Materials

Porcine pancreatic lipase type II, p-nitrophenyl laureate (pNPL), caffeic acid, p-coumaric acid, capsaicin, quercetin, Tris, bovine serum albumin (BSA) and Triton X-100 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Methanol and ethanol were purchased from J.T.Baker (Center Valley, PA, USA). Orlistat capsules from Redustat® (Liomont® Laboratories, Toluca, Mexico) were purchased in a local drugstore.

Extraction of polyphenolic compounds from dry smoked jalapeño pepper (chipotle)

Extraction of polyphenolic compounds was carried out according to Moreno-Escamilla et al. (13), the phenolic content was determined by the Folin-Ciocalteu’s method, and results were expressed in µg of gallic acid equivalents (GAE) per g of extract.

Pancreatic lipase activity assay

Commercial pancreatic lipase was prepared according to McDougall et al. (17) with some modifications. Pancreatic lipase was dissolved in 20 mM Tris buffer (pH=7.2) at the concentration of 2.0 mg/mL, centrifuged at 1411´g in Eppendorf 5810R centrifuge (Eppendorf AG, Hamburg, Germany) for 20 min, and the supernatant was filtered in a polyvinylidene fluoride (PVDF) membrane filter (0.22 µm; Durapore_®, Hercules, CA, USA). Purity of the lipase and co-lipase was estimated by SDS-PAGE analysis (93% purity; EDAS 290, Kodak®, New York, NY, USA). Protein concentration was determined by Bradford method (18), reading the absorbance of cationic form of the dye Coomassie Blue G250 at 595 nm, and results were expressed in mg/mL using bovine serum albumin (BSA; 1 mg/mL) as standard. Pancreatic lipase substrate (pNPL, 0.08% mass per volume) was dissolved in 5 mM sodium acetate solution (pH=5.0, 1% Triton X-100). Final pNPL concentrations were 11.7-124.2 μM. Caffeic acid, p-coumaric acid and capsaicin were dissolved separately at three different concentrations (25, 50 and 75 μg/mL) in methanol. Due to its lower solubility, quercetin was dissolved in ethanol at the same concentrations. An Orlistat solution was prepared at 1.2 mg/mL in methanol (19). Stability of the compounds in the solvent was tested for the time of assay by measuring their UV-Vis spectra (Agilent 8453 HP UV-Vis spectrometer G1103A; Agilent Technologies, Santa Clara, CA, USA).

The control assay contained 500 μL of Tris 200 mM working buffer solution (pH=8.2), 260 μL of distilled water, 140 μL of pancreatic lipase solution and 40 μL of substrate solution. For the inhibition studies, 60 μL of each polyphenolic compound (or extract) and 200 μL of distilled water were used. The substrate was added to start the reaction, then the absorbance was recorded at 400 nm during 1000 s at 37 °C. All samples were assayed in triplicate.

The inhibition percentage was calculated according to Dalar and Konczak (20) from the measured absorbances of control and sample (endpoint):
where A_cb is the absorbance of the control blank, A_c is the absorbance of the control, A_sb is the absorbance of the sample blank and A_s is the absorbance of the sample. Blanks were prepared without the pancreatic lipase solution and the substrate solution, respectively. IC₅₀ was calculated from a substrate concentration versus inhibition percentage plot.

Accurate K_i values of polyphenolic compounds were determined by calculating the apparent maximum reaction rate (v_max) and Michaelis-Menten constant (K_m) by both linear (Lineweaver-Burk) and non-linear (Michaelis-
-Menten) analyses First, apparent v_max and K_m values were obtained from the least-square regression lines by plotting the reciprocal of the substrate concentration [S] against the reciprocal of the enzyme reaction velocity as in the work by Alvarez-Parrilla et al. (21):

These values were used as initial values for the determination of the parameters by non-linear analysis using Sigma Plot v. 13.0 (22), by fitting the value of the experimental initial reaction rate (v₀) as a function of the initial inhibitor concentration to the Michaelis-Menten equation (Eq. 3). The v₀ values were obtained from a linear curve fitting (slopes) of a time-course plot of the enzymatic reaction:

K_i and K_i’ (dissociation constant of free enzyme and enzyme-substrate complex, respectively) values for a mixed-
-type inhibition were obtained from the following equations (23):

where apparent K_m’ and v_max’ are the values in the presence of the inhibitor, and [I] is the inhibitor concentration. These calculations were done for each inhibitor concentration.

Polyphenolic compound-pancreatic lipase interaction measured by fluorescence spectra

Different volumes of caffeic acid, p-coumaric acid, quercetin and capsaicin (or extract) at 0.0125 M were added into a 10-mL flask (final concentrations 0.025-0.25·10^-4 M). A volume of 1.0 mL of the pancreatic lipase solution (10^-4 M) was added, and the solution was diluted to 10 mL with the working buffer and incubated at 37 °C for 1 h in darkness. Emission fluorescence spectra were measured in the 300-550 nm interval, at a fixed excitation wavelength of 280 nm in a RF 5301-PC spectrofluorometer (Shimadzu, Columbia, MD, USA). Even though the intrinsic fluorescence of pancreatic lipase is attributed to three amino acid residues, phenylalanine (Phe), tyrosine (Tyr), and tryptophan (Trp), the fluorescence of Phe and Tyr residues at this wavelength is negligible, so intrinsic fluorescence of the enzyme is mainly due to Trp (24). The polyphenolic compounds do not exhibit Förster resonance energy transfer (FRET) at that excitation wavelength (25). The Stern-Volmer equation was used to determine the presence of a quenching mechanism, by plotting F₀/F against [Q]:
where F₀ and F are the fluorescence intensities at the maximum λ_em in the absence and presence of an interacting molecule (polyphenol), respectively, and [Q] is the molar concentration of the interacting molecule. The k_q is the bimolecular quenching constant, τ₀ is the fluorescence lifetime in the absence of the quencher, and K_sv is the Stern-Volmer quenching constant (26). A τ₀ of 1.59 ns was used for pancreatic lipase, according to previously published results (27).

Molecular docking

The crystal structure of pancreatic lipase (Protein Data Bank (PDB) (28) code 1ETH (29)) was used as template for the generation of pancreatic ligand complex models. Ligand structures: caffeic acid, p-coumaric acid and quercetin were obtained from crystallographic complexes with different proteins (PDB codes: 4EYQ (30), 1D7E (31) and 1E8W (32)). Capsaicin structure was generated and minimized using PyMOL (33). The docking analyses were performed with AutoDock Vina using the interphase installed in USCF Chimera (34) run with the default parameters and a search volume of approximately the same size as pancreatic lipase. The crystal structure of pancreatic lipase was considered rigid, and polyphenolic compounds and capsaicin structures were considered flexible during docking. The possible interactions of the best-scored solution of the models for each ligand were analyzed.

Statistical analysis of the enzyme kinetics and the binding experiments

All the experimental assays were run in triplicates. Results are expressed as mean values±standard deviation. Analysis of variance and Fisher’s least significant difference analysis were performed by SPSS v. 20 software (35) for the determination of statistical differences between treatments.

Pancreatic lipase activity

The effect of four compounds: caffeic acid, p-coumaric acid, quercetin and capsaicin (Fig. 1), commonly found in hot pepper, and the hot pepper extract, on the activity of pancreatic lipase were monitored at saturation conditions, using 62.1 μM pNPL. Enzyme inhibition percentage increased in a hyperbolic trend, as the concentration of the polyphenolic compounds increased. p-Coumaric acid and capsaicin appear to have weak inhibition power, since lower inhibition was observed even at higher concentrations (data not shown). The hot pepper extract showed no pancreatic lipase inhibition (36). To compare the inhibitory capacity of the analyzed compounds, IC₅₀ values were determined for caffeic acid, p-coumaric acid, quercetin and orlistat (Table 1). Caffeic acid had the highest IC₅₀, followed by p-coumaric acid and finally quercetin, which showed an IC₅₀ statistically similar to that of orlistat. In agreement with these results, You et al. (10) observed that quercetin had lower IC₅₀ (276.8 µM) than ellagic acid (342.4 µM). Both quercetin and ellagic acid have the same molecular mass of 302.2 g/mol, respectively; however, their molecular structures are different (Figs. 1c and d). Quercetin, which is a flavonoid, has two aromatic rings in its structure and one heterocyclic ring in a linear structure, while ellagic acid possesses two aromatic rings and two heterocyclic rings, presenting a more rigid structure. This emphasizes the relevance of the structure of flavonoid in comparison with other polyphenolic structures (37). For example, the differences between ellagic acid and quercetin might be associated with their different inhibitory activities against pancreatic lipase (17). Differences between phenolic acids and flavonoids such as quercetin may be responsible for our obtained different inhibitory capacities. Similar to our results, where phenolic acids had higher IC₅₀ values than quercetin, Sergent et al. (1) observed that quercetin IC₅₀ ((21.5±9.4) µM) was lower than that of ferulic acid ((123.9±13.4) µM; Fig. 1e) and not significantly different from Orlistat ((32.0±8.5) µM), detected using 4-methylumbelliferone as substrate.

Fig. 1

The chemical structures of polyphenolic compounds: a) caffeic acid, b) p-coumaric acid, c) quercetin, d) ellagic acid, and e) ferulic acid; and the compounds: f) capsaicin, and g) orlistat. Elements (C, H and O) of caffeic acid, p-coumaric acid, quercetin and capsaicin for the molecular docking analysis are indicated

In contrast, no significant effect of the extract on pancreatic lipase activity was observed. Similar results have been reported for black chokeberry (Aronia melanocarpa L.) extracts (38). However, other studies have shown high to moderate pancreatic lipase inhibition by plant extracts such as cumin (Cuminum cyminum) (39), green tea (Camellia sinensis) (40), muscadine grape (Vitis rotundifolia) (10), and other edible plants, for example Beijing grass (Murdannia loriformis) (41). Here, the lack of pancreatic lipase-inhibitory effect could be explained by the difference in the content of bioactive compounds compared to other extracts with high inhibitory activity, or by the synergistic/antagonistic effects among bioactive compounds in the extracts (38). According to Chen and Kang (15), hot pepper extracts contain high mass fractions (per dry mass) of p-coumaric acid ((4.56±0.11) mg/g) and capsaicin ((9.05±0.05) mg/g), both of them with low pancreatic lipase-inhibitory activity, which may explain the low effect of the extract. Another phenolic acid mainly found in hot peppers is caffeic acid ((1.32±0.08) mg/g), while quercetin is the major flavonoid ((8.86±1.03) mg/g). The observed IC₅₀ values of p-coumaric acid, caffeic acid and quercetin were compared to their content found by Chen and Kang (15). We found that only quercetin had a higher content than its IC₅₀ value (almost 5-fold), whereas phenolic acids had lower content than their IC₅₀ values (i.e. almost 6-fold for p-coumaric acid). In this way, the absence of the inhibitory effect of the hot pepper extract may be explained by the low content of capsaicin and phenolic acids (lower than their IC₅₀ values).

The effects of caffeic acid, p-coumaric acid, quercetin and capsaicin on the enzyme kinetics of pancreatic lipase using pNPL as substrate were evaluated, and results are shown in Fig. 2, where the initial velocity v₀, shown as absorbance change (λ=400 nm) in arbitrary units per second, versus substrate concentration (µM) was plotted. Even though all compounds had a significant effect (p<0.05) on v₀, compared to control, quercetin seemed to be the best inhibitor. To characterize the inhibitory pattern, both linear and non-linear analyses of inhibition curves of each compound and the extract were performed. Linearized plots indicated a mixed-type inhibition, in agreement with other authors (5, 10, 26), and suggested two inhibition components, competitive and non-competitive. Next, non-linear regression fit of data to Eqs. 3–5 was performed to obtain the catalytic parameters of pancreatic lipase in the presence and absence of polyphenolic compounds (Table 1). Significant differences in apparent K_m values with respect to control were observed among all compounds. The apparent v_max values showed no significant differences with respect to control. Similar apparent v_max and smaller apparent K_m values of quercetin than of control support the idea that a mixed-type inhibition regulates the process (42). Phenolic acids possessed significantly lower apparent K_m values than control. Quercetin had an apparent K_m higher than control, and capsaicin did not exhibit statistically significant differences compared to control. According to Zhang et al. (43), a higher molecular mass structure might be beneficial for decreasing enzymatic activity (higher apparent K_m value), which can be attributed for example to the increase of hydroxyl group in the structure of quercetin with respect to phenolic acids. Narita and Inouye (42) observed a mixed-
-type inhibition when using p-coumaric acid as inhibitor of pancreatic α-amylase activity and explained that lower apparent K_m values in the presence of inhibitors meant that the binding affinity of the substrate and the inhibitor for their respective sites increases when both are present (p-coumaric acid and substrate).

Fig. 2

Effect of caffeic acid, p-coumaric acid, quercetin and capsaicin on the hydrolysis of p-nitrophenyl laureate (pNPL) catalyzed by pancreatic lipase. Symbols represent experimental initial reaction rate (v₀ (expressed as ΔA_{400 nm}/s))±standard deviation. The lines are Michaelis-Menten curves fitted to experimental data. Asterisk represents statistical difference between treatments (Fisher’s least significant difference analysis, p≤0.05)

Lower K_i than K_i´ values of the polyphenolic compounds (caffeic acid, p-coumaric acid and quercetin) were observed. These lower K_i´ values indicate that the inhibitor has more affinity for the enzyme-substrate complex than for the free enzyme in a mixed inhibition mechanism (42). The competitive component of mixed inhibition prevailed in the systems. All tested compounds except capsaicin had higher K_i´ values than their respective K_i values, indicating that they bind tighter to the free enzyme than to the substrate-bound form of pancreatic lipase. These results are in agreement with the competitive inhibition of pancreatic lipase activity by polyphenolic extract observed by You et al. (10). Our results may suggest that capsaicin might be binding to a different site in the enzyme, instead of the substrate binding site. In accordance with IC₅₀ values, K_i and K_i´ values (Table 1) of quercetin were the lowest of all studied compounds. These results agree with those reported by You et al. (10). Liu (44) and Sergent et al. (1) also observed that phenolic acids were not as potent pancreatic lipase inhibitors as quercetin. Quercetin inhibitory activity on other pancreatic enzymes such as α-amylase has also been studied (45).

Binding of polyphenolic compounds to pancreatic lipase by fluorescence spectroscopy analysis

The fluorescence intensity and quenching of Trp residues from pancreatic lipase upon the incubation with polyphenolic compounds was monitored, and data were analyzed using Eq. 6. There are seven Trp residues in the pancreatic lipase that may be responsible for the fluorescence emission of this protein (24). According to crystal structure (PDB code 1ETH (28)), five of them are located in the N-terminal domain, while the other two are in the C-terminal domain, and three Trp residues are located near the active site of the enzyme. Pancreatic lipase also possesses 25 Phe residues, and 16 Tyr residues that may account for fluorescence. Changes in the fluorescence intensity of pancreatic lipase are attributed to the interaction of ligands, which results in a change in the polarity of the environment surrounding Trp residues, but changes in the environment of Phe and Tyr residues may also have minor contribution to pancreatic lipase fluorescence.

To evaluate the pancreatic lipase-polyphenolic compound interactions, series of emission spectra of pancreatic lipase at fixed concentration (2.0 mg/mL) were recorded upon the addition of different polyphenolic compound concentrations (2.5-25 μM). Fig. 3a shows that the addition of different concentrations of caffeic acid to pancreatic lipase decreased the fluorescence intensity. The same behaviour was observed of the rest of the polyphenolic compounds. A slight bathochromic shift of 12 nm was recorded in the presence of caffeic acid. The rest of the polyphenolic compounds showed similar bathochromic shifts (5-8 nm). This decrease in fluorescence intensity accompanied by a bathochromic shift in the maximum fluorescence would correspond to an increase in the polarity environment surrounding the Trp residues due to a conformational transition of the protein, which involves the exposure of Trp residues (46, 47). The small bathochromic displacement observed reflected that the environment of the Trp during the polyphenolic compounds-pancreatic lipase interaction did not change considerably (3, 26).

Fig. 3

Effect of polyphenolic compounds on fluorescence spectra of pancreatic lipase: a) emission fluorescence spectra of pancreatic lipase (2.0 mg/mL) in the presence of different concentrations of caffeic acid (final concentrations: 0, 0.025, 0.05, 0.10, 0.125, 0.15, 0.175, 0.20, 0.225 and 0.25·10^-4 M), λ_ex=280 nm, and b) Stern-Volmer plot (Eq. 6) for the experimental data of pancreatic lipase fluorescence (λ_em=354 nm) at different caffeic acid concentrations as quencher [Q]. F₀ and F=fluorescence intensities at the maximum of λ_em in the absence and presence of the polyphenolic compounds

Table 2 summarizes the K_sv and k_q values of polyphenolic compounds. The linearity observed by plotting F₀/F and [Q] (Fig. 3b) allowed to determine the existence of a quenching mechanism (Eq. 6). A linear Stern-Volmer plot is indicative of a single class of fluorophore equally accessible to the quencher (46). The sequence of calculated K_sv values for the polyphenolic compounds was quercetin, caffeic acid, p-coumaric acid and capsaicin. Our results agree with those of Li et al. (16), who studied the interaction of quercetin, isoquercetin and rutin with pancreatic lipase and observed that the sequence of K_sv values did not correspond to the inhibition sequence. Also, our K_sv value ((4.5±0.7)·10^-4 M^-1) for quercetin was similar to the value obtained by Li et al. (16) of 4.95·10^-4 M^-1. Contrarily to the calculated K_sv values observed by Gonçalves et al. (48) of pancreatic lipase and procyanidin fractions from grape seed, Li et al. (16) did not find a relation between the polyphenolic structures with higher molecular mass or higher number of hydroxyl groups, and higher K_sv values.

The calculated k_q values were higher than 2.0·10^-10 M^-1·s^-1, the maximum value for a dynamic mechanism in aqueous medium (16, 26). This indicates that the quenching process is an intrinsic and static process, involving only Trp residues (5, 16, 26). Static quenching has been previously reported for the interaction of pancreatic lipase with flavonoids, e.g. quercetin (16). This type of quenching is characterized by the formation of a non-fluorescent complex between the enzyme and the ligand, in which the enzyme segment contains the Trp residue (46). Thus, the observed changes in the maximum emission intensity reflected the direct interactions of the analyzed compounds with the areas located near the Trp, and the induced conformational changes of the enzyme (49).

It seems that the flavonoid quercetin, which has more hydroxyl groups in its structure than the phenolic acids (caffeic and p-coumaric acids) (37), was more effective in quenching the intrinsic fluorescence of pancreatic lipase (higher K_sv and k_q values), suggesting more affinity (50). Gonçalves et al. (48) observed that K_sv and k_q increased as the molecular mass of three different procyanidin fractions increased, while the IC₅₀ decreased (higher inhibitory capacity) with respect to pancreatic lipase activity. This behaviour can be explained considering that even though the main interactions between polyphenolic compounds and pancreatic lipase are recognized as non-covalent (5, 7, 16, 24), the polar interactions (determined by molecular docking) between quercetin and pancreatic lipase would explain the lowest IC₅₀ calculated for quercetin. Gonçalves et al. (3) described pancreatic lipase-polyphenolic compound interactions as reversible, mainly hydrogen bonds and π-stacking interactions; and others only mentioned these interactions as weak (26). The other non-polyphenolic compound, capsaicin, had K_sv and k_q values significantly lower than phenolic acids and quercetin (data not shown), which indicates that the observed inhibition may follow a different mechanism than polyphenolic compounds (16).

Molecular docking

To predict the possible binding sites of polyphenolic compounds with pancreatic lipase, docking studies were carried out with all tested compounds (Fig. 4, Table 3). All polyphenolic compounds had one binding site. Binding sites for caffeic acid, p-coumaric acid and quercetin were located close to the enzyme active site, serine¹⁵³ (Ser¹⁵³), aspartic acid¹⁷⁶ (Asp¹⁷⁶) and histidine²⁶³ (His²⁶³), as described by Kokotos (51), while capsaicin had a binding site far away from the enzyme active site. The same possible binding site was observed for orlistat (data not shown), in agreement with other authors who reported that Ser¹⁵³ is involved in the binding of this molecule with the enzyme (51, 52). The interaction of these compounds with residues near the active site may explain the competitive component of the observed mixed-inhibition of the polyphenolic compounds. The inhibitory mechanism might be explained considering that the inhibitor interacts with the substrate-enzyme binding. In other words, the inhibitor would bind either to free enzyme or substrate-enzyme complex, but near the substrate binding site. The non-competitive component for polyphenolic compound mixed-type inhibition could be explained considering that the binding to the active site occurs in a form that only affects the catalytic steps and not the substrate binding.

Fig. 4

Crystal structure of pancreatic lipase (Protein Data Bank code 1ETH (28)) model of the binding sites for caffeic acid, p-coumaric acid, quercetin and capsaicin achieved by docking experiments: a) surface representation of the overall three-dimensional structure of pancreatic lipase (shown in dark grey colour at www.ftb.com.hr) indicating the active site and the potential binding sites for polyphenolic compounds (b-e). Specific interactions of: b) caffeic acid, c) p-coumaric acid, d) quercetin, and e) capsaicin. Proposed binding amino acids are pointed out (shown in green colour at www.ftb.com.hr). Hydrophobic bonds and polar interactions are drawn in dotted lines (shown as blue and green dotted lines, respectively, at www.ftb.com.hr)

The pancreatic lipase binding site for polyphenolic compounds was mainly formed by four pancreatic lipase amino acid residues: Phe⁷⁸, Tyr¹¹⁵, His¹⁵² and Phe²¹⁶ (Figs. 4b-e).Residues Tyr¹¹⁵, Ser¹⁵³ and Proline¹⁸¹ (Pro¹⁸¹) were also involved in the caffeic acid-pancreatic lipase interaction (Fig. 4b). p-Coumaric acid also interacts with glutamic acid¹⁸⁰ (Glu¹⁸⁰) (Fig. 4c and Table 3; Fig. 1 for references). In the quercetin-pancreatic lipase complex, quercetin was found near Asp⁸⁰, arginine²⁵⁷ (Arg²⁵⁷), valine²⁶⁰ (Val²⁶⁰) and alanine²⁶¹ (Ala²⁶¹) (Fig. 4d). In the capsaicin-pancreatic lipase complex, capsaicin interacted mainly with leucine⁴¹ (Leu⁴¹), lysine⁴² (Lys⁴²), Arg⁶⁵, Glu³⁷¹ and Tyr⁴⁰⁴ (Fig. 4e). This different binding region for capsaicin, with respect to the polyphenolic compounds, could be attributed to the different structure of capsaicin (53), shown in Fig. 1. In this way, the non-competitive component of capsaicin could be in this site. Even though the polyphenolic compound binding sites were in the vicinity of the catalytic site, the small fluorescence shift may be attributed to the fact that no significant polarity changes were predicted around Trp residues.

He et al. (7) proposed that the binding forces between polyphenolic compounds and pancreatic lipase were mainly hydrophobic interactions. These hydrophobic and non-covalent interactions are mentioned in Table 3 and shown in Figs. 4b-e (blue doted lines; colour version available at www.ftb.com.hr). Among them, π-stacking interactions are formed between aromatic rings from polyphenolic compounds and pancreatic lipase aromatic amino acids, Phe and Tyr. Our results are in agreement with those reported by Kim et al. (54), which indicated that hydrophobic bonds (with residues like Phe) were the main interactions found between polyphenolic compounds and pancreatic lipase. Another hydrophobic bond predicted by the docking analysis was the hydrogen bond between the hydroxyl groups of polyphenolic compounds and polar groups of pancreatic lipase (5). Fewer polar interactions in Fig. 4 (shown as green dotted lines; see www.ftb.com.hr) were observed for the polyphenolic compound-pancreatic lipase complexes. Quercetin-pancreatic lipase complex presented more polar interactions compared to the rest of the polyphenolic compound-pancreatic lipase complexes.

The quercetin size and structural rigidity could be responsible for its binding characteristics (more polar interactions) and consequently its higher pancreatic lipase inhibitory potency. You et al. (10) reported a higher inhibitory potency (lower IC₅₀) of quercetin than of ellagic acid (Fig. 1). Böhl et al. (55) described that quercetin (among other flavonoids) presented higher affinity towards proteins due to structural characteristics. Quercetin possesses characteristics such as catechol structure in B ring and double bond between C2 and C3 in C ring. Other factors such as number of aromatic rings and number of free hydroxyl groups present in their structure are associated with the ability of polyphenolic compounds to interact with proteins (37, 56).

∆F values calculated from fluorescence studies for polyphenolic compound-pancreatic lipase interactions could be associated with their binding with three amino acids close to Trp residues. According to docking results, caffeic acid, p-coumaric acid and quercetin bound with Phe⁷⁸ and His²⁶⁴, found near Trp residues located at positions 86 and 252. In contrast, the capsaicin-pancreatic lipase complex was located in the vicinity of only one Trp residue (Trp³⁰, which is close to Leu⁴¹ and Lys⁴²). This suggested that polyphenolic compounds have a higher effect on protein conformation (Trp environment) than capsaicin and may explain the lower ΔF, K_sv and k_q values obtained for the latter in the fluorescence studies.

Caffeic acid showed high π-stacking interactions with Phe⁷⁸, Tyr¹¹⁵, Pro¹⁸¹ and Phe²¹⁶, and two polar bonds with Ser¹⁵³ and His²⁶⁴. A total of ten possible bonds between this polyphenolic compound and pancreatic lipase were observed. It seems that caffeic and p-coumaric acids showed an ability to form a complex with pancreatic lipase (no significant differences between K_sv and k_q values) similar to that observed for quercetin, but through less polar interactions (2 against 4 or 5 observed for the other studied compounds). This higher number of hydrophobic interactions present in the phenolic acid-pancreatic lipase complex, instead of the more polar interactions present in quercetin-pancreatic lipase complex, could explain the lower inhibitory activity of the phenolic acids. Notwithstanding, further studies regarding the phenolic acid-pancreatic lipase binding are needed.