Spectroscopic and Theoretical Investigation of β-Lactoglobulin Interactions with Hematoporphyrin and Protoporphyrin IX.

Hematoporphyrin (HP) and protoporphyrin IX (PPIX) are useful porphyrin photosensitizers with significant application values in photodynamic therapy. Currently, many strategies have been developed to improve their clinical performance, such as incorporating them with nanoparticle (NP) carriers. In this work, we studied the possibility of using β-lactoglobulin (BLG) as a potential NP carrier due to their hydrophobic affinity, pH sensitivity, and low cost of extraction and preservation. The interaction mechanisms of BLG with HP and PPIX were investigated using spectroscopic techniques and molecular docking methods. The molecular docking results agree well with the experimental results, which demonstrate that the formations of HP-BLG and PPIX-BLG complexes are endothermic processes and the main acting force is hydrophobic force. Furthermore, the opening-closure states of EF loop have a great influence on the HP-BLG complex formation, where the central hydrophobic cavity of β-barrel is available for HP binding at pH 7.4 but not available at pH 6.2. However, the formation of the PPIX-BLG complex is less dependent on the states of the EF loop, and the binding sites of PPIX are both located on the external surface of BLG under both pH 7.4 and 6.2 conditions. All of our results would provide new insight into the mechanisms of noncovalent interactions between BLG and HP/PPIX. It is believed that this work indicated the potential application values of BLG in designing pH-sensitive carriers for the delivery of HP and PPIX, as well as other poorly soluble drugs.

Introduction

Porphyrins and their derivatives, as one of the photosensitizers, offer prominent potential and scope either as imaging contrast or photodynamic therapy (PDT) agents due to their good biocompatibility, high intrinsic specificity for tumors, and easy generation of reactive oxygen species.[1−3] Among many porphyrin drugs, hematoporphyrin (HP) and protoporphyrin IX (PPIX) are two typical representatives based on the porphyrin structure. HP, as the first-generation photosensitizers, is employed in early PDT, which is widely used in the clinical treatment of a variety of cancers.[1−3] PPIX is a precursor of heme and shown to have a longer wavelength absorption, which is considered as the second-generation photosensitizers and employed in PDT.[2,3] Currently, nanotechnology, incorporating nanoparticle (NP) carriers with photosensitizers, is a cutting-edge strategy to improve the performance of PDT.[3] Among many kinds of NP carriers, protein molecules with good hydrophobic affinity, for example, blood plasma-derived proteins, have a great development prospect.[4−6] From a physiological point of view, once entering the bloodstream, porphyrins and their derivatives often combine with circulating proteins (e.g., low-density lipoproteins, serum albumin, and glycoproteins) to form porphyrin (or its derivatives)–protein complexes,[2] which changes the drugs’ hydrophobicity/hydrophilicity and determines their affinity and avidity for tumor tissue.[2,3] However, it is expensive to extract and preserve the blood plasma-derived proteins, which limits the large-scale application of such protein carriers. Then, it is of great significance to develop inexpensive and widely available proteins as carriers or part of carriers. Additionally, given that most porphyrins and their derivatives are site nonspecific drugs for tumor tissue[7,8] and that the pH value of the tumor microenvironment (pH 5.8–7.4) is different from that of the normal extracellular microenvironment (pH ≈ 7.4),[9,10] the pH-sensitive proteins might be one of the ideal carriers, which is conducive to control the release of these drugs in the tumor microenvironment.

Bovine β-lactoglobulin (BLG), containing 162 amino acid residues with a molecular mass of 18.3 kDa and a diameter of ∼6 nm, is soluble and constitutes 7–12% of total milk proteins.[11] The secondary structure of BLG consists of nine antiparallel β-strands (strands A–I) and one α-helix, in which strands A–H fold up into a flattened β-barrel with seven connection loops labeled as AB, BC, CD, DE, EF, FG, and GH, respectively.[12−14] BLG belongs to the lipocalin protein superfamily and usually contains three main regions for interacting with hydrophobic ligands, including the central hydrophobic cavity of the β-barrel, the surface hydrophobic pocket in a groove between the α-helix and the β-barrel, and the external surface near the bottom of the β-barrel.[12−14] Many studies have proven the excellent abilities of BLG in binding and transporting a great variety of amphiphilic and hydrophobic agents or drugs.[15−18] Moreover, among the binding regions of BLG, the accessibility of the central hydrophobic cavity of the β-barrel is dependent on pH values. EF loop works as a gate over the binding site, and its reversible opening–closure can prevent ligand binding below pH 7.1 but allows it at higher pH values.[13,17,19] Hence, BLG could be employed as a pH-sensitive transport vector for drugs, such as porphyrins and their derivatives, whose delivery and release is required to depend on physiological environments.[16,20,21] However, the mechanisms of noncovalent interactions between porphyrins (or their derivatives) and BLG, especially the possible mechanism differences under different pH conditions, remain to be investigated, which are critical to the potential of using BLG for drug delivery but often overlooked.

In this work, we sought to explore the possibility of using BLG as a carrier (or part of a carrier) for HP and PPIX, two kinds of classical porphyrins. To this end, various spectroscopic techniques and molecular docking methods were implemented to investigate the interactions of HP and PPIX with BLG at alkaline and acidic conditions (pH 7.4 and 6.2). Our results demonstrate that both HP and PPIX can combine with BLG to form stable complexes. Their formation processes are endothermic, and the main acting force is hydrophobic force. Furthermore, our data also show that BLG is pH-sensitive and pH values have a great influence on the combination mode of the HP-BLG complex, in which the central hydrophobic cavity of the β-barrel is available for HP binding at pH 7.4 but not available at pH 6.2. Unlike the HP-BLG complex, the formation of the PPIX-BLG complex is less dependent on the effects of pH values on the BLG structure, and the main hydrophobic binding sites are located on the external surface of BLG under pH 7.4 and 6.2 conditions. It is believed that our findings will contribute to an in-depth understanding of the binding mechanisms of BLG with HP and PPIX and provide a theoretical basis for the following BLG-based carrier design and its applications in the delivery of poorly soluble drugs.

Results and Discussion

Fluorescence Quenching Mechanism

Fluorescence quenching analysis is a powerful and accurate method to investigate molecular interactions.[19,22] Previous studies have shown that a BLG monomer has two tryptophan (Trp-19 and Trp-61) residues and four tyrosine (Tyr-20, Trp-42, Trp-99, and Trp-101) residues, which can possess intrinsic fluorescence.[19,23] Given that the 280 nm light can excite both Trp and Tyr residues,[19,23,24] herein, BLG solutions were excited at 280 nm and the emission spectra of wavelength at 300–500 nm were recorded in the absence and presence of different concentrations of HP and PPIX. As shown in Figure S1, with increasing concentrations of HP and PPIX, the fluorescence intensity of BLG decreases gradually, manifesting that the interactions of HP and PPIX with BLG lead to fluorescence quenching. To analyze the fluorescence quenching contributions of Trp and Tyr residues, synchronous fluorescence studies were applied by setting Δλ at 60 nm for Trp and 15 nm for Tyr.[25,26] As shown in Figure S2, the synchronous fluorescence intensity of each BLG sample at Δλ = 60 nm shows a stronger intensity than that at Δλ = 15 nm, indicating that Trp residues are dominant in the total fluorescence emission. Additionally, with increasing HP/PPIX concentrations, the quenching degree of fluorescence intensity of each BLG sample at Δλ = 60 nm is much more obvious than that at Δλ = 15 nm. These results suggest that the Trp residues are mainly responsible for the fluorescence quenching instead of Tyr residues. It is worth noting that, for the intrinsic fluorescence of Trp residues, the contribution from Trp-19 is superior to that from Trp-61[27] because Trp-19 is located on an apolar environment while Trp-61 is partly exposed to an aqueous solvent.[28]

To shed light on the quenching mechanism, the fluorescence emission spectra of BLG solutions with different concentrations of HP and PPIX at different temperatures (27, 37, and 47 °C) and different pH conditions (7.4 and 6.2) were recorded. Then, the modified Stern–Volmer equations (eqs and 2) were applied to calculate the fluorescence quenching parameters:[29,30]

F0 and F are the maximum fluorescence intensities in the absence and presence of a quencher (HP or PPIX), and [C] is the concentration of a quencher. Ksv is the Stern–Volmer quenching constant, and its value is the quotient of fa–1 and (faKsv)−1. Kq is the biomolecular quenching rate constant, which is a key index for determining whether the fluorescence quenching mechanism is static or dynamic.[14,28,30] τ0 is the fluorescence average lifetime of the fluorophore and has been reported to be 1.28 ns for the Trp residues of BLG.[23,31]

As shown in Figure , for all groups, the dependences of F0/(F0 – F) on the reciprocal value of the quencher concentration [C]−1 are linear. Then, the detailed Ksv and Kq values of each group could be calculated and the corresponding results are listed in Tables S1 and S2. For the interaction of HP with BLG, the Kq value is from 9.884 × 1012 to 45.642 × 1012 M–1 s–1 at pH 7.4 and from 30.881 × 1012 to 33.688 × 1012 M–1 s–1 at pH 6.2. Given that the maximal Kq value of dynamic quenching constant is just 2.0 × 1010 M–1 s–1,[14,28,30] the Kq value of the interaction of HP with BLG suggests that the dominant mechanism of fluorescence quenching is static rather than dynamic.[14,28,30] Namely, the fluorescence quenching is mainly caused by the formation of the HP-BLG complex instead of dynamic collision. For the interaction of PPIX with BLG, the Kq value is from 10.572 × 1012 to 61.885 × 1012 M–1 s–1 at pH 7.4 and from 27.641 × 1012 to 53.654 × 1012 M–1 s–1 at pH 6.2, which is also much higher than the maximal dynamic quenching constant. It suggests that static quenching is dominant, meaning that the formation of the PPIX-BLG complex is mainly responsible for the fluorescence quenching.

Figure 1

Modified Stern–Volmer plots of BLG (5 μM) interaction with HP or PPIX at temperatures of 27, 37, and 47 °C. (a) HP, at pH 7.4; (b) HP, at pH 6.2; (c) PPIX, at pH 7.4; (d) PPIX, at pH 6.2.

It is worth noting that, with increasing temperature, the Ksv value of the HP-BLG complex increases evidently from 12.651 × 103 to 58.422 × 103 M–1 at pH 7.4 but does not show noticeable change (just from 39.528 × 103 to 43.121 × 103 M–1) at pH 6.2 (Table S1). According to previous studies proposing that the EF loop of BLG can act as a gate over the central hydrophobic cavity of the β-barrel (Figure S3), which open at pH > 7.1 and close at pH < 7.1,[13,16,17,19] the Ksv variation degree of the HP-BLG complex under different pH conditions suggests that the formation efficiency of the HP-BLG complex could depend on the effects of pH values on the BLG structure. However, the Ksv variation degree of the PPIX-BLG complex under different pH conditions (Table S2) is different from that of the HP-BLG complex, which increases evidently both at pH 7.4 (from 13.533 × 103 to 79.213 × 103 M–1) and at pH 6.2 (from 35.380 × 103 to 68.677 × 103 M–1). This indicates that the effects of pH values on the BLG structure may have little influence on the formation efficiency of the PPIX-BLG complex.

Circular Dichroism (CD) and UV–vis Studies

According to previous studies, bindings of ligands to proteins usually lead to the changes in the secondary structure.[12,32,33] Then, to figure out the conformational changes of BLG after its interactions with HP and PPIX, CD analysis was applied to characterize the secondary structure of BLG with different ratios of HP as well as PPIX. The results of CD spectra presented in Figure imply that the interactions of both HP and PPIX with BLG can induce a few conformational changes in the structure of BLG. Moreover, the detailed contents of secondary structure elements of each sample were calculated using the recorded CD spectra data. For the interaction of HP with BLG (as listed in Table S3), the proportion of the α-helix decreases and the proportion of the β-sheet increases in the presence of HP under both pH 7.4 and 6.2 conditions. Furthermore, the content of the random coil structure increases slightly when the HP concentration exceeds 10 μM. For the interaction of PPIX with BLG (as listed in Table S4), in the presence of PPIX, the proportions of both the α-helix and the β-sheet seem to be unchanged at pH 7.4 but slightly decrease at pH 6.2. Additionally, the content of the random coil structure increases slightly when the HP concentration reaches to 30 μM at pH 7.4 or exceeds 10 μM at pH 6.2.

Figure 2

Circular dichroism spectra of BLG (5 μM) in the absence and presence of HP or PPIX at various concentrations (0, 5, 10, 20, and 30 μM) at 37 °C. (a) HP, at pH 7.4; (b) HP, at pH 6.2; (c) PPIX, at pH 7.4; (d) PPIX, at pH 6.2.

The UV–vis absorption spectrum is an easy and effective tool to evaluate the structural changes during the complex formation between ligands and protein. Then, to confirm the results of the fluorescence quenching and CD studies, UV–vis experiments were also performed. As shown in Figure , both HP and PPIX do not have obvious absorption peaks at 280 nm. After the addition HP or PPIX, the absorbance intensity of each BLG solution sample at 280 nm is significantly increased under both pH 7.4 and 6.2 conditions, suggesting conformational changes in the structure of BLG. Given that the UV–vis absorption spectrum is altered in a static quenching mechanism but not altered in a dynamic quenching mechanism,[25,34,35] it can be inferred that the dominant fluorescence quenching mechanism of BLG by HP/PPIX is a static quenching procedure. All the above results indicate that the complex formations are the outcomes of the interactions between BLG and HP/PPIX, resulting in fluorescence quenching via a static quenching mechanism.

Figure 3

UV–vis absorption spectra of BLG (5 μM) in the absence and presence of HP or PPIX at 37 °C. (a) 5 μM HP, at pH 7.4; (b) 5 μM HP, at pH 6.2; (c) 30 μM PPIX, at pH 7.4; (d) 30 μM PPIX, at pH 6.2.

Thermodynamic Analysis

According to previous reports, in the case of static quenching models, the binding parameters of the ligand–protein complexes could be calculated using eq .[36]Kb is the binding constant, and n is the average number of ligands binding to per protein molecule.

As shown in Figure , for all groups, the dependences of log[(F0 – F)/F] on the ligand concentration [C] are linear. The detailed Kb and n values of each group could be calculated, and the corresponding results are listed in Tables S5 and S6. It can be seen that, with increasing temperature, the Kb value of the HP-BLG complex markedly increases from 22.641 × 103 to 57.584 × 103 M–1 at pH 7.4 but only shows a little enhancement from 39.737 × 103 to 44.545 × 103 M–1 at pH 6.2 (Table S5). As listed in Table S6, the Kb value of the PPIX-BLG complex obviously increases from 18.574 × 103 to 67.437 × 103 M–1 at pH 7.4, which is similar to that of the HP-BLG complex at pH 7.4. However, the Kb value of the PPIX-BLG complex also shows evident enhancement from 32.211 × 103 to 58.304 × 103 M–1 at pH 6.2, which is different from that of the HP-BLG complex at pH 6.2. With increasing temperature, the enhanced Kb value reveals that the formations of both HP-BLG and PPIX-BLG are affected by heat, suggesting that their formation processes may be endothermic. Furthermore, under pH 6.2 conditions, the different Kb variation degrees between the HP-BLG complex and PPIX-BLG complex also indicate that the effects of pH values on the BLG structure should have different impacts on the formations of these two complexes.

Figure 4

Curves of log[(F0 – F)/F] vs log[C] of BLG interaction with HP or PPIX at temperatures of 27, 37, and 47 °C. (a) HP, at pH 7.4; (b) HP, at pH 6.2; (c) PPIX, at pH 7.4; (d) PPIX, at pH 6.2.

Given that the thermodynamic parameters are the main criteria to estimate binding modes,[32,37,38] the Gibbs free energy (ΔG), enthalpy change (ΔH), and entropy change (ΔS) of the HP-BLG and PPIX-BLG complex were evaluated using the van’t Hoff equation (eq )[34,35] in order to better figure out the forces acting between ligands and the BLG protein. T is the thermodynamic temperature, and R is the gas constant. Kb is the binding constant at a corresponding temperature, which is obtained from eq and listed in Tables S5 and S6.

The calculated ΔG value of each group is summarized in Tables and 2. For both HP-BLG and PPIX-BLG complexes, the negative sign for ΔG implies that their formation processes are spontaneous.[28] As shown in Figure , the dependences of ΔG on the thermodynamic temperature T are linear. Thus, the ΔH and ΔS of the HP-BLG and PPIX-BLG complex formations could be computed (Tables and 2), which can roughly predict the main four types of noncovalent forces stabilizing ligand–protein complexes: (a) ΔH > 0 and ΔS > 0, hydrophobic force; (b) ΔH < 0 and ΔS < 0, van der Waals force or hydrogen bonds; and (c) ΔH < 0 and ΔS > 0, electrostatic interactions.[32,37,38]

Table 1

Thermodynamic Parameters of the HP-BLG Complex ΔG (kJ mol–1)

	pH = 7.4				pH = 6.2
temp (°C)	ΔG (kJ mol–1)	ΔH (kJ mol–1)	ΔS (kJ mol–1)		ΔG (kJ mol–1)	ΔH (kJ mol–1)	ΔS (kJ mol–1)
27	–25.011	37.299	0.208		–26.412	4.537	0.103
37	–26.944				–27.508
47	–29.162				–28.478

Table 2

Thermodynamic Parameters of the PPIX-BLG Complex

	pH = 7.4				pH = 6.2
temp (°C)	ΔG (kJ mol–1)	ΔH (kJ mol–1)	ΔS (kJ mol–1)		ΔG (kJ mol–1)	ΔH (kJ mol–1)	ΔS (kJ mol–1)
27	–24.517	51.318	0.253		–25.890	23.638	0.165
37	–27.471				–27.668
47	–29.582				–29.194

Figure 5

ΔG vs T (thermodynamic temperature) plots of BLG interaction with (a) HP and (b) PPIX at pH 7.4 and 6.2.

For the HP-BLG complex, the ΔH and ΔS values are 37.299 and 0.208 kJ mol–1 at pH 7.4 and 4.537 and 0.103 kJ mol–1 at pH 6.2, respectively. These positive ΔH and ΔS values indicate that the main acting force of the HP-BLG complex is hydrophobic force. For the PPIX-BLG complex, the ΔH and ΔS values are 51.318 and 0.253 kJ mol–1 at pH 7.4 and 23.638 and 0.165 kJ mol–1 at pH 6.2, respectively. It means that hydrophobic force also plays a major role in the formations of the PPIX-BLG complex.

Furthermore, it is worth noting that the ΔH and ΔS values of the HP-BLG complex at pH 6.2 decrease obviously in comparison with those of the HP-BLG complex at pH 7.4, suggesting weakened hydrophobic force between HP and BLG.[38,39] This interesting phenomenon might be caused by the closed EF loop at pH < 7.1,[13,19] which hinders HP from entering the central hydrophobic cavity of the β-barrel. On the other hand, although the ΔH and ΔS values of the PPIX-BLG complex decrease with the shift of pH value from 7.4 to 6.2, the decrease degrees are less than those of HP-BLG. It means that the formation of the PPIX-BLG complex is less dependent on the accessibility of the central hydrophobic cavity of the β-barrel (i.e., closed/opened EF loop) than that of the HP-BLG complex.

Fluorescence Resonance Energy Transfer (FRET) Studies

As previously mentioned, for the intrinsic fluorescence of Trp residues, Trp-19 is the major contributor to the fluorescence emission.[27,28] According to Förster’s theory,[40] FRET is a useful method to estimate the distance (r0) between the donor (in this case Trp-19 of BLG) and the acceptor (in this case HP or PPIX), following eqs and 6. E is the efficiency of energy transfer between the donor and the acceptor, R0 is the critical distance when the transfer efficiency is 50%, K2 is the orientation factor, N is the refractive index of the medium, and φ is the fluorescence quantum yield of the donor. According to previous reports,[21,28] the values of K2, N, and φ are 2/3, 1.336, and 0.08, respectively. J is the spectral overlap integral between the emission spectrum of the donor and the absorption spectrum of the acceptor, which can be calculated by following eq .[25,26]F(λ) and ε(λ) represent the fluorescence intensity of the donor and the molar extinction coefficient of the acceptor, respectively.

The fluorescence emission spectra of BLG (5 μM) and the UV–vis absorption spectra of HP (5 μM) and PPIX (5 μM) at 37 °C are displayed in Figure S4. Then, the J value is calculated by integration of the overlapping spectra in a wavelength of 290–440 nm, and the R0 and r0 values could be computed. As listed in Table , the R0 and r0 of the HP-BLG complex are 3.47 and 4.58 nm at pH 7.4, and 3.52 and 4.80 nm at pH 6.2, respectively. Meanwhile, the R0 and r0 of the PPIX-BLG complex are 2.35 and 3.44 nm at pH 7.4 and 2.17 and 3.09 nm at pH 6.2, respectively. According to Förster’s theory and previous reports, the condition of non-radioactive energy transfer is 0.5 R0 < r0 < 1.5 R0,[41] and R0 < r0 < 8 nm is a strong evidence for static quenching.[42] Thus, the r0 and corresponding R0 values of each sample not only reveal that resonance energy transfer occurs between HP/PPIX and BLG but also confirm that static quenching is dominant in these quenching processes, which is in line with the hypothesis mentioned in the above section (i.e., fluorescence quenching mechanism).

Table 3

Overlap Integral (J), Critical Distance (R0), and Donor–Acceptor Distance (r0) of the HP-BLG Complex and PPIX-BLG Complex at 37 °C

	pH = 7.4				pH = 6.2
complex	J (×10–13 cm3 M–1)	R0 (nm)	r0 (nm)		J (×10–13 cm3 M–1)	R0 (nm)	r0 (nm)
HP-BLG	1.121	3.47	4.58		1.288	3.52	4.80
PPIX-BLG	0.114	2.35	3.44		0.070	2.17	3.09

Moreover, it is worth noting that, with the shift of pH value from 7.4 to 6.2, the r0 value of HP-BLG becomes larger (from 4.58 to 4.80 nm) but that of PPIX-BLG becomes smaller (from 3.44 to 3.09 nm). These results indicate that the exposed β-barrel caused by opening the EF loop is a closer and available site for HP binding but not for PPIX binding. Namely, the access to the central hydrophobic cavity of the β-barrel is a pivotal way for BLG to bind with HP but not with PPIX.

Molecular Docking Results

To detail binding sites of the HP-BLG complex and PPIX-BLG complex, molecular docking was performed using AutoDock Vina and the results were analyzed using Discovery Studio. Given that previous studies have reported three binding sites of BLG: the central hydrophobic cavity of the β-barrel, the surface hydrophobic pocket in a groove between the α-helix and the β-barrel, and the external surface near the bottom of the β-barrel,[20,23,43] herein, the molecular docking was performed by approaching HP and PPIX molecules to these three sites under the condition of pH 7.4 and 6.2, respectively.

For the HP-BLG complex, the molecular docking results show that the EF loop opens the entrance to the β-barrel at pH 7.4, and these three sites could produce stable binding configurations with diverse energy minima: −28.01 kJ mol–1 for the central hydrophobic cavity of the β-barrel, −28.42 kJ mol–1 for the external surface near the bottom of the β-barrel, and −22.41 kJ mol–1 for the surface hydrophobic pocket in a groove between the α-helix and the β-barrel. Obviously, the central hydrophobic cavity of the β-barrel and the external surface near the bottom of the β-barrel are two main sites for binding HP at pH 7.4, which are selected to explore the conformations of HP-BLG. As shown in Figure a and Figure S5a, the HP molecule located within the central hydrophobic cavity of the β-barrel is adjacent to Pro-38, Asn-90, Asn-88, Asn-109, Ser-116, Leu-39, Met-107, Ile-56, Val-41, Leu-58, Asp-85, Ile-71, Ile-72, Lys-69, Lys-70, and Phe-105 residues. Among them, a lot of amino acids are hydrophobic, such as Pro-38, Leu-39, Met-107, Ile-56, Val-41, Leu-58, Ile-71, Ile-72, and Phe-105. Meanwhile, Figure b and Figure S5b display that the HP molecule located on the external surface near the bottom of the β-barrel is surrounded by Lys-47, Leu-57, Gln-68, Glu-44, Gln-159, Glu-158, Gln-59, Glu-157, Tyr-20, and Thr-18 residues, in which Leu-57 is hydrophobic. On the other hand, the simulation analysis indicates that the central hydrophobic cavity of the β-barrel cannot yield energetically favorable conformation for HP binding at pH 6.2, resulting in only two stable binding sites. The main reason might be the case that the EF loop closes the access to the β-barrel under acidic conditions,[13,16,17,19] blocking the entrance of the HP. For these two available binding sites, the HP molecule located on the surface hydrophobic pocket in a groove between the α-helix and the β-barrel (−27.17 kJ mol–1) is more stable than that on the external surface near the bottom of the β-barrel (−25.92 kJ mol–1). Figure c and Figure S5c display that the HP molecule located on the surface hydrophobic pocket in a groove between the α-helix and the β-barrel is surrounded by Thr-4, Lys-8, Thr-6, Gln-5, Asp-96, Lys-135, Ala-139, Leu-104, and Lys-141 residues, in which Ala-139 and Leu-104 are hydrophobic.

Figure 6

Molecular docking analysis of the binding sites. The BLG is represented in a ribbon model, and the ligands are represented in a stick model. (a, b) HP-BLG complex at pH 7.4; (c) HP-BLG complex at pH 6.2; (d) PPIX-BLG complex at pH 7.4; (e) PPIX-BLG complex at pH 6.2.

For the PPIX-BLG complex, the molecular docking results show similar outcomes in docking PPIX under pH 7.4 and 6.2 conditions, in which the central hydrophobic cavity of the β-barrel cannot yield an energetically favorable conformation. For other two binding sites, the PPIX molecule located on the external surface near the bottom of the β-barrel (−29.26 kJ mol–1 at pH 7.4 and −29.678 kJ mol–1 at pH 6.2) is more stable than that on the surface hydrophobic pocket in a groove between the α-helix and the β-barrel (−26.33 kJ mol–1 at pH 7.4 and −25.08 kJ mol–1 at pH 6.2). Figure d and Figure S6a display that the PPIX molecule located on the external surface near the bottom of the β-barrel is surrounded by Lys-47, Leu-57, Gln-68, Gln-159, Gln-59, Glu-44, Glu-158, Leu-156, His-161, Glu-157, Tyr-20, and Thr-18 residues at pH 7.4, in which Leu-57 and Leu-156 are hydrophobic amino acids. Meanwhile, Figure e and Figure S6b show that PPIX is adjacent to Thr-18, Gly-17, Ala-16, Lys-14, Gln-13, Val-15, Tyr-99, Lys-100, Thr-125, and Tyr-20 residues at pH 6.2, in which Gly-17, Ala-16, and Val-15 are hydrophobic. All the above results indicated that hydrophobic interactions might contribute to the stability of the HP-BLG complex and PPIX-BLG complex, which is consistent with the results of the thermodynamic analysis (Figure and Tables and 2).

Moreover, the formation of the HP-BLG complex is affected by pH conditions, in which the opened EF loop (at pH 7.4) can provide entrance for HP to bind the central hydrophobic cavity of the β-barrel. On the other hand, the formation of the PPIX-BLG complex is less dependent on pH values (i.e., closed/opened EF loop) because the binding sites of PPIX to BLG are mainly located on the external surface sites of BLG (e.g., the external surface near the bottom of the β-barrel).

ANS (8-Anilino-1-naphthalenesulfonate) Binding Analysis

ANS is a molecule with both hydrophilic and hydrophobic properties, which is an excellent tool to reflect ligand–protein interactions via its fluorescence changes.[44,45] Usually, BLG can possess two binding sites for ANS: an external site that is close to a hydrophobic patch on the protein surface and an internal site that is located in the hydrophobic β-barrel of the protein.[30,44] Herein, ANS (120 μM) was used to further characterize the hydrophobic interactions in the formations of the HP-BLG complex and PPIX-BLG complex. The solution of each group was excited at 380 nm, and the emission spectra of wavelength at 400–600 nm were recorded. As shown in Figure , the ANS fluorescence intensity of BLG alone shows an obvious decrease with the shift of pH value from 7.4 to 6.2, suggesting that EF loop states would control the access to the central hydrophobic cavity of the β-barrel and impact the interaction of ANS with BLG.

Figure 7

ANS emission of BLG (5 μM) upon increasing the concentrations of HP or PPIX (A = 0, B = 10, C = 20, D = 30, and E = 40 μM) at 37 °C. (a) HP, at pH 7.4; (b) HP, at pH 6.2; (c) normalized results of the maximum ANS emission intensity in panels (a) and (b). (d) PPIX, at pH 7.4; (f) PPIX, at pH 6.2; (f) normalized results of the maximum ANS emission intensity in panels (d) and (e).

For the HP-BLG complex, the ANS fluorescence intensity shows an evident decrease with increasing the ratio of HP under both pH 7.4 and 6.2 conditions (Figure a,b). Furthermore, Figure c shows that the decrease degree of ANS fluorescence intensity at pH 7.4 conditions is much larger than that at the 6.2 condition (especially at 10 and 20 μM HP). These results indicate that hydrophobic force is the main acting force of the HP-BLG complex and HP can compete with ANS for hydrophobic site binding, especially for the central hydrophobic cavity of the β-barrel under the opened state of the EF loop (pH 7.4).

For the PPIX-BLG complex, under both pH 7.4 and 6.2 conditions, increasing the ratio of PPIX can also cause decreases in ANS fluorescence intensity (Figure d,e), indicating the pivotal role of hydrophobic force in the formations of the PPIX-BLG complex. Additionally, the decrease degree of ANS fluorescence intensity at pH 7.4 conditions is similar to that at pH 6.2 conditions (Figure f), suggesting that the competition between PPIX and ANS is less affected by the EF loop state (i.e., opened or closed). Namely, the main binding sites of PPIX with BLG are located on the external surface instead of the internal hydrophobic region (i.e., the central hydrophobic cavity of the β-barrel), which is in line with the results of molecular docking (Figure and Figure S6).

Conclusions

In summary, interactions of BLG with HP and PPIX were investigated under both alkaline and acidic conditions (pH 7.4 and 6.2) in order to explore the potential of BLG acting as the pH-sensitive carriers for these drugs. The fluorescence quenching analyses reveal that the static quenching is dominant, resulting in the formations of HP-BLG and PPIX-BLG complexes. In addition, the results of CD spectra and UV–vis studies indicate that the secondary structures of BLG are changed after binding HP or PPIX, and FRET analyses reveal that resonance energy transfer could occur between HP/PPIX and BLG, which also confirm the formations of complexes. The thermodynamic parameters show that the formations of these two complexes are spontaneous and endothermic processes, in which hydrophobic force is the main acting force. Furthermore, these data also display that the binding sites of HP to BLG might be different from that of PPIX to BLG, and a highly possible reason is the opening–closure states of the EF loop under different pH values, determining the accessibility to the central hydrophobic cavity of the β-barrel. This hypothesis is confirmed by molecular docking and ANS binding analyses. The results indicate that the central hydrophobic cavity of the β-barrel is available for HP binding only when the EF loop is opened (pH 7.4), while the binding sites of HP are just located on the external surface of BLG when the EF loop is closed (pH 6.2). Additionally, the binding sites of PPIX are both located on the external surface of BLG, which are less dependent on pH values. This work could provide a new insight into the mechanisms of noncovalent interactions between BLG and HP/PPIX. It also suggests the potential application values of BLG in designing pH-sensitive carriers for the delivery of poorly soluble drugs, in which BLG could be used as a carrier or part of a carrier.

Experimental Methods

Materials

Bovine milk BLG (purity >90%) and ANS (8-anilino-1-naphthalenesulfonate) were purchased from Sigma-Aldrich (USA). HP and PPIX were purchased from Shanghai Aladdin Biochemical Technology (China). Other reagents and materials were analytical grade and used without further purification.

Stock Solution Preparation

All the solutions were prepared using 20 mM sodium phosphate buffer solution (pH 7.4). BLG powder was dissolved into sodium phosphate buffer solution to prepare 50 μM BLG stock solution. HP was initially dissolved in pyridine to prepare 5 mM HP stock solution in pyridine, while PPIX was initially dissolved in dimethyl sulfoxide to prepare 5 mM HP stock solution in dimethyl sulfoxide. Then, the 5 mM HP stock solution in pyridine was dissolved into sodium phosphate buffer solution to yield the 50 μM HP stock solution with 1% v/v pyridine. Similarly, the 5 mM PPIX stock solution in dimethyl sulfoxide was dissolved into sodium phosphate buffer solution to yield the 50 μM PPIX stock solution with 1% v/v dimethyl sulfoxide.

Synthesis of HP-BLG and PPIX-BLG Complexes

The HP-BLG complex with different ratios of HP to BLG was prepared by adding appropriate amounts of the 50 μM BLG stock solution and 50 μM HP stock solution in sodium phosphate buffer solution. The final concentrations of BLG and HP were 5 and 0–40 μM, respectively. Similarly, the PPIX-BLG complex with different ratios of PPIX to BLG was also prepared using the same method as that for the HP-BLG complex. The HP-BLG and PPIX-BLG solutions were adjusted to pH 7.4 and 6.2 by 1 M NaOH or 1 M HCl, respectively, followed by incubating at different temperatures (27, 37, and 47 °C) for 20 min.

Characterization

Fluorescence studies were performed using a Hitachi F-4600 fluorescence spectrophotometer (Tokyo, Japan). UV–vis absorption spectra were recorded with a UV–vis spectrophotometer (UV2660, Shimadzu, Japan). Circular dichroism (CD) measurements were achieved using a Chirascan qCD spectrometer (Applied Photophysics, UK). CDNN software was utilized to analyze the secondary structure of the protein.

Molecular Docking Study

Molecular docking analyses were performed using AutoDock Vina software. The protein structures of BLG at pH 7.4 (ID: 2GJ5) and at pH 6.2 (ID: 3BLG) were obtained from the RCSB Protein Data Bank (http://www.rcsb.org/pdb). BLG molecules were dehydrated and hydrotreated using AutoDockTools before docking analysis. The molecular structures of HP and PPIX were modeled using ChemDraw Professional 19 and optimized to the minimized energy states. The docking processes were performed using a grid box (20 Å × 20 Å × 20 Å) with a grid spacing of 0.375 Å followed by approaching the ligands to different sites of the BLG structure (running 10 times). The docking results were analyzed using Discovery Studio 2020.