Highly Selective and Tunable Protein Hydrolysis by a Polyoxometalate Complex in Surfactant Solutions: A Step toward the Development of Artificial Metalloproteases for Membrane Proteins.

This study represents the first example of protein hydrolysis at pH = 7.4 and 60 °C by a metal-substituted polyoxometalate (POM) in the presence of a zwitterionic surfactant. Edman degradation results show that in the presence of 0.5% w/v 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) detergent, a Zr(IV)-substituted Wells-Dawson-type POM, K15H[Zr(α2-P2W17O61)2]·25H2O (Zr1-WD2), selectively hydrolyzes human serum albumin exclusively at peptide bonds involving Asp or Glu residues, which contain carboxyl groups in their side chains. The selectivity and extent of protein cleavage are tuned by the CHAPS surfactant by an unfolding mechanism that provides POM access to the hydrolyzed peptide bonds.

Introduction

On the basis of complete sequencing of several genomes, 30% of all proteins are estimated to be hydrophobic membrane proteins.[1,2] The study of the structure and function of this large class of proteins is essential as they are among the prime drug targets. In fact, 60% of the current drug targets are located on the cell surface.[3] The large size, structural complexity, and intrinsic hydrophobicity of these biomolecules often complicate their structural investigations by the existing experimental methods. Therefore, controlled fragmentation of membrane proteins into smaller fragments remains a key approach to facilitate their structural analysis.[4] However, this requires the hydrolysis of the peptide bond, which has an estimated half-life of between 350 and 600 years under physiological conditions.[5−7] In addition, as a result of their hydrophobic nature, membrane proteins are insoluble in aqueous solutions, and denaturation is often encountered. Therefore, the addition of surfactants is often essential to mimic a hydrophobic environment and stimulate solubilization of membrane proteins. Generally, the concentration of the surfactant should be maintained above the critical micelle concentration (CMC),[8] and therefore, relatively high surfactant concentrations (usually in the range of 0.5–1.0% w/v) are required.[9]

Proteolytic enzymes are often used to hydrolyze peptide bonds in proteins for many purposes in biochemistry, biotechnology, and proteomics. However, they are very sensitive to the experimental conditions and often undergo denaturation in the presence of surfactants, which results in the loss of catalytic activity. Therefore, they are not suitable for the hydrolysis of hydrophobic and membrane proteins. Consequently, there is an urgent need for new synthetic proteases that are compatible with surfactants and can eventually be used to hydrolyze membrane proteins in an efficient and selective manner under mild reaction conditions.

Several metal complexes have been shown to induce hydrolysis of peptide bonds in proteins;[10−16] however, their catalytic ability under the conditions pertinent to membrane proteins has been largely unexplored. PdII and PtII complexes have been studied for the selective hydrolysis of proteins in solutions containing surfactants;[9,16] however, very low pH conditions (2.5–2.9) were required to observe the reactivity, which is disadvantageous as this often leads to pronounced background cleavage and also results in denaturation of the protein under study. More recently, a Ni(II)-containing complex has been shown to cleave a terminal affinity tag from a recombinant fusion protein, thus providing an interesting alternative for currently used techniques.[17]

Polyoxometalates (POMs) are a class of inorganic metal–oxygen clusters with a range of tunable properties.[18−20] POMs are commercially used as catalysts in a wide range of chemical reactions.[21−24] Moreover, the biological activity of POMs has been reported with respect to their antiviral, antibacterial, and anticancer properties.[25−28] We have recently demonstrated that POMs exhibit reactivity toward a range of biologically relevant molecules and their model systems.[29−34] Recently, the ability of metal-substituted POMs to act as a novel class of artificial peptidases has been introduced by our group. Zr(IV)-substituted POMs have been proven to selectively cleave proteins, ranging from a small, flexible polypeptide system, such as oxidized insulin chain B,[35] to larger protein systems with a defined tertiary structure and surface charge, such as hen egg white lysozyme (HEWL),[36] human serum albumin (HSA),[37,38] myoglobin,[39] and cytochrome c.[40] The hydrolytic experiments were performed in aqueous solutions under physiological pH conditions; however, to develop Zr(IV)-substituted POMs as artificial proteases for hydrophobic proteins, their activity as catalysts in the presence of surfactants needs to be established. We recently demonstrated that in the presence of anionic, neutral, and zwitterionic surfactants, the Zr(IV)-substituted Wells–Dawson POM, K15H[Zr(α2-P2W17O61)2] (1), preserves its catalytic activity toward hydrolysis of the peptide bond in dipeptides.[41] Diffusion Ordered NMR Spectroscopy (DOSY) has shown that anionic and neutral surfactants do not exhibit any interaction with 1, whereas a tertiary system is formed between the dipeptide, POM, and the surfactant in the presence of zwitterionic surfactants such as 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS). The interaction between 1 and CHAPS occurs through the positively charged ammonium group, whereas, at the same time, the negative charge of the sulfonate group attenuates this electrostatic interaction. Thereby, precipitation that typically occurs in the presence of positively charged surfactants is prevented.[41,42] However, the applicability of Zr(IV)-substituted POMs toward protein hydrolysis in surfactant solutions is challenging because interactions between the POM and the surfactant may shield POM binding to the protein, which is largely electrostatic in nature. Moreover, the partial unfolding of the protein caused by surfactants may affect the selectivity of protein hydrolysis by altering POM/protein interactions, which are proposed to occur according to the lock and key principle.[25,43] Therefore, in this study, we explore the compatibility of 1 to act as a metalloprotease for the hydrolysis of HSA in the presence of the zwitterionic surfactant, CHAPS (see Figure ), which is widely used for solubilizing membrane proteins.[44,45] HSA consists of 585 amino acids and serves as an excellent model for studying the hydrolytic activity of Zr-POMs in surfactant solutions. The protein has been structurally well characterized, and its hydrolysis by the Zr-POMs has been studied in detail. This offers the possibility to elucidate the role of the detergents in the activity and selectivity of the catalyst.

Figure 1

(a) Equilibria between the 1:2, 1:1, and 2:2 species of the Zr(IV)-substituted Wells–Dawson POM. WO6 octahedrons are represented in blue, PO4 tetrahedrons in red, and Zr(IV) in green. (b) Chemical structure of the zwitterionic surfactant, CHAPS.

Results and Discussion

In the hydrolysis experiments, HSA (0.02 mM) was incubated with 1 in phosphate buffer (10 mM, pH 7.4) at 60 °C in the presence of different concentrations of CHAPS (0.0–2.0% w/v). To monitor the progress of the reaction, aliquots of the homogeneous reaction mixture were first taken at different time intervals. Then, the polypeptide fragments that were formed as a result of protein hydrolysis were separated by SDS-PAGE. The presence of multiple bands on SDS-PAGE (see Figure ) clearly indicate that 1 is able to hydrolyze HSA in the presence of up to 2.0% w/v CHAPS, which is well above its CMC (0.5% w/v) and above the range of surfactant concentrations that are typically used for solubilization of membrane proteins (0.5–1.0% w/v).[9] Interestingly, an increase in CHAPS concentration led to a gradual decrease in the number and intensity of the bands in SDS-PAGE. Longer incubation times resulted in more intense bands, which is consistent with the increasing yields of protein fragmentation (see Figure S1). Our previous studies have shown that no hydrolysis of HSA was observed in the absence of the Zr-POMs.[37,38] In the presence of the lacunary Wells–Dawson K10[α2-P2W17O61]·20H2O POM, no hydrolysis was observed either, indicating that the presence of Zr(IV) is essential for the hydrolytic activity. Moreover, when the ZrCl4 salt was used, limited hydrolysis yields and the formation of insoluble Zr(IV) gels were observed, emphasizing the need of the POM as a Zr(IV)-stabilizing and protein-recognizing ligand.

Figure 2

Silver-stained sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels of HSA hydrolysis by Zr1-WD2 POM in phosphate buffer (10 mM, pH 7.4) at 60 °C in the absence and presence of surfactants: Influence of increasing % w/v CHAPS after 2 days of incubation. From left to right: protein ladder; 0.0, 0.2, 0.5% CHAPS; protein ladder; 1.0, 2.0% CHAPS.

Edman degradation of polypeptide fragments observed in SDS-PAGE unambiguously confirmed that, in the presence of 0.5% CHAPS, HSA was hydrolyzed by 1 at seven distinct sites: Cys62–Asp63, Gly71–Asp72, Asp107–Asp108, Lys313–Asp314, His367–Glu368, Ser470–Asp471, and Ala511–Asp512 visible on the gel (see Figures S2 and S3 and Table S1). Remarkably, all of the hydrolyzed peptide bonds in HSA were upstream from either the Asp or the Glu amino acid, both of which contain a carboxyl group in their side chains. This is in full accordance with our previous findings, which showed that, in the absence of surfactants, proteins were mainly hydrolyzed next to Asp or Glu residues that were located in the positive patches of proteins. The high selectivity of Zr(IV) toward hydrolyzing peptide bonds that involve Asp and Glu residues can be rationalized by two explanations: (i) the anchoring of the carboxyl group side chain to Zr(IV), or (ii) by the ability of the deprotonated carboxyl group to accept a proton from the coordinated water molecule, making it a more effective nucleophile.[46] Interestingly, CHAPS plays a significant role in tuning the reactivity and selectivity of 1 toward HSA hydrolysis. In the absence of CHAPS, the cleavage was observed at four peptide bonds (Arg114–Leu115, Ala257–Asp258, Lys313–Asp314, Cys302–Glu303), whereas seven cleavage sites were observed in the presence of CHAPS. Interestingly, all cleaved peptide bonds are located in the vicinity of a positively charged surface patch that can electrostatically interact with the negatively charged POM surface (Figure ). With the exception of the hydrolysis site, Lys313–Asp314, which is observed both in the absence of CHAPS and in the presence of 0.5% CHAPS, the surfactant is clearly responsible for altering the selectivity and generating an increase in the number of cleavage sites. Surfactants are known to break noncovalent interactions in proteins,[47] and it is likely that this results in the opening of the three-dimensional (3D) structure of HSA, making the potential X–Asp cleavage sites more accessible for hydrolysis.

Figure 3

3D structure and surface charge distribution. Negatively charged surface areas are shown in red, and positively charged surface areas are shown in blue. Asp residues in Asp–X bonds that are hydrolyzed are shown in yellow. The image was created using PyMol molecular visualization system software.

Circular dichroism (CD) measurements shown in Figure show a clear indication that CHAPS interacts with HSA in such a way that the α-helical content of HSA gradually lowers from 1.0% CHAPS onward, indicating a partial unfolding of the protein. In addition, Figure S5 shows that, under a constant CHAPS concentration and increasing concentrations of 1, a slight decrease in α-helix takes place, indicating that interactions between 1 and HSA take place in the surfactant solution.

Figure 4

CD spectra of HSA (5 μM) in phosphate buffer (10 mM, pH 7.4, 10% D2O) in the absence and presence of increasing % CHAPS.

It has been known that 1 undergoes equilibria in solution and that different species can be present depending on the experimental conditions, such as concentration, pH, temperature, or incubation time. The monomeric 1:1 species, which is presumed to be the catalytically active species, is a very elusive species due to the fast equilibria between the different POMs (see Figure ). However, we were recently able to isolate a noncovalent complex formed between the structurally analogous Zr-Keggin POM and HEWL. The single crystal X-ray structure of the complex showed that the monomeric Zr-Keggin POM was exclusively present in the co-crystal with the protein, giving the first full structural evidence for the existence of the catalytically active form.[48] Furthermore, the structural model showed that the POM initially binds to positively charged patches on the surface of the protein. The interaction types vary from water-mediated hydrogen bonds to direct electrostatic interactions, but overall, the interaction seems to be directed by the electrostatic attraction between the negatively charged POM ligand and the positively charged amino acid side chains.[48]

To study the influence of increasing CHAPS concentrations on solution speciation of the POMs, 31P NMR spectroscopy was further applied (see Figure S6). In the absence of CHAPS, 1 is present as the dimeric species (Zr1-WD2), whereas upon an increase in % CHAPS, a gradual conversion takes place, in which 1 partially transforms into α2-P2W17. As α2-P2W17 is hydrolytically inactive, this conversion could explain the gradual decrease in reactivity observed upon an increase in % CHAPS. In addition, 31P NMR measurements were taken to understand the interaction behavior between 1 and HSA in the presence of 0.5% CHAPS (see Figure S7). The dimeric species of 1 (Zr1-WD2) is predominantly present in solution in the absence of HSA. However, upon increasing concentrations of HSA, a gradual disappearance of the signal corresponding to 1 is observed. This clearly demonstrates that in the presence of 0.5% CHAPS, 1 interacts with HSA to form a POM/HSA complex with a longer correlation time and consequently a faster T2 relaxation.[49]

Tryptophan (Trp) fluorescence quenching is a powerful tool to study binding of different molecules to a protein, and it was further used to study the interaction between 1 and HSA in the presence of CHAPS. HSA contains only one Trp residue at position Trp-214, and the overall protein emission is dominated by this residue, which absorbs at the longest wavelength and displays the largest extinction coefficient. Energy absorbed by phenylalanine (Phe) and Tyr residues is often efficiently transferred to the Trp residues in the same protein.[50] In a first step, the effect of the CHAPS on the Trp fluorescence of HSA was investigated (see Figure S8). In these steady-state fluorescence experiments, the concentration of HSA was kept constant, whereas the concentration of CHAPS was increased from 0 to 5% w/v. Interestingly, it was found that the emission wavelength at the maximum intensity decreased from 342 to 328 nm as a result of the added CHAPS. The more hydrophobic environment, which is created by the addition of CHAPS, results in the shift in the emission maximum to lower wavelengths.[50]

Subsequently, the effect of the presence of 0.5% w/v CHAPS on the quenching of the Trp fluorescence of HSA by 1 was examined (see Figure ). The concentrations of HSA and CHAPS were kept constant, whereas the concentration of 1 was increased stepwise.[51] Effective quenching of Trp fluorescence by 1 was observed in the presence of CHAPS, indicating that the presence of the surfactant does not hinder POM/protein interactions. Analysis of the data by the Tachiya model (see Experimental Section and Figure S10) revealed that quenching of Trp fluorescence was a result of the binding of multiple Zr1-WD2 molecules to HSA.[52] Although the value of the quenching constant (Kq) of Trp by 1 in the presence of the surfactant is slightly lower (3.0 × 105 M–1) than the Kq in the absence of the surfactant (5.1 × 105 M–1), the model revealed that more POM was bound per HSA unit when the surfactant was present (Table S2). This is in accordance with the more “open” 3D structure of the protein caused by the surfactant, making the interaction sites more accessible for POM binding.

Figure 5

Emission fluorescence spectra of HSA in the presence of different concentrations of Zr1-WD2 ([HSA] = 10–5 M, [CHAPS] = 0.5 wt %, pH = 7.4). From top to bottom, the concentration of Zr1-WD2 was increased stepwise from 0 to 10–5 M with increments of 10–6 M.

Conclusions

In conclusion, in this study, we have demonstrated that a metal-substituted POM, K15H[Zr(α2-P2W17O61)2] (1), selectively hydrolyzes a relatively large protein, HSA, in the presence of a zwitterionic surfactant, CHAPS, under physiological pH conditions. The POM catalyst preserves its hydrolytic activity in the presence of surfactant concentrations well above those that are typically used for solubilizing hydrophobic proteins. Most importantly, in the presence of the surfactant, 1 acts as a site-selective agent, hydrolyzing HSA exclusively at peptide bonds containing amino acid residues that contain carboxyl groups in their side chain. The hydrolyzed Asp and Glu peptide bonds are either located in the positive patches of the protein or contain positively charged amino acids such as Lys and His, which aid electrostatic interaction with the negatively charged POM surface. Interestingly, this study revealed that by partially unfolding the 3D structure of proteins, surfactants can be used as a means to tune the selectivity of hydrolysis by making the potential cleavage sites more accessible to the POM catalyst. We are currently investigating the hydrolysis of water-insoluble proteins in surfactant solutions. These findings may be an important step forward in developing Zr-substituted POMs as a potential class of metalloproteases for the hydrolysis of hydrophobic and membrane proteins.

Experimental Section

Chemicals

α-/β-K6P2W18O62·14/19H2O,[53] α2-K10P2W17O61·20H2O,[53] and K15H[Zr-(α2-P2W17O61)2]·25H2O[54] were prepared according to published procedures. HSA was purchased from Sigma–Aldrich in the highest available purity (≥99%) and was used without further purification.

Hydrolysis Study

Solutions containing HSA (0.02 mM) and K15H[Zr-(α2-P2W17O61)2]·25H2O (1 mM) with different % w/v of CHAPS (0–10% w/v) were prepared in phosphate buffer (10 mM, pH 7.4). Samples were incubated at 60 °C, and aliquots were taken at different time increments and analyzed by SDS-PAGE.

Electrophoresis

SDS-PAGE was performed on a 16% (w/v) polyacrylamide resolving gel (Tris–HCl buffer, 1.5 M, pH 8.8) and a 4% (w/v) polyacrylamide stacking gel (Tris–HCl, 0.5 M, pH 6.8). Each sample (15 μL) was mixed with the sample buffer (5 μL) and heated to 95 °C for 5 min. A total of 10 μL of the resulting solution was loaded onto the gel. A PageRuler prestained protein ladder (10–170 kDa) was used as a molecular mass standard. An OmniPAGE electrophoretic cell was used with an EV243 power supply (both produced by Consort). Two SDS-PAGE gels were run at the same time in a Tris–Glycine-SDS running buffer with the maximum voltage set to 200 V, a constant current set to 70 mA, and the maximum power set to 50 W. The total running time was approximately 1.5 h. SDS-PAGE gels were visualized with silver staining, and an image of each gel was taken using a GelDoc EZ Imager (BioRad).

Edman Degradation

SDS-PAGE gels were blotted onto a PVDF membrane (using a BioRad Trans-Blot Turbo RTA Transfer Kit) and stained with Coomassie blue. Coomassie-stained protein bands were cut from the membrane, destained in methanol, and washed with ultrapure water to remove any remaining salts before Edman degradation on a capillary Procise 491cLC protein sequencer (Applied Biosystems).[55]

31P NMR Spectroscopy

To study the stability of the Zr-POM with increasing % of the surfactant, solutions containing K15H[Zr-(α2-P2W17O61)2]·25H2O (1 mM) and increasing % w/v of CHAPS (0–2%) in phosphate buffer (10 mM, pH 7.4, 10% D2O) were prepared. To study the effect of increasing concentrations of HSA on the stability of Zr-POM in the presence of a specific w/v % of the surfactant, solutions containing K15H[Zr-(α2-P2W17O61)2]·25H2O (2 mM) and HSA (0, 0.2, 0.4, or 0.5 mM) in a certain w/v % of the surfactant (0.5% CHAPS) in phosphate buffer (10 mM, pH 7.4) were also prepared. 31P NMR spectra of the different samples were recorded after mixing and after incubation at 60 °C at different time increments by using a Bruker Avance 400 MHz spectrometer. As an external standard, 25% H3PO4 in D2O in a sealed capillary was used.

CD Spectroscopy

To study the effect of increasing % of the surfactant on the secondary structure of HSA, solutions containing HSA (5 μM) and increasing % w/v of CHAPS (0–5%) in phosphate buffer (10 mM, pH 7.4) were prepared. To study the effect of increasing concentrations of Zr-POM on the secondary structure of HSA in the presence of a specific % surfactant, solutions containing HSA (5 uM) and K15H[Zr-(α2-P2W17O61)2]·25H2O (0–25 μM) in a certain % surfactant (0.5% CHAPS) in phosphate buffer (10 mM, pH 7.4) were also prepared. CD measurements were performed at RT using a Jasco J-1500 spectropolarimeter, and each sample was separately measured in a quartz cell with an optical path length of 1 mm. Scans were recorded in the far-UV wavelength region (λ = 200–260 nm), where peptide bond absorption takes place. For each CD measurement, a baseline correction was performed by subtracting the spectrum of the respective buffer solution with its specific % surfactant from the spectrum of the protein.

Fluorescence Spectroscopy

Steady-state fluorescence experiments were recorded on a Photon Technology Quanta Master QM-6/2005 spectrofluorimeter. Quartz cuvettes with a 10.0 mm optical path length were used. Spectra were recorded in a buffered 10 μM protein concentration solution (phosphate buffer, pH = 7.4) at room temperature monitoring the emission from 305 to 400 or 420 nm, with a maximum at approximately 330 nm. Excitation of the sample took place at 295 nm to avoid excitation of tyrosine residues. The emission and excitation slit widths were opened at 0.37 mm (resolution of 1.0 nm). The following concentrations of the used surfactants in the absence of POM were measured: 0, 0.2, 0.5, 1.0, 2.0, and 5.0% w/v. The Zr1-WD2 POM concentration was increased stepwise from 0 to 10 μM with increments of 1.0 μM and in the presence of 0.5% w/v of surfactant. The analysis of the results for the Zr1-WD2 POM in the presence of SDS was carried out with the help of a derived Stern-Volmer equation[56]where F0 is the unquenched fluorescence intensity, F the fluorescence in the presence of the quencher, and [Q] the concentration of the quencher.

The fluorescence data of the Zr1-WD2 POM in the presence of CHAPS were analyzed using a different equation because we expect the number of bound molecules to be higher than 1. For this reason, the Tachiya model,[52]eq , is usedwhere F0 is the unquenched fluorescence intensity, F the fluorescence in the presence of the quencher, [Q] the concentration of the quencher, [M] the concentration of the protein, and m the number of binding sites.