Ellagitannins (Ellagic Acid, Urolithin A, Urolithin B) Inhibit the Catalytic Activity of Human Recombinant Metalloproteinase 9

3.1. Materials

All chemicals used were of analytical grade and were provided by Merck (Germany). Isopropyl-β-D-thiogalactopyranoside (IPTG) was obtained from Sigma-Aldrich (USA). Ellagic acid, Uro A, and Uro B were prepared in the laboratory of Dr. Iranshahi at Mashhad University of Medical Sciences. Ellagitannin powders were first dissolved in DMSO to achieve a concentration of 1 mg/mL and further diluted with phosphate buffer to prepare working dilutions for use in experiments.

3.2. Expression and Purification of Protein

Escherichia coli BL21 (DE3) cells containing the plasmid vector pet21a (+)-rhMMP-9 were used in this study. The cloned and expressed rhMMP-9, containing amino acid residues from 107 to 707, consisted of the catalytic domain (Phe107 to Gly223), a three fibronectin repeat domain (Asn224 to Cys388), a Zn²⁺-binding domain (Pro389 to Pro447), an O-glycosylated domain (Glu448 to Pro511), and a hemopexin domain (Val512 to Asp707) (Figure 1). The recombinant bacteria were used for the expression of the rhMMP-9 gene according to previously optimized conditions (19).

Figure 1.

Plasmid system encoding full-length recombinant human matrix metalloproteinase 9 (rhMMP-9) and amino acid sequence of rhMMP-9 divided by domains

A single colony of this microorganism was cultured in 10 mL of Luria-Bertani (LB) medium containing 100 µg/mL of ampicillin and incubated for 16 hours at 37°C with shaking to prepare a preculture. One milliliter of preculture was cultured into 100 mL of ampicillin-containing LB medium and incubated in a shaker incubator at 37°C until the optical density reached 0.5 - 0.6 at 600 nm. The MMP-9 expression was induced by adding 0.1 mM IPTG to the culture media and incubating for 4 hours. Cells were collected by centrifugation (2500 g, 5 min), lysed with 20 mM Tris buffer (pH 7), and sonicated twice (10 min, 4°C) with a 10-minute rest interval at 4°C. Sonication was performed using a professional ultrasonic cleaner Sonic 3MX by James Products Europe (160 W). Cells were centrifuged again (12000 g, 10 min, 4°C). The supernatant fraction containing the desired protein was purified by loading onto a Ni-NTA column. The Ni-NTA column was equilibrated with lysis buffer (pH 7). A linear gradient of washing buffer (pH 8.0) containing 50 mM Tris, 300 mM NaCl, and imidazole (30, 40, and 50 mM) was used to wash the column. rhMMP-9 protein was eluted with the same buffer containing 60 mM imidazole. Enzyme fractions were collected and dialyzed twice with 20 mM Tris buffer (pH 7) (20).

The rhMMP-9 expression and purification were validated by SDS-PAGE (12.5% polyacrylamide). The sample was diluted with 20 mM Tris (pH 7) buffer and 5x sample buffer. The mixture was maintained for 5 minutes in boiling water, and 15 µL of this solution was transferred into the lane. Gel staining was accomplished with Coomassie Brilliant Blue R-250 reagent (21).

3.3. Protease Activity Assay

To determine the hydrolyzing activity of rhMMP-9, casein was used as a substrate. A casein solution (1% w/v) was added to the extracted and purified rhMMP-9, and the final volume was adjusted to 1 mL with Tris buffer (pH 7). The expression product and substrate mix were incubated with shaking for 5 minutes at 37°C. To stop the process, 1 mL of cold trichloroacetic acid (TCA, 15%) was added to the mixture. This mixture was then kept at -20°C for 10 minutes and centrifuged twice (8000 g, 5 min, 4°C) to separate soluble peptides. The pellet was discarded, and the absorbance of the supernatant soluble peptides fraction was evaluated at 280 nm. All spectrophotometric measurements were accompanied by a blank assay in the absence of the enzyme. Enzyme activity was calculated from the tyrosine standard curve and expressed as mean ± SD in units of [µmol/min] after at least five repeats (22). SPSS software version 9 was used for data interpretation, and all graphs were drawn and kinetic parameters calculated using an Excel spreadsheet software program. The Quest Graph IC₅₀ Calculator was used for the calculation of IC₅₀ values.

3.4. Protein Concentration Measurement

Measurement of rhMMP-9 concentration was done using a routine Bradford test procedure (23). Several albumin standard solutions were prepared, and their absorbance at 595 nm was read after their reaction with Bradford reagent. The absorbance of the expression product of rhMMP-9 (supernatant fraction obtained in the process of protein expression before and after purification) after its incubation with Bradford reagent was also measured, and its concentration was obtained from the standard Bradford curve. The culture medium inoculated with recombinant E. coli BL21 (DE3) without stimulation of protein expression by IPTG was used as a control in all of the experiments.

3.5. Surface Plasmon Resonance Measurements

A dual-flow channel MP-SPR Navi 210A analyzer (BioNavis Ltd., Tampere-region, Finland) was used to evaluate SPR kinetic parameters on gold chips (BioNavis Ltd., Finland). A solution of both EDC/NHS ((N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride) (0.2 M) and NHS (N-hydroxysuccinimide) (0.05 M)) was employed for the activation of rMMP-9 enzyme carboxylic groups. Immobilization of the rMMP-9 enzyme on the CMD Au chip was accomplished by a standard amine coupling protocol. For blocking unreacted sites, ethanolamine–HCl (1 M, pH 8.5) reagent was used. Five different concentrations of Uro A, Uro B, and EA, including 5, 10, 20, 30, and 40 µM for each, were injected at a flow rate of 30 µL/min for 300 seconds (24).

3.6. Molecular Docking

The protein structure of human matrix metalloproteinase MMP-9 (gelatinase B) with PDB ID 1L6J was downloaded from the RSCB PDB database (25). The protein molecule was protonated, and water molecules were removed. Ligand structures were obtained from the PubChem site with code numbers 5281855 for EA, 5488186 for Uro A, and 5380406 for Uro B. Ligand structures were optimized using Gaussian View 03 software. Docking was performed with AutoDock 4.2 software. The blind dock was accomplished with a grid box having the number of grid points = 126 × 126 × 126 and with grid point spacing = 0.375 Å. Visualization of docking poses was done using Discovery Studio 2021 and Chimera UCSF.

4.1. Expression and Purification of Recombinant Human Matrix Metalloproteinase 9 from Escherichia coli

Active full-length rhMMP-9 was successfully expressed by E. coli BL21 (DE3) cells through the expression vector pET21a-MMP-9 (Figure 1). Expression of rhMMP-9 in the soluble fraction was revealed by SDS-PAGE (Figure 2) and through monitoring catalytic activity. The protein content of the supernatant after expression was compared with an equal sample that did not receive IPTG during expression. Protein concentration was measured by the Bradford method, and results confirmed that more than 40% of the total proteome of bacteria is rhMMP-9 protein. The specific activity of the recombinant enzyme was calculated as the enzyme's activity per milligram of total protein and constituted 1.86 ± 0.03 µmol min^-1 mg^-1. Purification with a Ni-Agarose column was done according to the His-tag of the recombinant protein, and the purified protein was used for kinetic experiments.

Figure 2.

SDS-PAGE analysis of recombinant matrix metalloproteinase 9 (rMMP-9) expression product: Lane 1, before isopropyl-β-D-thiogalactopyranoside (IPTG) induction (2.3 mg/mL); lane 2, after IPTG induction (8.45 mg/mL); lane 3, pure protein [matrix metalloproteinase 9 (MMP-9); lane 4, molecular marker

4.2. Inhibition of Recombinant Human Matrix Metalloproteinase 9 Activity by Ellagic Acid and Its Derivatives

A cell-free casein degradation-based assay was employed to evaluate the influence of ellagitannins on the catalytic activity of the rhMMP-9 enzyme. By adding various fixed quantities of casein as a substrate and a fixed concentration of enzyme, the enzyme activity was determined in the presence of different concentrations of ellagitannins (ranging from 0 to 70 µM). The rate of tyrosine production by the protease activity of rhMMP-9 was assessed based on a calibration curve over 1 minute, which refers to enzyme activity. As shown in Figure 3, ellagitannins could inhibit rhMMP-9 activity in a concentration-dependent manner. The concentrations of the inhibitor that could inhibit 50% of the enzyme’s activity, or its IC₅₀ values, were calculated as 13.17 µM, 33.29 µM, and 17.14 µM for Uro B, Uro A, and EA, respectively. Urolithin B showed the highest inhibition of rhMMP-9 based on protease activity measurements (Figure 3). Our results also demonstrated that Uro B and Uro A could entirely inhibit the catalytic activity of the enzyme at high concentrations, but the enzymatic function of rhMMP-9 does not reach zero in the presence of EA, which could reduce 90% of catalytic activity at a high concentration (60 µM).

Figure 3.

Dose-response curves of recombinant human matrix metalloproteinase 9 (rhMMP-9) activity inhibition at the presence of various concentrations of ellagic acid (EA), urolithin A (Uro A); and urolithin B (Uro B). Data were stated as mean ± SD (n = 6).

4.3. Mixed Inhibition Mechanism of rhMMP-9 Activity by Ellagitannins

Measurement of rhMMP-9 enzyme casein degradation ability at different concentrations of ellagitannins demonstrated classic Michaelis-Menten saturation kinetics. The Michaelis-Menten expression describes the behavior of many enzymes at different concentrations of substrates. To compare the behavior of the enzyme in the presence and absence of inhibitors (ellagitannins), the Lineweaver-Burk equation or double reciprocal plot was used in this research. With the increase of substrate amount, as shown in Figure 4, an escalation in enzyme activity was observed; nevertheless, a notable decrease in enzyme activity at higher substrate concentrations was detected.

Figure 4.

Michaelis-Menten curve [A for urolithin A (Uro A); C for urolithin B (Uro B); E for ellagic acid (EA)] and Lineweaver-Burk plot (B for Uro A; D for Uro B; F for EA) for recombinant human matrix metalloproteinase 9 (rhMMP-9) in the presence of various concentrations of ellagitannins. Values were expressed as mean ± SD (n = 5).

To examine the effects of inhibitors on the kinetic parameters of rhMMP-9, Michaelis-Menten saturation curves and Lineweaver-Burk plots were drawn in the presence of various concentrations of inhibitors (5 - 30 µM) and substrate range (0 - 0.3%) (Figure 4). It was shown that both the maximum velocity of the system and K_m decreased with increasing the concentration of Uro B from 5 to 15 µM. For Uro A, with the increase of inhibitor concentration from 5 to 15 µM, the maximum velocity of the system also decreased, but K_m increased. The EA also caused a reduction in V_max value, but the K_m constant decreased at the lower concentration of inhibitor (15 µM) and then remained constant at the higher concentration of inhibitor (30 µM). Table 1 summarizes the kinetic parameters for the rhMMP-9 enzyme in the presence and absence of inhibitors at different concentrations.

Comparison of the results and Lineweaver-Burk plots for rhMMP-9 in the presence of various concentrations of Uro A (Figure 4B), Uro B (Figure 4D), and EA (Figure 4F) demonstrated a mixed type of inhibition with different inhibitor constants. It is known that inhibitors can attach to the enzyme by two processes Equations 1 and 2.

Therefore, Equations 3 and 4:

Ki, or the dissociation constant for the EI complex, indicates how tightly the inhibitor is binding to the free enzyme [E] (Equation 3), while αKi (KI) represents the dissociation constant for the ESI complex, referring to how tightly the inhibitor is binding to the [ES] complex (Equation 4). Hence, two types of inhibitor constants, Ki and αKi, can be evaluated in each inhibition case according to secondary plots.

The first of these is a Dixon plot of 1/Vʹ_max against the concentration of inhibitor [I], from which the value of -αKi as the x-intercept can be determined (Figure 5A - C). The second is the slope of the primary plot (from the Lineweaver-Burk plot) against the concentration of inhibitor [I], from which the value of -Ki as the x-intercept can be determined (Figure 5D - F).

Figure 5.

Secondary plots of 1/Vʹ_max against [Io] [A for urolithin A (Uro A), B for urolithin B (Uro B), C for ellagic acid (EA)] and secondary plots of the slope of the primary plot against [Io] (D for Uro A, E for Uro B, F for EA). Data were stated as mean ± SD (n = 5).

Factor of alfa (α) can be determined from Equation 5.

Calculated values of Ki, KI, and α are given in Table 1 for each inhibitor. For all two urolithins tested KI < Ki and α is less than 1, so it could be concluded that these inhibitors demonstrated non-competitive-uncompetitive patterns of inhibition which is a specific type of mixed inhibition (27). This type of inhibition confirmed that urolithin molecules could attach to both free enzyme and enzyme-substrate complex but the inhibitor affinity to enzyme-substrate complex is more than to the free enzyme.

In the case of EA KI > Ki and α is more than 1, so it could be concluded that this inhibitor demonstrated competitive-non-competitive model of inhibition which is a specific type of mixed inhibition (27). This type of inhibition confirmed that EA could attach to both free enzyme and enzyme-substrate complex however the inhibitor affinity to the free enzyme is more than to enzyme-substrate complex.

4.4. Increased Affinity of Urolithin B to rhMMP-9 Based on SPR Results

The SPR technique, a current method for examining the interaction between ligand and protein, was employed in this research. As presented in Figure 6, association (Ka) and dissociation (Kd) constants were evaluated through sensorgrams (RU vs. time). Five identical concentrations for each of the three natural compounds were used (5, 10, 20, 30, 40 µM).

Figure 6.

Sensograms of A, urolithin A (UA); B, urolithin B (UB) and C, ellagic acid (EA) binding to recombinant human matrix metalloproteinase 9 (rhMMP-9) at different concentrations at 37^°C (5, 10, 20, 30, 40 µM)

The equilibrium constant KD, which evaluates the affinity of a ligand to a protein, was determined using the formula: ${\displaystyle =Kd/Ka}$ . A lower value of the KD constant indicates a higher interaction affinity between the protein and ligand. In this study, the KD values were measured for Uro A-rhMMP-9, Uro B-rhMMP-9, and EA-rhMMP-9 complexes to be 6.7 × 10^-5 M, 4.3 × 10^-5 M, and 11.3 × 10^-5 M, respectively, at 37°C, as shown in Table 2.

The achieved results indicate that Uro A-rhMMP-9 and Uro B-rhMMP-9 complexes have lower KD constant values compared to the EA-rhMMP-9 complex, which means that urolithins, especially Uro B, have more affinity for the MMP-9 enzyme, while this affinity is noticeably reduced for the EA-rhMMP-9 complex.

4.5. Molecular Docking Analysis

To understand and predict the inhibition mechanism and identify the attachment sites of the three inhibitor molecules on rhMMP-9, we conducted molecular docking. The results showed that there are two potential binding positions for Uro A, Uro B, and EA molecules on the MMP-9 protein: The catalytic domain and the fibronectin domain, which exhibit the lowest binding energy compared to other binding positions. Table 2 summarizes the docking results for both binding positions.

Molecular docking analysis confirmed that the three inhibitor molecules could bind to the fibronectin domain as an allosteric site, with EA showing more affinity to this domain compared to Uro A and Uro B, which could tightly bind to the catalytic site. Urolithin A (Figure 7A) and Uro B (Figure 7C), with binding energies of -7.44 kcal/mol and -6.72 kcal/mol, respectively, interacted with the loop connecting the βV sheet of the catalytic and the first repeat of fibronectin domains. Both Uro A (Figure 7B) and Uro B (Figure 7D) occupied approximately the same position (Gly215, Leu212, Lys214, Val217, His231, and Ala229), except for some additional interactions formed between Uro A and the receptor protein (MMP-9). The EA (Figure 7E and F) bound to the cavity formed between the catalytic and fibronectin domains, making connections with amino acids of both domains at a different allosteric site separate from the active site compared to urolithins, with a calculated binding energy of -7.35 kcal/mol.

Figure 7.

Molecular docking (35) poses [A, for urolithin A (Uro A); C, for urolithin B (Uro B); E, for ellagic acid (EA)]; and 3D interaction (B for Uro A; D for Uro B; F for EA) of ellagitannins with fibronectin domain of matrix metalloproteinase 9 (MMP-9)

The second docking position for ellagitannins, tested in this research, confirmed that they bound to the catalytic domain of the enzyme. Uro A (Figure 8A) and Uro B (Figure 8C), with binding energies of -8.41 kcal/mol and -8.54 kcal/mol, respectively, and EA (Figure 8E), with a binding energy of -6.19 kcal/mol, occupied an approximately similar binding position at the catalytic domain of the enzyme. Uro A, Uro B, and EA were fitted into the S'1 pocket of the active site, but EA had a more outward position compared to urolithins. Uro A (Figure 8B) and Uro B (Figure 8D) had similar interaction points except for one additional hydrogen bond for Uro A with residue Glu402, which plays a significant role in the catalytic mechanism of MMP-9. Matrix metalloproteinase 9 (MMP-9) enzyme amino acid residues interacting with the three ligands at the catalytic domain of the enzyme are shown in Table 2. The more outward position of EA in the S'1 pocket compared to the binding positions of Uro A and Uro B led to its interaction with different amino acid residues, except Arg 424 (Figure 8F). This outward position of EA can likely be explained by its more rigid and bulkier structure compared to urolithins, as well as its more hydrophilic structure compared to urolithins, which is less compatible with the hydrophobic nature of the S'1 pocket of the catalytic domain.

Figure 8.

Molecular docking poses [A for urolithin A (Uro A); C for urolithin B (Uro B); E for ellagic acid (EA)] and 3D interaction (B for Uro A; D for Uro B; F for EA) of ellagitannins with catalytic domain of matrix metalloproteinase 9 (MMP-9)