Unraveling the Energetics and Mode of the Recognition of Antibiotics Tetracycline and Rolitetracycline by Bovine Serum Albumin

Authors


Corresponding author: Nand Kishore, nandk@chem.iitb.ac.in

Abstract

An understanding of the detailed energetics and mechanism of the binding of drugs with target proteins is essential for devising guidelines to synthesize new drugs. Binding of the antibiotic drugs tetracycline and rolitetracycline with serum albumin has been studied by a combination of isothermal titration calorimetry, differential scanning calorimetry, steady-state and time-resolved fluorescence, and circular dichroism spectroscopies. Both tetracycline and rolitetracycline bind to bovine serum albumin in a sequential manner with first binding being the major binding event with an association constant of the order of 104 for tetracycline and 103 for rolitetracycline, respectively. Ionic strength dependence and binding in the presence of tetrabutylammonium bromide and sucrose indicate involvement of a mix of hydrophobic, ionic, and hydrogen bonding interactions. The isothermal titration calorimetry results for the binding of these drugs to bovine serum albumin in the presence of warfarin and in the presence of each other indicate that both these drugs share binding site 2 on bovine serum albumin. The differential scanning calorimetry results provide quantitative information on the effect of drugs on the stability of bovine serum albumin. A comparison of isothermal titration calorimetry and fluorescence results demonstrates that the former technique has been able to explain the sequential binding events that can be missed by the fluorescence measurements.

Abbreviations:
BSA

bovine serum albumin

DSC

differential scanning calorimetry

HTAB

hexadecyltrimethylammonium bromide

ITC

isothermal titration calorimetry

RTC

rolitetracycline

SDS

sodium dodecyl sulfate

TC

tetracycline

TX-100

triton X-100

It is known that the interactions between drug and plasma proteins are responsible for controlling the free drug concentration in blood which affects the drug distribution, metabolism, elimination, and pharmacological activity (1,2). Thus, the drug discovery and development process require a detailed understanding of the drug–plasma protein interactions both qualitatively and quantitatively. Owing to the binding capabilities of serum albumins, more research has been focused on plasma proteins for many years (1–7). It is known that serum albumin is the most abundant of the proteins in blood plasma, and because of its diverse binding properties, it serves as a transport protein for several endogenous and exogenous molecules (8). It is further reported that human serum albumin accounts for nearly 60% of the total protein, providing about 80% contribution to the osmotic pressure of the blood (9). There is a high level of sequence identity (approximately 76%) between bovine serum albumin (BSA) and human serum albumin, the major difference being two tryptophan residues (W131 and W214) present in the former while only one tryptophan (W214) in the latter (10,11).

Tetracyclines (TC) belong to a class of antibiotics drugs that have a common base chemical structure with different substituent attached to it (12). Such molecules belong to broad-spectrum antibiotics that are known to inhibit prokaryotic translation by interfering with binding of the aminoacyl-tRNA to the ribosomal A site (13). Tetracycline (Figure 1A) is an important member of antibiotics, and it was the first therapeutically superior drug to be made by chemical alteration of an antibiotic produced by microbial metabolism (14). Tetracyclines have been reported (15) to be effective against a wide range of micro-organisms including Gram-positive and Gram negative bacteria chlamydiae, micoplasmas, rickettsial, and protozoan. Besides this, it has wide applications in clinical practice of veterinary medicine, animal nutrition, and as feed additives (16). TC and rolitetracycline (RTC) have been reported (17) to be potent inhibitors of human immunodeficiency virus type I integrase, the latter being stronger.

Figure 1.

 Molecular structure of tetracycline (A) and rolitetracycline (B).

Rolitetracycline (Figure 1B), a derivative of TC, is reported (18) to be more soluble and possibly less irritating to tissue than TC. Therefore, it was suggested (18) that RTC may be given intravenously or intramuscularly in serious bacterial infections when oral administration is not practicable. Although it is an antibiotic, it has been described (19) that RTC can effectively inhibit Aβ fibril formation, which is associated with Alzheimer’s disease. RTC has also been reported (20) to significantly inhibit the dengue virus propagation.

To the best of our knowledge, the binding thermodynamics of TC and RTC with BSA has not been reported in the literature, although rare qualitative studies are available for TC (13,16). A detailed understanding of the binding of drugs to albumin has significant physiological importance as this protein is a major component in the body for transport, modulation and inactivation of metabolites, and various activities of the drug. It is further important to understand the nature of interactions and the strength of drug association with the protein, which are useful in drug delivery, pharmacokinetics, and pharmacodynamics of the drugs.

In this work, isothermal titration calorimetry (ITC) has been used to quantitatively determine the thermodynamic parameters accompanying the binding of TC and RTC with BSA. The nature of interactions involved in the binding process has been investigated by studying the binding as a function of temperature and in the presence of co-solutes having ionic, hydrophobic, and mixed hydrophilic–hydrophobic properties. Fluorescence spectroscopy has been used as an alternate method to determine the binding parameters and to understand the mechanism of binding. The conformational changes in the protein and its stability upon drug binding have been also been investigated.

Experimental Methodology

Materials

Bovine serum albumin, TC, RTC, triton X-100 (TX-100), and sodium chloride, of the best available purity grade, were purchased from Sigma-Aldrich Chemical Company, (Mumbai, India). Tetrabutylammonium bromide (TBAB) was purchased from Spectrochem Pvt. Ltd. (Mumbai, India) and sucrose was purchased from Merk Specialities Pvt. Ltd. (Mumbai, India). The water used to prepare the solutions was double distilled and further deionized using a Cole–Parmer mixed-bed ion-exchange column. All the experiments were performed at pH 7.4 in 20 mm phosphate buffer. The protein stock solutions were prepared by extensive overnight dialysis at 4 °C against the phosphate buffer. The pH of the solutions was measured on a Standard Control Dynamics pH meter at ambient temperature. The concentration of protein was determined on a Jasco V-550 UV–visible double-beam spectrophotometer, using a value of inline image = 6.8 at 280 nm (21). The values of the binding parameters depend on the accuracy of the concentrations of the protein and the ligand used in the analysis of the data. An error of 10% in the protein concentration can affect the values of the association constant (K) and enthalpy of binding (ΔH) up to 20%. The drug solutions were also prepared in the dialysate buffer.

Isothermal titration calorimetry

The thermodynamic parameters for the BSA–drug binding interactions were determined using an isothermal titration calorimeter (VP-ITC; Microcal, Northampton, MA, USA). The ligand solution was titrated with the sample cell in aliquots using a 250-μL rotating stirrer–syringe, and the reference cell contained the buffer. Each experiment consisted of 25 consecutive injections of 10 μL of 4.516 mm TC solution into 0.09 mm BSA or 4.516 mm RTC solution into 0.15 mm BSA solution in the cell. Each injection had a duration of 6 seconds, and the time between the consecutive injections was kept at 4-min interval. To analyze the titration heat profiles, the Origin 7.0 software, provided by Microcal, was used. The total heat content Q of the solution contained in the active cell volume V0 (determined relative to zero for the unliganded species) at fractional saturation Θ after the ith injection is determined (5) from

image(1)

where Mt is the total concentration of the macromolecule, n is the number of binding sites in the macromolecule, and ΔH is the molar heat of ligand binding. The heat released ΔQ (i) from the ith injection for an injection volume dVi is then given (5) by the following equation:

image(2)

The thermodynamic parameters associated with the binding process – binding constant (K), enthalpy (ΔH), entropy (ΔS), and stoichiometry (n) of the binding – were determined by the best fit of the chosen model to the experimental data points.

Differential scanning calorimetry

The differential scanning calorimetry (DSC) experiments were performed on a Nano DSC (TA Instruments, New Castle, DE, USA) at scan rate of 1 °C/min. All the solutions were degassed using Barnstead/Thermolyne degassing unit from Nuova before loading into the calorimetric vessels. The excess heat capacity scans for the protein transitions were obtained by subtracting the corresponding scans of buffer verses buffer from that of protein solutions verses buffer. When the experiments were conducted in the presence of a drug, an equivalent amount of the drug was added to the reference cell as that present in the protein solution of the sample cell. The DSC data were analyzed by Nanoanalyser according to single or multiple transitions. To check the reversibility of the thermal transitions, the sample in the first scan was heated to a little over the complete denaturation temperature, cooled immediately, and reheated at the same scan rate. It was observed that all the thermal denaturations of BSA in the absence and the presence of drugs are irreversible.

Fluorescence spectroscopy

Steady-state fluorescence measurements

The fluorescence quenching experiments were carried out on a Cary Eclipse fluorescence Spectrophotometer. The excitation and emission slit width were fixed at 5 nm, and the scan speed used was kept at medium. The concentration of BSA was kept at 0.1 mg/mL in the steady-state fluorescence experiments. For the selective excitation of the tryptophan molecules, the excitation wavelength was 295 nm. The reference spectra containing the same amount of ligand as in the sample were subtracted from that of the complex.

Lifetime measurements

Time-resolved fluorescence measurements were performed at the magic angle using a pulsed Nano-LED-based time-correlated single-photon counting fluorescence spectrometer from IBH, UK, with λex = 295 nm and λem = 340 nm. The full width at half maximum of the instrument response function is 250 ps, and the resolution is 56 ps per channel. The data were fitted to bi-exponential function after deconvolution of the instrument response function by an iterative re-convolution technique. This was achieved by the IBH DAS 6.0 data analysis software, using reduced chi-squared and weighted residuals as parameters for goodness of the fit.

Circular Dichroism spectroscopy

The spectra in the far-UV CD (190–260 nm) and the near-UV CD (260–320 nm) regions were obtained on a JASCO-810 spectropolarimeter for observing the alterations in the secondary and tertiary structure of the protein. The concentration of BSA and path lengths used was 5 μm and 0.2 cm, respectively, for the far-UV CD experiments, and 15 μm and 1 cm, respectively, for near-UV CD experiments. The spectropolarimeter was thoroughly purged with nitrogen before starting the experiments. Each spectrum was baseline corrected and was taken as an average of three accumulations at a scan rate of 500 nm/min and a response time of 1 seconds. The molar ellipticity inline image was calculated from the observed ellipticity θ as

image(3)

where c is the concentration of the protein in mol/dm3 and l is the path length of the cell in centimeters.

Results and Discussion

Isothermal titration calorimetry of the binding of TC and RTC with BSA

A representative ITC heat profile for the titration of 4.516 mm TC with 0.09 mm BSA at pH 7.4 and 298.15 K is shown in Figure 2A. The upper panel of Figure 2A shows the raw data for 25 sequential injections of TC into the protein solution, and the lower panel shows the integrated binding heats as a function of the [TC]/[BSA] molar ratio. The data fitted best to a two sequential binding site model. The smooth solid line shown in the Figure 2 is the best fit to the experimental data. The binding parameters thus measured in the temperature range of 15–37 °C are reported in Table 1. Each value in this table is an average of two to three independent measurements. The reported standard deviation with each value in the tables is equal to one standard deviation obtained from those associated with the data of independent measurements fitted to the chosen binding model. If there are no protons released or absorbed during the binding, buffer ionization can contribute to the experimentally determined enthalpies of binding. As phosphate has a small value of the enthalpy of ionization (3.6 kJ/mol) (22), the observed values of enthalpy are practically the binding enthalpies of the drug to the protein. Analysis of the experimental data by alternate binding models yielded poor fits (Figure S1).

Figure 2.

 Raw data for the titration of (A) 0.415 mm tetracycline (TC) with 0.09 mm bovine serum albumin (BSA) and (B) 0.415 mm rolitetracycline (RTC) with 0.15 mm BSA at pH 7.4 and 25 °C, showing the calorimetric response as successive injections of the ligand that are added to the sample cell (upper panel), and the integrated heat profile of the calorimetric titration shown in panel A (lower panel). The solid line represents the best non-linear least-squares fit to the experimental data points.

Table 1.   Values of binding constant (K), enthalpy (ΔH), and entropy (ΔS) of binding of tetracycline and rolitetracycline with bovine serum albumin at various temperatures (t) according to sequential binding model
t/°CK1K2ΔH1ΔH2ΔS1ΔS2
per mkJ/molJ/K/mol
Tetracycline
 15(2.33 ± 0.04) × 104151 ± 43−7.5 ± 0.4−45.6 ± 3.856.1434.3
 20(2.34 ± 0.25) × 104188 ± 45−4.3 ± 0.2−53.6 ± 0.469.00.7
 25(2.58 ± 0.22) × 104272 ± 47−6.2 ± 0.3−61.9 ± 0.463.6−194.1
 37(2.18 ± 0.74) × 104281 ± 66−12.9 ± 1.5−128.5 ± 2.149.8−366.4
Rolitetracycline
 15(3.77 ± 0.73) × 103160 ± 40−6.2 ± 0.2−67.4 ± 14.247.716.5
 20(3.50 ± 0.73) × 103160 ± 20−13.8 ± 1.1−7.8 ± 2.520.786.2
 25(3.36 ± 0.67) × 103490 ± 10−6.4 ± 0.8−3.4 ± 0.846.057.7
 30(2.70 ± 0.42) × 103410 ± 60−17.2 ± 1.914.6 ± 2.98.8117.6
 37(1.59 ± 0.42) × 103560 ± 10−25.5 ± 5.47.9 ± 3.65.878.2

The integrated heat profile shown in Figure 2A has been corrected for the dilution effects of TC and the protein. The dilution-corrected titration profile shows that the binding is an exothermic process. Fitting of these data to the sequential model with two binding sites described earlier provides a value of binding constant at the first site of the order of 104, which is stronger than the binding at the second site at all of the studied temperatures as in Table 1. The binding at both the sites is observed to be exothermic but more entropically controlled at the first site. The value of binding constant at the first site remains almost the same in the temperature range of 288.15–303.15 K. The value of heat capacity of binding calculated from the plot of enthalpy of binding versus temperature was found to be small (−0.31 kJ/K per mol).

The representative ITC profiles for the titration of 4.516 mm RTC with 0.15 mm BSA at pH 7.4 are shown in Figure 2B. Here, the data fitted best to a three sequential binding site model. In this case also, analysis of the data according to alternate binding models yielded poor fits (Figure S2). As albumin has a capacity to offer multiple binding sites to ligands, we have tried fitting the experimental data to other binding models also, but a large error was found to be associated with thermodynamic parameters of binding of drugs to BSA. Therefore, Figure 2 represents the best fitting to the experimental data points according to sequential model with two or three binding sites for the binding of TC and RTC, respectively. The binding of RTC at the first site is accompanied with an association constant of the order of 103, which is stronger than the binding at the second and third sites at all of the studied temperatures (Table 1). The variation in the value of binding constant with rise in temperature at both first and second site is small. Here also, the value of heat capacity of binding for the first site calculated from the plot of enthalpy of binding versus temperature was found to be small (−0.80 kJ/K per mol).

The value of ΔHo obtained from the temperature dependence of the binding constant for the first binding site is −2.12 and −27.0 kJ/mol for TC and RTC, respectively, which are lower than those obtained by calorimetry. The deviations in the values of enthalpy of binding obtained by calorimetry from those obtained by temperature dependence of the binding constant can be due to the sequential binding nature of the ligand to the protein.

Unraveling the binding interactions

Effect of salt on binding

To know the contribution of charge–charge interactions in the biomolecular association, an analysis of the effect of salts on the interactions is a useful approach (5,23). To understand the role of electrostatic interactions in the binding process, the ionic strength dependence of the binding of TC and RTC with BSA was studied. The experiments were performed in the presence of 0.01, 0.05, and 0.1 m NaCl at pH 7.4 and 298.15 K (Figure 3).

Figure 3.

 Isothermal titration calorimetry (ITC) profile for the titration of 0.415 mm tetracycline (TC) to 0.09 mm bovine serum albumin (BSA) in the presence of (A) 0.01 m NaCl and (B) 0.05 m NaCl at pH 7.4 and 25 °C.

For TC, the presence of 0.01 m NaCl reduces the value of K1 from (2.58 ± 0.22) × 104 to (0.56 ± 0.10) × 104 at 25 °C (Table 2). Further increase in NaCl concentration to 0.05 and 0.10 m in solution leads to the total loss of the binding. For RTC also, an ionic strength of 0.05 and 0.10 m leads to loss of binding (Figure S3). These results indicate that the ions of NaCl interfere in the binding of TC and RTC with BSA suggesting significant involvement of electrostatic interactions in the binding of TC and RTC with BSA.

Table 2.   Values of binding constant (K), enthalpy (ΔH), and entropy (ΔS) of binding of tetracycline and rolitetracycline with bovine serum albumin in the presence of additives at 298.15 K
AdditiveConc. (m)K1K2ΔH1 (kJ/mol)ΔS1 (J/K/mol)
per m
  1. TRAB, tetrabutylammonium bromide; TX-100, triton X-100.

Tetracycline
 Sucrose0.5(1.21 ± 0.06) × 10449 ± 3−23.4 ± 0.46.3
1.0(0.50 ± 0.07) × 104650 ± 367−12.9 ± 1.328.5
 NaCl0.01(0.56 ± 0.10) × 104273 ± 127−5.0 ± 0.455.2
0.05No binding   
0.10No binding   
 TBAB0.02(0.52 ± 0.05) × 104260 ± 38−16.3 ± 0.817.2
0.05No binding   
0.10No binding   
 TX-1000.35(7.51 ± 0.39) × 10335 ± 5−13.4 ± 0.329.3
Rolitetracycline
 Sucrose0.50(4.66 ± 0.50) × 103407 ± 140−9.6 ± 0.437.2
1.00No binding   
 NaCl0.01(2.11 ± 0.40) × 103331 ± 65−5.9 ± 0.451.9
0.05No binding   
0.10No binding   
 TBAB0.02No binding   
 TX-1000.35(8.80 ± 0.80) × 10396 ± 6−29.1 ± 2.3−41.0

Effect of co-solute having mixed hydrophilic–hydrophobic character on binding

To understand the role of hydrophobic interactions in the binding process, experiments were carried out on the binding of TC and RTC with BSA in the presence of TBAB that has four bulky hydrophobic groups and hence can interfere in these interactions. The thermodynamic parameters of binding of TC and RTC to BSA in the presence of TBAB are given Table 2, and the binding profiles are shown in Figure 4 and Figure S4, respectively.

Figure 4.

 Isothermal titration calorimetry (ITC) profile for the titration of 0.415 mm tetracycline (TC) to 0.09 mm bovine serum albumin (BSA) in the presence of (A) 0.02 m TBAB and (B) 0.05 m TBAB at pH 7.4 and 25 °C.

For the binding of TC, the value of the binding constant decreased by almost 50% in the presence 0.02 m TBAB (Table 2). However, with further increase in the concentration of TBAB from 0.02 to 0.05 m, no binding pattern was observed (Figure 4). TBAB is a molecule with mixed hydrophobic and ionic character and hence can interfere in the binding of TC with both the charged and hydrophobic amino acid residues. The enhanced effects of TBAB at higher concentration resulting in no binding pattern are most likely due to larger interference in both hydrophobic and ionic interactions. With RTC, no binding pattern was observed even at 0.02 m concentration of TBAB (Figure S4). This is because of higher hydrophobic nature of RTC that offers greater hydrophobic interaction with TBAB, hence leading to loss of binding affinity.

Effect of co-solute having hydroxyl groups on binding

Sucrose possesses eight hydroxyl groups and hence can interfere in hydrogen bonding interactions in aqueous solutions containing the drug and the protein. The ITC experiments were conducted in the presence of sucrose as co-solute to understand the role of hydrogen bonding in the binding process. The values of the binding affinity for TC decreased from (1.21 ± 0.60) × 104 to (0.50 ± 0.07) × 104 with an increase in the concentration of sucrose from 0.5 to 1.0 m (Figure 5A, Table 2). This suggests that hydrogen bonding plays a role in the interaction of TC with BSA.

Figure 5.

 Isothermal titration calorimetry (ITC) profile for the titration of 0.415 mm tetracycline (TC) to 0.09 mm bovine serum albumin (BSA) in the presence of (A) 0.5 m sucrose and (B) 0.35 mm triton X-100 (TX-100) at pH 7.4 and 25 °C.

The values of the binding affinity for RTC also decreased and totally vanished with an increase in the concentration of sucrose from 0.5 to 1.0 m (Figure S5). This suggests involvement of hydrogen bonding in the interaction of RTC also with BSA. The data indicate stronger interference of sucrose in the interaction of RTC with BSA compared with that of TC. Although TC can participate in stronger hydrogen bonding with sucrose, the loss of binding affinity in case of RTC appears to arise because of combined effect of altered solvent structure and hydrogen bonding.

Effect of surfactants on binding

Surfactants at low concentration can interfere in the binding of drugs with the target sites of the protein via hydrophobic interactions and ionic interference depending on its cationic, anionic, or non-ionic nature. As both TC and RTC lead to precipitation in the presence of anionic surfactant sodium dodecyl sulfate and cationic surfactant hexadecyltrimethylammonium bromide, ITC experiments could not be performed.

To understand the contribution of non-columbic interactions in the binding, experiments were conducted in the presence of the non-ionic surfactant TX-100. Figure 5B shows the dilution-corrected integrated heat profile accompanying the titration of 4.516 mm TC with 0.09 × 10−3 mm BSA containing 0.35 mm TX-100, where the surfactant is in the postmiceller form. It is seen in Table 2 that the presence of TX-100 in solution slightly lowers the binding affinity of the drug for the protein. As TX-100 is a neutral surfactant capable of forming H bonds and interfering in the hydrophobic interactions through hydrophobic groups, the change in binding affinity suggests the involvement of non-columbic interactions in the binding process. Similar results were obtained with RTC (Figure S6, Table 2) suggesting involvement of non-columbic interactions in this system also.

Differential scanning calorimetry

Generally, when a ligand binds to a protein in its native state, it imparts thermal stability to the biological macromolecule. To understand the effect of the binding of drugs on the thermal stability of BSA, DSC was employed. Figure 6A is representative of the DSC scans for the thermal denaturation of BSA in 20 mm phosphate buffer at pH 7.4 in the absence and the presence of varying concentrations of TC. The heat absorption curves shown in this figure correspond to denaturation of BSA solutions of 6 mg/mL protein.

Figure 6.

 Differential scanning calorimetry (DSC) scans of the thermal denaturation of 0.09 mm bovine serum albumin (BSA) at pH 7.4 in the presence of (A) tetracycline (TC) at drug to protein molar ratio 5:1 and (B) rolitetracycline (RTC) at drug to protein molar ratio 5:1 at a scan rate of 1 °C/min.

The transition temperature of BSA in the absence of the drug is observed at 55.1 ± 0.1 °C with a transition enthalpy of 338 ± 8 kJ/mol. In the presence of TC and RTC at drug to protein molar ratios from 1:1 to 10:1, broad endotherms are obtained (Figure 6), the deconvolution of which yields a best fit to two components unfolding at different temperatures (Table 3). At TC to BSA molar ratio from 1:1, the two components unfold at 55.9 ± 0.1 °C and 62.8 ± 0.1 °C, respectively. With increase in the TC to BSA molar ratio from 1:1 to 10:1, the transition temperature of the first component increases by 2.3 ± 0.1 °C, whereas the second component of the main transition remains at an average transition temperature of 62.9 ± 0.1 °C. The transition enthalpy associated with the first transition shows a slow decrease, whereas that for the second transition increases significantly with increase in TC to BSA molar ration. Upon further increase in TC to BSA molar ratio to 20:1, only one endotherm is obtained with a transition temperature of 62.8 ± 0.1 °C.

Table 3.   Transition temperature (tm)a,b and enthalpy change (ΔH)a,b associated with the thermal unfolding of 0.09 mm BSA in presence of different concentrations of tetracycline and rolitetracycline at pH 7.4
[Drug]/[BSA] t m(1) t m(2)ΔH(1)ΔH(2)
°CkJ/mol
  1. BSA, bovine serum albumin; DSC, differential scanning calorimetry.

  2. a(1) and (2) represent first and second component obtained on deconvolution of the main DSC endotherm.

  3. bThe values of tm and ΔH have an uncertainty of 0.1 °C and 2%, respectively.

Drug: Tetracycline
 055.1 ± 0.1339 ± 8
 155.9 ± 0.162.8 ± 0.1408 ± 9236 ± 5
 557.4 ± 0.163.0 ± 0.1401 ± 8256 ± 6
 1058.4 ± 0.163.4 ± 0.1388 ± 8297 ± 6
 2062.8 ± 0.1329 ± 7
Drug: Rolitetracycline
 156.0 ± 0.160.3 ± 0.1493 ± 10212 ± 4
 557.8 ± 0.164.2 ± 0.1392 ± 8238 ± 4
 1059.8 ± 0.165.2 ± 0.1323 ± 3260 ± 6
 2063.7 ± 0.1359 ± 8

The binding of TC with BSA can be described by the following equation:

image(4)
image(5)

The ITC results suggest that K1 ≫ K2, hence the thermal unfolding of BSA will essentially be dictated by eqn 4. The value of K1 at 25 °C is (2.58 ± 0.22) × 104/m, which is a moderate binding strength. The thermal unfolding of the TC–BSA complex under these conditions will be according to the following equations:

image(6)
image(7)

Here, BSA(N) and BSA(D) represent the native and thermally denatured forms of the protein, respectively. Thus, the main endotherm observed in the DSC scan will have contributions from the thermal unfolding of the TC–BSA complex and unligated BSA in the solution.

The increase in the transition enthalpy corresponding to the second transition is because of increased ligation of the protein at higher drug to protein molar ratio. In a similar manner, as explained earlier, the thermal unfolding of BSA in the presence of RTC also yields a broad endotherm, the deconvolution of which leads to two components fitting the main endotherm. Here, with increase in the molar ratio of RTC to BSA, the transition temperature of the first component increases by 3.8 ± 0.1 °C. However, the transition enthalpy decreases from 493 ± 10 to 323 ± 3 kJ/mol. The transition temperature of the second transition also shows a slow increase with a significant increase in the transition enthalpy 48 ± 7 kJ/mol with increase in the drug to protein molar ratio from 1 to 10. The thermal denaturation of the RTC–BSA complex can also be expressed similar to the eqns 6 and 7 in view of K1 being much larger than K2 and K3. Here also, at RTC to BSA molar ratio of 20:1, only one endotherm corresponding to fully complexed protein is observed.

The values of binding constant for the binding of TC and RTC with BSA are of the order of 104 and 103, respectively. As the binding affinity is not very strong, a solution of the drug and the protein is expected to contain significant amount of the complexed and uncomplexed forms of the protein at equilibrium. As a result of this, two components of the main transition observed belong to the complexed and uncomplexed forms of the protein. These results suggest that TC and RTC stabilize both the components of the main transition. The Tm of the first component increases by 4 °C in case of TC, whereas this increase is by 5.4 °C in case of RTC at drug to protein molar ratio of 10:1. The second component of the main DSC transition also shows stabilization at higher drug to protein molar ratio. At drug to protein molar ratio of 20:1, only one component of main transition is observed, which belongs to fully ligated protein. All the thermal transitions were observed to be completely irreversible, thereby restricting the application of equilibrium thermodynamics for the further analysis of the DSC data (24,25).

Fluorescence spectroscopy

Steady-state fluorescence measurements

To understand the effect of TC on the tryptophan environment of the protein and to determine the binding constant of the interaction by another method, the intrinsic fluorescence of BSA (1.5 μm) in the presence of increasing concentration of drugs was studied. The fluorescence emission spectra of the protein in the presence of TC and RTC are shown in Figure 7A and Figure S7, respectively.

Figure 7.

 (A) Intrinsic fluorescence emission spectra of 1.5 μm bovine serum albumin (BSA) in the presence of increasing tetracycline (TC) [TC]/[BSA] and (B) plot of F0/(F0F) against 1/[TC] at pH 7.4.

The addition of TC at increasing molar ratio of the [drug]/[BSA] from 0 to 25 leads to appreciable quenching in the maximum intensity of emission. This indicates that TC binds to BSA, and the binding of the drug is happening in the close vicinity of tryptophan residues, which are located at positions 134 and 212 in subdomains I A and II A, respectively (26,27).

The values of binding constant and binding stoichiometry were calculated (17,28) using the following equation:

image(8)

where F0 and F are the fluorescence intensities of the tryptophan residues of the protein in the absence and the presence of the drug. The values of the binding constant and binding stoichiometry were obtained from the intercept and slope of the plot of inline image versus log [Q]. These plots for the binding of TC and RTC with BSA are shown in Figure 7B and Figure S7B, respectively. The value of K and n obtained for the binding of TC with BSA is 4.33 × 106/m and 1.3, respectively. The fluorescence results clearly indicate that the binding is not 1:1 in nature, but involves more than one molecule of the drug binding to the protein. The ITC data show that TC binds to BSA according to two sequential binding site model with K1 = (2.58 ± 0.22) × 104 and K2 = (2.72 ± 0.47) × 103/m at 25 °C. The overall binding constant K, which is the product of K1 and K2, is 7.2 × 106, which is close to the value of the binding constant obtained from the fluorescence measurements. The plot of F0/F and τ0/τ versus concentration of TC (Figure S8) shows that the contribution of the dynamic quenching to the overall process is insignificant. This is also seen in the lifetime measurements using time-resolved fluorescence spectroscopy of BSA in the presence of TC.

Time-resolved fluorescence measurements

Fluorescence lifetime measurements were carried out for BSA in the absence and the presence of TC and RTC (Figure 8). A sample containing 15 μm BSA was excited at 295 Nm, and emission was monitored at 344 nm. The decay curves fitted well to a bi-exponential function in the entire concentration range of the drugs studied. It is seen in Table 4 that the average lifetime of tryptophan changes from 5.75 to 4.74 ns upon addition of TC at molar ratio 1:5, which is very small. The change in the lifetime of 1.01 ns suggests that the contribution of the excited state complexation to the overall fluorescence quenching is not significant. Therefore, the quenching of the observed fluorescence can be assigned to the association of the drug with the fluorophore, and hence, the quenching constant obtained by Stern–Volmer eqn 8 can be considered as binding constant.

Figure 8.

 Fluorescence decay profiles of 15 μm bovine serum albumin (BSA) in the presence of tetracycline (TC) at different drug to protein molar ratios. The inset in this figure represents the fluorescence decay profiles for rolitetracycline (RTC).

Table 4.   Lifetimes of fluorescence decay of 0.015 mm BSA in the presence of varying [TC]/[BSA] and [RTC]/[BSA] molar ratios at pH 7.4a
[D]/[P]Lifetime/(ns)AmplitudeAvg. lifetime/(ns)χ2
τ1τ2a1a2<τ>
  1. BSA, bovine serum albumin.

  2. aI(t) = I(0) [a1e−t/τ1 + a2e−t/τ2].

  3. Where a1 = b1/(b1 + b2); a2 = b2/(b1 + b2); b1 and b2 are determined experimentally.

  4. Average lifetime: <τ> = a1τ1 + a2τ2; chi-square represents the goodness of the fit.

Tetracycline
 03.266.470.2260.7745.751.09
 13.997.100.4720.5285.631.09
 22.316.170.2490.7515.211.07
 31.985.990.2410.7595.020.97
 42.106.090.3040.6964.891.08
 52.026.040.3230.6774.741.01
Rolitetracycline
 04.458.170.6000.4005.941.07
 14.267.340.5010.4995.791.09
 23.296.610.3110.6895.581.10
 31.956.070.1940.8065.271.09
 41.745.960.2130.7875.051.09
 51.466.170.2790.7214.851.11
 71.705.810.2920.7084.611.12

The use of eqn 8 for RTC gives an overall binding constant of 2.63 × 106 and binding stoichiometry of 1.25. Here also, the fluorescence lifetime of tryptophan showed a small change of 1.33 ns at a [drug]/[protein] molar ratio of 7 suggesting minimum contribution of dynamic quenching. The ITC results fitted to three sequential binding model with K1 = (3.36 ± 0.37) × 103 and K2 = (0.49 ± 0.01) × 103 at 25 °C. The value of K3 is very small with large error, hence not considered. Here also, the overall binding constant K obtained from ITC (K = K1.K2 = 1.65 × 106) is very close to that observed from fluorescence measurements (Kfluorescence = 2.63 × 106).

Conformational characterization by circular dichroism spectroscopy

The far-UV CD and near-UV CD spectra of BSA in the presence of TC and RTC are shown in Figure 9 and Figure S9, respectively. It is seen that the ligand binding does not cause an appreciable change in the tertiary structure of BSA. However, the secondary structure is slightly reduced upon binding with TC and enhanced upon binding with RTC. Nevertheless, these small changes in the secondary structure do not cause appreciable changes in the thermal stability of BSA (DSC Table 3). The fluorescence results also do not show any appreciable change in the value of λmax upon addition of the drug to the protein.

Figure 9.

 Circular dichroism (CD) spectra of bovine serum albumin (BSA) in the presence of tetracycline (TC): (A) far-UV CD spectrum and (B) near-UV CD spectrum at different drug to protein molar ratios. The concentrations of BSA in the far-UV CD and near-UV CD experiments were 5 and 15 μm, respectively.

Possible binding sites for the drug

The crystal structure of human serum albumin is reported in the literature at a resolution 2.5 × 10−10 m (29). Because of 76% structural homology between BSA and human serum albumin, the crystal structure of HSA can be used to address the binding sites for the ligands. It has two major binding sites in the subdomains II A and III A known as warfarin and diazepam binding sites, respectively (29). Site 1, which is believed to be a binding site for salicylates, sulfonamides, and several other similar types of drugs, includes binding pocket formed by hydrophobic side chains, and the entrance of the pocket is surrounded by Arg257, Arg222, Lys199, His242, Arg218, and Lys195 (30). Site 2 corresponding to pocket of subdomain III A, which is known to be a binding site for tryptophan, thyroxin, octanoate, and other similar drugs, is lined by hydrophobic side chains and double disulfide bridge of helix III A-h3 (31).

To determine the possible binding site for the drugs, fluorescence experiments were conducted in the presence of warfarin (Figure 10A), which is known to bind at site 1. Addition of warfarin to BSA at a molar ratio of 4:1 causes a 20-nm red shift (Inset, Figure 10A) in the fluorescence emission spectra suggesting a major conformational change in the protein. Addition of TC to this warfarin–BSA complex leads to an approximate 20 nm blue shift (Figure 10A), thereby bringing the BSA conformation back to the native state of the protein.

Figure 10.

 (A) Intrinsic fluorescence emission spectra of bovine serum albumin (BSA)–warfarin complex in the presence of increasing [TC]/[BSA] molar ratios, and inset shows a red shift of 20 nm upon warfarin–BSA complexation (dashed line) from uncomplexed (solid line) protein. (B) Intrinsic fluorescence emission spectra of BSA–TC complex in the presence of increasing [warfarin]/[BSA] molar ratios. TC, tetracycline.

Reverse fluorescence experiments were also performed to know more about the binding site. Figure 10B shows the changes in the fluorescence emission spectra when warfarin is titrated with the BSA–TC complex. The value of binding constant for the association of warfarin with BSA under these conditions is 4.44 × 104/m, which is close to the value of the binding constant for the warfarin–BSA complex reported in the literature (32) in the absence of any other drug or additive. These observations suggest that TC and warfarin do not share the same binding site. As warfarin is reported (33) to bind at site 1, and its binding affinity is not affected by the presence of TC, it can be inferred that TC binds to BSA at site 2 in the subdomain III A, which is lined by hydrophobic residues (Figure 11). TC and RTC having mixed hydrophobic–hydrophilic character bind to BSA with lesser affinity in the presence of TBAB, NaCl, sucrose, and TX-100. These observations suggest that TC and RTC bind to BSA through a mix of hydrophobic, ionic, and hydrogen bonding interactions.

Figure 11.

 Representation of tetracycline (TC) or rolitetracycline (RTC) binding to bovine serum albumin (BSA) at site 2 (domain III A). Although the diagram has been interpreted in terms of binding of these drugs to BSA, the structural representation of protein was considered for human serum albumin because of unavailability of crystal structure of BSA. The cartoon structure of HSA was generated in VMD (v1.8.7) from PDB ID-1GNJ. Color codes: violet color represents bound TC/RTC at site 2 and red color represents the basic amino acids Arg257, Arg222, Lys199, His242, Arg218, and Lys195 surrounding the entrance of the pocket at site 1.

Similar results were obtained when combinatorial experiments were carried out with RTC, warfarin, and BSA (Figure S10). The values of binding constant for the association of RTC with BSA–warfarin complex and of warfarin with BSA–rolitetracyclin complex are 7.67 × 106 and 3.49 × 104, respectively. This further suggests that RTC also binds to BSA at site 2.

Binding of RTC and TC to BSA in the presence of each other

To further understand whether these drugs share common binding site on the protein, the binding constant of TC with BSA was measured in the presence of RTC and vice versa. Figure 12 shows the ITC profile for the titration of RTC with BSA–TC complex. In this experiment, the BSA–TC complex was in 1:3 molar ratio to ensure saturation of TC binding sites on the protein. As seen in Figure 11A, the ITC profile did not show any binding pattern, indicating complete loss of binding affinity. Similar titrations were also performed with TC–BSA complex in the presence of RTC. Here, the integrated heat profiles fit best to a two sequential binding site model. The values of the binding parameters show that TC binds to BSA with slightly reduced binding affinity (K = 8.7 ± 1.3) × 103, Figure 11B) suggesting partial pre-occupancy of the binding site by RTC. These results indicate that TC and RTC bind to the same site on BSA.

Figure 12.

 Isothermal titration calorimetry (ITC) profile for the titration of (A) 0.415 mm rolitetracycline (RTC) to bovine serum albumin (BSA)–tetracycline (TC) complex and (B) 0.415 mm TC to BSA–RTC complex at pH 7.4 and 25 °C.

Conclusions

The antibiotic drugs TC and RTC bind to BSA in a sequential manner with a predominate binding constant of the order of 104 and 103, respectively. The values of binding constant, binding enthalpy, binding entropy, and binding stoichiometry have been determined using ITC. A comparison of the ITC and fluorescence results demonstrates that the former technique has been able to explain the sequential binding events that can be missed by the fluorescence measurements. The effect of ionic strength, TBAB, TX-100, and sucrose suggests involvement of a combination of hydrophobic, electrostatic, and hydrogen bonding interactions in the complexation of TC and RTC with BSA. Combinatorial study of the binding of these drugs with BSA in the presence of each other and in the presence of warfarin suggests that TC and RTC share site 2 on BSA. The differential scanning calorimetric results have provided a quantitative measure of the thermal stability of BSA in the presence of TC and RTC. A quantitative understanding of the binding thermodynamics of the interaction of TC and RTC with BSA has enabled evaluation of binding strength, mode of interaction, effect on protein stability, and identification of the groups responsible in the recognition of these drugs by the protein. Such studies provide guidelines to improve and synthesize target-oriented drugs.

Acknowledgments

The authors are grateful to the Department of Science and Technology and the Council of Scientific and Industrial Research, New Delhi, for financial support.

Ancillary