Received: 11 June 2016
Revised: 1 December 2016
Accepted: 16 December 2016
DOI 10.1002/bio.3280
RESEARCH ARTICLE
Binding of hydroxylated polybrominated diphenyl ethers with human serum albumin: Spectroscopic characterization and molecular modeling Lulu Yang
|
Wu Yang
|
Zhiwei Wu
Guangxi Colleges and Universities Key Laboratory of Food Safety and Detection, Collaborative Innovation Center for Water Pollution Control and Water Safety in Karst Area College of Chemistry and Bioengineering, Guilin University of Technology, Guilin, China Correspondence Zhongsheng Yi, College of Chemistry and Bioengineering, Guilin University of Technology, Guilin 541004, China. Email:
[email protected]
|
Zhongsheng Yi
Abstract Three hydroxylated polybrominated diphenyl ethers (OH‐PBDEs), 3‐OH‐BDE‐47, 5‐OH‐BDE‐47, and 6‐OH‐BDE‐47, were selected to investigate the interactions between OH‐PBDEs with human serum albumin (HSA) under physiological conditions. The observed fluorescence quenching can be attributed to the formation of complexes between HSA and OH‐PBDEs. The thermodynamic parameters at different temperatures indicate that the binding was caused by hydrophobic forces and hydrogen bonds. Molecular modeling and three‐dimensional fluorescence spectrum showed conformational and microenvironmental changes in HSA. Circular dichroism analysis showed that the addition of OH‐PBDEs changed the conformation of HSA
Funding information National Natural Science Foundation of China, Grant/Award Number: 21267008; Guangxi Natural Science Foundation of China, Grant/ Award Number: 2013GXNSF AA019034
with a minor reduction in α‐helix content and increase in β‐sheet content. Furthermore, binding distance r between the donor (HSA) and acceptor (three OH‐PBDEs) calculated using Förster's nonradiative energy transfer theory was 0 and ΔS > 0, hydrophobic interactions dominate the
gated by following the time evolution of structural quantities[32]:
binding process; (2) when ΔH < 0 and ΔS < 0, van der Waals interaction and hydrogen bonding dominate the reaction; (3) when ΔH < 0 and ΔS > 0, electrostatic/ionic interactions dominate.
(1) To confirm the stability of the simulations, the root‐mean‐square deviation (RMSD) was analyzed during the 20‐ns simulation time. (2) The stability of the backbone atoms of HSA during the MD simulations was assessed from the radius of gyration (Rg).
Table 2 shows a negative value for free energy (ΔG), indicating that the interaction is spontaneous. A positive entropy change occurred
Initially, RMSD values were calculated to assess the conforma-
because the water molecules were arranged in an orderly manner
tional stability of each complex over the entire simulation.[33] The
around the OH‐PBDEs and HSA, acquiring a more random configura-
RMSD plots show that these complexes reached equilibrium at 10 nsec,
The positive ΔH and ΔS
as shown in Figure 3. The Rg of the proteins in the system is a criterion
values of the interactions indicate that binding is mainly entropy
of the compactness of protein arrangement and can be used to mea-
driven, and the change in enthalpy is detrimental to the spontaneous
sure the protein aggregates during MD simulation. The plot of the Rg
process. The driving force for these processes is mainly the typical
of the protein vs. time is also shown in Figure 3. The Rg of the back-
hydrophobic interactions, and the binding site may be located in the
bone atoms increased upon the binding of 5‐OH‐BDE‐47 and 6‐OH‐
hydrophobic pocket.
BDE‐47, indicating a more loose structure for HSA after the simula-
[31]
tion because of hydrophobic interactions.
tion, whereas 3‐OH‐BDE‐47–HSA complex mainly maintained a
3.4 | Conformational analysis of HSA after the OH‐ PBDEs binding
preserved compact conformation.
3.4.1.1 |
3D fluorescence
After establishing strong binding of 3‐OH‐BDE‐47, 5‐OH‐BDE‐47, and
3D fluorescence contour maps have become a popular analytical tech-
6‐OH‐BDE‐47 on HSA, conformational changes in the protein on
nique.
FIGURE 3
They
provide
more
detailed
information
The RMSD (a). Rg (b) of OH‐PBDEs–HSA complexes with respect to their initial sstructures as a function of time
about
the
6
YANG
ET AL.
FIGURE 4 Contour map of HSA (a), the 3‐OH‐BDE‐47–HSA system (b), the 5‐OH‐BDE‐47–HSA system (c), and the 6‐OH‐BDE‐47–HSA system (d). c(HSA) = 1 × 10−6 M, c(OH‐PBDEs) = 5 × 10−9 M. pH = 7.40, T = 298 K
conformational changes in proteins.[34] The 3D fluorescence contour maps of the protein and OH‐PBDEs–HSA complex are shown in Figure 4, and the corresponding characteristic parameters are listed in Table 3. The fluorescence intensity of the Rayleigh scattering peak (λex = λem) increased with the addition of OH‐PBDEs. This is probably because the formation of OH‐PBDEs–HSA complexes increased the diameters of the macromolecules, thus enhancing the scattering effect.[27] Peak a relates to the fluorescence spectral behavior of polypeptide backbone structures, whereas the strong peak b mainly shows the spectral characteristics of tryptophan and tyrosine residues. The maximum emission wavelength and fluorescence intensity of the residue are related to microenvironment polarity. Figure 4 clearly shows that the intensity of peak b decreased, but to a different extent after the addition of 3‐OH‐BDE‐47, 5‐OH‐BDE‐47, and 6‐OH‐BDE‐47, indicating that the hydrophobic microenvironment near the Trp and Tyr residues was altered. The fluorescence intensity of peak a increased with the addition of OH‐PBDEs because of the
FIGURE 5
CD spectra of HSA alone (a) or in the presence of 3‐OH‐ BDE‐47 (b), 5‐OH‐BDE‐47 (c), or 6‐OH‐BDE‐47 (d). c(HSA) = 1.0 × 10−6 M, c(HSA)/c(OH‐PBDEs) = 2:1 conformational changes in the polypeptide backbone structures in
TABLE 3
Characteristic parameters of 3D fluorescence spectra Peak a
Compound
λex/λem (nm/nm)
Free HSA
HSA. These results are consistent with the MD analysis.
Peak b
Intensity
λex/λem (nm/nm)
Intensity
225/350
755.2
280/355
751.3
3‐OH‐BDE‐47
225/355
812.2
280/350
650.9
5‐OH‐BDE‐47
225/355
851.0
280/355
728.9
6‐OH‐BDE‐47
223/355
825.0
280/355
658.8
3.4.2
|
CD spectroscopy
The CD spectra of HSA exhibits two negative bands in the UV region at ~208 nm and 222 nm, characteristic of the typical α‐helix structure of a protein; they can be attributed to the n → π* transition of the peptide bond of the α‐helix.[5] The CD spectra of HSA in the absence or presence of three OH‐PBDEs at room temperatures are shown in
YANG
7
ET AL.
TABLE 4 Secondary structure of free HSA and OH‐PBDEs systems (CD spectra) at pH 7.4
Compound
α‐helix (%) β‐sheet (%) β‐turn (%) Random coil (%)
respectively. These results clearly show that binding of OH‐PBDEs changed the secondary structure of the protein with a slight loss in helical content, but the β‐sheet content showed the opposite trend.
Free HSA
48.4
18.9
13.7
19.5
However, changes in the secondary structure are not enough to
3‐OH‐BDE‐47
47.1
19.6
14.0
19.7
destabilize the HSA spatial structure.
5‐OH‐BDE‐47
46.5
19.8
14.3
19.3
6‐OH‐BDE‐47
46.6
21.0
13.5
19.3
3.5
|
Molecular modeling analysis
The interaction models for OH‐PBDEs and HSA are shown in Figure 6; Figure 5. The CD signal intensity of HSA decreased at all wavelengths
only the residues around 6 Å of the ligand are shown. Clearly, the OH‐
with the addition of OH‐PBDEs, indicating an induced perturbation
PBDEs were adjacent to hydrophobic residues such as Trp214,
of the secondary structure of the protein.[35] The contents of the
Leu219, Phe223, Leu238, Val241, and Ile290 of HSA subdomain IIA.
HSA secondary structures, e.g. α‐helix, β‐sheet, β‐turn, and random
This indicates that hydrophobic force was the main interaction force
coil, were analyzed using the online CDPro program; the results are
in the binding of OH‐PBDEs to HSA, consistent with the results of
shown in Table 4. When the molar ratio of HSA to OH‐PBDEs was
the thermodynamic analysis. Moreover, hydrogen bonds were also
2:1, the α‐helix content decreased from 48.4 to 47.1%, 46.5% and
observed near the probe molecule. For 3‐OH‐BDE‐47, the oxygen
46.6% for 3‐OH‐BDE‐47, 5‐OH‐BDE‐47, and 6‐OH‐BDE‐47,
atom of the hydroxyl with Lys195 and the hydrogen atom with
FIGURE 6 Binding model for OH‐PBDEs and HSA, 3‐OH‐BDE‐47–HSA (a), 5‐OH‐BDE‐47–HSA (b), and 6‐OH‐BDE‐47–HSA (c). Hydrophobic and hydrophilic amino acids are shown by different colors: hydrophobic (orange), hydrophilic (blue). The hydrogen bond between OH‐PBDEs and HSA is represented using a yellow solid line
8
YANG
ET AL.
Glu292, hydrogen bond distances of 2.193 and 1.953 Å, respectively. For 5‐OH‐BDE‐47 and 6‐OH‐BDE‐47, three hydrogen bonds were formed between each molecule and Try150, Lys199, and Arg257. The distances of the hydrogen bonds between the donor and receptor are shown in Figure 6. The formation of hydrogen bonds decreased the hydrophilicity and increased the stability in the OH‐PBDEs–HSA system. The docking results match well with the thermodynamic analysis and indicated that hydrophobic interactions and hydrogen bonds play an important role in the binding of OH‐PBDEs to HSA.
3.6
|
Analysis of binding free energy
Combined with MD simulations, MM‐PBSA analysis provided information about the binding free energy in a thermodynamic sense and evaluated the favorable/unfavorable contributions of individual residues for binding. The MM‐PBSA approach combines three energetic terms to evaluate the change in the free energy on binding, potential energy in the vacuum, polar and non‐polar solvation energies, and configurational entropy associated with complex formation. ΔGpolar arises from electrostatic interaction, and ΔGnonpolar arises from hydrophobic interaction.[36] As shown in Table 5, both negative ΔGvdW and ΔGnp were observed in the three OH‐PBDEs–HSA complexes, leading to a negative ΔGnonpolar of −109.84 kJ·mol−1, −116.39 kJ·mol−1, −109.45 kJ·mol−1, indicating a favorable hydrophobic interaction between OH‐PBDEs and HSA. However, the electrostatic solvation energy was strongly unfavorable to complex formation, with a positive value for ΔGPB, i.e. the major driving force for binding of OH‐PBDEs with HSA is hydrophobic interaction rather than electrostatic interaction, consistent with the experimental and docking results as mentioned above. Moreover, the binding free energy was decomposed on each residue to estimate the contribution per residue. The residues with significant contributions to complex formation were identified as shown in Figure 7 with a criterion of ±2.5 kJ·mol−1 to the binding free energy.[37] In the 3‐OH‐BDE‐47–HSA system, seven residues (Arg257, Arg218, Arg222, Ile290, Ala291, Lys199, and Glu292) in HSA subdomain IIA were identified as the hot spots. In the 5‐OH‐
FIGURE 7
Binding free energy contribution of each residue in the complex. Residues with |ΔGbind| ≥2.5 kJ/mol are shown
BDE‐47–HSA system, six residues (Lys199, Arg257, Arg218, Ile290, Ala291, and Arg222) showed significant contributions to binding. In
Ile290, Ala291, and Arg218 made the main favorable contributions
the 6‐OH‐BDE‐47–HSA system, Arg257, Lys199, Glu292, His242,
for initial binding between HSA and 6‐OH‐BDE‐47. Moreover, the formation of hydrogen bonds between the hydrophilic amino acid
Binding free energies (kJ·mol−1) for the three OH‐PBDEs– HSA complexes
TABLE 5
3‐OH‐BDE‐ 47–HSA
5‐OH‐BDE‐ 47–HSA
6‐OH‐BDE‐ 47–HSA
ΔGvdW
‐93.63
‐99.66
‐93.17
ΔGelec
‐2.23
‐0.81
‐2.02
ΔGPB
29.35
32.16
48.84
ΔGnp
‐16.21
‐16.73
‐16.28
Energy term
ΔGpolara ΔGnonpolarb ΔGbindc
OH‐PBDEs was important for binding and stabilized the complexes, whereas Tyr150 was unfavorable with a positive ΔGbind of 0.72 kJ·mol−1 for 6‐OH‐BDE‐47.
3.7
|
Energy transfer between OH‐PBDEs and HSA
27.11
31.34
46.82
Förster's resonance energy transfer (FRET) theory is a good tech-
‐109.84
‐116.39
‐109.45
nique to determine the distance between the amino acid residues
‐82.73
‐85.05
‐62.63
and drug in the binding site. When the efficiency of transfer was
( ΔG polar = ΔG elec + ΔG PB , ΔG nonpolar = ΔG vdW + ΔG np , ΔG bind = ΔG polar + ΔG nonpolar .) a
residues (Tyr150, Lys195, Lys199, Arg257, and Glu292) of HSA and
b
c
50%, the condition 1:1 donor‐to‐acceptor concentration prevailed. The overlap of the absorbance spectra of 3‐OH‐BDE‐47, 5‐OH‐
YANG
9
ET AL.
and molar absorption coefficient of the acceptor, respectively, at wavelength λ. The data show that the values of r are