Sample preparation

Guava (Psidium guajava L.) and soursop (Annona muricata L.) fruits were acquired from a local market in Tepic, Nayarit. The fruits were used in maturity stage they were transported to the Laboratorio Integral de Investigación en Alimentos, and were processed immediately to produce the purees. The guava was used only seedless, while soursop fruit was used the pulp without peel and seeds. The samples were homogenized (Ultraturrax, T25, IKAworks, Wilmington, NC) and were freeze-dried (Labconco Model 77522020, Kansas, USA) ground, sieved and stored hermetically for later use in analysis.

In vitro digestion model and bioaccessibility percentage (%) in guava and soursop pulp

Freeze-dried guava and soursop pulps were submitted to an in vitro digestion model adapted from the methodology proposed by ^{Saura-Calixto, García-Alonso, Goñi, & Bravo (2000}) with some modifications (^{Blancas-Benítez, Pérez-Jiménez, Montalvo-González, González-Aguilar & Sáyago-Ayerdi, 2018}) (Figure 1). Firstly, the samples were subjected to an enzymatic hydrolysis process with pepsin (300 mg/mL, P-7000, Sigma Aldrich) (37 °C, 1 h) (Step 1, gastric fraction-GasF), pancreatin (5 mg/mL, P-1750, Sigma Aldrich) (37 °C, 6 h) and α-amylase (120 mg/mL, A-6255, Sigma Aldrich) (37 °C, 16 h) (Step 2, intestinal fraction-IntF). After the hydrolysis, samples were centrifuged (Step 3) to separate the soluble and insoluble indigestible fractions. The supernatants were dialyzed (D9652, 12-14 KDa, Sigma Aldrich, 48 h) to simulate passive absorption (Step 4). After dialysis, the PC associated with the soluble indigestible fraction (SIF) was determined (Step 5). The residue, after samples were centrifuged was used to determine the PC associated with insoluble indigestible fraction (IIF) after an organic extraction (^{Pérez-Jiménez et al., 2008}) (Step 6). Both the PC associated with the SIF and IIF correspond to the non-bioaccesible PC fraction. The total soluble polyphenols (TSP) content of the different fractions GasF, IntF, SIF, and IIF were determined, and the samples were analyzed by HPLC-DAD as described below. The in vitro bioaccessibility percentage (%BA) of PC was determined using Eq. 1:

Figure 1 Bioaccesibility of phenolic compounds in guava and soursop pulp samples: Step 1 gastric fraction, Gas-F (Pepsine), Step 2 intestinal fraction, Int-F (Pancreatine and amylase), Step 3 centrifugation to separate supernatants and residues, Step 4 dialysis 24-48 h, Step 5 non-bioaccesible PC, associated to soluble indigestible fraction (SIF), Step 6 non-bioaccesible PC, associated to insoluble indigestible fraction (IIF). 

Where PC-IntF= the PC released on the intestinal fraction, PC-SIF= the PC associated with the IIF, and PC-IIF= the PC associated with the IIF.

In vitro release kinetics of PC in guava and soursop pulp

The in vitro released kinetics of PC from guava and soursop pulp were determined according to an in vitro digestion method (^{Blancas-Benítez et al., 2015}b) Briefly, 300 mg of dried sample was weighed and combined with 10 mL of phosphate buffer (0.05 M, pH 1.5) and 0.2 mL of pepsin solution (300 mg/mL, P-7000, >250 units/mg, Sigma Aldrich). The solution was incubated at 37 °C for 1 h, afterwards, phosphate buffer (4.5 mL, 0.05 M, pH 6.9) was added, and the samples were transferred to cellulose dialysis bags (D9652, 12-14 KDa, Sigma Aldrich). One milliliter of pancreatic α-amylase (120 mg/mL, A-6255, 110 units/mg, A6255, Sigma) was added to each dialysis bag, the samples were adjusted to a volume of 30 mL, and the dialysis tubes were sealed. The tubes were placed in a glass vessel with 200 mL of phosphate buffer (0.05 M, pH 6.9) that had previously been stabilized at 37 °C. The samples were incubated for 3 h with continuous stirring. At 30 min intervals, 1 mL extract of the liquid containing the dialyzed compounds were taken and used for the analysis of the TSP using HPLC-DAD, as described in the following section. To calculate the kinetic parameters of PC release during the in vitro digestion, the final rate (Vf ) of PC release was determined according to Eq. 2:

Where ΔC is the difference in concentration between the final and the initial PC concentration, At is the time difference between a specific time and the initial time, and Vf is the final rate of PC release during the in vitro digestion (mg GAE/min).

Total soluble polyphenols (TSP) content in the guava and soursop pulp digestion fractions

TSP contents were quantified in all digested fractions of guava and soursop from the in vitro digestion assay and at each point in the release kinetics assay; this analysis was conducted according to the method described by ^{Montreau (1972}) with slight modifications, 250 μL aliquots of each fraction (in vitro digestion or kinetics assay) were mixed with 1000 μL of sodium carbonate (7.5% w/v) and 1250 μL of Folin-Ciocalteu's reagent. Afterward, the absorbance of each sample was measured at 750 nm using a 96-well microplate reader (Bio-Tek®, Synergy HT, Winooski, VT, USA) with Gen5 software. Gallic acid was used as the standard (0.0125-0.2 mg/mL, R² > 0.9997), and the results were expressed as mg of gallic acid equivalents (g GAE/100 g DW).

Identification of phenolic compounds (PC) by HPLC-DAD analysis on digested fractions of guava and soursop pulp

The identification of the PC was carried out using an HPLC Agilent 1260 series (Agilent Technologies, Santa Clara, CA, USA) equipped with a UV-Vis Diode Array Detector (DAD). Samples were injected (10 μL, flow rate 0.4 mL/min in a Poroshell 120 EC-C18 (4.6 mm χ 150 mm, particle size 2.7 μm) (Agilent Technologies). The gradient elution was carried out using water containing 0.1% trifluoracetic acid (Sigma Aldrich) as solvent A and acetonitrile (Sigma Aldrich) as solvent B applied as follows: 0 min, 5% B; 10 min, 23% B; 15 min, 50% B; 20 min, 50% B; 23 min, 100% B; 25 min, 100% B; 27 min, 5% B, 30 min, 5% B. DAD detection was performed at 280-320 nm. The data analysis was performed using OpenLab CDS, ChemStation Edition software (Agilent Technologies, Santa Clara, CA, USA). Characterization of the PC was based on retention time.

Antioxidant capacity (AOX)

All analyses of AOX were slightly modified to adjust in a microplate reader. Ferric reducing antioxidant power (FRAP) assay was performed as was described by ^{Benzie & Strain (1996}), modified by ^{Álvarez-Parrilla, de la Rosa, Amarowicz & Shahidi (2010}). FRAP solution 10:1:1 (v/v/v) dissolved in a sodium acetate buffer (0.3 M; pH 3.6), TPTZ-HCl (10 mM, 40 mM), and ferric chloride hexahydrated (20 mM) was warmed to 37 °C before mixing it with the samples. Briefly, 24 μL of sample from the aqueous-organic extraction was mixed with 180 μL of FRAP solution and the absorbance was measured at 595 nm after 30 min using microplate reader (Biotek, Synergy HT). Results were expressed as millimol of trolox (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic) equivalents per gram DW (mmoL TE/g DW). Radical-scavenging activity was determined by 2,2'-azinobis- 3-ethylbenzotiazoline-6-sulphonic acid (ABTS) radical assay as was described by ^{Re et al., (1999)}; with some modifications. For this determination, ABTS (7 mM) was dissolved in potassium persulphate (2.42 mM) and kept in the dark at room temperature for 14 h. The solution was adjusted with phosphate buffer at an absorbance of 0.70 (±0.02). Trolox was used as a standard and methanol as a blank. Samples of 10 μL of extract were added in a microplate reader (Biotek, Synergy HT, Winooski, VT, USA) of 300 μL of capacity and 280 μL of ABTS radical was added. Then, the mixture was incubated at 37°C in the dark and the absorbance was measured after 6 min, at 734 nm. A calibration curve was prepared using an aqueous solution of trolox as standard. The results are reported in mmoL TE/g DW.

Statistical analysis

All analyses were performed in triplicate; means values and standard deviations from each determination were calculated. Statistical significance between guava and soursop pulp was analyzed by t-student. Data were analyzed using the software Statistic 8.0 Release for Windows (Stat Soft. Inc., Tulsa, OK, USA) with a significance level of α=0.05.

In vitro digestion and bioaccessibility percentage (%) in guava and soursop pulp

Table I shows that the TSP of guava and soursop pulp in gastric and intestinal stages of an in vitro digestion. The PC in Gas-F from soursop pulp were higher than guava pulp. Nevertheless in the IntF of the digestion an increase in the content of TSP respect the GasF was observed from both pulps. This could be due to the partial release of PP bound to the cell wall material of the plant food (^{Chandrasekara & Shahidi, 2012}). It has been reported that some PC present in plant foods, can be associated mainly with carbohydrates which can decrease their bioaccesibility (^{Saura-Calixto, 2010}). During intestinal digestion, a series of enzymes, mainly some hydrolases responsible for the hydrolysis of carbohydrates, could react with some links or interactions that exist between carbohydrates and PC (^{González-Aguilar, Blancas-Benítez, & Sáyago-Ayerdi, 2017}; ^{Blancas-Benítez et al., 2015b}), which would result in an increase in the concentration of PC in the samples, after intestinal hydrolysis, a process that would occur when consuming any plant food, and that can be observed in the guava and soursop pulps analyzed in this study.

The PC profile of guava and soursop pulp found on in vitro digestion showed that gallic acid and chlorogenic acid were mainly detected in both samples, although caffeic acid was detected only in soursop pulp. Also, gallic acid and chlorogenic acid were the PC that was mostly found in SIF and IIF, respectively from guava and soursop pulp. Although these two main compounds are potentially absorbed in the intestine, a substantial percentage of them were also found on the indigestible fractions, which can reach the colon. In the colon this phenolic acids, because of its simple structure, may still be absorbed on the large intestine, it has been previously documented on various in vitro and in vivo models (^{Rui et al., 2014}; ^{Vetrani et al., 2016}). All these compounds have been previously identified in both guava and soursop fruit, and they have been attributed beneficial effects for health (^{Coria-Téllez, Montalvo-González, Yahia & Obledo Vázquez, 2018}; ^{Jiménez-Escrig, Rincón, Pulido & Saura-Calixto, 2001}; ^{Onyechi, Ibeanu, Nkiruka, Eme & Madubike, 2012}; ^{Rojas-Garbanzo, Zimmermann, Schulze-Kaysers & Schieber, 2017}; ^{Soares, Pereira, Marques & Monteiro., 2007}).

For the bioaccessibility determination was considerate the overall distribution of the TSP content, the data showed that 67.69% of the TSP from the guava pulp are bioaccessible, while 71.30% of TSP from soursop pulp. It is apparent that soursop pulp has a higher bioaccessibility of PC than the guava pulp. This suggests that soursop pulp PC are not bound to the food matrix, making them more available for intestinal absorption than PC present in guava pulp (^{Bohn, 2014}; ^{Cuervo et al., 2014}).

In vitro release kinetics of PC in guava and soursop pulp

Figure 2 shows the release kinetics of PC from guava and soursop pulp. The PC released did not showed statistical difference with similar rate in the soursop pulp (0.10 mg GAe/ min) and the guava pulp (0.09 mg GAe/min), during the first 30 min of digestion. The release rate was similar after 150 min of digestion, without subsequent changes in either pulps, These results may be related to those from Table I, where a higher bioaccessibility was found for soursop pulp, it is apparent that the soursop pulp PC might have a low interaction with the food matrix and they could be highly bioaccessible than the PC presents in guava pulp.

Figure 2 In vitro release kinetics of total soluble polyphenols of guava and soursop pulps 

Figure 3 shows the different PC that was identified through time of the in vitro digestion on each fruit pulp. Gallic acid was the main compound found in the guava and soursop pulp. The PC from both pulps remained constant throughout the 180 min of digestion. As previously mentioned, gallic acid was the main compound detected, and its percentage of detection increased during the in vitro digestion. The same behavior was observed for chlorogenic, coumaric and hydroxycinnamic acids in guava pulp; and for chlorogenic and caffeic acids in soursop pulp. These results not only revealed the potential bioaccessibility of PC, but the ratio with which these compounds could be released and absorbed in the small intestine (^{Blancas-Benítez et al., 2015}b).

Figure 3 Phenolic compounds profile (HPLC-DAD) of guava (a) and soursop (b) pulp digestion fractions. a) Gastric fraction, b) Intestinal fraction c) After dialysis process Peak annotations; 1: Gallic acid, 2: Chlorogenic acid, 3: Coumaric acid, 4: Hydroxycinnamic acid, 5: Caffeic acid.