Soursop (Annona muricata L.) fruits at physiological maturity were harvested according to the ripening index reported by Balois-Morales et al. [19], corresponding to 160 days after anthesis. The harvest took place in the ejido Venustiano Carranza, Tepic, Nayarit, Mexico (21°32′2.77″ N, 104°58′37.73″ W; 893 m a.s.l.).

Polysaccharides were extracted from soursop fruits using conventional and ultrasound-assisted wet-milling methods. Fruit grinding and homogenization were performed in an industrial blender (International LI-5a, Mexico) for 3 min using a 1% citric acid solution in a 1:4 (w/v) ratio.

Conventional extraction

The extraction of starch and pectin followed a previously reported methodology [20]. The ground pulp was washed with distilled water and sequentially sieved through 100 mesh (150 μm), 200 mesh (75 μm), and 270 mesh (53 μm) sieves. The remaining bagasse was subjected to a second milling with distilled water in a 1:2 (w/v) ratio for 3 min, followed by the same filtration sequence.

Ultrasound-assisted extraction

The ultrasound-assisted extraction was carried out according to a previously reported method [9,20]. Sonication was performed using an ultrasonic processor (Model CPX750, Cole-Parmer Instruments, Vernon Hills, IL, USA) with a maximum nominal power of 750 W, operated at 40% amplitude (300 W) and a frequency of 20 kHz for 10 min at 25°C. Temperature was monitored using a mercury thermometer. The sonicated bagasse was filtered through 100, 200, and 270 mesh sieves. Subsequently, the residue was milled again with distilled water (1:2 w/v) for 3 min and refiltered.

Separation and drying

For both extraction methods, the liquid phase was allowed to stand for 24 h. The precipitate was recovered by decantation and centrifuged at 1537× g for 5 min at 25°C (Hermle Z326K, Wehingen, Germany) to remove residual liquid. As a result, two pastes were obtained from each extraction method, differing in color (white and brown). These pastes were dried in a recirculation oven (LSIS-B2V/VC 55, Germany) for 24 h at 35°C, ground, and sieved (100 mesh). Samples were then weighed for yield quantification and stored for further analysis.

The dried pastes containing starch and pectin were characterized by Fourier Transform Infrared Spectroscopy (FT-IR) to identify their chemical groups and confirm the presence of polysaccharides.

Methanolic extracts were prepared from the dried starch- and pectin-containing pastes, and acetogenins were identified and quantified by High-Performance Liquid Chromatography (HPLC).

The dried pastes containing starch and pectin were characterized by Fourier transform infrared spectroscopy (FT-IR) to identify functional chemical groups and confirm the presence of polysaccharides. FT-IR spectra were obtained using an ATR-FT-IR spectrophometer (Cary 630 FT-IR, Agilent Technologies, Santa Clara, CA, USA) equipped with a zinc selenide (ZnSe) Crystal. Spectra were recorded in attenuated total reflection (ATR) mode by accumulating 64 scans at resolution of 4 cm⁻¹ over the wavenumber range of 600–4000 cm⁻¹. Commercial corn starch (Sigma-Aldrich, S-4126) and citrus pectin (Genu Pectin, Rapid set, CP Kelco, G51669; DE = 71%) were used as reference standards to identify characteristic functional groups. Data acquisition and analysis were performed using SigmaPlot 14.0 software.

The degree of esterification (DE) was calculated to distinguish low- and high-methoxyl pectins using a previously reported equation [21]: (1) D E ( % ) = A b s 1745 A b s 1745 + A b s 1630 × 100 where DE = degree of esterification; Abs = absorbance.

For nomenclature, the following codes were used: AEC, AEU, PEC, and PEU, where A = starch, P = pectin, C = conventional extraction, and U = ultrasound-assisted extraction (e.g., AEU = starch extracted by ultrasound).

For chromatographic analysis, starch and pectin extracts were prepared to identify and quantify acetogenins. Twenty grams of starch or pectin paste were dispersed in 100 mL of methanol. The suspension was sonicated in an ultrasonic bath (KYMEN® Digital Ultrasonic Cleaner, Model JP-4820, Shenzhen, China) at 35 kHz for eight cycles of 8 min each, with agitation between cycles. The mixture was centrifuged at 1537× g for 5 min at 4°C to recover the liquid phase, followed by a second centrifugation at 6147× g for 10 min at 4°C. The supernatant was filtered through filter paper and concentrated under reduced pressure using a rotary evaporator (IKA®, C-MAG HS7, Wilmington, NC, USA) at 40°C, 80 rpm, and 65 cm Hg [3].

Identification and quantification of acetogenins were performed following the methodology of López-Romero et al [27]. Purified acetogenic fractions were resuspended in HPLC-grade methanol and filtered through 0.22 μm nylon membrane filters (Thermo Scientific, Shanghai, China).

Samples were injected into an HPLC system (Agilent Technologies 1260 Infinity, Waldbronn, Germany®) equipped with a diode-array detector (DAD) and a reverse-phase C18 column (5 μm particle size, 4.6 mm × 250 mm, Agilent Zorbax SB-C18) at 20°C. The mobile phase consisted of methanol (Eluent A) and water (Eluent B) under a linear gradient: 0–40 min, 85:15 (v/v, methanol:water), 40–60 min, 95:5 (v/v, methanol:water) at a flow rate of 1 mL min⁻¹.

Acetogenins were detected at 220 nm. Quantification was based on calibration curves prepared with two standards (anonacin and pseudoannonacin). Results were expressed as micrograms of anonacin/pseudoannonacin equivalents per gram of dry sample (μg anonacin/pseudoannonacin g⁻¹ d.b.).

A completely randomized design a 2 × 2 factorial arrangement (extraction method x polysaccharide phase) was used. Four treatments were evaluated: starch and pectin obtained by conventional and ultrasound-assisted extraction. All analyses were performed in triplicate (n = 3 independent extractions per treatment), and the experimental unit corresponded to each independent extraction. Data were analyzed using a two-way analysis of variance (ANOVA), and mean comparisons were performed using Tukey’s multiple comparison test (p ≤ 0.05). Statistical analyses were conducted using SAS® software (version 9.2).

FT-IR analysis of the starch samples revealed characteristic absorption bands corresponding to the stretching vibrations of OH groups (3275.2 cm⁻¹) and CH groups (2908.5 cm⁻¹), as well as CH bending vibrations (1329.1 cm⁻¹). The fingerprint region showed bands at 1148.5, 1072.2, 987.7, 930, and 855.4 cm⁻¹, corresponding to polysaccharide structures consistent with the standard starch spectrum (Fig. 1A). For pectin, absorption bands were observed corresponding to OH stretching (3267.1 cm⁻¹), symmetric and asymmetric stretching of CH, CH₂, and CH₃ groups (2916.6 cm⁻¹), and esterified and free C=O stretching vibrations (1732.2 and 1636 cm⁻¹, respectively). The fingerprint region exhibited bands at 1147.1, 1074.0, 1036.3, 968.4, and 855.4 cm⁻¹, consistent with the standard pectin spectrum (Fig. 1B). Additionally, the identified pectin presented a degree of esterification (DE) of 39.9%, classifying it as low-methoxyl pectin.

TLC results of starch and pectin extracts (PEU, PEC, AEU, and AEC) showed pink bands upon reaction with Kedde reagent, indicating the presence of acetogenins. Pectin extracts exhibited seven bands, with Rf values ranging from 0.11 to 0.80 (PEU) and from 0.11 to 0.82 (PEC). In addition to acetogenins, other pigments were observed that may have influenced the reduction of Rf values (Fig. 2A,B). Starch extracts presented Rf values from 0.17 to 0.89 (AEC) and 0.17 to 0.91 (AEU), with no visible presence of other pigments (Fig. 2C,D). Among all samples, PEU and PEC extracts exhibited lower Rf values compared to AEU and AEC.

From the pectin extracts, a total of 70 fractions were obtained and grouped according to Rf values into 14 (PEU) and 9 (PEC) fraction groups: PEU: G1 (F1–F2), G2 (F3), G3 (F4), G4 (F5–F8), G5 (F9–F14), G6 (F15–F28), G7 (F29–F70) (Fig. 3A), PEC: G1 (F1–F2), G2 (F3–F4), G3 (F5), G4 (F6–F7), G5 (F8–F10), G6 (F11–F31), G7 (F32–F70) (Fig. 3B), For starch extracts, 20 fractions were obtained from AEU and 31 from AEC: AEU: G1 (F1), G2 (F2), G3 (F3), G4 (F4–F7), G5 (F8), G6 (F9–F16), G7 (F17–F20) (Fig. 3C), AEC: G1 (F1–F2), G2 (F3), G3 (F4), G4 (F5–F6), G5 (F7), G6 (F8–F26), G7 (F27–F31) (Fig. 3D). Fractions G3 (PEU), G4 (PEC), G4 (AEU), and G3 (AEC) showed the highest color intensity in the Kedde reaction, confirming a greater concentration of acetogenins (Fig. 3). The presence of acetogenins was detected in both starch and pectin extracts.

Two acetogenins—anonacin and pseudoannonacin—were identified in the G3 (PEU), G4 (PEC), G4 (AEU), and G3 (AEC) fractions (Fig. 4). The concentration of anonacin in pectin extracts was 397.23 μg g⁻¹ d.b. (PEU) and 302.29 μg g⁻¹ d.b. (PEC), while in starch extracts it was 3.12 μg g⁻¹ d.b. (AEU) and 1.63 μg g⁻¹ d.b. (AEC) (Table 1). The concentration of pseudoannonacin in pectin extracts was 452.81 μg g⁻¹ d.b. (PEU) and 340.07 μg g⁻¹ d.b. (PEC), whereas in starch extracts it was 21.61 μg g⁻¹ d.b. (AEU) and 20.73 μg g⁻¹ d.b. (AEC) (Table 2).

The absorption bands in the 3.000–3.600 cm⁻¹ region are attributed to the stretching of OH groups, which are related to intra- and intermolecular hydrogen bonds, whereas the stretching vibrations observed between 2.891–2.925 cm⁻¹ correspond to CH groups [28]. Pourfarzad et al. [29] reported that the wavenumber at 1.640 cm⁻¹ is associated with moisture content and arises from the stretching vibrations of water molecules absorbed by amylose. Starches exhibit characteristic peaks (fingerprint region) in the 800–1.300 cm⁻¹ range, which are sensitive to conformational and crystalline order. These peaks are assigned to stretching vibrations of CO, CC, and COH bonds, as well as COH bending [30,31]. In soursop starches, in addition to the typical functional groups of starch, additional bands between 1.200 and 1.700 cm⁻¹ were observed, which, according to Warren et al. [31], correspond to minor components such as protein and lipid residues. Regarding soursop pectin, peaks in the 3.140–3.400 cm⁻¹ region are associated with OH stretching, while those near 2.900 cm⁻¹ correspond to CH and CH₂ stretching [28]. Santos et al. [32], studying pectins extracted from industrial residues, reported that the key peaks in pectin structure are those representing the degree of esterification, appearing in the 1.600–1.750 cm⁻¹ region, indicative of free and esterified carboxyl groups. Wavenumbers below 1.500 cm⁻¹ correspond to the fingerprint region, which includes complex vibrational interactions of functional groups present in pectin (800–1.200 cm⁻¹) [28]. The pectin identified in this study was classified as low-methoxyl, with a degree of esterification (DE) below 50%, consistent with the classification proposed by Rascón-Chu et al. [33].

The different retention factors (Rf) observed in the chromatographic bands of pectin and starch extracts are likely due to the higher content of phytochemicals in pectins compared to starches [8]. These authors reported high levels of phenols, phytosterols, and alkaloids in soursop fruits, which could interfere with acetogenin migration on the TLC plate, reducing their Rf values. Furthermore, since soursop peel contains a high amount of pectin, pigment residues may remain in this polysaccharide. Meléndez-Martínez et al. [34], reported that pigments present in fruit peels are highly soluble in organic solvents such as chloroform (a low-polarity compound), resulting in higher Rf values that can interfere with acetogenin band migration. Variations in Rf values observed among TLC bands can also be attributed to differences in the functional groups of the acetogenin molecules present in the extracts, such as hydroxyl (–OH), keto (C=O), and ether (–O–) groups [25]. Fernández [35] stated that such structural characteristics confer intermediate polarity to acetogenins, which are soluble in chloroform, dichloromethane, and ethyl acetate, and highly soluble in methanol and ethanol. Additionally, differences in band color intensity—where PEU and PEC extracts exhibited lower intensity than starch extracts (Fig. 2) may be associated with the higher content of total lipids (4.03% and 2.4%), phytosterols (3.55 and 2.09 mg EC g⁻¹ pectin d.b.), and acetogenins (1.5 and 1.8 mg EAn g⁻¹ pectin d.b.) in pectin phases [8]. The high content of these compounds may interfere with the colorimetric Kedde reaction, since the alkaline medium (KOH) promotes lipid saponification [36].

Open-column chromatography allows the separation of different acetogenic fractions. These molecules are polyketide-derived compounds that are highly soluble in organic solvents such as methanol, acetone, ethanol, ethyl acetate, chloroform, and dichloromethane, but insoluble in hexane, acetonitrile, petroleum ether, and water [30].

Thus, mobile phases composed of mixed solvents can be used to isolate acetogenins according to their polarity. The different Rf values obtained in this study suggest the presence of acetogenins with distinct structural features, depending on the type and number of functional groups—such as hydroxyl (OH), lactone, unsaturation, and furan rings [25]. Acetogenins containing hydroxyl groups are more polar and can form hydrogen bonds with the SiO₂ of the chromatographic plate, resulting in lower Rf values. In contrast, those containing carbonyl (C=O) groups are less polar and yield higher Rf values [26,37]. The presence of alcohol, epoxide, and alkene groups modifies molecular polarity and, consequently, the retention factor, while solvent composition during purification can also affect migration [38,39]. Differences among acetogenin Rf groups may further relate to the fruit tissues of origin (peel, seed, or pulp) [26]. Based on these findings, it is likely that PEU and PEC correspond to peel- and seed-derived structures, whereas AEC and AEU originate mainly from pulp tissues [8,40]. In this study, the acetogenic fraction groups that exhibited greater pink color intensity in the Kedde reaction were confirmed by HPLC-DAD analysis. According to Aguilar-Hernández et al. (2022), a stronger pink coloration corresponds to higher acetogenin content, as the lactone rings of acetogenins react with 3,5-dinitrobenzoic acid in alkaline medium (KOH), forming a stable pink complex. The higher acetogenin content in pectins likely reflects their origin from seed and peel tissues [8], consistent with results reported by Aguilar-Hernández et al. [26] and López-Romero et al. [27] for soursop seeds and pulp. The ultrasound-assisted extraction method increased acetogenin yield, likely due to ultrasonic cavitation, which causes cell wall disruption through microbubble formation, growth, and collapse, releasing energy into the medium. This process enhances solvent penetration and contact surface area, improving extraction efficiency [41]. The lower acetogenin concentrations observed in starches (AEU and AEC) may result from washing during purification and from their origin in the pulp, which contains fewer acetogenins [15]. In contrast, acetogenins are synthesized mainly in the peel and seed tissues [26,42]. Recent studies have highlighted the potential of ultrasound-assisted extraction and advanced chromatographic techniques for improving the recovery and characterization of acetogenins and other bioactive compounds from plant-derived polysaccharides [26,27,41].

Annonacin and pseudoannonacin were the predominant acetogenins identified in both starch and pectin fractions. These structurally similar molecules differ at the C15 carbon atom [26]. Annonacin is a mono-THF α,α′-dihydroxylated acetogenin, whereas pseudoannonacin is a ketonic tetra- or trihydroxylated acetogenin [43]. A key feature of these acetogenins is their long nonpolar chain, which may allow them to form amylose–lipid complexes in starch, since amylose adopts a helical structure capable of incorporating free fatty acids [44]. Such complexes may form during extraction, particularly under ultrasound-assisted conditions, where the localized temperature increase caused by cavitation may promote complex formation and enhance acetogenin recovery. The acetogenin concentrations (μg mg⁻¹) obtained in this study were lower than those reported by López-Romero et al. [27], who quantified pseudoannonacin (350 mg g⁻¹) and annonacin (15 mg g⁻¹) from soursop seeds. This difference likely arises because, in polysaccharides such as pectins and starches, acetogenins are bound as polar lipids (via hydrogen bonding) that require polar solvents (e.g., methanol–water) for extraction, as applied in this study. In contrast, higher concentrations occur in specialized idioblast cells, where acetogenin biosynthesis takes place, primarily in the seeds and peel of soursop fruits [45].