Biochemical characterization and inhibition of thermolabile hemolysin from Vibrio parahaemolyticus by phenolic compounds

Cloning the VpTLH gene and recombinant protein expression

The nucleotide sequence of VpTLH used in this study was obtained from the gene bank accession number AB012596.1. VpTLH was obtained as a synthetic gene (Atom®) and cloned into pET-28b (+) plasmid, adding the C-terminal 6x-His tag for the purification process. Chemically competent E. coli BL-21 strain rosetta II cells were transformed with VpTLH plasmid by thermal shock and incubated in SOC media (tryptone 2% w/v, yeast extract 0.5% w/v, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM MgSO4 and 20 mM glucose) at 37 °C by 4 h. Bacterial cells were plated in Luria-Bertani agar plates supplemented with kanamycin (25 µg/ml) at 37 °C overnight for the plasmid selection. After that, a single colony was inoculated in 5 ml of the antibiotic-LB medium by four h at room temperature; this culture was scaled up to 50 ml under the same conditions and incubated overnight. Subsequently, a Fernbach flask containing 1 l of LB medium added with kanamycin (25 µg/ml) was inoculated with the 50 ml culture and incubated at 37 °C and 220 rpm. When the optical density reached ≈ 0.6 units (λ = 600 nm), we added IPTG (Isopropyl β-D-1-thiogalactopyranoside) to a final concentration of 1 mM, inducing the overexpression of VpTLH. The Fernbach flask was maintained in an orbital shaker (200 rpm) for 16 hours at 25 °C. Bacterial cells were recovered by centrifugation at 7,000 rpm for 20 minutes at 4 °C, and the pellet was washed using 0.7% NaCl and spun as before. The supernatant was discarded, and the bacterial cell pellet was stored at −80 °C until use.

Protein purification and in vitro refolding

VpTLH was recovered from frozen pellet, which was resuspended in lysis buffer (50 mM Tris base, 1 mM DTT, 5 mM Benzamidine, 5 mM EDTA, 100 mM NaCl; pH 7.0) at ratio 1:8 (w/v). Bacterial cells were lysed by sonication on an ice bed with six pulses of 5-s and 5-s rest at 30 % amplitude. The homogenate was clarified at 12,000 rpm for 20 min at 4 °C, and SDS-PAGE (12%) stained with blue-coomassie was used to analyze the protein expression in soluble and insoluble fractions (Laemmli, 1970). Several overexpression conditions were analyzed, but in all cases, VpTLH was obtained as inclusion bodies. Therefore, inclusion bodies were isolated from insoluble cellular debris; which was resuspended by sonication (as before) using buffer 1 (50 mM Tris base, DTT 1 mM, 5 mM EDTA, 2% Triton x-100; pH 7.0) at ratio 1:4 (w/v), and the homogenate was centrifugated at 12,000 rpm for 20 min. The precipitate was recovered, and centrifugation was repeated three times. After that, the pellet was 2-times washed in buffer 2 (buffer 1, without Triton X-100) under the same conditions. Finally, the recovered inclusion bodies were solubilized using urea (50 mM Tris base, DTT 1 mM, urea 8M, pH 7.0) by sonication and incubated overnight at 4 °C with constant stirring. The homogenate was clarified (12;000 rpm at 4 °C for 30 min), and the soluble fraction containing the VpTLH in urea 8M was recovered. The protein concentration was quantified at 280 nm in a nanodrop®equipment (ϵ ≈ 96,510 M⁻¹ cm-1). VpTLH was purified by Immobilized Nickel Affinity Chromatography on an Äkta Prime (GE) under denaturing conditions at 25 °C. Therefore all buffers used during this process contained 8 M urea. Briefly, 1 ml HiTrap®IMAC-HP column was equilibrated with five volumes of buffer A (50 mM Tris base, 500 mM NaCl, 8 M urea; pH 7.4), then 4 ml (15 mg denatured protein) of VpTLH solution was loaded into the column. Non-bounded proteins were washed with buffer A until absorbance (λ = 280 nm) reached the base-line. Bound proteins were eluted with a linear gradient of buffer B (buffer A + 500 mM imidazole), eluted fractions were collected, and SDS-PAGE analyzed. Fractions containing a single band with a molecular weight of 48.3 kDa were pooled and refolded by dialysis with a cut-off membrane of 12–14 kDa at 4 °C. Refolding was carrying out by sequentially decreasing urea concentration (4, 1 M and without urea) in buffer 50 mM Tris-Base buffer at pH 7.5 (250 µl by 6 hours each). Buffer without a chaotropic agent was changed twice (12 hours each), then the protein solution was removed from membrane dialysis, centrifuged, and stored at 4 °C.

Enzymatic and hemolytic activity assay of the VpTLH

Enzymatic activity was measured spectrophotometrically using the lipase/esterase assay described by Nawani, Dosanjh & Kaur (1998), which was modified to measure the activity of TLH in the presence of egg yolk lecithin as an enzyme activator (Shinoda et al., 1991). Each reaction was conducted in a final volume of 1 ml containing: 50 mM Tris-HCl pH 7.5, 0.0001% egg yolk lecithin, 200 µM p-nitrophenyl laurate (PNPL), and the reaction was started by addition of 10 µL of refolded VpTLH (6.2 µM final concentration). PNPL hydrolysis was measured at a wavelength of 410 nm (PNPLε_410nm = 11.8 cm⁻¹ M⁻¹) for 5 min at a temperature of 37 °C in a Cary 50®UV-VIS spectrophotometer (Varian). Negative control assay consisted of the same reaction without VpTLH. Specific activity was calculated using the equation: (1)${\displaystyle U∕mgprotein=\left(\mathrm{m}\cdot \mathrm{V}\right)∕\left(\epsilon \cdot \mathrm{p}\cdot \mathrm{l}\right)}$

where m is the slope of the reaction; V, the reaction volume; p the protein concentration (mg/ml) and l, cell path-length in cm.

Hemolytic activity of VpTLH was quantified using human erythrocytes as a substrate (Malagoli, 2007). Erythrocytes were extracted from human blood, which was kindly donated by one male volunteer, who was previously informed about the extraction procedure, the use, and the proper disposal of blood samples, according to the established protocol by the Institutional program of environmental health and safety of the Universidad de Sonora. Also, the volunteer signed an informed consent about her/his participation in this research. Briefly, erythrocytes were isolated from Human blood by centrifugation at 700 rpm for 5 minutes at 4 °C. The plasma was discarded, and erythrocytes were washed three times (0.9% saline solution); they were then carefully resuspended in saline phosphate buffer (PBS; 8 mM Na2HPO4, 2 mM KH2PO4, 37 mM NaCl, 2.7 mM KCl, pH 7.4) to 1:200 ratio. Hemolytic activity was measured by the release of hemoglobin at 540 nm and expressed as a percentage of hemolysis using Tween 20 (0.2%) as a positive control (100% hemolysis). Erythrocytes autolysis (without enzyme or detergent) was also recorded and subtracted in each assay. Each assay was conducted with 300 µL of the erythrocyte suspension, 0.0001% egg yolk lecithin, and VpTLH (0.23 mM final concentration), PBS was added to a final volume of 1 ml. Reaction tubes were carefully homogenized and incubated at 37  °C for 60 min after tubes were centrifuged (2 min at 700 rpm), and the supernatant was recovered, and absorbance was recorded at wavelength of 540 nm. All measurements were carried out by triplicate.

Biochemical parameters of VpTLH

To determine the pH effect on VpTLH enzymatic activity, we used several buffer solutions varying pH values (5.5, 6, 7, 7.5, 8, 8.5, 9, 10, and 11) under the standard assay mentioned before. PNPL molar extinction coefficient (ε) was used for each of pH values: 1.97, 6.52, 11.8, 12.75, 13.9, 14.3, 14.6 and 14.6 (cm µ⁻¹ M⁻¹), respectively (Kademi et al., 2000). Sodium citrate (pH 6), Tris-HCl for pH 7 to 9, and sodium carbonates (pH 10 and 11). All buffer’s concentration was kept constant at 100 mM. The temperature on VpTLH activity was assayed in standard conditions at pH 8.0 by varying temperature assay from 10–80 °C, increasing by 10 °C. Results were expressed as a percentage of residual activity. The temperature with the highest Vp TLH activity was used as 100% of residual activity. Furthermore, VpTLH activation energy (Ea) was calculated by plotting linearized Arrhenius equation: (2)${\displaystyle \ln \left(\mathrm{k}\right)=\ln \left(\mathrm{A}\right)-\left(\mathrm{Ea}∕\mathrm{R}\right)∕\left(\mathrm{T}\right)}$

where slope m = −Ea/R; k, initial velocities; T, temperature (kelvin) and R, universal gas constant (J/molK). A negative control without enzyme in each pH and temperature condition was assayed to discard substrate precipitation or chemical hydrolysis,

For the thermostability test, the enzyme was incubated 15 min at temperatures from 10 to 80 °C with 10 °C intervals. Then enzymatic activity was measured under standard conditions. Residual activity was calculated as a percentage at VpTLH showed the highest activity (100%). Additionally, VpTLH melting-temperature (Tm) was obtained by Boltzmann sigmoidal analysis using Prism 5 software (GraphPad®).

Determination of the VpTLH Michaelis-Menten parameters

The kinetics parameters km and V_max of VpTLH were determined from the initial velocities by varying PNPL concentrations from 20 to 400 µM using the enzymatic standard conditions mentioned before. Initial velocities were recorded for 2 minutes and were adjusted to the Michaelis-Menten non-linear regression model using the Prism 5 program (GraphPad®). All measurements were carried out by triplicate. Additionally, the Michaelis-Menten constant (K_m), V_max, and turnover number (k_cat) of the enzyme were calculated (Michaelis et al., 2011).

Enzymatic and hemolytic activity inhibition assays

The following phenolic acids were used: gallic acid (GA), vanillic acid (VA), protocatechuic acid (PR), and chlorogenic (CL). The following flavonoids were used: quercetin, morin, rutin, and epigallocatechin gallate (EGCG) added to the standard enzymatic assay. All flavonoids were used to a final concentration range: 1–20 µM, while phenolics acids were evaluated at 30 µM and 100 µM. Meanwhile, quercetin, morin, and EGCG were evaluated as an inhibitor for VpTLH hemolytic activity, which was added to the standard hemolytic assay to a final concentration range of 1–20 µM. VpTLH residual activity (enzymatic or hemolytic) in the presence of phenolic compounds was calculated as a percentage of VpTLH activity in the absence of phenolic compounds. Flavonoids showed the highest inhibitory effect (see results section), were selected to calculate inhibitor concentration required to reduce 50% of enzyme activity (IC₅₀) using a concentration range of 0.1–50 µM. The data were normalized using VpTLH in absence of inhibitor as 100% of enzymatic activity; then analyzed with the dose-response variable slope model based in Hill Slope equation as described in Prism-5 software (GraphPad®).

Comparison of the TLH sequences from different Vibrio species and homology modeling of the VpTLH

The amino acid sequence of the VpTLH, with access code Q99289 in the UniProt database, was compared with that of other Vibrio species, including Vibrio cholerae (Q9KMV0), Vibrio alginolyticus (C7EWQ8), Vibrio harveyi (Q2XPT2), Vibrio anguillarum (A0A191WA34) and Vibrio vulnificus (A0A1V8MSL8). These were compared using the ClustalW algorithm (File S1). Based on the alignment obtained, the conserved regions of the sequences were identified, and the presence of the GDSL and SGNH motif, which are specific domains of this type of enzymes. The VpTLH homology model was built using free access algorithms such as PHYRE2, I-TASSER, SWISS-MODEL, and commercial software MOE (File S2). As we expected, all the predicted structures based on 6JL1 crystal structure (TLH from V. vulnificus), showed that the overall structure of predicted models is quite similar in RMSD = 0.324 Å. PHYRE2 model (intensive mode) showed the lowest RMSD (0.223 Å). Therefore this model was used for structural analysis and molecular docking simulations. Structural analysis was performed in UCFS Chimera 1.13.1 (Pettersen et al., 2004) and CCP4-MG programs (McNicholas et al., 2011).

Molecular Docking of the VpTLH substrates and Inhibitors

Both natural (phosphatidylcholine, PC), synthetic (PNPL) substrate and best IC50 inhibitors were docked into Vp TLH active site using AutoDock Vina algorithm in UCFS Chimera 1.13.1 program (McNicholas et al., 2011; Trott & Olson, 2010). Before docking experiments, both ligands and protein were analyzed in DockPrep function to minimize structure, partial charges calculation, and hydrogen atoms were also added. The 3D structure files of the PNPL, quercetin, morin, and EGCG compounds were obtained from the PubChem database with access codes 74778, 5280343, 5281670, and 65064, respectively. In contrast, PC structural data were obtained from the MOE database. The fpocket2 algorithm predicted a putative ligand pocket cavity in the Phyre2 investigator program (Kelley et al., 2015; Le Guilloux, Schmidtke & Tuffery, 2009) and the Site Finder function of the MOE program. Sequence alignments located VpTLH active site amino acids and spatial coordinates were established by superposition with Vv TLH crystal structure (RMSD = 0.400 Å). VpTLH final docking area was established in coordinates X = 24.4143 Å, Y =  − 2.51601 Å and Z =  − 35.3974 Å (volume = 28,802.47 ų). For each ligand, 20 poses were generated using Iterated Local search method supported in AutodDocvina (Trott & Olson, 2010); models with low free energy binding force function (ΔG) were analyzed in Discovery Studio 2019 program (Biova®).

Purification and refolding of VpTLH

All the over-expression conditions produced the VpTLH (≈ 47 kDa) protein in the insoluble fraction (Fig. 1A, lanes 1–5). It was in agreement with no phospholipase activity detected on soluble protein fraction by an agar-plates assay using egg yolk lecithin as substrate (0.1 % egg yolk lecithin, 3% agar). The higher production of VpTLH was using 0.4 mM IPTG for 16 hours at 25 °C. After that, the inclusion bodies were solubilized in 8 M urea and purified by a single chromatographic step (IMAC) under denaturing conditions. As a result, we obtained a single peak elution at 115 mM imidazole (Fig. 1B), corresponding to a unique SDS-page band of VpTLH molecular weight (Fig. 1A, lanes 6 and 7). The VpTLH was refolded by dialysis, and the esterase activity using PNPL as a substrate on the soluble fraction confirmed that protein was active (Fig. 1C) with a specific activity of 0.47 U/mg of protein (Shinoda et al., 1991). Figure 1D showed that VpTLH enzymatic activity gradually increased, with the highest PNPL hydrolysis at 10 µg lecithin, suggesting that VpTLH could hydrolyze PNPL in lecithin, as early noticed. Furthermore, the enzymatic activity decreased by <50% in the presence of >50 µg lecithin. A quantitative hemolytic activity assay showed that 2–10 µg of purified VpTLH elicits 76–85 % relative activity (100% hemolysis was tween-20) (File S3). In this sense, we assayed both enzymatic and hemolytic activity using 10 μg of lecithin and confirmed that VpTLH is a lecithin-dependent protein.

pH and temperature effect on VpTLH activity

VpTLH maximum activity was detected at pH 8.0 (100%), slightly decreasing at pH 8.5 (97% residual activity). Enzymatic activity suddenly decreased at pH <7.5 and >9.0. Contrary, the VpTLH was inactive at acidic medium (pH 6.0) while showed low activity (<20%) at alkaline pH (Fig. 2A). Amino acid sequence analysis indicated that VpTLH was thermolabile hemolysin (Nishibuchi et al., 1989). Interestingly, to our knowledge, there is no report evaluating the effect of temperature on VpTLH activity. First, we found that activity increased from 10 °C (10% residual activity) to a maximum level at 50 °C. At higher temperatures (>50 °C), the VpTLH enzymatic activity rapidly decreased, and at 80  °C, no esterase activity was found (Fig. 2B). These data were analyzed using the linearized Arrhenius equation by plotting natural logarithm (initial velocities) against the inverse of each temperature in Kelvin degrees (Fig. 2C). This analysis showed a single inflection point (50 °C) with linear activity decreasing as a temperature increase (until 80  °C), suggesting that temperature above 50 ^∘C drastically affects VpTLH enzymatic activity. The Arrhenius equation calculates the activation energy (Ea), which is the energy that VpTLH requires to hydrolyze PNPL, resulting in Ea = 26, 688 kJ/mol. Additionally, we evaluated the temperature stability by incubating the enzyme from 10 to 80 °C for 20 min. After that, the enzyme retained >80% residual activity at temperatures below 40 °C. At 60 °C the enzyme only retained 6 % of residual activity, being inactive at 70 and 80 °C. The data were adjusted to the Boltzmann sigmoid model (R² = 0.9985), calculating VpTLH Tm = 50.94 °C (Fig. 2D).

VpTLH Michalelis-Menten parameters

Initial velocities were measured using PNPL as a substrate from 20 to 400 µM (Fig. 3) and showed a typical Michaelis-Menten profile by plotting initial velocities vs. [PNLF]. Data were adjusted to non-linear regression analysis with a correlation factor R² = 0.9851. After that, we obtained the VpTLH kinetics parameters, a V_max = 0.7736 U/mg (± 0.041) and a k_m = 0.151 mM (±0.017) (Fig. 3). Also, the enzyme turnover number (k_cat) was 37.37 s-1 (±1.97 SD) (Table 1). TLHs Michalelis-Menten kinetic parameters reports are scarce; recently Vv TLH kinetics constants were determined using a fluorogenic substrate (Red/Green BODIPY PC-A2) (Wan, Liu & Ma, 2019) and other authors reports PLA2 activity (from snake venom) using 4N3OBA as substrate (Pereanez et al., 2011). In spite of differences in substrates chemical composition used in each report, we observe that VpTLH has the highest Vmax and turn over compared with other TLH and other PLA2 enzymes, lower substrate affinity (k _m) than V. vulinificus TLH (Wan, Liu & Ma, 2019). Table 1, shows that substrate affinity have variable magnitude among compared enzymes, which is closely-related to substrate differences used (Eisenthal, Danson & Hough, 2007). Meanwhile, Wicka et al. (2016), reported similar substrate affinity and k_cat for cold-adapted GDSL-lipase from Pseudomonas sp. S9 using p-nitrophenyl butyrate which has short carbon chain in fatty acids substituent than PNPL (Wicka et al., 2016).

Polyphenols inhibited both VpTLH enzymatic and hemolytic activity

As shown in Fig. 4, phenolic acids GA, PR, and VA inhibited the VpTLH activity by 20% at 30 µM, while CL does not affect activity compared to control (assay without phenolics acids). Increasing phenolics compounds to 100 µM did not affect the VpTLH activity (p < 0.05) (Fig. 4). Also, rutin did not affect the activity at all evaluated concentrations; while, quercetin, morin, and EGCG at 20 µM (the highest concentration) decreased the activity by 70%, 65%, and 67%, respectively (Table 2). At low concentration (1 µM), morin was the most effective to inhibit VpTLH (30% inhibition), whereas, at 10 µM, both quercetin and EGCG were also able to reduce activity by 60%. These results suggest that flavonoids were more suitable to inhibit VpTLH phospholipase activity. Additionally, the dose-response analysis showed that quercetin was the best-evaluated inhibitor (IC₅₀=4.51 µM). EGCG and morin also exhibited similar IC₅₀ values: 6.290 µM and 9.914 µM, respectively (Fig. 5).

The phospholipids are abundant in erythrocytes’ membrane cells; thus, VpTLH hemolytic activity in flavonoids at the same concentrations used in enzymatic inhibition experiments (1–20 µM) was assayed. However, the effect of phenolic compounds on VpTLH hemolytic activity was not evaluated. All flavonoids gradually diminished hemolysis, increasing its concentration, compared to control without flavonoids (Fig. 6). Low concentration (1 and 5 µM) did not significantly affect erythrocytes lysis, but quercetin and EGCG at 10 and 20 µM inhibited the VpTLH hemolytic activity 15% and 30%, respectively. Morin achieved only 15 % inhibition at the highest evaluated concentration. We could not evaluate flavonoids at concentration >20 µM because precipitation of hemolytic assays components was observed.

Homology modeling and docking analysis indicates that VpTLH has conserved folding and an active site cavity suitable to bind both substrates and inhibitors

We aligned the VpTLH amino acid against TLHs of other pathogenic Vibrio species such as Vibrio alginolyticus (Va), Vibrio harveyi (Vh), Vibrio campbelli (Vc), Vibrio cholerae (Vch), Vibrio diabolicus (Vd), and Vibrio anguillarum (Van). After that, the VpTLH maintained a high-sequence identity >80 % with Vd, Va, Vh, and Vc; while Vv, Van, and Vch showed a lower sequence identity being 73%, 65%, and 64 %, respectively (File S1). This analysis showed that VpTLH has hydrolase/esterase superfamily well-conserved GDSL and SGNH motifs (Akoh et al., 2004; Upton & Buckley, 1995), as was previously reported in other TLHs (Jang et al., 2017; Jia, Woo & Zhang, 2010; Wang et al., 2007). Two main domains comprised the TLH sequence, the N-terminal domain included from amino acid residue 24 to 133 (signal peptide 1–23), and C-terminal (also called SGNH domain) comprised 134–418 amino acids (numbering were according to the VpTLH sequence). The N-terminal’s biological function is not well defined, while SGNH-domain is directly related to enzymatic function divided into four blocks that contain invariable catalytic residues (Akoh et al., 2004). GDSL motif is located in block I (139-158) and contained catalytic serine residue (Ser153), while that Gly204, Asn248, and His393 are found in blocks II, III, and V, respectively. SGNH hydrolases have conserved catalytic triad His-Ser-Asp. This last amino acid residue was also found in VpTLH block V (Asp390), and Gly’s substitutions were found in Va and Vv TLHs (Jang et al., 2017; Li, Mou & Nelson, 2013).

VpTLH homology model was built in PHYRE2 using Vv TLH crystal structure (PDB: 6JL1) as a template since its share >74% sequence identity with VpTLH, and the resulting model showed an excellent superposition with the template (RMSD = 0.256 Å) (Fig. 7A). N-terminal domain (109 amino acid residues) was composed of β-sheets and three small α-helices exposed to solvent while the sizeable C-terminal domain (274 a.a) adopts a typical SGNH α/ β/ α folding related to phospholipase function. Ser-Asp-His catalytic triad and other active site amino acids were located in this domain and distributed in four well-conserved blocks of the SGNH superfamily (Fig. 7A) (Akoh et al., 2004). β-sheet central core flanked by α-helices compose this domain; all four blocks converge to form ligand pocket cavity as was predicted by the fpocket2 algorithm in the Phyre2 investigator program (Kelley et al., 2015; Le Guilloux, Schmidtke & Tuffery, 2009) and Site Finder function of the MOE program. We superposed the VpTLH model to the Vv TLH structure, and the catalytic triad was located in the pocket, suggesting it as the VpTLH active site (Figs. 7A and 7B). Although N-terminal domains remain close to the active site, no catalytic function was previously reported for this domain. Nucleophile Ser153 interacted with both Asp390 and His393 that comprised the conserved catalytic triad of serine hydrolases family (Fig. 7C). However, structural and biophysical experimental techniques are necessary to get a more precise model; crystallization assays are in progress in our laboratory with both apo and holo Vp TLH.

Finally, we used molecular docking simulations of possible interactions with substrates and flavonoids into the C-terminal domain with area = 28.36 Å × 32.65 Å × 31.09 Å  following AutoDockVina protocols with centered predicted active site cavity (Fig. 7D). For each ligand, 20 interaction-models were constructed, and we selected the best solution based on the free-energy binding and the position inner the active site. The most favorable interacting-coupling score for substrates were: Phosphatidylcholine (PC) = −3.9 kcal/mol and PNPL= −4.5 kcal/mol. PC is a natural substrate for TLH and one of the most abundant phospholipids, including phosphatidylserine, in the cellular membrane.

Figure 8 shows the molecular docking and interaction diagram of PC and PNPL into Vp TLH active site. PC properly accommodates with the aliphatic chain (glycerol sn-1) buried inside the active site, while fatty acid (glycerol sn-2) was located in the active site surface and polar substituent (choline) is more exposed to the solvent. PC interacted with several active site amino acids by hydrogen bonds between Gln292 and OH-groups (both fatty acids) and Asn254 with phosphate group; also, Lys303 stabilize phosphate group by a saline bridge. Asn252, Ala206, and Tyr253 stabilize fatty acids chains and choline methyl groups by aliphatic C-H interactions. Non-canonical substrate interaction prediction showed an embedded PNPL into VpTLH active site. The p-nitrophenol ring shows hydrophobic π-alkyl- interaction to Ala206 (coordination), and the carboxylate was stabilized by hydrogen bonds with Tyr360 lateral chain OH-. 14-C fatty acid chain (laurate) was located in a similar arrangement as the second PC fatty acid substituent (sn-2). Interestingly, catalytic residues (SGNH) did not interact with both substrates in docking experiments under the used conditions. Similar results were also reported in the crystallographic model of Vv TLH single mutant (Gly389Asn) in complex with hexamethylene glycol, concluding that TLHs could elicit conformational flexibility upon substrate binding (Wan, Liu & Ma, 2019).

Flavonoids that inhibited both enzymatic and hemolytic activity (quercetin, morin, and EGCG) were also docked into the predicted VpTLH active site (Fig. 9A). Interactions diagram showed that free energy binding was favorable to EGCG > morin > quercetin with scores: −7.9, −7.2, and −6.4 kcal/mol, respectively. EGCG (gallic acid substituent) and quercetin (ring A) were buried into the active site as observed with both substrates. Ala206 π-alkyl stabilizes EGCG gallic acid and quercetin ring A, while hydrogen bridge Gln292 with quercetin 1′-oxygen and EGCG 4′-hydroxyl group (Fig. 9B). Such interactions were not observed in morin docking simulations, positioned near active site surface throw π-anion interaction with Glu300 (ring B and C) and hydrogen bridge to Thr297 and Van der Waals forces with oxygen groups located in morin ring B (Fig. 9B). Morin and quercetin have identical chemical formula and molar mass but differ in hydroxyl; morin is 2′, 4′ while quercetin is 3′, 4′ orientation in ring B; such could be related to differences in ΔG scores and active site interactions (Xiao et al., 2012). EGCG and quercetin displayed similar disposition into active site VpTLH, as observed with substrates evaluated. These results are consistent with the inhibition experiment; therefore, both flavonoids could be suitable compounds for chemical modification for structure/function studies and evaluate inhibition/attenuation capacity during V. parahaemolyticus infection.