; VM: void volume; [I]: competing agent concentration; mL,app: apparent number of moles; VA: breakthrough volume; [A]: analyte concentration; μeff: effective mobility of the analyte; μf: mobility of the free analyte; μc: the mobility of the complex; A: protein absorbance in presence of drug; A0: protein absorbance in absence of drug; Amax: protein saturation absorbance; F: protein fluorescence intensity in presence of drug; F0: protein fluorescence intensity in absence of drug; [Q]: quenchant concentration; Δ[θ]: ellipticity change: ε: proportionality constant; Q: total heat change; △H: binding enthalpy; V: active volume; [A]tot: total analyte concentration; [T]: free titrant concentration; TM: protein transition midpoint temperature in presence of drug; T0: protein transition midpoint temperature in absence of drug; ΔH◦: enthalpy change at T0; ΔCp: heat capacity change at T0; R: gas constant; [L]TM: free ligand concentration at TM.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 4 where r is the number of moles of bound drug per mole of total protein, m is the total number of different types of binding sites, ni is the number of binding sites of the i-th class and the binding sites do not interfere with each other, Kai is the association constant for the i-th binding site class, and [Df] is the concentration of the free drug. r can be obtained indirectly through [Df], using Equation (4) [28,29]: r = [Db] [Pt] = [Dt] − [

Df ] [Pt] (4) where [Db] is the concentration of the bound drug, [Dt] is the total concentration of the drug, and [Pt] is the total concentration of protein.

A nonlinear fit of the r vs. [Df] curve is then performed to obtain Ka and n. Rabbani-Chadegani et al. [43] used spectroscopy and ED to determine the binding affinity of berberine to nucleosomes and histone H1. The association constant between berberine and nucleosomes, calculated by the Scatchard equation, was Ka = 5.57 × 103 L mol−1, whereas that between berberine and histone H1 was Ka = 3.61 × 103 L mol−1. These results indicate that berberine has a high affinity for both nucleosomes and histone H1.

ED is easy to operate and inexpensive [44]. It can maintain the interacting substances in the real balance after the dialysis [3]. How­ ever, it takes a long time for the system to reach equilibrium (typically more than 6 h [45]), and additional quantitative methods are needed to determine the concentration of the free drug [43,46,47]. In addition, the nonspecific adsorption and Donnan effects on the semipermeable membrane can further affect the accuracy of the measurements [4,48].

The Donnan effect results in small molecules being unevenly distributed on both sides of the semipermeable membrane, due to the presence of macromolecular ions during osmotic equilibrium [49,50]. A 96-well plate ED device was developed to save time and improve the analysis throughput [47,51].

2.2. High-performance affinity chromatography HPAC was first proposed by Ohlson et al., in 1978 [52]. This method combines affinity chromatography (AC) and HPLC-grade support ma­ terials (such as silica) [53,54]. The HPAC stationary phase is generally composed of three parts: matrix, spacer arm, and affinity ligand [55–57]; the latter interacts with the target protein based on specific and

Fig. 1. Schematic illustration of the equilibrium dialysis device. The figure was created with this review.

Fig. 2. HPAC modes of zonal elution and frontal analysis for studying drug–protein interactions [54]. Reprinted with permission from Elsevier.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 5 reversible interactions [58].

Zonal elution (ZE) and frontal analysis (FA) are common HPAC modes for investigating drug–protein interactions, as shown in Fig. 2 [54,59]. In both modes, the drug is injected into a chromatographic column containing the protein stationary phase, under specific mobile phase conditions. The main difference between the two modes is the amount of sample injected. ZE involves the same operating procedure as most chromatographic analyses. A small-plug drug is injected into the chromatographic column, and the quantitative analysis is mainly based on the retention time [31,54,59]. The advantages of ZE include fast analysis and low sample consumption. FA is generally used to inject a large-plug drug into the chromatographic column, and the following quantitative analysis is mainly based on the peak intensity [31,53,59,

60]. Although the FA analysis requires a long time and a large number of samples, the corresponding measurements are relatively accurate [31,

54,60].

2.2.1. Zonal elution The relevant variables in the ZE experiment include mobile phase parameters (pH, ionic strength, solvent polarity, temperature, and competing agent), target protein type, and affinity ligand type in the column [31,57]. Quantitative information on the binding between the target and the immobilized binding agent can be obtained by monitoring changes in target retention time [31,61]. Therefore, the drug–protein association constant can be estimated from the retention time or retention factor of the drug injected into the affinity column. The retention factor can be calculated from the retention time using Equa­ tion (5) [31]: k = tR −tM tM (5) where k is the retention factor, tR is the retention time of the drug in the stationary phase, and tM is the dead time of the column. When the drug and protein have relatively fast binding/dissociation kinetics, under linear elution conditions (the apparent value of k is not affected by the injection volume or flow rate) the retention factor and association constant are connected by Equation (6) [30]: k = (Ka1n1 + Ka2n2 + ⋯+ Kannn)mL

VM (6) where Ka1–Kan are the association constants of the different classes of binding sites, n1–nn are the fractions of the number of binding sites of a given class relative to the total number of binding sites, mL is the total number of moles of all active binding sites, and VM is the column void volume. Equation (6) shows that the retention factor of the drug is directly related to the drug–protein association constant and the amount of active protein present in the column. When the drug interacts inde­ pendently with the different binding sites of the target protein, the total equilibrium constant nKa is proportional to k, where nKa = Ka1n1 + Ka2n2

+ ⋯ + Kannn. If there is only one binding site, then the multisite Equa­ tion (6) can be simplified to Equation (7) [30,31]: k = KamL

VM (7) where Ka is the association constant when only one binding site exists.

The Hage group [62] prepared a HPAC column containing HSA by physical entrapment, and compared the binding of site-specific probes and sulfonylurea drugs with HSA and glycosylated HSA using the ZE method. In their study, the association constants for multisite binding (sulfonylurea drugs) and single-site binding (site-specific probes) were calculated by Equations (6) and (7), respectively [62]. The association constants calculated by this method were consistent with those obtained in other studies [63]. As the original activity of the protein in the HPAC column prepared by the entrapment method is essentially retained, the obtained binding parameters are more accurate. In the case of HPAC columns with proteins immobilized by covalent bonds, the accuracy of the binding parameters needs further improvement, owing to the decreased activity of the immobilized proteins. Competitive experi­ ments can help solve this problem.

Dunn and Chaiken first applied the idea of competition and displacement to study enzyme–inhibitor interactions in 1974 [64]. The competition and displacement method involves the continuous passage of a mobile phase with a known concentration of competing agent into the chromatographic column until saturation, followed by the injection of a small-plug target sample into the chromatographic column [31,53,

65,66]. Changing the concentration of the competing agent in the mo­ bile phase results in different retention times of the target. If the competing agent and the target directly compete for the same binding site in the protein, then the target retention time and concentration of the competing agent are linked by Equation (8) [30,31]:

1 k = Ka,ILVM[I] Ka,ALmL + VM Ka,ALmL (8) where Ka,AL is the association constant of the interaction between ana­ lyte (A) and affinity ligand (L), Ka,IL is the association constant of the interaction between competing agent (I) and L, and [I] is the concen­ tration of the competing agent in the mobile phase. The association constant Ka,IL can thus be obtained from Equation (8) by plotting 1/k vs. [I], and the corresponding change in slope can also be used to determine the type of competition in the interaction [54,65]. A linear plot with a positive slope indicates that the analyte and the competing agent directly compete for the same binding site of the protein, as in the case of the interaction of digitoxin as analyte and glimepiride as a competitive agent in an HSA column [67]. A slope close to zero indicates no competitive relationship between the two components, as found for the interaction of L-tryptophan (analyte) and glimepiride (competitive agent) in an HSA column [67]. A nonlinear response with a negative slope indicates the occurrence of a positive allosteric effect or multisite binding; this is the case of the interaction between R-warfarin (analyte) and glimepiride (competitive agent) in an HSA column [67]. Finally, a nonlinear response with a positive slope denotes a negative allosteric effect or multisite binding, as in the interaction of tamoxifen (analyte) and glimepiride (competitive agent) in an HSA column [67].

2.2.2. Frontal analysis In FA, a drug of known concentration is continuously passed into a chromatographic column containing the affinity ligand [53,57,60]. As the drug binds to the affinity ligand, the chromatographic column gradually reaches saturation. At this time, the amount of drug flowing out of the column continues to increase and finally forms a breakthrough curve [53,59]. From 1992 to 1994, HPAC in FA mode was used to study the interaction between proteins and various analytes [68,69]. Initially, these studies focused on stereoselective binding such as that of R- or

S-warfarin with HSA and D- or L-tryptophan with HSA [68,69]. This method has become a routine tool for characterizing drug–protein in­ teractions [53,57,60,70,71]. In the presence of a rapid bind­ ing/dissociation kinetics between the drug and protein and of only one binding site, the association constant and total number of moles of binding sites can be obtained from the breakthrough time and concen­ tration of the drug, as shown in Equations (9) and (10) [30,31]: mL, app = (VA −VM) × [A] (9)

1 mL,app = 1 KamL[A] + 1 mL (10) where mL,app is the apparent number of moles of analyte required to reach the average value of the breakthrough curve at a given concen­ tration of analyte [A]. In Equation (9), VA is the breakthrough volume of the analyte, which can be obtained from the breakthrough time and flow rate of the analyte. The 1/mL,app vs. 1/[A] curve can be fitted by Equation

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 6 (10) to obtain a straight line with slope = 1/KamL and intercept = 1/mL.

The ratio of the intercept to the slope then gives the Ka value. The calculation for the multisite binding model is similar to that of the single-site model. Several studies have reported the multisite binding interaction of a variety of drugs with normal and glycosylated HSA [67,

71,72].

The ZA and FA modes of HPAC share several advantages in the study of drug–protein interactions. The HPAC column can be reused many times: for example, a column containing HSA immobilized on silica particles was used for 500–1000 injections [73]. HPAC consumes a small amount of protein and the measurements have good repeatability. In addition, HPAC can be combined with a variety of detection systems, with UV and MS detectors being the most common [74–76]. These two detection technologies are typically used for interaction studies in label-free experiments. However, HPAC also suffers from some limita­ tions. The preparation and condition optimization of the affinity column can be very time-consuming, with even up to 21–82 h in sol-gel chem­ istry [77,78] or 16–36 h in polymer reaction [79,80] for the preparation of affinity monolithic columns. At the same time, it is necessary to maintain the activity of the protein in the column as much as possible.

Phosphate buffered saline at pH = 7.4 is often used as a physiological solution for proteins to remain the activity. If the solution is not changed frequently, the inactivated protein will directly affect the correct acquisition of binding constants. The binding of the protein to the spacer arm in the preparation of the affinity column should keep the native conformation of the protein as much as possible. HPAC studies of drug–protein interactions are mostly based on known binding sites. This is mainly because the protein needs to avoid the active binding area when immobilized in the column.

2.3. Capillary electrophoresis CE separation is based on the change in the mobility of the analyte in an electric field [81–83]. The mobility is mainly determined by the inherent physical properties (size, shape, and charge) of the analyte and the chemical additives in the electrolyte solution [33,81,84]. CE was first used to study drug–protein interactions in 1992 [85,86]. Owing to advantages such as fast analysis and high separation efficiency, this technique has become a powerful tool for investigating drug–protein interactions [2–4,81].

Five main CE modes are used in drug–protein interaction studies: capillary zone electrophoresis (CZE) [33,87–89], affinity capillary electrophoresis (ACE) [32,33,81,82,89–91], FA [33,82,89,92–94],

Hummel–Dreyer (HD) [33,82,89,95], and vacancy peak (VP) [33,82,

89]. Galievsky et al. [89] made a detailed summary of the various modes of CE, and the five main CE modes are shown in Fig. 3. These modes mainly differ in terms of applied injection conditions and running buffer [33,96]. The methods used for obtaining the binding parameters can be broadly divided into two categories: those based on mobility changes to obtain association constants (such as ACE) [81,82], and those based on peak intensities or peak areas to obtain information on the association constants and the number of binding sites (such as CZE, FA, HD, and VP) [2,3,33].

2.3.1. Affinity capillary electrophoresis ACE is often used to study drug–protein interactions [32,33,81,82,

90,91]. The basic ACE operation introduces an additive with different concentrations as a running buffer and another component as a sample [81]. The current example is the addition of protein in the buffer as an additive. When only one binding site is present, the association constant can be calculated from the change in the peak mobility of the drug [33,

81]. This mode is more suitable for the analysis of weakly interacting systems [87]. However, partial-filling ACE (pf-ACE) also can evaluate stronger interactions, which is a mode of the ACE approach [90,97].

Both ACE configurations require binders with fast equilibrium. The ACE analysis must meet two prerequisites: the free drug should have different mobility from the complex [32], and the protein concentration after equilibration should be close to that added in the buffer; in other words, the concentration of protein added in the buffer should be much higher than that of the injected drug [32].

In 1992, Wren and Rowe [34] proposed a mathematical model for the ACE separation of enantiomers. The model was supported by the experimental results using the propranolol enantiomer with β-cyclo­ dextrin, where β-cyclodextrin acted as a chiral selector. A complex approximately as large as the selector would be formed based on the interaction between β-cyclodextrin and propranolol enantiomers. Since

Fig. 3. Schematic and signal diagrams of the five main CE modes. L is detectable a ligand; T is a target; EM is the equilibrium mixture of L and T [89]. Reprinted with permission from ACS Publications.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 7 the complexes carry a different charge from the enantiomers, the effective mobility of the enantiomers is changed, resulting in the sepa­ ration of the enantiomers. Furthermore, the chiral separation of pro­ pranolol enantiomers in buffer systems containing organic reagents (methanol or acetonitrile) was also proved to be suitable for the math­ ematical model [98]. The mathematical model is suitable for 1:1 stoi­ chiometry. Although not all complex formation equilibria follow this stoichiometric ratio, 1:1 is the preferred stoichiometric ratio for many complexes, especially analyte–β-cyclodextrin complexes [99–101]. If this condition is satisfied, the effective mobilities of enantiomers are shown in Equation (11) [98]. μeff = μf + μcKa[Pt]

1 + Ka[Pt] (11) where μeff is the effective mobility of the enantiomer, Ka is the associa­ tion constant of the enantiomer to the chiral selector, μf is the mobility of the free enantiomer, μc is the mobility of the enantiomer-selector com­ plex, and [Pt] is the concentration of the chiral selector. The concen­ tration of the chiral selector is much higher than the concentration of the analyte, so the concentration of the selector is actually the same as the concentration of the enantiomer-selector complex regardless of the presence or absence of the analyte [34,99,102].

Applying this mathematical model to drug–protein interactions, proteins resemble chiral selectors and drugs resemble enantiomers.

When the concentration of protein and the free drug is known, the drug–protein association constant can be obtained by fitting Equation (11). However, the model is limited to the interaction between one analyte and one selector and does not include acid-base balance [34]. In

1993, Vigh et al. [103] extended the model by using pH as a separation parameter and found that chiral separation is not only affected by the type and concentration of the selector, but also by the pH in the buffer system. Specifically, pH affects the charge of the analyte, the ionization degree of the complex, and the electroosmotic flow, so the pH also needs to be chosen carefully. Dubský et al. introduced a multiple complexation equilibrium theory [104] and developed the all-in-one open-source software CEval (http://echmet.natur.cuni.cz/download) for evaluating

ACE measurement data [105]. The CEval program processes raw detector signal/time data into effective mobility, including automatic detection and evaluation of peak parameters, apparent mobility calcu­ lated by the Haarhoff-Van der Linde function, and viscosity correction.

In addition, the program can determine association constants by nonlinear regression and provides statistical modules to evaluate the resulting binding data.

Farcas¸ et al. [90] reported a robust and selective pf-ACE method for screening small fragments. The accuracy of the method was validated by determining the complex dissociation constants of three known thrombin inhibitors (benzamidine, p-aminobenzamidine, and nafamo­ stat). Compared with traditional spectrophotometric activity assays, the pf-ACE method has higher discriminative power for weak binding [97], which proves the application value in the field of fragment-based drug discovery. In pf-ACE, only a portion of the capillary is pre-filled with protein-containing buffer, known as the target plug. The rest of the capillary is filled with neat background electrolyte [81,90,102]. When a voltage is applied, the analyte migrates through the target plug and the neat background electrolyte to reach the detector, as shown in Fig. 4 [90]. Since the capillary was partially filled with protein, the observed analyte migration time was the sum of the migration time with the target plug and the migration time with the neat background electrolyte, as shown in Equation (12) [90,97]. To observe shifts in analyte mobility, the mobility of the complex needs to be significantly different from the free analyte.

Mt = MT＿Pr + MT＿free (12) where MT is the observed migration time of the analyte, MT_Pr is the migration time of the analyte in contact with the protein-containing plug, and MT_free is the migration time of the analyte in contact with the neat background electrolyte.

To determine the association constant, the analyte was kept at a constant concentration while the concentration of protein added to the background electrolyte was varied. Taking into account partial filling, the equations described by Tanaka [106] and Weber [107] were adop­ ted, and the association constant was determined by nonlinear regres­ sion using JMP® 12.1.0 software (SAS), as shown in Equation (13).

Fig. 4. Principle of the pf-ACE [97]. The migration time of the probe ligand is monitored in three conditions: neat background electrolyte (A), background electrolyte containing the target injected in 2/3 of the effective length of the capillary (B), and background electrolyte with target and interacting fragment (C). Reprinted with permission from Elsevier.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 8 μep＿Pr = ( μc −μfree

) Ka[P] 1 + Ka[P] (13) where μep_Pr is the electrophoretic mobility of the analyte in the presence of protein, μc is the mobility of the complex, μfree is the mobility of the analyte in the neat background electrolyte, Ka is the association con­ stant, and [P] is the target plug protein concentration. μep_Pr can be ob­ tained by Equation (14) [90,106,107]: μep＿Pr = L⋅1Pr

VMT＿Pr (14) where L is the total capillary length, lPr is the length of the protein plug,

V is the applied voltage, and MT_Pr is the migration time of the analyte in contact with the plug containing protein. A nonlinear fitting of a μep-pr vs. [P] plot with Equation (13) can be used to obtain Ka.

2.3.2. Other modes of capillary electrophoresis CZE uses a background electrolyte as the running buffer solution and a small-plug mixed solution of drug and protein injected into the capillary [33,87,88,108]. This approach is most suitable for the analysis of systems involving strong drug–protein interactions [3,87], i.e., sys­ tems in which the complex does not dissociate within the electrophoretic separation time. When mixed solutions containing different concentra­ tions of the drug are injected into the capillary, the peak intensity cor­ responding to the free drug changes; this change can be used to estimate the concentration of the free drug. CZE relies on the quantitative determination of the free drug concentration to obtain the drug–protein association constant and the number of binding sites [33], using the

Scatchard equation [Equation (3)] [28]. Despite the small sample con­ sumption of CZE, the instability of the peak intensities results in a large error in the measured value of the free drug concentration. The FA mode was developed to obtain more accurate measurements. FA and CZE operate roughly the same except for the injection amount. In FA, a large-plug mixed solution of drug and protein is injected into the capillary [93,94]. To study the drug–protein interaction, FA requires that the protein and the complex have approximately the same mobility, which should also be completely different from that of the drug [33,92].

Frontal analysis continuous capillary electrophoresis (FACCE) is an improved FA mode [94], in which the sample solution is continuously injected into the capillary. This mode helps maintain the equilibrium between drug and protein, and the obtained binding parameters are also more accurate. However, the disadvantage of this mode is that it con­ sumes a large sample volume. Malburet et al. [94] studied the anti­ gen–adjuvant interaction in vaccines using FACCE. In their study, information on the binding between adjuvant and antigen was obtained by determining the concentration of the free antigen.

The HD and VP modes involve the same sample preparation as FA. In

HD mode, the ligand is dissolved in a buffer as the running buffer, and the equilibrium mixture is injected into the capillary as the sample [33,

82,95]. VP uses a small-plug pure buffer as a sample and injects it into a capillary filled with an equilibrated drug–protein mixed solution [33,

82]. The FA, HD, and VP modes are all suitable for the analysis of sys­ tems involving weak interactions, and the binding parameters are calculated by the Scatchard equation based on the peak height or peak area [3,28]. Michalcova et al. [109] used three CE methods (ACE, FA, and HD) to study the interaction between bovine serum albumin (BSA) and salicylic acid. By comparing the association constants, they found that the CE–FA method achieved good reproducibility and provided a good nonlinear fitting of the obtained curve [109].

CE has advantages such as fast analysis, high separation efficiency, and easy automation [2,4,84], but also some shortcomings. For example, proteins easily adsorb to the bare capillary wall, which may cause a decrease in peak area or peak tailing [110,111]. In addition, CE is usually combined with UV detectors [108,112], whose low sensitivity limits the study of trace samples. In addition, when CE is combined with a fluorescence detector, the protein generally needs fluorescent labeling.

During the labeling process, on the one hand, the native conformation of the protein should be maintained, and on the other hand, it should be ensured that the binding site of the fluorescent label and the drug does not compete with the protein.

2.4. Spectroscopy methods Spectroscopic methods are based on the changes in electronic and spectral energy levels induced by the interaction between drugs and proteins [3–5]. These methods can provide a more comprehensive characterization of the drug–protein binding mechanism. Fig. 5 shows spectroscopy applied to the interaction of trifluoperazine with BSA [113]. The spectroscopic techniques used for studying drug–protein interaction mainly include UV–Vis absorption, fluorescence, and circu­ lar dichroism (CD) [11,13,114]. Among them, UV–Vis absorption, fluorescence spectroscopy and CD allow calculating the drug–protein association constant from changes in spectral characteristics.

2.4.1. UV–Vis absorption spectroscopy The basis of drug–protein interaction studies by UV–Vis absorption spectroscopy is that protein molecules contain ultraviolet chromophores that can absorb specific wavelengths in the ultraviolet region [115].

Generally, the 170–230 and 240–300 nm absorption windows provide information on the main protein chains and aromatic amino acid resi­ dues, respectively [116]. After the drug binds to the protein, a change in the 170–230 nm region of the spectrum indicates that the peptide chain has changed [116]. A blue or red shift in the peaks at 240–300 nm in­ dicates that the polarity of the aromatic amino acid residues has become higher or lower, respectively [116]. This can be used to determine the effect of the drug on the secondary structure of the protein, to a certain extent. The association constant can be obtained by substituting the absorbance of the UV–Vis absorption spectrum before and after the drug–protein interaction into the Benesi-Hildebrand equation (shown in

Equation (15)) [13,117]. The Benesi-Hildebrand equation generally requires that the initial concentration of protein is much larger than the initial concentration of the drug. For the 1:2 interaction system, whether it is a weak or strong interaction system, the equation curve may show two situations of linearity or nonlinearity, which significantly interferes with the accurate determination of the association constant [118]. Due to this fact, the equation is more suitable for a 1:1 interaction system.

1 A −A0 = 1 Amax −A0 + 1 (Amax −A0)Ka[Dt] (15) where A0 and A are the absorbances of the protein in the absence and presence of the drug, Amax is the saturated absorbance of the protein, and [Dt] is the concentration of the drug. A linear fit of the (1/(A – A0) vs. 1/ [Dt]) curve using Equation (15) can then provide the association con­ stant. By analyzing the corresponding UV–Vis absorption spectrum,

Yasmeen et al. [13] found that the interaction between cromolyn so­ dium (CS) and BSA was strong, with a Ka value as high as 6.9 × 104 L mol−1.

Job’s method of continuous variation can also be used to determine the number of drug–protein binding sites [113,119,120]. During the experiment, the total concentration of drug and protein was kept con­ stant, and the mole fraction of drug or protein in the mixture was changed, generally increasing from 0.0 to 1.0. The change value of the fluorescence intensity (ΔF) of the mixture is plotted against the mole fraction, and the number of binding sites between the drug and the protein can be obtained according to the mole fraction value corre­ sponding to the maximum ΔF in the figure [113,119,120]. Raghav et al. [113] determined the number of binding sites of trifluoperazine and BSA by Job’s method of continuous mutation, as shown in Fig. 6. An inter­ section plot was observed at a mole fraction value [TFP/(TFP + BSA)] of

0.68, indicating that the number of TFP binding sites on BSA is 2.0.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 9 2.4.2. Fluorescence quenching

Fluorescence spectroscopy has become the most widely used spec­ troscopic technique to study drug–protein interactions, owing to ad­ vantages such as good accuracy, rapidity, and wide applicability [13,

121,122]. In this method, fluorescence quenching can provide infor­ mation on the drug–protein association constant and the number of binding sites; in particular, the drug is used to reduce the fluorescence intensity of proteins containing fluorescent residues [122]. Zhang et al. [35] proposed the following equation to calculate the association con­ stant and the number of binding sites [Equation (16)]. The equation is only valid for static quenching. Usually, fluorescence quenching data are first analyzed by the Stern-Volmer equation to check the static quenching process [123,124]. log [F0 −F

F ] = logKa + nlog[Q] (16) where F0 and F are the fluorescence intensities of the protein in the absence and presence of a quencher, respectively, [Q] is the concen­ tration of the quencher, and n is the number of binding sites. Ka and n can thus be obtained by a linear fit of the log [(F0 −F)/F] vs. log [Q] plot.

2.4.3. Circular dichroism CD is often used to estimate protein secondary structure in solution [13,125–128]. Therefore, CD can be used to determine whether there is

Fig. 5. Spectroscopy for studying trifluoperazine–BSA interactions [113]. Reprinted with permission from Elsevier.

Fig. 6. The number of binding sites of trifluoperazine and BSA by Job’s method of continuous mutation [113]. Reprinted with permission from Elsevier.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 10 a change in protein conformation upon drug–protein interaction, which is described in detail in the “Spectroscopic Methods” of “Protein conformation changes” in this review (see 6.1). Since changes in CD spectra are proportional to the number of complexes formed after drug–protein interactions, CD can also be used to estimate association constants for drug–protein interactions [129].

If the binding of the drug to the protein is relatively weak, under the condition that the dissociation constant of the complex is greater than

100 times the protein concentration, it can be assumed that the con­ centration of the free drug is approximately equal to the total concen­ tration of the drug added to the protein. In this case, the drug–protein association constant can be determined by using the transformed form of the Scatchard equation, which involves a change in ellipticity, as shown in Equation (17) [28,129].

Δ[θ] [Pt] [ Df ] = Ka ( ε −Δ[θ] [Pt] ) (17) where Δ[θ] is the ellipticity change, [Pt] is the initial total concentration of protein, [Df] is the concentration of a free drug, Ka is the association constant, and ε is the proportionality constant. Ka is obtained by linearly fitting the Δ[θ]/[Df] vs. Δ[θ] curves. Ka is the negative number of the slope of the curve. The Scatchard equation is suitable for the binding of drugs and proteins with one or more equivalent binding sites. However, estimating association constants from spectral data is much more diffi­ cult when there are multiple binding sites. The apparent association constant can usually be estimated by the Hill equation [shown in

Equation (18)] [87,129,130]. In experiments, proteins need to be titrated with drugs.

Δ[θ] = Δ[θ]maxKa h[Dt]h 1 + Ka h[Dt]h + C (18) where Δ[θ] is the ellipticity change, Δ[θ]max is the maximum ellipticity change, Ka is the association constant, h is the apparent cooperativity constant of the interaction, [Dt] is the initial total concentration of the drug, and C is a constant for baseline offset correction. When h = 1, the binding is non-cooperative. Ka is obtained by nonlinear fitting of the

Δ[θ] vs. [Dt] curves.

If the binding of the drug to the protein is relatively strong and the dissociation constant is closer to that of the protein concentration, it cannot be assumed that the concentration of the free drug is the same as the initial total concentration of the drug. The concentration of the free drug needs to be adjusted according to the concentration of the drug added to the protein solution. Equation (19) can be used to calculate association constants for drugs with high affinity to proteins [129,131].

Δ[θ] = Δ[θ]max ⋅ ( 1 + Ka[Dt] + Ka[Pt] 2Ka[Pt] −1 + Ka[Dt] + Ka[Pt]

2Ka[Pt]2 −[Dt] [Pt] 1 2 (19) where Δ[θ] is the ellipticity change, Δ[θ]max is the maximum ellipticity change, Ka is the association constant, [Dt] is the initial total concen­ tration of the drug, and [Pt] is the initial total concentration of protein.

Ka was obtained by fitting the data to the Levenburg Marquardt algo­ rithm in the commercial program SigmaPlot [129,131].

In addition to providing the association constant and number of binding sites, spectroscopic methods can also reveal the changes in the protein structure caused by the drug. This approach is fast and simple to operate and is generally suitable for systems where drugs and proteins have high affinity [3–5,122]. However, its widespread use is limited by its large sample consumption and low sensitivity of the UV–Vis ab­ sorption spectrum. Fluorescent labeling of proteins is sometimes required when fluorescent techniques are used to measure drug–protein binding.

2.5. Calorimetry Calorimetry techniques can directly obtain thermodynamic infor­ mation from the heat exchanges associated with physical, chemical, and biological processes, as well as indirect insight into the binding in­ teractions [36,132]. Isothermal titration calorimetry (ITC) and differ­ ential scanning calorimetry (DSC) are commonly used to study drug–protein interactions [132–134].

2.5.1. Isothermal titration calorimetry ITC can measure the heat released and absorbed in drug–protein binding [134–137]. The ITC instruments are based on a power compensation design [36]. The device is composed of a sample cell filled with protein and a reference cell filled with buffer, as shown in Fig. 7 [16]. During the experiment, an equal volume of the drug is added to the sample cell containing the protein. Because the reaction process is accompanied by the release or absorption of heat, the temperatures of the sample and reference cells are unbalanced. This imbalance can be compensated by adjusting the feedback power applied to the sample cell heater [132,138]. When the drug–protein binding is an exothermic process, the power of the sample cell is reduced, whereas the power is increased when an endothermic reaction occurs [138].

ITC records the variation in power required to maintain the same temperature over time during each titration [132]. The plot of these changes consists of a series of peaks, whose individual areas represent the heat generated by the drug–protein interaction during each titration [36,138]. When the binding sites of the interaction are gradually satu­ rated, the peak area gradually decreases. Only the heat generated by the dilution of the titrant allows the peak shape to stabilize [36]. The power change vs. time data measured by ITC are then converted into a binding isotherm [36,138]. The association constant and the number of binding sites can be obtained by fitting the binding isotherm to a specific binding model, such as that shown in Equation (20) [36]:

Q = ΔH⋅V⋅[A]tot⋅n⋅KaT 1 + Ka[T] (20) where Q is the total heat change in the interaction, △H is the binding enthalpy, V is the active volume, [A]tot is the total concentration of the analyte, and [T] is the concentration of the free titrant. Zia et al. [16] used ITC to study the interaction between dutasteride and serum α2M antiprotease. The changes in enthalpy, entropy, and free energy ob­ tained in the experiment indicated that the binding reaction was spon­ taneous, and these thermodynamic properties were consistent with fluorescence quenching and molecular docking results. R`afols et al. [139] studied the binding of anti-inflammatory drugs and serum pro­ teins by combining ITC with FA/CE.

2.5.2. Differential scanning calorimetry ITC cannot be used to study binding reactions between species with very high affinity. DSC can overcome the shortcomings of ITC because it can be used for a wide range of affinity studies [3]. The experimental device used in this method is very similar to the ITC, but the program settings are different, as shown in Fig. 8 [140]. In a DSC experiment, the protein solution in the sample cell is heated at a controlled rate [133].

When the protein changes from the natural folded to the thermally unfolded state, the differential power vs. temperature diagram changes significantly [133]. The position of the transition midpoint in the dif­ ferential diagram corresponds to 50% of the protein in the native conformation and the other 50% in the denatured conformation [133,

141]. When placing a pure protein solution and a mixture of drug and protein in a sample cell, the temperature of the transition midpoint of the mixture is higher than that of the pure protein solution [38]. The unfolding enthalpy and heat capacity changes of the protein can be obtained from the differential diagram of the pure protein solution [38].

Therefore, DSC indirectly evaluates drug–protein interactions through

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 11 changes in protein conformation [37,142,143].

For a two-state reversible transition, the association constant and the number of binding sites can be derived from the transition midpoint temperature, non-folding enthalpy change, and non-folding heat ca­ pacity change, as shown in Equation (21) [37,38].

Ka(TM) = exp { −ΔH0 nR ( 1 TM − 1 T0 ) + ΔCP nR ( ln TM

T0 + T0 TM −1 )} −1 [L]TM (21) where Ka(TM) is the association constant at the temperature TM, T0 and

TM are the transition midpoint temperatures of the protein solution in the absence and presence of the drug, respectively, ΔH◦and ΔCp are the enthalpy and heat capacity changes of the transition at T0, respectively,

R is the gas constant (8.314 J K−1 mol−1), and [L]TM is the concentration of unbound ligand at TM. [L]TM can be calculated by Equations (22) and (23) [38]: [L]Tm = [Ltot] −n[Ptot]

2 [Ltot] ≥n[Ptot] (22) [L]Tm = n[Ptot] 2 [Ltot] ≤n[Ptot] (23) where [Ptot] and [Ltot] are the total concentrations of protein and ligand in the solution, respectively. [Ltot] ≥n[Ptot] indicates that the binding

Fig. 7. Schematic illustration of isothermal titration calorimetry approach. (left) Illustration of isothermal titration calorimeter device; (top right) plot of raw titration data; (bottom right) binding isotherm for titration [16]. Reprinted with permission from Elsevier.

Fig. 8. Schematic illustration of differential scanning calorimetry approach [140]. Reprinted with permission from Elsevier.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 12 site of the ligand on the natural protein is fully saturated, whereas [Ltot]

≤n[Ptot] means that no free ligand is present until the temperature of protein deconstruction is reached. Berrio Escobar et al. [144] used DSC to study the binding mechanism of uridine derivatives and phospholipid biomembranes.

Calorimetry does not require fixation, modification, or labeling, and can directly provide the thermodynamic properties [132,133].

Currently, ITC is more popular in the study of drug–protein interactions in calorimetry [13,145,146], because the thermodynamic parameters directly obtained by ITC can help the rational design of drugs. In contrast, there is little literature on DSC to study drug–protein in­ teractions [147,148], but DSC has a certain potential in disease screening and monitoring. In addition, DSC can also be used to estimate drug–protein interactions between components with very strong affinity (association constant up to 1015 L mol−1 [3]). However, the results of a calorimetry test mainly depend on the instrument. Because heat is a common signal, each part of the measurement process may influence the overall thermal effect. Any inaccuracy in the thermal effect test leads to errors in the evaluation of the binding parameters. In addition, the method also suffers from limitations such as large sample consumption, long testing time (ITC takes at least 2.5 h to run a complete titration experiment and often even longer [3]), low throughput, and high sample purity requirements [3,5].

3. Thermodynamic properties and binding force type

The determination of thermodynamic properties is crucial for the analysis of drug–protein interactions. On the one hand, they allow judging whether the interaction can proceed spontaneously; on the other hand, they can support the further determination of the type of binding force involved. The latter directly determines the strength of the binding affinity; drug–protein binding mainly relies on noncovalent forces [149]. Generally, thermodynamic properties can be obtained through direct measurements and calculations.

3.1. Direct measurements Calorimetry is the most direct technique to obtain thermodynamic information [132–134]. Drug–protein interactions create a thermody­ namic system composed of proteins, ligands, and solvents. The binding of drugs and proteins is controlled by the interaction and the exchange of energy between these components [138].

The binding enthalpy △H reflects the energy exchange between all system components [138]. Its value is the net result of the formation and breaking of single bonds caused by many individual molecular in­ teractions. When the system gains energy by increasing the bond strength, it releases heat, and the △H value becomes negative. This exothermic process [150] leads to the formation of noncovalent bond interactions. When the system gains energy by breaking chemical bonds, it absorbs heat from the environment, resulting in a positive enthalpy change. This endothermic reaction does not result in the formation of noncovalent bond interactions [150]. ITC can directly provide the △H value. DSC employs the measured (experimental) and estimated (ex­ pected) heat to obtain △H, as shown in Equation (24) [38,151]:

−nfΔH = ΔH(TM) − [ ΔH0 + ΔCP(TM −T0) ] (24) where f is the fraction of occupied sites, TM and T0 are the Kelvin tem­ peratures defined in Equation (21), △H(TM) is the enthalpy change at

TM, and ΔCp is the heat capacity change at T0. The first and second terms on the right side of the equation are the experimental heat measured when the compound is formed, and the estimated (expected) heat, respectively.

The Gibbs free energy change (△G) represents the ability of a thermodynamic system to perform maximum or reversible work under constant temperature and pressure. This parameter determines whether a reaction can proceed spontaneously [152]. Under thermodynamic control, a negative △G value indicates that the binding reaction be­ tween the drug and protein proceeds spontaneously until thermody­ namic equilibrium is reached. A more negative ΔG denotes that the binding reaction between the drug and the protein proceeds more easily; therefore, ΔG determines the probability of the drug–protein interaction resulting in the formation of a complex. In the standard state, △G can be obtained from the association constant, as shown in Equation (25) [36,

132,153]:

ΔG = −RT ln Ka (25) where T is the Kelvin temperature.

The entropy change (ΔS) is a measure of the disorder in the distri­ bution of atoms and molecules in a system [138]. When ΔS increases, the disorder increases, which causes ΔG to decrease, favoring the binding between drugs and proteins. ΔS can be calculated by Equation (26) [132,

153]:

ΔG = ΔH −TΔS (26) The heat capacity change (△CP) reflects the heat change with the temperature under constant pressure [36,138]. It measures the ability of a solution to absorb heat. In drug–protein interactions, △CP has a strong correlation with the surface burial upon complex formation. When nonpolar surface burial occurs, the desolvation of proteins and drugs contributes negatively to ΔCP [138], whereas the contribution is positive in the case of polar surface burial [138]. After measuring △H by ITC at different temperatures, △CP can be obtained from the slope of a linear fit using Equation (27) [36].

ΔH(T) = ΔH(T0) + ΔCp(T −T0) (27) where T is the experimental temperature, T0 is the reference tempera­ ture, and △H(T) is the binding enthalpy at temperature T. Because △CP can be directly obtained by DSC, a linear fitting is not needed. Bychkova et al. [148] investigated the effects of pH and ionic strength on the interaction of HSA with magnetic iron oxide nanoparticles (MNPs). In the experiment, MNPs titrated HSA, and the enthalpy changes and denaturation temperature of HSA, and the changes on the particle sur­ face under different buffer conditions were determined by DSC. Ac­ cording to the DSC measurement, the protein adsorption process was accompanied by the loss of thermodynamic stability. It was found that with the increase of pH value, the sensitivity of protein to thermal denaturation increased, while with the increase of NaCl concentration, the sensitivity of protein to thermal denaturation decreased. Further­ more, the differences in thermodynamic properties of the interaction under different buffer conditions were revealed by ITC, which were mainly manifested in the association constant (Ka), the number of binding sites (n), binding enthalpy (ΔH), and entropy change (ΔS) of

HSA binding to MNPs. Cotrina et al. [146] studied the interactions be­ tween binary and ternary molecules between transthyretin, Aβ peptides, and small-molecular chaperones by ITC. ITC provides Gibbs free energy change (ΔG), binding enthalpy (ΔH), entropy change (ΔS), association constant (Ka), and the number of binding sites (n) from the complete thermodynamic profile of a single experiment. By comparing the changes of thermodynamic properties in binary and ternary molecular systems before and after the addition of small-molecular chaperones, good TTR stabilizers were screened.

In the direct measurement method, thermodynamic information is directly obtained through ITC and DSC. This method can provide a more accurate estimate of the thermodynamic properties of drug–protein in­ teractions. However, the thermodynamic system may be affected by interferences from heat sources other than the binding reaction, which causes errors in the measurements [36,154].

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 13 3.2. Calculation methods

Calculation methods can be used for the indirect determination of thermodynamic properties; the van’t Hoff equation is commonly used for this purpose. By substituting the association constants obtained at different temperatures in the van’t Hoff equation, the binding enthalpy and entropy can be obtained as shown in Equation (28) [153]: ln Ka = −ΔH

RT + ΔS R (28) The △H and △S values can then be obtained from the slope and intercept of a linear fit of the ln Ka vs. 1/T plot, respectively. Paiva et al. [155] studied the interaction of the adsorbent with BSA and used the van’t Hoff equation to determine the thermodynamic properties during the adsorption process, including binding enthalpy (ΔH), entropy change (ΔS), and Gibbs free energy change (ΔG).

This method does not require the experimental evaluation of ther­ modynamic properties, which greatly reduces the experimental costs and time. However, the thermodynamic properties are greatly affected by Ka [156–158].

3.3. Binding force type Drug–protein interactions mainly involve noncovalent forces such as hydrogen bond, hydrophobic, electrostatic, and van der Waals forces [12,13,16,17]. The thermodynamic properties can be used to determine the binding force type. Ross et al. [149] summarized the thermodynamic laws controlling the binding force between small molecules and bio­ logical macromolecules through a large number of experiments. When

△G > 0, △S < 0, and △CP < 0, there is no interaction between the drug and the protein. △G < 0, △H > 0, △S > 0, and △CP < 0 indicate a hydrophobic interaction, whereas △G < 0, △H ≈0, and △S > 0 show that the drug–protein interaction involves electrostatic forces. Finally, when △G < 0, △H < 0, △S < 0, and △CP < 0, the drug–protein interaction is based on hydrogen bond or van der Waals forces.

4. Binding sites The binding of drugs to different protein domains achieves different therapeutic effects; thus, the identification of the binding sites is of high importance for the early screening and development of new drugs. The methods used to obtain information on the binding sites mainly include competition and displacement and MS approaches.

4.1. Competition and displacement experiments The competition and displacement method is widely performed in the ZE mode of affinity chromatography [31,53,65,66], ACE (a component of the analyte dissolved in a background electrolyte) [32,33,

81,82,91], and recently pf-ACE for fragment binding site characteriza­ tion [90,97,102]. This approach can not only provide the association constant through Equation (8) but also identify the specific binding site of the interaction according to the relationship between the competing sites [30,31,159]. The method involves replacing the analyte with a probe, generally exhibiting specific binding to the protein. The binding site of the drug can be determined by examining the influence of different concentrations of the drug in the mobile phase on the retention time of the probe [31,53,65,66]. Noctor et al. [159] employed the warfarin probe at the HSA site I to determine the binding site of octanoic acid in HSA through competition and displacement experiments. In addition, the Hage group [160] utilized R-warfarin and L-tryptophan to study the competition between acetohexamide and toluene sulfamide on

HSA. They showed that acetohexamide and tolbutamide bind at HSA sites I and II, respectively [160].

The competition and displacement method uses a small-plug competing agent and analyte to determine the binding site of the drug in the protein. Because the study of HSA domain probes is relatively mature, this technique has been widely applied for investigating the binding sites of HSA.

4.2. Mass spectrometry The core components of an MS device are an ionization source, a mass analyzer, and a detector [161,162]. The sample molecules are first ionized into charged particles in the ionization source, then separated based on the m/z ratio in the mass analyzer, and finally measured in the detector to generate a mass spectrum [161–164]. In 2003, Zhang et al. [165] first applied electrospray ionization (ESI)-MS to the quantitative determination of noncovalent interactions between ligands and pro­ teins. The automatic hydrogen–deuterium exchange (HDX) method developed in 2004 has led to further advances in the study of drug–­ protein interactions [166].

The advent of ESI-MS has contributed greatly to the retention of weakly bound drug–protein complexes transferred from the solution state to the gas phase. MS for studying biomolecules is often referred to as “Non-denaturing mass spectrometry” or “Native mass spectrometry”, and is now commonly referred to as “Native MS” [167–171]. This concept was first proposed by Heck’s group in 2004 [170]. It refers to the general term for the ESI-MS analysis of biomolecules obtained under non-denaturing solvent conditions (usually using volatile ammonium acetate or bicarbonate buffers) [168,169,171]. The purpose of Native

MS using the non-denaturing solvent is to preserve the native confor­ mation of the proteins as much as possible, but it must not satisfy the conformation in the physiological environment. Leney et al. [168] compared mass spectra of protein mixtures in denaturing and non-denaturing solvent conditions. The protein mixtures were found to have narrower charge distributions under non-denaturing solvent con­ ditions and appeared in high m/z regions with less charge. In the non-denaturing solvent condition, the protein retains the compact configuration and can be clearly distinguished in the mass spectrum.

However, under denaturing conditions, most proteins carried more charge and appeared in the low m/z region, which is not easy to distinguish. Based on the high m/z region of native protein spectra observed under non-deforming solvent conditions, mass analyzers co-developed with Native MS technology are also required [169,172].

Time-of-flight mass analyzers theoretically believe that the m/z range is infinite, and it also has the advantages of high sensitivity, fast analysis speed, and high resolution, so these mass analyzers have been widely used in the field of Native MS research [169,171,172].

Native MS can accurately provide the m/z of drug–protein com­ plexes, the type of protein subunits, and the number of binding sites for drug–protein interactions [169,171,172]. The number of binding sites can be obtained by comparing the mass spectra of drug–protein com­ plexes under denaturing and non-denaturing solvent conditions, as shown in Equation (29) [173,174]. n = MPDn −MP

MD (29) where n is the number of binding sites; MPDn is the molecular weight of the complex formed by the binding of the drug and the protein under non-denaturing solvent conditions; MP is the molecular weight of the protein obtained by the complex under deformation solvent conditions;

MD is the molecular weight of the drug obtained by the complex under deformation conditions. Rose et al. [175] studied the interaction of proteins with molecular weights up to 800 KDa and small molecules of

ADP/ATP. The obtained mass spectrum can clearly calculate the binding ratio of protein and small molecules in the complex. When Native MS is used for drug–protein interaction studies, it is necessary to ensure the compatibility of the sample solution with the MS instrument so as not to interfere with the ionization efficiency of the complex [169,171,172].

The pretreatment of water-soluble complex samples can use size

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 14 exclusion chromatography to purify the samples, and then exchange the samples with MS-compatible buffer solution systems for analysis [176–178]. To obtain more information about protein conformation or drug–protein binding site, etc., the chemical treatment of sample H-D exchange was further developed in MS analysis [166,179,180].

In HDX-MS, deuterium is used to replace the hydrogen atoms of the main amide chain in the protein, in order to study the corresponding structure and dynamics [180–184]. The binding sites of drugs and proteins can then be identified by comparing the mass spectra, as shown in Fig. 9 [185]. The steric hindrance and conformational changes of the protein in the interaction prevent the H–D exchange in the specific re­ gion of the protein involved in binding the drug [166,179–181]. This produces a mass spectrum that is different from the H–D exchanged spectrum of the protein alone. To minimize the reverse exchange, the

H–D exchange reaction must be quenched by lowering the pH [166,

179]. After the quenching is completed, the protein and complex are enzymatically digested, and the peptide fragment after enzymatic hy­ drolysis is then transferred to the mass spectrometer for analysis [185].

In the analysis of the spectrum, the peptides signals are compared with the primary amino acid sequence of the protein, and the difference in deuterium uptake between complex and protein is then used to deter­ mine the binding sites of the drug–protein interaction [166,179,180].

Although HDX-MS cannot provide high-resolution information on the protein structure [186], the key advantage of this method is that it is not limited by the size and complexity of the analyzed system. At the same time, it consumes a lower sample amount and can target protein systems that are challenging for traditional structural analysis methods.

In contrast to other techniques, MS does not require labels and can directly provide the stoichiometry interactions from the measured m/z values [173,174]. But it needs to be demonstrated that the equilibrium is not changed during the ESI process when the interacting species are transferred to the gas phase. Nonetheless, this has been proven reliable by many researchers for protein interaction systems [169,171,172]. MS also has other advantages such as high sensitivity, low sample con­ sumption, high throughput, and rapid analysis. In ESI mode, the sample is not affected by the applied energy; therefore, this soft ionization technology does not dissociate poorly stable drug–protein complexes [187]. However, to ensure compatibility with mass spectrometers, this method is not suitable for buffer salts with high ionic strength and low volatility [5], which are necessary for simulating physiological conditions.

5. Binding distance determination Fluorescence resonance energy transfer (FRET) experiments can quantify the drug–protein binding distance based on changes in the protein fluorescence intensity. This information can contribute to the identification of binding sites.

5.1. Fluorescence resonance energy transfer In 1948, F¨orster established the theory behind the FRET technique [188]. FRET can determine the distance between donor and acceptor molecules during drug–protein interactions [11,189,190]. In this approach, the excited donor fluorophore does not emit fluorescence, mainly because it transfers nonradiative energy to the nearby acceptor fluorophore, which causes the donor fluorescence to be quenched and the acceptor to emit longer-wavelength fluorescence, as shown in Fig. 10 [189]. According to F¨orster’s FRET theory, the donor and acceptor molecules must satisfy three conditions [191]: (1) the donor emission spectrum must have a certain degree of overlap with the absorption spectrum of the acceptor; (2) the donor molecule should be able to emit fluorescence, and the distance between the donor and acceptor mole­ cules should be short enough and no longer than 10 nm; and (3) the dipoles of the donor and acceptor fluorophores should be parallel to each other.

According to the formula for the distance between donor and acceptor molecules derived by F¨orster [see Equation (30)], the energy transfer efficiency (E) is related not only to the distance between donor and acceptor but also to the critical energy transfer distance (R0) [188,

192].

E = 1 − F F0 = R6 0 ( R6 0 + r6) (30) where F0 and F are the fluorescence intensities of the protein in the absence and presence of the ligand, R0 is the critical distance when the transfer efficiency is 50%, and r is the average distance between donor and acceptor. R0 can be obtained, as shown in Equation (31) [188,192]:

R6 0 = 9.79 × 10−25K2n−4φJ (31) where K2 is the spatial orientation factor of the dipole transition be­ tween donor and acceptor, n is the refractive index of the medium, Φ is the fluorescence quantum yield of the donor, and J is the spectral overlap between the donor emission spectrum and the acceptor

Fig. 9. Schematic illustration of hydrogen–deuterium exchange mass spectrometry [185]. Reprinted with permission from JoVE.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 15 absorption spectrum. The determination of the K value has not yet found a reliable experimental method to measure it. Fortunately, the research on the mobility and statistical dynamics of the dye-labeled system has been shown and confirmed by a lot of experience that the value of K2 takes 2/3 is more reasonable in various biological systems [193]. The value of n is typically 1.4 in the aqueous environment of biological systems [193]. Φ is defined as the ratio of the number of photons emitted to the number of photons absorbed [194]. The Φ value is generally measured by the reference method in experiments. Under the condition of the same excitation wavelength, the fluorescence integral area of the measured fluorescent solution and the reference fluorescent solution were determined, wherein the reference fluorescent solution (usually quinine sulfate) had a known fluorescent quantum yield [195,196]. At the same time, the absorbance of incident light of the same excitation wavelength was also measured. The Φ value of the measured sample can be obtained by substituting the values of fluorescence integral area and absorbance into Equation (32) [194]:

Φ = Φr⋅Ar⋅F⋅n2 A⋅Fr⋅nr2 (32) where Φ and Φr are the fluorescence quantum yield of the measured fluorescent and the reference fluorescent solution, respectively, A and Ar are the absorbance of the measured fluorescent solution and the refer­ ence fluorescent solution at the same excitation wavelength, respec­ tively, F and Fr are the area under the corrected emission curve (expressed in the number of photons) of the measured fluorescent so­ lution and the reference fluorescent solution, respectively, n and nr are the refractive index of the measured fluorescent solution and the reference fluorescent solution, respectively. J can be obtained from

Equation (33) [188,192]:

J = ∫∞ 0 F(λ)ε(λ)λ4dλ ∫∞ 0 F(λ)dλ (33) where F(λ) and ε(λ) are the fluorescence intensities of the donor and acceptor molar extinction coefficient at wavelength λ, respectively. The ε(λ) can be obtained according to the Beer-Lambert law, as shown in

Equation (34) [197]. Beer-Lambert law shows that the substance’s absorbance depends on the optical pathlength and the concentration of the absorbing substance [197].

A = ε(λ)cl (34) where c is the molar concentration of the absorbing substance and l is the optical pathlength. The distances of azelastine, eperisone hydrochloride, and vitamin from HSA tryptophan residues obtained by FRET were 2.34,

2.18, and 3.05 nm, respectively [11,12,192].

An important task in the design of a FRET experiment is the selection of an appropriate donor–acceptor pair. In addition, the donor and acceptor molecules must be present in sufficient concentrations to induce FRET.

6. Protein conformation changes Drug–protein interactions may lead to changes in the spatial conformation of proteins, which affects some of their active functions.

We can establish whether the drug interacts with the protein by evalu­ ating conformational changes in the latter. Current methods to deter­ mine conformational changes in proteins mainly spectroscopy.

6.1. Spectroscopy methods Synchronous fluorescence and CD are commonly used to investigate protein conformations [13,127,128,198–200]. Synchronous fluores­ cence spectrometry is measured by simultaneously scanning the wave­ lengths of the excitation and emission of two monochromators [201] and minimizing Rayleigh scattering. Compared with conventional fluorescence, the method can simplify the spectrum, narrow the spectral band, and reduce the spectral overlap in the complex system of various fluorescent substances, so it has higher selectivity [202,203]. In syn­ chronous fluorescence experiments, synchronous scanning is carried out after determining the difference (△λ = λEm - λEx = constant) between the excitation and emission wavelengths [200], and the fluorescence spectrum of amino acid residues is obtained. When △λ = 15 nm, syn­ chronous fluorescence only shows the spectral characteristics of tyrosine residues [200,204], whereas for △λ = 60 nm the technique only shows the spectral characteristics of tryptophan residues [200,204]. The maximum fluorescence emission wavelength and intensity of amino acid residues are sensitive to the polarity of the environment. A red shift in

Fig. 10. Schematic illustration of FRET process during ligand interaction with protein [189]. Reprinted with permission from CellPress.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 16 the emission wavelength indicates an increased polarity of the micro­ environment in which the amino acid residues are located [198,205,

206], whereas a blue shift indicates an increased hydrophobicity [206].

Ariga et al. [198] studied the interaction between epinastine hydro­ chloride (EPN) and BSA. EPN causes a red shift in the maximum emis­ sion wavelengths of tryptophan and tyrosine residues, indicating that the polarity of the environment increases. This shows that the addition of EPN alters the conformation of BSA.

Vo-Dinh et al. [207] proposed the basic theory of synchronous fluorescence spectrometry. The theory suggests that the simultaneous fluorescence intensity is proportional to the concentration of the ana­ lyte. The theory is based on a limited absorbance range. This is mainly due to the inner filter effect. The effect alters the excitation and emission spectrum, thus distorting the synchronized maximum wavelength values [208–211]. The inner filter effect arises from the fact that, on the one hand, the excitation intensity of the cuvette center in the fluorescence spectrometer has a lower fluorescence emission due to sample absorp­ tion [208]; on the other hand, the excitation spectrum and emission spectrum of the analyte overlap significantly, which leads to the light emitted at the center can be reabsorbed by the analyte itself [211].

When the absorbance of the solution is lower than 0.3, a mathematical formula can be used to further correct the internal filtering effect [212].

CD can estimate the general content of α- and β-helices in protein [13,125–128]. If the changes in protein secondary structure are more accurately reflected, specialized algorithms which include datasets of reference spectra are needed for deconvolution [213–215]. The CD study of protein conformational changes is mainly based on the differ­ ence in the absorption of left and right circularly polarized light by the optically active groups of the protein [216]. Serum albumin shows two negative peaks in the far-ultraviolet region, at 208 and 222 nm, which are the characteristic absorption peaks of α-helices [13,217,218].

Changes in the protein conformation can be assessed by examining changes in the mean residual ellipticity (MRE) at the characteristic wavelength, which can be calculated from Equation (35) [13,217,219]:

MRE = θobs 10 × n × c × l (35) where θobs is the ellipticity, n is the number of amino acid residues, c is the molar concentration, and l is the path length of the light. Tian et al. [220] used CD to study the interaction between mononaphthalimide spermidine (MINS) and BSA, and found that the CD spectrum of BSA did not change significantly after the addition of MINS, as shown in Fig. 11.

Therefore, they inferred that MINS only caused a slight change in the conformation of BSA, but did not damage the stability of the BSA helical structure [220].

Synchronous fluorescence and CD experiments are easy to perform and the results can be obtained in a short time. CD is often the preferred method to study protein conformational changes [13,125–128]. How­ ever, synchronous fluorescence and CD experiments have requirements to be meaningful. For example, accurate determination of protein con­ centration in experimental manipulations is important for reliable determination of secondary structure by far-ultraviolet CD. In addition, the impact of potential interferences effects (such as the inner filter ef­ fect) and key factors of the technique are critical for accurate interpre­ tation of the data. The sensitivity of this method is generally low, and its sample consumption is relatively large. Therefore, this approach is not suitable for the study of trace proteins.

7. Protein stability changes The common methods used for the detection of biologically active targets in recent years include cellular thermal shift assay (CETSA) and thermal proteome profiling (TPP) [21–25]. CETSA and TPP are based on the analysis of thermal stability shifts and can be used to study protein stability changes in combination with the immunoblotting blot or MS techniques, respectively [221].

7.1. Thermal stability shift analysis In CETSA, the drug is added to multiple aliquots of cell lysates or intact cells, followed by heating to different temperatures and cooling down. Then, the precipitated protein is separated from the soluble protein by centrifugation. Finally, the soluble protein is quantified by immunoblotting [189,220,221]. Generally, the protein hydrophobic cores are increasingly exposed to increasing temperature, leading to protein precipitation at high temperatures [189]. A more stable protein has a higher resistance to heat-induced precipitation, and the amount of soluble protein increases [189,221]. Therefore, protein stability can be evaluated by measuring the amount of soluble protein. When the ther­ mal melting curve of the soluble protein amount measured by immu­ noblotting at different temperatures shows a temperature shift (△Tm) after binding, this confirms that the drug induces a stability change in the protein [23]. CETSA requires prior knowledge of the target protein, and cannot be applied on the proteomic scale [221]. To overcome these limitations, Savitski et al. [22] introduced TPP. This method generally employs isobaric tandem mass spectrometry, which can achieve protein quantification under ten different experimental conditions [22].

CETSA and TPP have some limitations: they are suitable for the detection of soluble proteins but have some shortcomings for the detection of hydrophobic proteins such as membrane proteins. In addition, some proteins, especially those with small molecular weights, do not aggregate within the recommended temperature range, and cannot be studied with the above methods [25,189,221]. Finally, drug-induced changes in protein stability may occur only on a small scale in large proteins and do not affect their overall stability [189,221].

8. Conclusion This review provides a systematic analysis of methods used for determining various binding parameters of drug–protein interactions.

Spectroscopic [13,117,121,122] and calorimetric [36–38,132–134] methods are simple to operate and do not require the separation of drugs and proteins. These two methods are used in drug–protein interaction studies to obtain the association constant and the number of binding sites by changing the physical and chemical properties of the substance itself [13,36–38,117]. However, these two methods have high re­ quirements for sample purity, and the sample consumption volume is large [3]. ED, HPAC, and CE require separation of analytes, and the association constants and the number of binding sites can be obtained from retention time, peak height, or peak area of the analytes [43,54,90,

Fig. 11. CD method for studying MINS–BSA interaction [220]. Reprinted with permission from Elsevier.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 17 108]. These methods require low sample purity and less volume con­ sumption of the sample [2]. Among them, CE has great potential in the study of trace protein binding [108]. The ITC and DSC in the direct measurements can realize the direct acquisition of thermodynamic properties [132–134]. The calculation methods require the indirect calculation of mathematical formulas to obtain thermodynamic pa­ rameters, so it has a large deviation [156–158]. In addition, changes in thermodynamic properties can be used to determine the type of binding force [149]. In terms of binding sites, competition and displacement [160] and MS [185] consume less sample volume, and MS can also enable proteomic studies. FRET is a common method for evaluating binding distances [189]. Synchronous fluorescence [198] and CD [220] in spectroscopy are often used in the study of protein conformational changes, and CD can also estimate changes in protein secondary struc­ ture [13,125–128]. In thermal stability shift analysis, CETSA and TPP can evaluate changes in protein stability in cells using immunoblotting or MS, respectively [21–25]. Among them, TPP is often used as a high-throughput screening technology for drug discovery [22].

Because the parameters obtained by each method are different, in most cases the combination of different methods can provide a more comprehensive understanding of the interaction mechanism. However, most methods are not suitable for cell studies. In addition to thermal stability shift approaches, other methods mainly involve in vitro condi­ tions. Therefore, in vivo experiments require more effective techniques such as X-Ray crystallography [222,223], intracellular NMR [224,225] and cryo-electron microscopy [226,227]. Future studies should aim to combine different new technologies for a more accurate determination of the potential mechanisms of drug–protein binding in vivo.

Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments This work was supported by the Beijing Outstanding Young Scientist

Program (BJJWZYJH01201910005017) and the National Natural Sci­ ence Foundation of China (Nos. 21936001 and 22127805).

References [1] X. Li, X. Wang, M. Snyder, Systematic investigation of protein-small molecule interactions, IUBMB Life 65 (2013) 2–8, https://doi.org/10.1002/iub.1111. [2] K. Wanat, E. Brzezinska, A.W. Sobanska, Aspects of drug-protein binding and methods of analyzing the phenomenon, Curr. Pharmaceut. Des. 24 (2018)

2974–2985, https://doi.org/10.2174/1381612824666180808145320. [3] K. Vuignier, J. Schappler, J.L. Veuthey, P.A. Carrupt, S. Martel, Drug-protein binding: a critical review of analytical tools, Anal. Bioanal. Chem. 398 (2010)

53–66, https://doi.org/10.1007/s00216-010-3737-1. [4] B. Seyfinejad, S.A. Ozkan, A. Jouyban, Recent advances in the determination of unbound concentration and plasma protein binding of drugs: analytical methods,

Talanta 225 (2021) 122052, https://doi.org/10.1016/j.talanta.2020.122052. [5] X. Liu, M. Fang, F. Xu, D. Chen, Characterization of the binding of per- and poly- fluorinated substances to proteins: a methodological review, Trends Anal. Chem.

116 (2019) 177–185, https://doi.org/10.1016/j.trac.2019.05.017. [6] Q. Jiao, R. Wang, Y. Jiang, B. Liu, Study on the interaction between active components from traditional Chinese medicine and plasma proteins, Chem. Cent.

J. 12 (2018) 1–20, https://doi.org/10.1186/s13065-018-0417-2. [7] J. Drews, Drug discovery: a historical perspective, Science 287 (2000)

1960–1964, https://doi.org/10.1126/science.287.5460.1960. [8] M. Muttenthaler, G.F. King, D.J. Adams, P.F. Alewood, Trends in peptide drug discovery, Nat. Rev. Drug Discov. 20 (2021) 309–325, https://doi.org/10.1038/ s41573-020-00135-8. [9] M.M. Attwood, D. Fabbro, A.V. Sokolov, S. Knapp, H.B. Schi¨oth, Trends in kinase drug discovery: targets, indications and inhibitor design, Nat. Rev. Drug Discov.

20 (2021) 839–861, https://doi.org/10.1038/s41573-021-00252-y. [10] A. Sharma, A. Pathania, Diverse practices applicable for exploring the drug binding with protein–A review, Mater. Today Proc. 48 (2022) 1575–1581, https://doi.org/10.1016/j.matpr.2021.09.476. [11] F.M. Almutairi, M.R. Ajmal, M.K. Siddiqi, M. Amir, R.H. Khan, Multi- spectroscopic and molecular docking technique study of the azelastine interaction with human serum albumin, J. Mol. Struct. 1201 (2020), 127147, https://doi. org/10.1016/j.molstruc.2019.127147. [12] G. Rabbani, M.H. Baig, E.J. Lee, W.-K. Cho, J.Y. Ma, I. Choi, Biophysical study on the interaction between eperisone hydrochloride and human serum albumin using spectroscopic, calorimetric, and molecular docking analyses, Mol. Pharm.

14 (2017) 1656–1665, https://doi.org/10.1021/acs.molpharmaceut.6b01124. [13] S. Yasmeen, Riyazuddeen, S. Khatun, F. Abul Qais, Characterization of interactions between cromolyn sodium and bovine serum albumin by spectroscopic, calorimetric and computational methods, J. Biomol. Struct. Dyn.

38 (2020) 722–732, https://doi.org/10.1080/07391102.2019.1586588. [14] R.T. Recio, N.P. Guerra, A. Torrado, L.H. Skibsted, Interaction between calcium and casein hydrolysates: stoichiometry, binding constant, binding sites and thermal stability of casein phosphopeptide complexes, Int. Dairy J. 88 (2019)

25–33, https://doi.org/10.1016/j.idairyj.2018.08.009. [15] S. Siddiqui, F. Ameen, S. ur Rehman, T. Sarwar, M. Tabish, Studying the interaction of drug/ligand with serum albumin, J. Mol. Liq. 336 (2021), 116200, https://doi.org/10.1016/j.molliq.2021.116200. [16] M.K. Zia, T. Siddiqui, S.S. Ali, H. Ahsan, F.H. Khan, Deciphering the binding of dutasteride with human alpha-2-macroglobulin: molecular docking and calorimetric approach, Int. J. Biol. Macromol. 133 (2019) 1081–1089, https:// doi.org/10.1016/j.ijbiomac.2019.04.180. [17] B.K. Paul, N. Ghosh, S. Mukherjee, Interplay of multiple interaction forces: binding of norfloxacin to human serum albumin, J. Phys. Chem. B 119 (2015)

13093–13102, https://doi.org/10.1021/acs.jpcb.5b08147. [18] A. Kuzmanic, G.R. Bowman, J. Juarez-Jimenez, J. Michel, F.L. Gervasio,

Investigating cryptic binding sites by molecular dynamics simulations, Accounts

Chem. Res. 53 (2020) 654–661, https://doi.org/10.1021/acs.accounts.9b00613. [19] Y. Wang, Z. Yu, W. Xiao, S. Lu, J. Zhang, Allosteric binding sites at the receptor–lipid bilayer interface: novel targets for GPCR drug discovery, Drug

Discov. Today 26 (2021) 690–703, https://doi.org/10.1016/j. drudis.2020.12.001. [20] M. Pignataro, M. Herrera, V. Dodero, Evaluation of peptide/protein self-assembly and aggregation by spectroscopic methods, Molecules 25 (2020) 4854, https:// doi.org/10.3390/molecules25204854. [21] D.M. Molina, R. Jafari, M. Ignatushchenko, T. Seki, E.A. Larsson, C. Dan,

L. Sreekumar, Y. Cao, P. Nordlund, Monitoring drug target engagement in cells and tissues using the cellular thermal shift assay, Science 341 (2013) 84–87, https://doi.org/10.1126/science.1233606. [22] M.M. Savitski, F.B. Reinhard, H. Franken, T. Werner, M.F. Savitski, D. Eberhard,

D. Martinez Molina, R. Jafari, R.B. Dovega, S. Klaeger, B. Kuster, P. Nordlund,

M. Bantscheff, G. Drewes, Tracking cancer drugs in living cells by thermal profiling of the proteome, Science 346 (2014) 1255784, https://doi.org/

10.1126/science.1255784. [23] J.M. Dziekan, G. Wirjanata, L. Dai, K.D. Go, H. Yu, Y.T. Lim, L. Chen, L.C. Wang,

B. Puspita, N. Prabhu, R.M. Sobota, P. Nordlund, Z. Bozdech, Cellular thermal shift assay for the identification of drug–target interactions in the Plasmodium falciparum proteome, Nat. Protoc. 15 (2020) 1881–1921, https://doi.org/

10.1038/s41596-020-0310-z. [24] S. Attarha, A. Reithmeier, S. Busker, M. Desroses, B.D.G. Page, Validating signal transducer and activator of transcription (STAT) protein–inhibitor interactions using biochemical and cellular thermal shift assays, ACS Chem. Biol. 15 (2020)

1842–1851, https://doi.org/10.1021/acschembio.0c00046. [25] A. Mateus, N. Kurzawa, I. Becher, S. Sridharan, D. Helm, F. Stein, A. Typas, M.

M. Savitski, Thermal proteome profiling for interrogating protein interactions,

Mol. Syst. Biol. 16 (2020), e9232, https://doi.org/10.15252/msb.20199232. [26] Y. Kwon, Protein binding, in: Y. Kwon (Ed.), Handbook of Essential

Pharmacokinetics, Pharmacodynamics and Drug Metabolism for Industrial

Scientists, Springer US, Boston, MA, 2002, pp. 105–120, https://doi.org/

10.1007/0-306-46820-4_7. [27] M.P. Williamson, Using chemical shift perturbation to characterise ligand binding, Prog. Nucl. Magn. Reson. Spectrosc. 73 (2013) 1–16, https://doi.org/

10.1016/j.pnmrs.2013.02.001. [28] G. Scatchard, The attractions of proteins for small molecules and ions, Ann. N. Y.

Acad. Sci. 51 (1949) 660–672, https://doi.org/10.1111/j.1749-6632.1949. tb27297.x. [29] L. Michalcov´a, Z. Glatz, Comparison of various capillary electrophoretic approaches for the study of drug–protein interaction with emphasis on minimal consumption of protein sample and possibility of automation, J. Separ. Sci. 38 (2015) 325–331, https://doi.org/10.1002/jssc.201400914. [30] D.S. Hage, S.A. Tweed, Recent advances in chromatographic and electrophoretic methods for the study of drug-protein interactions, J. Chromatogr. B Biomed. Sci.

Appl. 699 (1997) 499–525, https://doi.org/10.1016/S0378-4347(97)00178-3. [31] X. Zheng, Z. Li, S. Beeram, M. Podariu, R. Matsuda, E.L. Pfaunmiller, C.J.W. Ii,

N. Carter, D.S. Hage, Analysis of biomolecular interactions using affinity microcolumns: a review, J. Chromatogr. B 968 (2014) 49–63, https://doi.org/

10.1016/j.jchromb.2014.01.026. [32] I.O. Neaga, S. Hambye, E. Bodoki, C. Palmieri, E. Ansseau, A. Belayew, R. Oprean,

B. Blankert, Affinity capillary electrophoresis for identification of active drug candidates in myotonic dystrophy type 1, Anal. Bioanal. Chem. 410 (2018)

4495–4507, https://doi.org/10.1007/s00216-018-1107-6. [33] X. He, Y. Ding, D. Li, B. Lin, Recent advances in the study of biomolecular interactions by capillary electrophoresis, Electrophoresis 25 (2004) 697–711, https://doi.org/10.1002/elps.200305727. [34] S.A.C. Wren, R.C. Rowe, Theoretical aspects of chiral separation in capillary electrophoresis: I. Initial evaluation of a model, J. Chromatogr. A 603 (1992)

235–241, https://doi.org/10.1016/0021-9673(92)85366-2.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 18 [35] Y. Zhang, G. Zhang, Y. Wang, Studies of interaction of mitomycin C and serum albumin by spectrum method, J. Anal. Sci. 16 (2000) 445–449, https://doi.org/

10.3969/j.issn.1006-6144.2000.06.002. [36] A. Thanassoulas, G. Nounesis, Isothermal titration calorimetry: a powerful tool for the characterization of molecular interactions, in: C. Demetzos, N. Pippa (Eds.), Thermodynamics and Biophysics of Biomedical Nanosystems: Applications and Practical Considerations, Springer Singapore, Singapore, 2019, pp. 63–103, https://doi.org/10.1007/978-981-13-0989-2_4. [37] G. Bruylants, J. Wouters, C. Michaux, Differential scanning calorimetry in life science: thermodynamics, stability, molecular recognition and application in drug design, Curr. Med. Chem. 12 (2005) 2011–2020, https://doi.org/10.2174/

0929867054546564. [38] V. Plotnikov, A. Rochalski, M. Brandts, J.F. Brandts, S. Williston, V. Frasca, L.- N. Lin, An autosampling differential scanning calorimeter instrument for studying molecular interactions, Assay Drug Dev. Technol. 1 (2002) 83–90, https://doi. org/10.1089/154065802761001338. [39] Y.L. Toh, Y.Y. Pang, M. Shwe, R. Kanesvaran, C.K. Toh, A. Chan, H.K. Ho, HPLC- MS/MS coupled with equilibrium dialysis method for quantification of free drug concentration of pazopanib in plasma, Heliyon 6 (2020), e03813, https://doi. org/10.1016/j.heliyon.2020.e03813. [40] J. Li, W. Li, X. Dai, D. Zhong, Y. Ding, X. Chen, Bioequivalence of paclitaxel protein-bound particles in patients with breast cancer: determining total and unbound paclitaxel in plasma by rapid equilibrium dialysis and liquid chromatography-tandem mass spectrometry, Drug Des. Dev. Ther. 13 (2019)

1739–1749, https://doi.org/10.2147/DDDT.S200679. [41] L. Chemlal, J. Akachar, S. Makram, B. Zoubir, Y. Cherrah, R. Eljaoudi,

A. Ibrahimi, M.A. Faouzi, The displacement study of 99mTc-DTPA-human serum albumin binding in presence of furosemide and metformin by using equilibrium dialysis and molecular docking, IUBMB Life 71 (2019) 2003–2009, https://doi. org/10.1002/iub.2167. [42] Z.M. Herbst, S. Shibata, E. Fan, M.H. Gelb, An inexpensive, in-house-made, microdialysis device for measuring drug-protein binding, ACS Med. Chem. Lett. 9 (2018) 279–282, https://doi.org/10.1021/acsmedchemlett.7b00316. [43] A. Rabbani-Chadegani, H. Mollaei, J. Sargolzaei, Investigation of the interaction between berberine and nucleosomes in solution: spectroscopic and equilibrium dialysis approach, Spectrochim. Acta Mol. Biomol. Spectrosc. 173 (2017)

418–424, https://doi.org/10.1016/j.saa.2016.09.052. [44] Y.-C. Chen, J.R. Kenny, M. Wright, C.E.C.A. Hop, Z. Yan, Improving confidence in the determination of free fraction for highly bound drugs using bidirectional equilibrium dialysis, J. Pharmaceut. Sci. 108 (2019) 1296–1302, https://doi.org/

10.1016/j.xphs.2018.10.011. [45] S. Ahmad, D. Baker, D. Murnane, N. Spooner, U. Gerhard, Solid-phase microextraction for assessment of plasma protein binding, a complement to rapid equilibrium dialysis, Bioanalysis 13 (2021) 1101–1111, https://doi.org/10.4155/ bio-2021-0109. [46] N. Fabresse, I. Uteem, E. Lamy, Z. Massy, I.A. Larabi, J.-C. Alvarez, Quantification of free and protein bound uremic toxins in human serum by LC-MS/MS: comparison of rapid equilibrium dialysis and ultrafiltration, Clin. Chim. Acta 507 (2020) 228–235, https://doi.org/10.1016/j.cca.2020.04.032. [47] R. Yang, S. Ren, X. Jin, Y. Sun, Y. Dong, J. Zhang, W. Liang, L. Chen,

Determination of unbound fraction of dorzagliatin in human plasma by equilibrium dialysis and LC-MS/MS and its application to a clinical pharmacokinetic study, J. Pharmaceut. Biomed. Anal. 195 (2021), 113854, https://doi.org/10.1016/j.jpba.2020.113854. [48] K.S. Roy, E. Nazdraji´c, O.I. Shimelis, M.J. Ross, Y. Chen, H. Cramer, J. Pawliszyn,

Optimizing a high-throughput solid-phase microextraction system to determine the plasma protein binding of drugs in human plasma, Anal. Chem. 93 (2021)

11061–11065, https://doi.org/10.1021/acs.analchem.1c01986. [49] G.R. Bolton, A.W. Boesch, J. Basha, D.P. LaCasse, B.D. Kelley, H. Acharya, Effect of protein and solution properties on the donnan effect during the ultrafiltration of proteins, Biotechnol. Prog. 27 (2011) 140–152, https://doi.org/10.1002/ btpr.523. [50] T. Briskot, N. Hillebrandt, S. Kluters, G. Wang, J. Studts, T. Hahn, T. Huuk,

J. Hubbuch, Modeling the Gibbs–Donnan effect during ultrafiltration and diafiltration processes using the Poisson–Boltzmann theory in combination with a basic Stern model, J. Membr. Sci. 648 (2022), 120333, https://doi.org/10.1016/ j.memsci.2022.120333. [51] M. Soethoudt, U. Grether, J. Fingerle, T.W. Grim, F. Fezza, L. de Petrocellis,

C. Ullmer, B. Rothenh¨ausler, C. Perret, N. van Gils, D. Finlay, C. MacDonald,

A. Chicca, M.D. Gens, J. Stuart, H. de Vries, N. Mastrangelo, L. Xia,

G. Alachouzos, M.P. Baggelaar, A. Martella, E.D. Mock, H. Deng, L.H. Heitman,

M. Connor, V. Di Marzo, J. Gertsch, A.H. Lichtman, M. Maccarrone, P. Pacher,

M. Glass, M. van der Stelt, Cannabinoid CB2 receptor ligand profiling reveals biased signalling and off-target activity, Nat. Commun. 8 (2017) 1–14, https:// doi.org/10.1038/ncomms13958. [52] S. Ohlson, L. Hansson, P.-O. Larsson, K. Mosbach, High performance liquid affinity chromatography (HPLAC) and its application to the separation of enzymes and antigens, FEBS Lett. 93 (1978) 5–9, https://doi.org/10.1016/0014- 5793(78)80792-3. [53] L. Lecas, V. Dugas, C. Demesmay, Affinity chromatography: a powerful tool in drug discovery for investigating ligand/membrane protein interactions, Separ.

Purif. Rev. 50 (2021) 315–332, https://doi.org/10.1080/

15422119.2020.1749852. [54] Z. Li, D.S. Hage, Analysis of stereoselective drug interactions with serum proteins by high-performance affinity chromatography: a historical perspective,

J. Pharmaceut. Biomed. Anal. 144 (2017) 12–24, https://doi.org/10.1016/j. jpba.2017.01.026. [55] M. Aksoy, A new insight into purification of polyphenol oxidase and inhibition effect of curcumin and quercetin on potato polyphenol oxidase, Protein Expr.

Purif. 171 (2020), 105612, https://doi.org/10.1016/j.pep.2020.105612. [56] Z. Li, E. Rodriguez, S. Azaria, A. Pekarek, D.S. Hage, Affinity monolith chromatography: a review of general principles and applications, Electrophoresis

38 (2017) 2837–2850, https://doi.org/10.1002/elps.201700101. [57] E.L. Rodriguez, S. Poddar, S. Iftekhar, K. Suh, A.G. Woolfork, S. Ovbude,

A. Pekarek, M. Walters, S. Lott, D.S. Hage, Affinity chromatography: a review of trends and developments over the past 50 years, J. Chromatogr. B 1157 (2020),

122332, https://doi.org/10.1016/j.jchromb.2020.122332. [58] G.E. Premetis, N.E. Labrou, Reduce, reuse and recycle in protein chromatography: development of an affinity adsorbent from waste paper and its application for the purification of proteases from fish by-products, Biomolecules 10 (2020) 822, https://doi.org/10.3390/biom10060822. [59] Y. Bai, Y. Fan, G. Ge, F. Wang, Advances in chromatography in the study of drug- plasma protein interactions, Chin. J. Chromatogr. 39 (2021) 1077–1085, https:// doi.org/10.3724/sp.j.1123.2021.06028. [60] K. Kasai, Frontal affinity chromatography: an excellent method of analyzing weak biomolecular interactions based on a unique principle, Biochim. Biophys. Acta

Gen. Subj. 1865 (2021), 129761, https://doi.org/10.1016/j. bbagen.2020.129761. [61] C. Wang, X. Wang, X. Xu, B. Liu, X. Xu, L. Sun, H. Li, J. Zhang, Simultaneous high- throughput determination of interaction kinetics for drugs and cyclodextrins by high performance affinity chromatography with mass spectrometry detection,

Anal. Chim. Acta 909 (2016) 75–83, https://doi.org/10.1016/j.aca.2015.12.026. [62] A.J. Jackson, J. Anguizola, E.L. Pfaunmiller, D.S. Hage, Use of entrapment and high-performance affinity chromatography to compare the binding of drugs and site-specific probes with normal and glycated human serum albumin, Anal.

Bioanal. Chem. 405 (2013) 5833–5841, https://doi.org/10.1007/s00216-013- 6981-3. [63] X. Zheng, R. Matsuda, D.S. Hage, Analysis of free drug fractions by ultrafast affinity extraction: interactions of sulfonylurea drugs with normal or glycated human serum albumin, J. Chromatogr. A 1371 (2014) 82–89, https://doi.org/

10.1016/j.chroma.2014.10.092. [64] B.M. Dunn, I.M. Chaiken, Quantitative affinity chromatography. Determination of binding constants by elution with competitive inhibitors, Proc. Natl. Acad. Sci.

Unit. States Am. 71 (1974) 2382, https://doi.org/10.1073/pnas.71.6.2382. [65] P. Tao, S. Poddar, Z. Sun, D.S. Hage, J. Chen, Analysis of solute-protein interactions and solute-solute competition by zonal elution affinity chromatography, Methods 146 (2018) 3–11, https://doi.org/10.1016/j. ymeth.2018.01.020. [66] C. Zhang, E. Rodriguez, C. Bi, X. Zheng, D. Suresh, K. Suh, Z. Li, F. Elsebaei, D.

S. Hage, High performance affinity chromatography and related separation methods for the analysis of biological and pharmaceutical agents, Analyst 143 (2018) 374–391, https://doi.org/10.1039/c7an01469d. [67] R. Matsuda, Z. Li, X. Zheng, D.S. Hage, Analysis of multi-site drug-protein interactions by high-performance affinity chromatography: binding by glimepiride to normal or glycated human serum albumin, J. Chromatogr. A 1408 (2015) 133–144, https://doi.org/10.1016/j.chroma.2015.07.012. [68] B. Loun, D.S. Hage, Chiral separation mechanisms in protein-based HPLC columns. 1. Thermodynamic studies of (R)- and (S)-warfarin binding to immobilized human serum albumin, Anal. Chem. 66 (1994) 3814–3822, https:// doi.org/10.1021/ac00093a043. [69] J. Yang, D.S. Hage, Characterization of the binding and chiral separation of d- and l-tryptophan on a high-performance immobilized human serum albumin column,

J. Chromatogr. A 645 (1993) 241–250, https://doi.org/10.1016/0021-9673(93)

83383-4. [70] A.C. Moser, B. White, F.A. Kovacs, Measuring binding constants of His-tagged proteins using affinity chromatography and Ni-NTA immobilized enzymes, in: N.

E. Labrou (Ed.), Protein Downstream Processing: Design, Development, and

Application of High and Low-Resolution Methods, Springer US, New York, NY,

2021, pp. 405–416, https://doi.org/10.1007/978-1-0716-0775-6_26. [71] P. Tao, Z. Li, A.G. Woolfork, D.S. Hage, Characterization of tolazamide binding with glycated and normal human serum albumin by using high-performance affinity chromatography, J. Pharmaceut. Biomed. Anal. 166 (2019) 273–280, https://doi.org/10.1016/j.jpba.2019.01.025. [72] J. Anguizola, K.S. Joseph, O.S. Barnaby, R. Matsuda, G. Alvarado, W. Clarke, R.

L. Cerny, D.S. Hage, Development of affinity microcolumns for drug-protein binding studies in personalized medicine: interactions of sulfonylurea drugs with in vivo glycated human serum albumin, Anal. Chem. 85 (2013) 4453–4460, https://doi.org/10.1021/ac303734c. [73] D.S. Hage, High-performance affinity chromatography: a powerful tool for studying serum protein binding, J. Chromatogr. B 768 (2002) 3–30, https://doi. org/10.1016/S0378-4347(01)00482-0. [74] V. D’Atri, A. Murisier, S. Fekete, J.-L. Veuthey, D. Guillarme, Current and future trends in reversed-phase liquid chromatography-mass spectrometry of therapeutic proteins, Trends Anal. Chem. 130 (2020), 115962, https://doi.org/

10.1016/j.trac.2020.115962. [75] E. Vasconcelos Soares Maciel, A.L. de Toffoli, E. Sobieski, C.E. Domingues

Naz´ario, F.M. Lanças, Miniaturized liquid chromatography focusing on analytical columns and mass spectrometry: a review, Anal. Chim. Acta 1103 (2020) 11–31, https://doi.org/10.1016/j.aca.2019.12.064. [76] A.F. Asensio, E. Alvarez-Gonz´alez, A. Rodríguez, L.M. Sierra, E. Blanco-Gonz´alez,

Chromatographic methods coupled to mass spectrometry for the determination of

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 19 oncometabolites in biological samples-a review, Anal. Chim. Acta 1177 (2021),

338646, https://doi.org/10.1016/j.aca.2021.338646. [77] J.-c. Zhao, Q.-y. Zhu, L.-y. Zhao, H.-z. Lian, H.-y. Chen, Preparation of an aptamer based organic–inorganic hybrid monolithic column with gold nanoparticles as an intermediary for the enrichment of proteins, Analyst 141 (2016) 4961–4967, https://doi.org/10.1039/C6AN00957C. [78] Z. Wang, J.-c. Zhao, H.-z. Lian, H.-y. Chen, Aptamer-based organic-silica hybrid affinity monolith prepared via “thiol-ene” click reaction for extraction of thrombin, Talanta 138 (2015) 52–58, https://doi.org/10.1016/j. talanta.2015.02.009. [79] X. Yu, S. Lai, L. Wang, Y. Chen, X. Lin, Z. Xie, Preparation of aptamer-bound polyamine affinity monolithic column via a facile triazine-bridged strategy and application to on-column specific discrimination of ochratoxin A, J. Separ. Sci. 42 (2019) 2272–2279, https://doi.org/10.1002/jssc.201900175. [80] Y. Chen, M. Chen, J. Chi, X. Yu, Y. Chen, X. Lin, Z. Xie, Aptamer-based polyhedral oligomeric silsesquioxane (POSS)-containing hybrid affinity monolith prepared via a “one-pot” process for selective extraction of ochratoxin A, J. Chromatogr. A

1563 (2018) 37–46, https://doi.org/10.1016/j.chroma.2018.05.044. [81] Y. Wang, D.I. Adeoye, E.O. Ogunkunle, I.A. Wei, R.T. Filla, M.G. Roper, Affinity capillary electrophoresis: a critical review of the literature from 2018 to 2020,

Anal. Chem. 93 (2021) 295–310, https://doi.org/10.1021/acs. analchem.0c04526. [82] C. Zhang, A.G. Woolfork, K. Suh, S. Ovbude, C. Bi, M. Elzoeiry, D.S. Hage, Clinical and pharmaceutical applications of affinity ligands in capillary electrophoresis: a review, J. Pharmaceut. Biomed. Anal. 177 (2020), 112882, https://doi.org/

10.1016/j.jpba.2019.112882. [83] S. Krait, M.-L. Konjaria, G.K.E. Scriba, Advances of capillary electrophoresis enantioseparations in pharmaceutical analysis (2017-2020), Electrophoresis 42 (2021) 1709–1725, https://doi.org/10.1002/elps.202000359. [84] S. Bernardo-Bermejo, E. S´anchez-L´opez, M. Castro-Puyana, M.L. Marina, Chiral capillary electrophoresis, Trends Anal. Chem. 124 (2020), 115807, https://doi. org/10.1016/j.trac.2020.115807. [85] Y.H. Chu, L.Z. Avila, H.A. Biebuyck, G.M. Whitesides, Use of affinity capillary electrophoresis to measure binding constants of ligands to proteins, J. Med.

Chem. 35 (1992) 2915–2917, https://doi.org/10.1021/jm00093a027. [86] J.C. Kraak, S. Busch, H. Poppe, Study of protein-drug binding using capillary zone electrophoresis, J. Chromatogr. A 608 (1992) 257–264, https://doi.org/10.1016/

0021-9673(92)87132-R. [87] N. Li, S. Zeng, L. He, W.W. Zhong, Probing nanoparticle-protein interaction by capillary electrophoresis, Anal. Chem. 82 (2010) 7460–7466, https://doi.org/

10.1021/ac101627p. [88] T. Krizek, K. Molnarova, V. Pavlu, B. Filounova, E. Martinkova, Interaction of heparin and tetraarginine in capillary electrophoresis: implication for analytical applications, Electrophoresis 41 (2020) 1826–1831, https://doi.org/10.1002/ elps.202000011. [89] V.A. Galievsky, A.S. Stasheuski, S.N. Krylov, Capillary electrophoresis for quantitative studies of biomolecular interactions, Anal. Chem. 87 (2015)

157–171, https://doi.org/10.1021/ac504219r. [90] E. Farcas¸, C. Bouckaert, A.C. Servais, J. Hanson, L. Pochet, M. Fillet, Partial filling affinity capillary electrophoresis as a useful tool for fragment-based drug discovery: a proof of concept on thrombin, Anal. Chim. Acta 984 (2017) 211–222, https://doi.org/10.1016/j.aca.2017.06.035. [91] S.E. Deeb, H. W¨atzig, D.A. El-Hady, Capillary electrophoresis to investigate biopharmaceuticals and pharmaceutically-relevant binding properties, Trends

Anal. Chem. 48 (2013) 112–131, https://doi.org/10.1016/j.trac.2013.04.005. [92] E.F. Romano Jr., J.P. Quirino, Frontal analysis capillary electrophoresis: recent advances and future perspectives, Bioanalysis 10 (2018) 1143–1159, https://doi. org/10.4155/bio-2018-0051. [93] L. Michalcov´a, H. Nevídalov´a, Z. Glatz, Toward an automated workflow for the study of plasma protein-drug interactions based on capillary electrophoresis- frontal analysis combined with in-capillary mixing of interacting partners,

J. Chromatogr. A 1635 (2021), 461734, https://doi.org/10.1016/j. chroma.2020.461734. [94] C. Malburet, L. Leclercq, J.F. Cotte, J. Thiebaud, H. Cottet, Study of interactions between antigens and polymeric adjuvants in vaccines by frontal analysis continuous capillary electrophoresis, Biomacromolecules 21 (2020) 3364–3373, https://doi.org/10.1021/acs.biomac.0c00782. [95] G. Ramírez-García, F. d’Orly´e, S. Guti´errez-Granados, M. Martínez-Alfaro,

N. Mignet, C. Richard, A. Varenne, Electrokinetic Hummel-Dreyer characterization of nanoparticle-plasma protein corona: the non-specific interactions between PEG-modified persistent luminescence nanoparticles and albumin, Colloids Surf. B Biointerfaces 159 (2017) 437–444, https://doi.org/

10.1016/j.colsurfb.2017.08.012. [96] H. Nevídalov´a, L. Michalcov´a, Z. Glatz, Capillary electrophoresis-based approaches for the study of affinity interactions combined with various sensitive and nontraditional detection techniques, Electrophoresis 40 (2019) 625–642, https://doi.org/10.1002/elps.201800367. [97] E. Farcas¸, J. Hanson, L. Pochet, M. Fillet, Capillary electrophoretic mobility shift displacement assay for the assessment of weak drug-protein interactions, Anal.

Chim. Acta 1034 (2018) 214–222, https://doi.org/10.1016/j.aca.2018.06.024. [98] S.A.C. Wren, R.C. Rowe, Theoretical aspects of chiral separation in capillary electrophoresis: II. The role of organic solvent, J. Chromatogr. A 609 (1992)

363–367, https://doi.org/10.1016/0021-9673(92)80181-S. [99] M. Dovhunov´a, M. Malý, P. Dubský, G.S. Gerlero, P.A. Kler, Generalized model of the linear theory of electromigration and its application to electrokinetic chromatography: capillary zone electrophoretic systems with complex-forming equilibria, J. Chromatogr. A 1610 (2020), 460595, https://doi.org/10.1016/j. chroma.2019.460595. [100] A. Gogolashvili, K. Lomsadze, L. Chankvetadze, N. Takaishvili, P. Peluso,

R. Dallocchio, A. Salgado, B. Chankvetadze, Separation of tetrahydrozoline enantiomers in capillary electrophoresis with cyclodextrin-type chiral selectors and investigation of chiral recognition mechanisms, J. Chromatogr. A 1643 (2021) 462084, https://doi.org/10.1016/j.chroma.2021.462084. [101] R. Zhou, S. Wang, K. Du, Y. Shang, J. He, J. Li, Y. Yao, Y.-x. Chang, Stereoselective enrichment and determination of citalopram enantiomers by cation-selective exhaustive injection and sweeping coupled with cyclodextrin modified electrokinetic chromatography, J. Chromatogr. A 1669 (2022), 462951, https:// doi.org/10.1016/j.chroma.2022.462951. [102] C. Davoine, M. Fillet, L. Pochet, Capillary electrophoresis as a fragment screening tool to cross-validate hits from chromogenic assay: application to FXIIa, Talanta

226 (2021) 122163, https://doi.org/10.1016/j.talanta.2021.122163. [103] Y.Y. Rawjee, R.L. Williams, G. Vigh, Capillary electrophoretic chiral separations using β-cyclodextrin as resolving agent: II. Bases: chiral selectivity as a function of pH and the concentration of β-cyclodextrin, J. Chromatogr. A 652 (1993)

233–245, https://doi.org/10.1016/0021-9673(93)80664-T. [104] P. Dubský, M. Dvoˇr´ak, M. Ansorge, Affinity capillary electrophoresis: the theory of electromigration, Anal. Bioanal. Chem. 408 (2016) 8623–8641, https://doi. org/10.1007/s00216-016-9799-y. [105] P. Dubský, M. ¨Ord¨ogov´a, M. Malý, M. Riesov´a, CEval: all-in-one software for data processing and statistical evaluations in affinity capillary electrophoresis,

J. Chromatogr. A 1445 (2016) 158–165, https://doi.org/10.1016/j. chroma.2016.04.004. [106] Y. Tanaka, S. Terabe, Estimation of binding constants by capillary electrophoresis,

J. Chromatogr. B 768 (2002) 81–92, https://doi.org/10.1016/S0378-4347(01)

00488-1. [107] Z. Chen, S.G. Weber, Determination of binding constants by affinity capillary electrophoresis, electrospray ionization mass spectrometry and phase-distribution methods, Trends Anal. Chem. 27 (2008) 738–748, https://doi.org/10.1016/j. trac.2008.06.008. [108] L. Wang, L. Zhao, W. Zhang, Y. Wu, Y. Liang, G. Guo, X. Wang, Investigation of metformin hydrochloride–bovine serum albumin interaction by narrow-bore capillary zone electrophoresis, Chem. Commun. 58 (2022) 2926–2929, https:// doi.org/10.1039/D1CC07067C. [109] L. Michalcov´a, Z. Glatz, Comparison of various capillary electrophoretic approaches for the study of drug-protein interaction with emphasis on minimal consumption of protein sample and possibility of automation, J. Separ. Sci. 38 (2015) 325–331, https://doi.org/10.1002/jssc.201400914. [110] C.L. Crihfield, C.J. Kristoff, L.M. Veltri, W.M. Penny, L.A. Holland, Semi- permanent cationic coating for protein separations, J. Chromatogr. A 1607 (2019)

460397, https://doi.org/10.1016/j.chroma.2019.460397. [111] J. Xing, M.-H. Wang, S. Wang, H.-L. Cong, Y.-Q. Shen, B. Yu, Analysis of proteins by capillary electrophoresis with a novel diazoresin/β-Cyclodextrin covalent capillary coating method, Ferroelectrics 563 (2020) 45–51, https://doi.org/

10.1080/00150193.2020.1760608. [112] P. Hemwech, A. Obma, S. Detsangiamsak, S. Wirasate, P. Chaiyen, P. Wilairat,

R. Chantiwas, Capillary electrophoresis-UV analysis using silica-layer coated capillary for separation of seven phenolic acids and caffeine and its application to tea analysis, SN Appl. Sci. 3 (2021) 872, https://doi.org/10.1007/s42452-021- 04849-1. [113] D. Raghav, S. Mahanty, K. Rathinasamy, Characterizing the interactions of the antipsychotic drug trifluoperazine with bovine serum albumin: probing the drug- protein and drug-drug interactions using multi-spectroscopic approaches,

Spectrochim. Acta Mol. Biomol. Spectrosc. 226 (2020), 117584, https://doi.org/

10.1016/j.saa.2019.117584. [114] M.S. Ali, H.A. Al-Lohedan, Spectroscopic and molecular docking investigation on the noncovalent interaction of lysozyme with saffron constituent “safranal,”, ACS

Omega 5 (2020) 9131–9141, https://doi.org/10.1021/acsomega.9b04291. [115] K. Nienhaus, G.U. Nienhaus, Probing heme protein-ligand interactions by UV/

Visible absorption spectroscopy, in: G. Ulrich Nienhaus (Ed.), Protein-ligand

Interactions: Methods and Applications, Humana Press, Totowa, NJ, 2005, pp. 215–241, https://doi.org/10.1385/1-59259-912-5:215. [116] L.O. Narhi, C.H. Li, R. Ramachander, J. Svitel, Y. Jiang, Optical spectroscopic methods for the analysis of biological macromolecules, in: N. Allewell, L.O. Narhi,

I. Rayment (Eds.), Molecular Biophysics for the Life Sciences, Springer New York,

New York, NY, 2013, pp. 33–90, https://doi.org/10.1007/978-1-4614-8548-3_3. [117] H.A. Benesi, J.H. Hildebrand, A spectrophotometric investigation of the interaction of iodine with aromatic hydrocarbons, J. Am. Chem. Soc. 71 (1949)

2703–2707, https://doi.org/10.1021/ja01176a030. [118] R. Wang, Z. Yu, Validity and reliability of Benesi-Hildebrand method, Acta Phys.

Chim. Sin. 23 (2007) 1353–1359, https://doi.org/10.1016/S1872-1508(07)

60071-0. [119] L. Condict, A. Hung, J. Ashton, S. Kasapis, High-temperature binding parameters and molecular dynamics of 4-hydroxybenzoic acid and β-casein complexes, determined via the method of continuous variation and fluorescence spectroscopy, Food Hydrocolloids 114 (2021), 106567, https://doi.org/10.1016/ j.foodhyd.2020.106567. [120] A.M. Abu-Dief, N.M. El-Metwaly, S.O. Alzahrani, F. Alkhatib, H.M. Abumelha,

T. El-Dabea, M.A.E.A.A. Ali El-Remaily, Structural, conformational and therapeutic studies on new thiazole complexes: drug-likeness and MOE- simulation assessments, Res. Chem. Intermed. 47 (2021) 1979–2002, https://doi. org/10.1007/s11164-020-04380-9.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 20 [121] L. Yong, M. Huang, Y. Wei, J. Xu, Z. Yi, Investigating the interaction between three perfluorinated carboxylic acids and the G protein-coupled estrogen receptor: spectroscopic analyses and computational simulations, Anal. Methods

12 (2020) 3944–3953, https://doi.org/10.1039/D0AY01052A. [122] M. Makarska-Bialokoz, Interactions of hemin with bovine serum albumin and human hemoglobin: a fluorescence quenching study, Spectrochim. Acta Mol.

Biomol. Spectrosc. 193 (2018) 23–32, https://doi.org/10.1016/j. saa.2017.11.063. [123] M.H. Gehlen, The centenary of the Stern-Volmer equation of fluorescence quenching: from the single line plot to the SV quenching map, J. Photochem.

Photobiol. C Photochem. Rev. 42 (2020), 100338, https://doi.org/10.1016/j. jphotochemrev.2019.100338. [124] E. Ciotta, P. Prosposito, R. Pizzoferrato, Positive curvature in Stern-Volmer plot described by a generalized model for static quenching, J. Lumin. 206 (2019)

518–522, https://doi.org/10.1016/j.jlumin.2018.10.106. [125] N.J. Greenfield, Using circular dichroism spectra to estimate protein secondary structure, Nat. Protoc. 1 (2006) 2876–2890, https://doi.org/10.1038/ nprot.2006.202. [126] F. Collin, O. Cerlati, F. Couderc, B. Lonetti, J.D. Marty, A.F. Mingotaud,

Multidisciplinary analysis of protein-lipid interactions and implications in neurodegenerative disorders, Trends Anal. Chem. 132 (2020), 116059, https:// doi.org/10.1016/j.trac.2020.116059. [127] D.J. Townsend, B. Mala, E. Hughes, R. Hussain, G. Siligardi, N.J. Fullwood, D.

A. Middleton, Circular dichroism spectroscopy identifies the β-adrenoceptor agonist salbutamol as a direct inhibitor of tau filament formation in vitro, ACS

Chem. Neurosci. 11 (2020) 2104–2116, https://doi.org/10.1021/ acschemneuro.0c00154. [128] J.A. Francis, M. Shalauddin, N.F.W. Ridzwan, S.B. Mohamad, W.J. Basirun,

S. Tayyab, Interaction mechanism of an antimalarial drug, sulfadoxine with human serum albumin, Spectrosc. Lett. 53 (2020) 391–405, https://doi.org/

10.1080/00387010.2020.1764588. [129] N.J. Greenfield, Circular dichroism analysis for protein-protein interactions, in:

H. Fu (Ed.), Protein-protein Interactions: Methods and Applications, Humana

Press, Totowa, NJ, 2004, pp. 55–77, https://doi.org/10.1385/1-59259-762-9:

055. [130] A.V. Hill, The possible effects of the aggregation of the molecules of hemoglobin on its dissociation curves, J. Physiol. 40 (1909) 4–7. [131] D.L. McClain, J.P. Binfet, M.G. Oakley, Evaluation of the energetic contribution of interhelical coulombic interactions for coiled coil helix orientation specificity11Edited by C. R. Matthews, J. Mol. Biol. 313 (2001) 371–383, https:// doi.org/10.1006/jmbi.2001.5044. [132] M.J. Swamy, R.S. Sankhala, B.P. Singh, Thermodynamic analysis of protein-lipid interactions by isothermal titration calorimetry, in: J.H. Kleinschmidt (Ed.), Lipid- protein Interactions: Methods and Protocols, Springer New York, New York, NY,

2019, pp. 71–89, https://doi.org/10.1007/978-1-4939-9512-7_4. [133] M.-L. Jobin, I.D. Alves, The contribution of differential scanning calorimetry for the study of peptide/lipid interactions, in: E. Ennifar (Ed.), Microcalorimetry of

Biological Molecules: Methods and Protocols, Springer New York, New York, NY,

2019, pp. 3–15, https://doi.org/10.1007/978-1-4939-9179-2_1. [134] V. Paketuryt˙e, A. Zubrien˙e, J.E. Ladbury, D. Matulis, Intrinsic thermodynamics of protein-ligand binding by isothermal titration calorimetry as aid to drug design, in: E. Ennifar (Ed.), Microcalorimetry of Biological Molecules: Methods and

Protocols, Springer New York, New York, NY, 2019, pp. 61–74, https://doi.org/

10.1007/978-1-4939-9179-2_5. [135] L. Baranauskiene, T.-C. Kuo, W.-Y. Chen, D. Matulis, Isothermal titration calorimetry for characterization of recombinant proteins, Curr. Opin. Biotechnol.

55 (2019) 9–15, https://doi.org/10.1016/j.copbio.2018.06.003. [136] E. Daems, G. Moro, R. Campos, K. De Wael, Mapping the gaps in chemical analysis for the characterisation of aptamer-target interactions, Trends Anal. Chem. 142 (2021), 116311, https://doi.org/10.1016/j.trac.2021.116311. [137] R.J. Falconer, B. Schuur, A.K. Mittermaier, Applications of isothermal titration calorimetry in pure and applied research from 2016 to 2020, J. Mol. Recogn. 34 (2021), e2901, https://doi.org/10.1002/jmr.2901. [138] X. Du, Y. Li, Y.L. Xia, S.M. Ai, J. Liang, P. Sang, X.L. Ji, S.Q. Liu, Insights into protein-ligand interactions: mechanisms, models, and methods, Int. J. Mol. Sci. 17 (2016) 144, https://doi.org/10.3390/ijms17020144. [139] C. R`afols, S. Zarza, E. Bosch, Molecular interactions between some non-steroidal anti-inflammatory drugs (NSAID׳s) and bovine (BSA) or human (HSA) serum albumin estimated by means of isothermal titration calorimetry (ITC) and frontal analysis capillary electrophoresis (FA/CE), Talanta 130 (2014) 241–250, https:// doi.org/10.1016/j.talanta.2014.06.060. [140] S. Wang, S. Yu, M. Siedler, P.M. Ihnat, D.I. Filoti, M. Lu, L. Zuo, A power compensated differential scanning calorimeter for protein stability characterization, Sensor. Actuator. B Chem. 256 (2018) 946–952, https://doi. org/10.1016/j.snb.2017.10.034. [141] S. Wang, S. Yu, M. Siedler, P.M. Ihnat, D.I. Filoti, M. Lu, L. Zuo, A power compensated differential scanning calorimeter for protein stability characterization, Sensor. Actuator. B Chem. 256 (2018) 946–952, https://doi. org/10.1016/j.snb.2017.10.034. [142] K. Bowers, N. Markova, Value of DSC in characterization and optimization of protein stability, in: E. Ennifar (Ed.), Microcalorimetry of Biological Molecules:

Methods and Protocols, Springer New York, New York, NY, 2019, pp. 33–44, https://doi.org/10.1007/978-1-4939-9179-2_3. [143] M. Amani, A.A. Moosavi-Movahedi, B.I. Kurganov, What can we get from varying scan rate in protein differential scanning calorimetry? Int. J. Biol. Macromol. 99 (2017) 151–159, https://doi.org/10.1016/j.ijbiomac.2017.02.010. [144] J.F. Berrío Escobar, D.M. M´arquez Fern´andez, C. Giordani, F. Castelli, M.

G. Sarpietro, Anomalous interaction of tri-acyl ester derivatives of uridine nucleoside with a l-α-dimyristoylphosphatidylcholine biomembrane model: a differential scanning calorimetry study, J. Pharm. Pharmacol. 71 (2019)

329–337, https://doi.org/10.1111/jphp.13038. [145] L. Villo, R. Risti, M. Reimund, K. Kukk, N. Samel, A. Lookene, Calorimetric approach for comparison of Angiopoietin-like protein 4 with other pancreatic lipase inhibitors, Biochim. Biophys. Acta Mol. Cell Biol. Lipids 1865 (2020)

158553, https://doi.org/10.1016/j.bbalip.2019.158553. [146] E.Y. Cotrina, A. Gimeno, J. Llop, J. Jim´enez-Barbero, J. Quintana, G. Valencia,

I. Cardoso, R. Prohens, G. Arsequell, Calorimetric studies of binary and ternary molecular interactions between transthyretin, Aβ peptides, and small-molecule chaperones toward an alternative strategy for alzheimer’s disease drug discovery,

J. Med. Chem. 63 (2020) 3205–3214, https://doi.org/10.1021/acs. jmedchem.9b01970. [147] M. Liu, T. Liu, Y. Shi, Y. Zhao, H. Yan, B. Sun, Q. Wang, Z. Wang, J. Han,

Comparative study on the interaction of oxyresveratrol and piceatannol with trypsin and lysozyme: binding ability, activity and stability, Food Funct. 10 (2019) 8182–8194, https://doi.org/10.1039/c9fo01888c. [148] A.V. Bychkova, M.V. Lopukhova, L.A. Wasserman, Y.N. Degtyarev, A.L. Kovarski,

S. Chakraborti, V.A. Mitkevich, The influence of pH and ionic strength on the interactions between human serum albumin and magnetic iron oxide nanoparticles, Int. J. Biol. Macromol. 194 (2022) 654–665, https://doi.org/

10.1016/j.ijbiomac.2021.11.110. [149] P.D. Ross, S. Subramanian, Thermodynamics of protein association reactions: forces contributing to stability, Biochemistry 20 (1981) 3096–3102, https://doi. org/10.1021/bi00514a017. [150] E. Schenckbecher, B. Meyer, E. Ennifar, ITC studies of ribosome/antibiotics interactions, in: E. Ennifar (Ed.), Microcalorimetry of Biological Molecules:

Methods and Protocols, Springer New York, New York, NY, 2019, pp. 89–98, https://doi.org/10.1007/978-1-4939-9179-2_7. [151] S. Nakamura, S.-i. Kidokoro, Protocols of IATC, DSC, and PPC: the multistate structural transition of cytochrome C, in: E. Ennifar (Ed.), Microcalorimetry of

Biological Molecules: Methods and Protocols, Springer New York, New York, NY,

2019, pp. 17–32, https://doi.org/10.1007/978-1-4939-9179-2_2. [152] M.K. Gilson, H.-X. Zhou, Calculation of protein-ligand binding affinities, Annu.

Rev. Biophys. Biomol. Struct. 36 (2007) 21–42, https://doi.org/10.1146/ annurev.biophys.36.040306.132550. [153] E.C. Lima, A.A. Gomes, H.N. Tran, Comparison of the nonlinear and linear forms of the van’t Hoff equation for calculation of adsorption thermodynamic parameters (ΔS◦and ΔH◦), J. Mol. Liq. 311 (2020), 113315, https://doi.org/

10.1016/j.molliq.2020.113315. [154] K. Gao, R. Oerlemans, M.R. Groves, Theory and applications of differential scanning fluorimetry in early-stage drug discovery, Biophys. Rev. 12 (2020)

85–104, https://doi.org/10.1007/s12551-020-00619-2. [155] T.S. Paiva, R.G. Nascimento, L.A.A. Veríssimo, R.C.F. Bonomo, C.M. Veloso, J.

V. de Resende, R.d.C.I. Fontan, Adsorption isotherms and thermodynamic properties of a butyl functionalized hydrophobic macroporous cryogel, Braz. J.

Chem. Eng. (2021) 1–9, https://doi.org/10.1007/s43153-021-00173-8. [156] R.B. Gudhka, D.J. Roush, S.M. Cramer, A thermodynamic evaluation of antibody- surface interactions in multimodal cation exchange chromatography,

J. Chromatogr. A 1628 (2020) 461479, https://doi.org/10.1016/j. chroma.2020.461479. [157] M. Bagheri, M.H. Fatemi, Fluorescence spectroscopy, molecular docking and molecular dynamic simulation studies of HSA-Aflatoxin B1 and G1 interactions,

J. Lumin. 202 (2018) 345–353, https://doi.org/10.1016/j.jlumin.2018.05.066. [158] J.C. Simoes-Cardoso, N. Yoshimoto, S. Yamamoto, Thermodynamic analysis of polyphenols retention in polymer resin chromatography by van’t Hoff plot and isothermal titration calorimetry, J. Chromatogr. A 1608 (2019), 460405, https:// doi.org/10.1016/j.chroma.2019.460405. [159] T.A.G. Noctor, I.W. Wainer, D.S. Hage, Allosteric and competitive displacement of drugs from human serum albumin by octanoic acid, as revealed by high- performance liquid affinity chromatography, on a human serum albumin-based stationary phase, J. Chromatogr. B Biomed. Sci. Appl. 577 (1992) 305–315, https://doi.org/10.1016/0378-4347(92)80252-L. [160] K.S. Joseph, D.S. Hage, Characterization of the binding of sulfonylurea drugs to

HSA by high-performance affinity chromatography, J. Chromatogr. B 878 (2010)

1590–1598, https://doi.org/10.1016/j.jchromb.2010.04.019. [161] G. Famiglini, P. Palma, V. Termopoli, A. Cappiello, The history of electron ionization in LC-MS, from the early days to modern technologies: a review, Anal.

Chim. Acta 1167 (2021) 338350, https://doi.org/10.1016/j.aca.2021.338350. [162] B. Hu, Z.-P. Yao, Electrospray ionization mass spectrometry with wooden tips: a review, Anal. Chim. Acta (2021), 339136, https://doi.org/10.1016/j. aca.2021.339136. [163] A. Tailor, J.C. Waddington, X. Meng, B.K. Park, Mass spectrometric and functional aspects of drug–protein conjugation, Chem. Res. Toxicol. 29 (2016) 1912–1935, https://doi.org/10.1021/acs.chemrestox.6b00147. [164] G. Zhu, Y. Shao, Y. Liu, T. Pei, L. Li, D. Zhang, G. Guo, X. Wang, Single-cell metabolite analysis by electrospray ionization mass spectrometry, Trends Anal.

Chem. 143 (2021), 116351, https://doi.org/10.1016/j.trac.2021.116351. [165] S. Zhang, C.K. Van Pelt, D.B. Wilson, Quantitative determination of noncovalent binding interactions using automated nanoelectrospray mass spectrometry, Anal.

Chem. 75 (2003) 3010–3018, https://doi.org/10.1021/ac034089d. [166] A. Sinz, Investigation of protein-ligand interactions by mass spectrometry,

ChemMedChem 2 (2007) 425–431, https://doi.org/10.1002/cmdc.200600298.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 21 [167] Q. Zheng, Y. Tian, X. Ruan, H. Chen, X. Wu, X. Xu, G. Wang, H. Hao, H. Ye,

Probing specific ligand-protein interactions by native-denatured exchange mass spectrometry, Anal. Chim. Acta 1036 (2018) 58–65, https://doi.org/10.1016/j. aca.2018.07.072. [168] A.C. Leney, A.J.R. Heck, Native mass spectrometry: what is in the name? J. Am.

Soc. Mass Spectrom. 28 (2017) 5–13, https://doi.org/10.1007/s13361-016-1545- 3. [169] S. Tamara, M.A. den Boer, A.J.R. Heck, High-resolution native mass spectrometry,

Chem. Rev. (2021), https://doi.org/10.1021/acs.chemrev.1c00212. [170] R.H.H.v.d. Heuvel, A.J.R. Heck, Native protein mass spectrometry: from intact oligomers to functional machineries, Curr. Opin. Chem. Biol. 8 (2004) 519–526, https://doi.org/10.1016/j.cbpa.2004.08.006. [171] J.L. Bennett, G.T.H. Nguyen, W.A. Donald, Protein–small molecule interactions in native mass spectrometry, Chem. Rev. (2021), https://doi.org/10.1021/acs. chemrev.1c00293. [172] I.K. Webb, Recent technological developments for native mass spectrometry,

Biochim. Biophys. Acta Protein Proteonomics 1870 (2022), 140732, https://doi. org/10.1016/j.bbapap.2021.140732. [173] K.J. Pacholarz, R.A. Garlish, R.J. Taylor, P.E. Barran, Mass spectrometry based tools to investigate protein-ligand interactions for drug discovery, Chem. Soc.

Rev. 41 (2012) 4335–4355, https://doi.org/10.1039/c2cs35035a. [174] S. Sanglier, C. Atmanene, G. Chevreux, A. Van Dorsselaer, Nondenaturing mass spectrometry to study noncovalent protein/protein and protein/ligand complexes: technical aspects and application to the determination of binding stoichiometries, in: J.D. Thompson, M. Ueffing, C. Schaeffer-Reiss (Eds.),

Functional Proteomics: Methods and Protocols, Humana Press, Totowa, NJ, 2008, pp. 217–243, https://doi.org/10.1007/978-1-59745-398-1_15. [175] R.J. Rose, E. Damoc, E. Denisov, A. Makarov, A.J. Heck, High-sensitivity orbitrap mass analysis of intact macromolecular assemblies, Nat. Methods 9 (2012)

1084–1086, https://doi.org/10.1038/nmeth.2208. [176] E. Lubomirsky, A. Khodabandeh, J. Preis, M. Susewind, T. Hofe, E.F. Hilder, R.

D. Arrua, Polymeric stationary phases for size exclusion chromatography: a review, Anal. Chim. Acta 1151 (2021) 338244, https://doi.org/10.1016/j. aca.2021.338244. [177] Y. Yan, T. Xing, S. Wang, T.J. Daly, N. Li, Coupling mixed-mode size exclusion chromatography with native mass spectrometry for sensitive detection and quantitation of homodimer impurities in bispecific IgG, Anal. Chem. 91 (2019)

11417–11424, https://doi.org/10.1021/acs.analchem.9b02793. [178] I.K. Ventouri, D.B.A. Malheiro, R.L.C. Voeten, S. Kok, M. Honing, G.W. Somsen,

R. Haselberg, Probing protein denaturation during size-exclusion chromatography using native mass spectrometry, Anal. Chem. 92 (2020)

4292–4300, https://doi.org/10.1021/acs.analchem.9b04961. [179] J.T.B. Stariha, R.M. Hoffmann, D.J. Hamelin, J.E. Burke, Probing protein- membrane interactions and dynamics using hydrogen-deuterium exchange mass spectrometry (HDX-MS), in: T. Daviter, C.M. Johnson, S.H. McLaughlin, M.

A. Williams (Eds.), Protein-ligand Interactions: Methods and Applications,

Springer US, New York, NY, 2021, pp. 465–485, https://doi.org/10.1007/978-1- 0716-1197-5_22. [180] D. Narang, C. Lento, D.J. Wilson, HDX-MS: an analytical tool to capture protein motion in action, Biomedicines 8 (2020) 224, https://doi.org/10.3390/ biomedicines8070224. [181] G.R. Masson, M.L. Jenkins, J.E. Burke, An overview of hydrogen deuterium exchange mass spectrometry (HDX-MS) in drug discovery, Expet Opin. Drug

Discov. 12 (2017) 981–994, https://doi.org/10.1080/17460441.2017.1363734. [182] I. Oganesyan, C. Lento, D.J. Wilson, Contemporary hydrogen deuterium exchange mass spectrometry, Methods 144 (2018) 27–42, https://doi.org/10.1016/j. ymeth.2018.04.023. [183] G.R. Masson, J.E. Burke, N.G. Ahn, G.S. Anand, C. Borchers, S. Brier, G.M. Bou- Assaf, J.R. Engen, S.W. Englander, J. Faber, R. Garlish, P.R. Griffin, M.L. Gross,

M. Guttman, Y. Hamuro, A.J.R. Heck, D. Houde, R.E. Iacob, T.J.D. Jørgensen, I.

A. Kaltashov, J.P. Klinman, L. Konermann, P. Man, L. Mayne, B.D. Pascal,

D. Reichmann, M. Skehel, J. Snijder, T.S. Strutzenberg, E.S. Underbakke,

C. Wagner, T.E. Wales, B.T. Walters, D.D. Weis, D.J. Wilson, P.L. Wintrode,

Z. Zhang, J. Zheng, D.C. Schriemer, K.D. Rand, Recommendations for performing, interpreting and reporting hydrogen deuterium exchange mass spectrometry (HDX-MS) experiments, Nat. Methods 16 (2019) 595–602, https://doi.org/

10.1038/s41592-019-0459-y. [184] O. Ozohanics, A. Ambrus, Hydrogen-deuterium exchange mass spectrometry: a novel structural biology approach to structure, dynamics and interactions of proteins and their complexes, Life 10 (2020) 286, https://doi.org/10.3390/ life10110286. [185] N. Au - Hentze, M.P. Au - Mayer, Analyzing protein dynamics using hydrogen exchange mass spectrometry, JoVE (2013), e50839, https://doi.org/10.3791/

50839. [186] E. Trabjerg, Z.E. Nazari, K.D. Rand, Conformational analysis of complex protein states by hydrogen/deuterium exchange mass spectrometry (HDX-MS): challenges and emerging solutions, Trends Anal. Chem. 106 (2018) 125–138, https://doi.org/10.1016/j.trac.2018.06.008. [187] S. Indelicato, D. Bongiorno, L. Ceraulo, Recent approaches for chemical speciation and analysis by electrospray ionization (ESI) mass spectrometry, Front. Chem. 8 (2021), https://doi.org/10.3389/fchem.2020.625945, 625945-625945. [188] T. F¨orster, Zwischenmolekulare energiewanderung und fluoreszenz, Ann. Phys.

437 (1948) 55–75, https://doi.org/10.1002/andp.19484370105. [189] M. Schürmann, P. Janning, S. Ziegler, H. Waldmann, Small-molecule target engagement in cells, Cell Chem. Biol. 23 (2016) 435–441, https://doi.org/

10.1016/j.chembiol.2016.03.008. [190] P.R. Gleason, P.I. Kelly, D.W. Grisingher, J.H. Mills, An intrinsic FRET sensor of protein-ligand interactions, Org. Biomol. Chem. 18 (2020) 4079–4084, https:// doi.org/10.1039/D0OB00793E. [191] J.A. Broussard, B. Rappaz, D.J. Webb, C.M. Brown, Fluorescence resonance energy transfer microscopy as demonstrated by measuring the activation of the serine/threonine kinase Akt, Nat. Protoc. 8 (2013) 265–281, https://doi.org/

10.1038/nprot.2012.147. [192] T. Chatterjee, A. Pal, S. Dey, B.K. Chatterjee, P. Chakrabarti, Interaction of virstatin with human serum albumin: spectroscopic analysis and molecular modeling, PLoS One 7 (2012), https://doi.org/10.1371/journal.pone.0037468 e37468-e37468. [193] K.E. Sapsford, L. Berti, I.L. Medintz, Materials for fluorescence resonance energy transfer analysis: beyond traditional donor–acceptor combinations, Angew.

Chem. Int. Ed. 45 (2006) 4562–4589, https://doi.org/10.1002/anie.200503873. [194] S. Fery-Forgues, D. Lavabre, Are fluorescence quantum yields so tricky to measure? A demonstration using familiar stationery products, J. Chem. Educ. 76 (1999) 1260, https://doi.org/10.1021/ed076p1260. [195] K. Nawara, J. Waluk, Improved method of fluorescence quantum yield determination, Anal. Chem. 89 (2017) 8650–8655, https://doi.org/10.1021/acs. analchem.7b02013. [196] D.M. Arvapalli, A.T. Sheardy, K.C. Alapati, J. Wei, High quantum yield fluorescent carbon nanodots for detection of Fe (III) ions and electrochemical study of quenching mechanism, Talanta 209 (2020) 120538, https://doi.org/

10.1016/j.talanta.2019.120538. [197] D.F. Swinehart, The Beer-Lambert law, J. Chem. Educ. 39 (1962) 333, https:// doi.org/10.1021/ed039p333. [198] G.G. Ariga, P.N. Naik, S.A. Chimatadar, S.T. Nandibewoor, Interactions between epinastine and human serum albumin: investigation by fluorescence, UV-vis, FT- IR, CD, lifetime measurement and molecular docking, J. Mol. Struct. 1137 (2017)

485–494, https://doi.org/10.1016/j.molstruc.2016.12.066. [199] H. Dezhampanah, R. Firouzi, Z. Moradi Shoeili, R. Binazir, Intermolecular investigation on interaction of two ternary copper(II) Schiff base complexes with bovine serum albumin, J. Mol. Struct. 1205 (2020), 127557, https://doi.org/

10.1016/j.molstruc.2019.127557. [200] X. Yu, X. Cai, S. Li, L. Luo, J. Wang, M. Wang, L. Zeng, Studies on the interactions of theaflavin-3,3′-digallate with bovine serum albumin: multi-spectroscopic analysis and molecular docking, Food Chem. 366 (2022), 130422, https://doi. org/10.1016/j.foodchem.2021.130422. [201] J.B.F. Lloyd, Synchronized excitation of fluorescence emission spectra, Nat. Phys.

Sci. (Lond.) 231 (1971) 64–65, https://doi.org/10.1038/physci231064a0. [202] J.C. Andre, P. Baudot, M. Niclause, Application of synchronous excitation spectrofluorimetry to drug analysis, Clin. Chim. Acta 76 (1977) 55–66, https:// doi.org/10.1016/0009-8981(77)90118-8. [203] B.O. Boateng, S. Elcoroaristizabal, A.G. Ryder, Development of a rapid polarized total synchronous fluorescence spectroscopy (pTSFS) method for protein quantification in a model bioreactor broth, Biotechnol. Bioeng. 118 (2021)

1805–1817, https://doi.org/10.1002/bit.27694. [204] W. Cheng, J. Ma, S. Wang, R. Lou, S. Wu, J. He, H. Kang, L. Liu, F. Xiao,

Interaction mechanism between resveratrol and ovalbumin based on fluorescence spectroscopy and molecular dynamic simulation, LWT - Food Sci. Technol. 146 (2021), 111455, https://doi.org/10.1016/j.lwt.2021.111455. [205] H. Zheng, P. Wan, S. Qi, H. Chen, H. Zhai, Investigating the interaction between

DNA-templated gold nanoclusters and HSA via spectroscopy, New J. Chem. 44 (2020) 14060–14066, https://doi.org/10.1039/D0NJ02075C. [206] X. Li, R. Xu, Z. Cheng, Z. Song, Z. Wang, H. Duan, X. Wu, T. Ni, Comparative study on the interaction between flavonoids with different core structures and hyaluronidase, Spectrochim. Acta Mol. Biomol. Spectrosc. 262 (2021), 120079, https://doi.org/10.1016/j.saa.2021.120079. [207] T. Vo-Dinh, Multicomponent analysis by synchronous luminescence spectrometry, Anal. Chem. 50 (1978) 396–401, https://doi.org/10.1021/ ac50025a010. [208] M.E. Pacheco, L. Bruzzone, Synchronous fluorescence spectrometry: conformational investigation or inner filter effect? J. Lumin. 137 (2013) 138–142, https://doi.org/10.1016/j.jlumin.2012.12.056. [209] A. Jouyban, E. Rahimpour, Using constant-wavelength synchronous fluorescence spectroscopy in nanoparticle-based sensors: a minireview, Anal. Methods 13 (2021) 968–973, https://doi.org/10.1039/D0AY02096F. [210] M.S. Ali, J. Muthukumaran, M. Jain, T. Santos-Silva, H.A. Al-Lohedan, N.S. Al- Shuail, Molecular interactions of cefoperazone with bovine serum albumin: extensive experimental and computational investigations, J. Mol. Liq. 337 (2021),

116354, https://doi.org/10.1016/j.molliq.2021.116354. [211] C. Ma, L. Li, J. Gu, C. Zhu, G. Chen, Inflection point of the fluorescence excitation spectra induced by secondary inner filter effect, Spectrosc. Lett. 51 (2018)

319–323, https://doi.org/10.1080/00387010.2018.1461655. [212] E.P. Kirby, Fluorescence instrumentation and methodology, in: R.F. Steiner,

I. Weinryb (Eds.), Excited States of Proteins and Nucleic Acids, Springer US,

Boston, MA, 1971, pp. 31–56, https://doi.org/10.1007/978-1-4684-1878-1_2. [213] A.J. Miles, S.G. Ramalli, B.A. Wallace, DichroWeb, a website for calculating protein secondary structure from circular dichroism spectroscopic data, Protein

Sci. 31 (2022) 37–46, https://doi.org/10.1002/pro.4153. [214] J.C. Ezerski, P. Zhang, N.C. Jennings, M.N. Waxham, M.S. Cheung, Molecular dynamics ensemble refinement of intrinsically disordered peptides according to deconvoluted spectra from circular dichroism, Biophys. J. 118 (2020)

1665–1678, https://doi.org/10.1016/j.bpj.2020.02.015.

L. Wang et al.

Analytica Chimica Acta 1219 (2022) 340012 22 [215] J. De Meutter, E. Goormaghtigh, FTIR imaging of protein microarrays for high throughput secondary structure determination, Anal. Chem. 93 (2021)

3733–3741, https://doi.org/10.1021/acs.analchem.0c03677. [216] Y. Li, Y. Ren, Y. Bai, M. Ikram, Y. Xu, Y. Wang, Z. Zhang, Double-layer chiral system with induced circular dichroism by near-field coupling, J. Phys. Chem. C

125 (2021) 25851–25858, https://doi.org/10.1021/acs.jpcc.1c08049. [217] D.-i. Kim, S.-j. Han, Y.-b. Lim, Unique behaviour of the α-helix in bending deformation, Chem. Commun. 58 (2022) 4368–4371, https://doi.org/10.1039/

D2CC00008C. [218] A. Kefala, M. Amprazi, E. Mylonas, D. Kotsifaki, M. Providaki, C. Pozidis,

M. Fotiadou, M. Kokkinidis, Probing protein folding with sequence-reversed α-helical bundles, Int. J. Mol. Sci. 22 (2021), https://doi.org/10.3390/ ijms22041955. [219] M. Sasmal, A.S.M. Islam, R. Bhowmick, D. Maiti, A. Dutta, M. Ali, Site-selective interaction of human serum albumin with 4-Chloro-7-nitro-1,2,3-benzoxadiazole modified olanzapine derivative and effect of β-Cyclodextrin on binding: in the light of spectroscopy and molecular docking, ACS Appl. Bio Mater. 2 (2019)

3551–3561, https://doi.org/10.1021/acsabm.9b00429. [220] Z. Tian, F. Zang, W. Luo, Z. Zhao, Y. Wang, X. Xu, C. Wang, Spectroscopic study on the interaction between mononaphthalimide spermidine (MINS) and bovine serum albumin (BSA), J. Photochem. Photobiol. B Biol. 142 (2015) 103–109, https://doi.org/10.1016/j.jphotobiol.2014.10.013. [221] J. Lyu, K. Wang, M. Ye, Modification-free approaches to screen drug targets at proteome level, Trends Anal. Chem. 124 (2020), 115574, https://doi.org/

10.1016/j.trac.2019.06.024. [222] A.A. Kermani, A guide to membrane protein X-ray crystallography, FEBS J. 288 (2021) 5788–5804, https://doi.org/10.1111/febs.15676. [223] N. Sepay, P.C. Saha, Z. Shahzadi, A. Chakraborty, U.C. Halder, A crystallography- based investigation of weak interactions for drug design against COVID-19, Phys.

Chem. Chem. Phys. 23 (2021) 7261–7270, https://doi.org/10.1039/

D0CP05714B. [224] L. Pontoriero, M. Schiavina, M.G. Murrali, R. Pierattelli, I.C. Felli, Monitoring the interaction of α-synuclein with calcium ions through exclusively heteronuclear nuclear magnetic resonance experiments, Angew. Chem. Int. Ed. 59 (2020)

18537–18545, https://doi.org/10.1002/anie.202008079. [225] R.B. Gudhka, C.L. Bilodeau, S.A. McCallum, M.A. McCoy, D.J. Roush, M.

A. Snyder, S.M. Cramer, Identification of preferred multimodal ligand-binding regions on IgG1 FC using nuclear magnetic resonance and molecular dynamics simulations, Biotechnol. Bioeng. 118 (2021) 809–822, https://doi.org/10.1002/ bit.27611. [226] S. Sheibani, K. Basu, A. Farnudi, A. Ashkarran, M. Ichikawa, J.F. Presley, K.H. Bui,

M.R. Ejtehadi, H. Vali, M. Mahmoudi, Nanoscale characterization of the biomolecular corona by cryo-electron microscopy, cryo-electron tomography, and image simulation, Nat. Commun. 12 (2021) 573, https://doi.org/10.1038/ s41467-020-20884-9. [227] M.J. Robertson, J.G. Meyerowitz, G. Skiniotis, Drug discovery in the era of cryo- electron microscopy, Trends Biochem. Sci. 47 (2022) 124–135, https://doi.org/

10.1016/j.tibs.2021.06.008.

Dongtang Zhang received his Ph.D. degree from Beijing Uni­ versity of Technology. Now he is working as a research assis­ tant at Beijing University of Technology. His research interests are microfluidic systems for chemical and biological applica­ tions, materials synthesis and processing.

Guangsheng Guo received his Ph.D. from Beijing University of

Chemical Technology. He is currently a professor in Beijing

University of Technology. He was awarded the 2014 Special

Government Allowances of the State Council of China. His research interests include nanotechnology/nanomaterials, micro/nano fluidics, and micro/nanoscale analysis.

Lijuan Wang is a Ph.D. student, currently studying in the

Department of Chemistry and Biology, Beijing University of

Technology. Her Ph.D. supervisor is professor Xiayan Wang.

Her research interests are in the development of methods for the study of drug-protein interactions.

Wenmei Zhang received her Ph.D. degree in Engineering from

Beijing University of Technology in 2020, and now is a post- doctoral fellow at Beijing University of Technology. Her doctoral research involved developing methods for microRNAs detection directly based on micro/nano channels. The current research interests include single-cell transcriptomics and proteomics.

Xiayan Wang received her Ph.D. degree from the University of

Science and Technology of China. She is currently a professor at the Beijing University of Technology. She was awarded the

National Funds for Distinguished Young Scholars in 2016 and

Beijing Outstanding Young Scientist in 2019. Her research in­ terests include micro/nanoscale analysis, micro/nano capillary chromatography, micro/nano fluidics, and the development of new micro/nano analytical instruments.

Yunlong Shao received his Ph.D. degree in Engineering from

Beijing University of Technology in 2020, and now is a post- doctoral fellow at Beijing University of Technology. His research interest is single-cell metabolomics mass spectrometry.

L. Wang et al.