First Comprehensive Characterization and Ultrasound-Assisted Extraction Optimization of an Unexplored Amazonian Berry (Chondrodendron tomentosum): A Novel Source of Anthocyanins and Phenolic Compounds

1. Introduction

Chondrodendron tomentosum Ruiz & Pav. (Menispermaceae), locally known in northeastern Perú as "uva de montaña" (mountain grape), is an Amazonian woody liana that produces dark purple fleshy drupes arranged in infructescences of 3–6 units [1,2]. The fruits, traditionally consumed in Andean-Amazonian communities, have a sweet-tart flavor and ripen between July and September in montane forests at 1500–2000 m a.s.l. [3]. Despite its traditional use, the phytochemical composition of C. tomentosum fruits remains uncharacterized in scientific literature, classifying it as an understudied and underexploited wild species.

Wild Amazonian and Andean berries constitute rich sources of phenolic compounds, anthocyanins, organic acids, and vitamins, with recognized antioxidant, anti-inflammatory, and anticancer properties [4,5,6,7]. These secondary metabolites confer nutraceutical value and position wild species as candidates for functional food development [8]. Additionally, their resistance to adverse climatic conditions and natural immunity to pathogens represent significant agronomic advantages for sustainable production systems [9].

C. tomentosum remains absent from indexed scientific literature: as of November 2025, no studies on fruit composition, bioactive compounds, or extraction methodologies have been reported in databases such as Scopus or Web of Science. This absence in scientific literature simultaneously represents a conservation risk and an unexplored opportunity for functional food development from underutilized tropical biodiversity.

Ultrasound-assisted extraction (UAE) is gaining in importance as a green technology for extracting bioactive compounds from plant matrices by enhancing extraction time, yield, and selectivity over traditional techniques [10,11]. The acoustic cavitation of the ultrasonic probe system, breaking down cell walls, increases mass transfer and reduces thermal degradation of labile metabolites [12,13]. Various berries have shown promising efficiency in collecting anthocyanins and total phenols: Prunus spinose [14], Aronia melanocarpa [15], Myrciaria cauliflora [16], Myrtus communis [17], Morus nigra [18].

The goal of this analysis was to analyze the biometric and physicochemical properties of C. tomentosum at 3 maturity stages (green, turning, and ripe) as well as to assess them along with the proximate characterization of ripe fruits. Ultrasound-assisted extraction (UAE) of phenolic compounds, anthocyanins, and antioxidant capacity from ripe fruits was optimized through Box-Behnken design, evaluating the effect of methanol concentration (25–75%), ultrasonic amplitude (30–70%), and extraction time (1–15 min). Once optimized, the optimal UAE conditions were validated and applied on green and turning stages to evaluate the variation of bioactive compounds during ripening. This study constitutes the first comprehensive scientific report addressing the biometric, physicochemical, proximate, and phytochemical characterization of this species, together with the optimization of green extraction technologies for bioactive compound recovery.

2. Materials and Methods

2.1. Plant Material

Wild fruits of C. tomentosum Ruiz & Pav. were collected at green (0), turning (1), and ripe (2) stages (Figure 1) in a montane forest in the village of Villa Rica, San José de Lourdes district, San Ignacio province, Cajamarca, Perú (−5.059457°, −78.892105°; 1626 m a.s.l.) during July–August 2025. Only healthy fruits without mechanical damage, bruise, or signs of insect or pathogen attack were selected. Taxonomic identification was verified, and the reference specimen (Botanical Determination Certificate No. 05-2024) was deposited in the Isidoro Sánchez Vega Herbarium (ISV), Universidad Nacional de Jaén, Perú. Samples were transported under refrigeration (4 ± 1 °C) to the Nutrition and Toxicology Laboratory of the Instituto de Investigación Tecnológico Agroindustrial (IITA), Universidad Nacional del Santa, Chimbote, Perú, for immediate conditioning and processing.

2.2. Chemical Reagents

Folin-Ciocalteu reagent, 2,2-diphenyl-1-picrylhydrazyl (DPPH), (±)-6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Trolox), HPLC-grade methanol, potassium chloride, sodium acetate trihydrate, hydrochloric acid (37%), and glacial acetic acid were supplied by Sigma-Aldrich (St. Louis, MO, USA). Anhydrous sodium carbonate was purchased from J.T. Baker (Phillipsburg, NJ, USA), and gallic acid from Merck KGaA (Darmstadt, Germany). Sodium hydroxide and phenolphthalein indicator were of analytical grade.

2.3. Classification of Maturity Stages and Sample Preparation

Fruits were washed with abundant running water, disinfected with sodium hypochlorite solution (80 mg/L) for 2 min, and rinsed with distilled water. Classification by maturity stage was established through a visual colorimetric scale adapted from the maturity index (MI) of olives Olea europaea L. [19], defining three categories according to epicarp pigmentation: green (index 0) - uniform green epidermis without pigmentation; turning (index 1) - epidermis with 25–75% purple-reddish pigmentation; ripe (index 2) - uniform dark purple epidermis with >95% pigmented surface (Figure 1).

2.4. Morphological Characterization

Visual observations were performed to describe the C. tomentosum plant, as well as the shape, color, and texture of the fruit and seed at three ripening stages [20,21].

2.5. Biometric Characterization

Fifty fruits per maturity stage were evaluated (n = 50). The recorded variables were: fresh weight (g), polar diameter (cm), and equatorial diameter (cm). Fresh weight was determined on a precision balance (Precisa, LX-4200C, Switzerland). Diameters were measured with a stainless steel vernier caliper (Pretul, escolar, Mexico) [22]. Polar diameter corresponded to the longitudinal peduncle-style axis; equatorial diameter was measured perpendicularly at the widest zone of the fruit.

2.6. Physicochemical Characteristics

Fruits from each maturity stage were homogenized in a mortar to obtain purée. Vitamin C content was determined by titration using the AOAC 967.21 (2019) [23] method. pH (AOAC 981.12), total soluble solids (AOAC 932.12), and titratable acidity (AOAC 942.15) were developed following standard AOAC methods (2023) [24]. pH was measured with a benchtop pH meter (Hanna, HI2211, Romania) previously calibrated with pH 4.0 and 7.0 buffer solutions. A manual refractometer (Wmeters, REF108, China) was used to estimate total soluble solids (TSS), and values were expressed in °Brix at 20 °C. Titratable acidity (TA) was determined by titrating the sample with an acid-base solution with 0.1 N NaOH to pH 8.8, and results were expressed as percentage of citric acid. The maturity index (MI) was calculated as the TSS/TA ratio [25].

2.7. Proximate Composition

Proximate characterization was performed only on ripe fruits (n = 3). Moisture was quantified by gravimetry after drying samples in an oven (POL-EKO/SLW 240, Poland) following the AOAC 945.15 (2023) [24] method. Ash was determined by incineration in a muffle furnace (Protherm Furnaces, PLF 110/6, France) at 550 °C using AOAC 223.03 (2023) [24]. Crude protein was estimated using a Kjeldahl apparatus (Labconco, 6030001, USA) following AOAC 954.01 (2019) [26] with 6.25 as the nitrogen conversion factor. Crude fiber was determined by acid-base hydrolysis following NTP 205.003 (2011) [27], and crude fat by Soxhlet extraction (Marconi, MA-491, Brazil) with petroleum ether as solvent using AOAC 930.09 (2019) [28]. Carbohydrate content was estimated by difference according to the AOAC 986.25 (2023) [24] method using Equation 1.

$$Carbohidratos (\%)=100-(humedad+proteína+grasa+fibra+cenizas)$$

(1)

2.8. Ultrasound-Assisted Extraction (UAE)

The UAE protocol was adapted from Christou et al. [29] and Razola et al. [30]. Seeds were manually removed, and the pericarp (pulp and peel) was homogenized in a mortar. 1.00 g of sample was weighed into a 50 mL beaker, adding methanolic solution at a 1:30 (w/v) ratio. Extraction was performed using a USCG 1800 ultrasonic homogenizer (Infitek, China) equipped with a titanium sonotrode (Ø 18 mm, frequency 24 kHz, nominal power 2500 W). The probe was immersed to approximately 50% of the volume without contact with the vessel walls, operating in continuous mode. Methanol (%, v/v), amplitude (%), and time (min) conditions varied according to the Box-Behnken design. Subsequently, extracts were centrifuged (Sigma 4-16KS, Osterode, Germany) at 4200 rpm, 4 °C for 20 min. Supernatants were filtered (Whatman No. 42 paper), protected from light in aluminum foil-wrapped Falcon tubes, and stored at −20 °C until respective analyses.

2.9. Quantification of Total Phenols

Total phenols were determined by the Folin-Ciocalteu spectrophotometric method [31] with modifications. A calibration curve was prepared with gallic acid (0–300 μg/mL in 50% methanol). In amber microtubes, 80 µL of extract was mixed with 160 µL of Folin-Ciocalteu reagent (10% v/v), incubating for 5 min in darkness. 640 µL of Na₂CO₃ (7.5% w/v) was added, and the mixture was incubated for 60 min in darkness at room temperature. Absorbance was measured at 765 nm in a microplate reader (Model Synergy H1, USA). Results were expressed as mg gallic acid equivalents (GAE)/100 g fresh weight basis.

2.10. Quantification of Monomeric Anthocyanins

Total anthocyanins were quantified by the pH differential method [32]. Buffer solutions were prepared: potassium chloride (pH 1.0, 0.025 M) adjusted with concentrated HCl and sodium acetate (pH 4.5, 0.4 M) adjusted with glacial acetic acid. 500 µL of extract was mixed with 500 µL of each buffer independently, homogenizing in a vortex (Model MS 3 digital, Germany). Absorbances were measured at 510 nm and 700 nm in the microplate reader. Monomeric anthocyanin concentration was calculated using Equation 2 and Equation 3:

$$\mathrm{A} = {(A510- A700nm)}_{\mathrm{pH} 1.0} – {(A510 - A700nm) }_{\mathrm{pH} 4.5}\dots$$

(2)

$$\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{A}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{o}\mathrm{c}\mathrm{y}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{s}\left({\displaystyle \frac{\mathrm{mg}}{\mathrm{L}}}\right)={\displaystyle \frac{\mathrm{A}\times \mathrm{PM}\times \mathrm{FD}\times 1 000}{(Ԑ\times 1)}}\dots$$

(3)

Where A is the sample absorbance, MW = 449.2 g/mol (cyanidin-3-glucoside), DF = dilution factor, ε = 26900 L·mol⁻¹·cm⁻¹, and l = path length in cm. Results were expressed as mg cyanidin-3-glucoside equivalents (C3G)/L.

2.11. Antioxidant Capacity by DPPH

Antioxidant activity was evaluated by the DPPH method [33]. A Trolox stock solution (1 mM) was prepared by dissolving 0.0125 g in 50 mL of methanol. A calibration curve was constructed with Trolox (5–500 µM). A methanolic DPPH solution (1 mM) was prepared, adjusting absorbance to 0.800 ± 0.020 at 515 nm with methanol. In a 96-well microplate, 10 µL of diluted extract (1:5) was mixed with 190 µL of DPPH solution (final volume 200 µL). The mixture was incubated for 30 min in darkness at room temperature, and absorbance was measured at 515 nm in a microplate reader (Model Synergy H1, USA). Results were expressed as µmol Trolox equivalents (TE)/100 g fresh weight basis.

2.12. Experimental Design and Statistical Analysis

RSM using a three-factor Box-Behnken design (15 experimental runs) was applied to optimize UAE. Independent variables were: methanol concentration (A: 25, 50, 75%, v/v), ultrasonic amplitude (B: 30, 50, 70%), and extraction time (C: 1, 8, 15 min). Response variables were: anthocyanins (mg C3G/L), total phenols (mg GAE/100 g), and antioxidant capacity (µmol TE/100 g). Data were fitted by least squares to a quadratic polynomial model in coded variables (Equation 4):

$$y={\beta }_0+{\beta }_{A}A+{\beta }_{B}B+{\beta }_{C}C+{\beta }_{AA}{A}^2+{\beta }_{BB}{B}^2+{\beta }_{CC}{C}^2+{\beta }_{AB}AB+{\beta }_{AC}AC+{\beta }_{BC}BC\dots$$

(4)

Where y is the predicted response, A, B y C are coded independent variables, β₀ is the intercept, β_A, β_B y β_C are linear coefficients, β_AA, β_BB y β_CC are quadratic coefficients, and β_AB, β_AC y β_BC are interaction coefficients. Optimal conditions were determined using Derringer-Suich's desirability function, simultaneously maximizing the three responses. Analyses were performed using Statgraphics Centurion software (version XVI.I, USA).

Biometric and physicochemical data were analyzed by one-way ANOVA and Tukey's multiple comparison test (α = 0.05) using Minitab software (version 14, USA).

Multivariate exploration was performed through principal component analysis (PCA) on a standardized data matrix (mean-centered, scaled to unit variance) comprising vitamin C, TSS, pH, TA, MI, anthocyanins, total phenols, and antioxidant capacity. A z-score heatmap was generated with annotation by maturity stage. Analyses were executed in R software (version 4.x, USA) with factoextra and ggplot2 packages.

3. Results and Discussion

3.1. Morphological, Biometric, and Physicochemical Characterization During Ripening

3.1.1. Plant and Fruit Morphology

C. tomentosum is a climbing woody liana characterized by large cordate leaves and small whitish-green flowers arranged in axillary inflorescences [1,2] (Figure 2A). The fruits are fleshy ovoid drupes of dark purple color, grouped in clusters of 3-6 units, with scarce mesocarp and sweet-tart flavor [3] (Figure 2B). Each drupe contains a single lunate seed with hard testa, characteristic of Menispermaceae (Figure 2C).

The ontogenic transition from green to ripe involved pronounced changes in epicarp pigmentation and mesocarp firmness (Table 1). The epicarp evolved from intense green and rough (green) to smooth yellowish-red (turning) and glossy blackish-violet (maduro), reflecting chlorophyll degradation and anthocyanin synthesis via the phenylpropanoid pathway [34,35]. The mesocarp transitioned from pale green and firm (green) to semi-soft red (turning) and juicy garnet (ripe), indicating pectin hydrolysis and accumulation of soluble sugars and organic acids [35,36]. The seed maintained constant lunate morphology and hardness, with testa evolving from whitish yellowish to light yellow.

Visual classification by colorimetric index (green = 0, turning = 1, ripe = 2) is used in drupes such as Olea europaea and Prunus spinosa, where epicarp pigmentation correlates with anthocyanin content, firmness, and physicochemical parameters [14,36]. In, C. tomentosum, the chromatic transition green → yellowish-red → blackish-violet represents a phenotypic marker of physiological ripening (Table 2).

3.1.2. Biometric Parameters

Fruit dimensions increased progressively during ripening (Table 2). Polar diameter increased from 2.34 ± 0.15 cm (green) to 2.49 ± 0.17 cm (ripe) (+6.4%), while equatorial diameter increased from 1.71 ± 0.15 cm to 1.97 ± 0.14 cm (+15.2%), accompanied by fresh weight gain from 3.89 ± 0.86 g to 5.74 ± 1.07 g (+47.7%; p < 0.0001). This growth pattern reflects cell expansion and accumulation of soluble solids characteristic of climacteric ripening in drupes [35,36].

Vitamin C content doubled from green to ripe phase (10.22 ± 0.01 vs. 5.12 ± 0.01 mg/100 g FW.; +99.6%), in line with activation of ascorbate biosynthesis during late ripening observed in Vaccinium spp. and other Andean berries [5,35,37]. TSS increased from 7.00 to 8.83 °Brix (+26.1%), which corresponded to the starch-to-sugar conversion and the building up of soluble carbohydrates [36]. At the same time, pH decreased from 6.28 to 4.27 (−32.0%) and titratable acidity increased from 0.25 to 0.94% (+276%), reflecting organic acid accumulation induced by berry ripening [36,38]. The maturity index (TSS/acidity ratio) decreased from 27.82 to 9.42, illustrating that acidity growth was greater than sugar production, which was contrary to common temperate berries and was in line with ripening fruit of tropical species such as Rubus and Vaccinium [35,39]. The acidic pH and higher TSS at maturity phase are positive for the stability of anthocyanins whose characteristic flavylium cation predominates below pH 5.0, improving extractability properties and chromatic intensity [36,38]. This biochemical framework indicates that ripe fruits maximize pigment content and technological acceptability for extraction of biologically active compounds.

3.2. Proximate Composition of Ripe Fruits

Ripe fruits presented moisture of 90.45 ± 0.47%, ash of 2.04 ± 0.14%, crude protein of 1.70 ± 0.02 g/100 g, crude fiber of 2.49 ± 0.01 g/100 g, crude fat of 0.05 ± 0.01 g/100 g, and carbohydrates by difference of 3.01 ± 0.49 g/100 g (dry basis). The high water content classifies C. tomentosum as an aqueous matrix comparable to grumixama (Eugenia brasiliensis, 86.2–89.9%) [40] and superior to cauchao (Amomyrtus luma, 75.3%) [37], favoring hydroalcoholic extraction of water-soluble metabolites [41].

The ash fraction exceeded 3.6–6.0 times the values of cauchao (0.57%) [37] and grumixama (0.34–0.49%) [40], suggesting high mineral content associated with montane volcanic soil [42]. Protein content was comparable to cauchao (1.73 g/100 g) [37] but higher than grumixama (0.65–0.83 g/100 g) [40]. The extremely low lipid content (2.3% of cauchao's value) confirms the characteristic profile of pigmented drupes where phenolic metabolites predominate over structural lipids [40,41]. This high-moisture, low-lipid matrix minimizes apolar interferences and maximizes solvent diffusion toward epidermal vacuoles rich in anthocyanins [37,41].

3.3. Ultrasound-Assisted Extraction Optimization

3.3.1. Box-Behnken Experimental Design

The Box-Behnken design evaluated 15 combinations of methanol (25–75%, v/v), ultrasonic amplitude (30–70%), and time (1–15 min) on anthocyanins, total phenols, and antioxidant capacity (Table 3). Response ranges were 42.64–120.82 mg C3G/L, 372.05–677.67 mg GAE/100 g, and 1869.09–5799.38 µmol TE/100 g, respectively, demonstrating variability of 2.8, 1.8, and 3.1-fold between extreme conditions. This second-order factorial design allows modeling linear, quadratic, and interaction effects with a reduced number of experiments compared to full factorial designs [15,16,43].

The quadratic models were statistically significant (p-model ≤ 0.0037) with R² of 97.62% (anthocyanins), 98.28% (phenols), and 96.57% (antioxidants), confirming robust predictive fit comparable to studies in Aronia melanocarpa (R² = 0.99) [15], Myrciaria cauliflora (R² = 0.98) [16], and Prunus domestica (R² = 0.97) [43]. The absence of residual autocorrelation (Durbin-Watson: 1.16–2.64; p > 0.05) validated independence of experimental errors [44,45]. The predictive equations in coded variables were:

$$\mathit{A}\mathit{n}\mathit{t}\mathit{h}\mathit{o}\mathit{c}\mathit{y}\mathit{a}\mathit{n}\mathit{i}\mathit{n}\mathit{s}=75.755+21.635\mathit{A}+7.862\mathit{B}+19.544\mathit{C}-8.466{\mathit{A}}^2+8.1{\mathit{B}}^2+4.387{\mathit{C}}^2+1.857\mathit{A}\mathit{B}+8.016\mathit{A}\mathit{C}+4.515\mathit{B}\mathit{C}\dots$$

(5)

$$\mathit{T}\mathit{o}\mathit{t}\mathit{a}\mathit{l}\mathit{p}\mathit{h}\mathit{e}\mathit{n}\mathit{o}\mathit{l}\mathit{s}=541.251+32.871\mathit{A}+32.302\mathit{B}+110.217\mathit{C}-27.183{\mathit{A}}^2-33.302{\mathit{B}}^2-7.067{\mathit{C}}^2-21.512\mathit{A}\mathit{B}-4.495\mathit{A}\mathit{C}+11.454\mathit{B}\mathit{C}\dots$$

(6)

$$\mathit{A}\mathit{n}\mathit{t}\mathit{i}\mathit{o}\mathit{x}\mathit{i}\mathit{d}\mathit{a}\mathit{n}\mathit{t}\mathit{c}\mathit{a}\mathit{p}\mathit{a}\mathit{c}\mathit{i}\mathit{t}\mathit{y}=3736.68+543.55\mathit{A}+453.656\mathit{B}+1049.47\mathit{C}-303.831{\mathit{A}}^2+89.198{\mathit{B}}^2-2.787{\mathit{C}}^2+158.884\mathit{A}\mathit{B}+142.16\mathit{A}\mathit{C}+464.109\mathit{B}\mathit{C}$$

(7)

Linear effects of methanol (A), amplitude (B), and time (C) were significant for all three responses (p ≤ 0.015), indicating monotonic increase toward upper levels like UAE optimizations in berries [16,43,44]. Quadratic terms lacked significance (p ≥ 0.17), suggesting absence of interior optimum and confirming that maximum conditions are located at the boundaries of the experimental space [15,45]. Two interactions were relevant: A×C increased anthocyanins (p = 0.048) and B×C increased antioxidant capacity (p = 0.0319), demonstrating synergy between solvent-time and amplitude-time respectively, consistent with ultrasonic cavitation mechanisms where prolonged times potentiate methanol and mechanical energy effects on cell disruption [16,44].

3.3.2. ANOVA Fit and Regression Coefficients

The quadratic models were significant (p ≤ 0.0037) with R² of 97.62% (anthocyanins), 98.28% (phenols), and 96.57% (antioxidants), without residual autocorrelation (Durbin-Watson: 1.16–2.64; p > 0.05). Regression coefficients (see Table 4) demonstrated the linear effect of methanol (A), amplitude (B), and time (C) on all (p ≤ 0.015) of the 3 responses, showing the monotonic increase. In summary, the non-significant terms showed no significance in quadratic variables (p ≥ 0.17), indicative of optimal at experimental space limits consistent with Fernández et al. [16] and Vázquez et al. [15]. The A×C interaction enhanced anthocyanins (p = 0.048) and B×C was found to provide the enhanced antioxidant capacity (p = 0.032), confirming solvent-time and amplitude-time synergy associated with ultrasonic cavitation mechanisms [43,44].

3.3.3. Response Surfaces and Mechanistic Interpretation

The 3D response surfaces (Figure 3, Figure 4, Figure 5 and Figure 6) illustrate the combined effects of methanol and amplitude at fixed time of 15 min on the three responses. The anthocyanin surface (Figure 3) shows progressive increase from 60 mg C3G/L (25% methanol, 30% amplitude) to 151 mg C3G/L (75% methanol, 70% amplitude), with color gradient transitioning from blue (low values) to red (high values). The elongated diagonal contours toward the upper right corner indicate synergistic effect between solvent concentration and ultrasonic energy. This pattern confirms that elevated methanol concentrations improve solubilization of glycosylated anthocyanins by reducing medium polarity, while higher amplitudes intensify acoustic cavitation that disrupts epidermal vacuoles rich in pigments [16,43].

The antioxidant capacity surface (Figure 4) exhibits analogous behavior, increasing from 3200 µmol TE/100 g (minimum conditions) to 6200 µmol TE/100 g (maximum conditions), representing a 2.0-fold increase. The surface shows parallel contours without pronounced interior curvature, consistent with the absence of significant quadratic terms (Table 4; p ≥ 0.17) and confirming additive effects of methanol and amplitude. The linear increase reflects that reducing capacity measured by DPPH increases proportionally with the extraction of total phenolic compounds and anthocyanins, main contributors to antioxidant activity [15,44].

Total phenols (Figure 5) present a surface with increase from 500 mg GAE/100 g to 670 mg GAE/100 g, showing greater influence of methanol (coefficient β = 32.87; p = 0.006) than amplitude (β = 32.30; p = 0.007). The more widely spaced contours in the direction of the methanol axis evidence its dominant effect on phenolic recovery. The optimal region is located at the upper boundary of the experimental space (≥ 65% methanol, ≥ 60% amplitude), confirming absence of interior maximum. Mechanistically, methanol facilitates rupture of ester bonds between phenolics and cell wall polysaccharides, while ultrasound accelerates mass transfer by localized microagitation in bubble collapse zones of cavitation [43,45].

3.3.4. Multi-Response Optimization and Validation

The desirability surface (Figure 6) identifies the optimal region at 57% methanol, 70% amplitude, and 15 min (desirability = 0.9996), represented by the red zone in the upper right corner where the three optimal responses converge. Desirability transitions from 0.4 (blue, minimum conditions) to 1.0 (red, optimal conditions), demonstrating simultaneous maximization without trade-offs between objectives. Experimental validation (Table 5) confirms 120.71 ± 1.89 mg C3G/L, 672.46 ± 5.84 mg GAE/100 g, and 5857.31 ± 60.20 µmol TE/100 g, with deviations of 1.69%, 0.47%, and 2.07% from predicted values, within acceptable limits (< 5%) for RSM models [44].

3.4. Application of Optimal Conditions Across Maturity Stages

Under optimized UAE conditions (57% methanol, 70% amplitude, 15 min), maturity stage determined significant differences in all three responses (ANOVA, p < 0.0001; Table 6). Monomeric anthocyanins were undetectable in green fruits, reached 46.01 ± 0.70 mg C3G/L in turning fruits, and 120.71 ± 1.89 mg C3G/L in ripe fruits (+162% vs. turning; p < 0.0001). This progressive increase reflects activation of the phenylpropanoid biosynthetic pathway during ripening, responsible for anthocyanin synthesis through sequential action of chalcone synthase, flavanone 3-hydroxylase, and UDP-glucose:flavonoid 3-O-glucosyltransferase [34,35,38]. Additionally, cell wall softening mediated by pectinases and cellulases in ripe stages facilitates mechanical disruption by ultrasonic cavitation, increasing release of vacuolar anthocyanins [14,16].

Total phenols exhibited a nonlinear pattern during ripening. The turning stage registered the minimum value (458.86 ± 5.66 mg GAE/100 g), while ripe reached the maximum (672.46 ± 5.84 mg GAE/100 g; +46.7% vs. turning; p < 0.0001), slightly exceeding green (647.83 ± 3.55 mg GAE/100 g; +3.8% vs. green). The transient depression in turning coincides with the rapid cell expansion phase, where dilution by water accumulation and metabolic reconfiguration toward anthocyanin biosynthesis temporarily reduce the total phenolic pool [5,34,35]. In contrast, the ripe stage combines net accumulation of structural phenols (hydroxycinnamic acids) with newly synthesized anthocyanins, resulting in maximum phenolic content [38,41].

Antioxidant capacity followed a trend analogous to total phenols: turning < green < ripe (3581.11 ± 39.80 < 5492.63 ± 54.30 < 5857.55 ± 60.20 µmol TE/100 g). The +63.6% increase between turning and ripe (p < 0.0001) correlates with synchronized accumulation of anthocyanins (+162%) and total phenols (+46.7%), evidencing additive contribution of both metabolite families to reducing activity measured by DPPH [15,34]. The acidic pH of the ripe stage (4.27 ± 0.02; Table 1) additionally favors stability of the flavylium cation (predominant form of anthocyanins at pH < 5), maximizing its antioxidant activity through electronic resonance of the oxonium ring [36,38].

3.5. Principal Component Analysis

PCA on eight physicochemical and bioactive variables explained 90.6% of total variability (PC1 = 67.4%, PC2 = 23.2%; Figure 7). PC1 ordered samples according to ontogenic gradient: green fruits in negative PC1, ripe in positive PC1, turning in intermediate positions. This separation reflects metabolic transition during ripening [5,35].

TSS, vitamin C, anthocyanins, total phenols, antioxidant capacity, and acidity presented positive loadings on PC1, associated with ripe fruits. pH and maturity index showed negative loadings, associated with green fruits. This configuration agrees with Table 1 data: TSS increased +26.1% (7.00 → 8.83 °Brix), vitamin C +99.6%, acidity +276% (0.25 → 0.94%), while pH decreased 32.0% (6.28 → 4.27) and maturity index −66.1% (27.82 → 9.42). This pattern, where acidity exceeds sugar accumulation, is characteristic of Andean and tropical berries of Vaccinium and Rubus, differing from temperate fruits [34,39,41].

PC2 differentiated the turning stage, which projected separately from green and ripe. This separation captures the transient metabolic minimum in turning: total phenols (458.86 mg GAE/100 g; −29.2% vs. green) and antioxidant capacity (3581.11 µmol TE/100 g; −34.8% vs. green; Table 5). This valley coincides with rapid cell expansion, where water dilution and biosynthetic reorientation toward anthocyanins temporarily reduce the extractable phenolic pool [5,35]. Vaccinium corymbosum and Rubus ulmifolius exhibit analogous minima at intermediate stages [34,38].

The standardized z-score heatmap confirmed differentiation by maturity (Figure 8). Ripe fruits presented positive z-scores in anthocyanins, total phenols, antioxidant capacity, vitamin C, TSS, and acidity, negative in pH and maturity index. This pattern indicates co-accumulation of primary (sugars, ascorbate) and secondary (phenols, anthocyanins) metabolites, characteristic of pigmented berries where ripening integrates shikimate, phenylpropanoid, and ascorbate-glutathione pathways [38,41]. Turning showed minimum z-scores exclusively in phenols and antioxidant capacity, confirming its metabolic transition.

Vacuolar acidification in ripe fruits (pH 4.27; Table 1) optimizes bioactive compounds under UAE. Acidic pH stabilizes the flavylium cation (predominant form at pH < 5), preventing conversion to colorless forms and maximizing chromophoric content and antioxidant activity [36,38]. Furthermore, the acidic environment protonates phenolic hydroxyl groups in hydroalcoholic mixtures allowing greater solubility in them and also in optimizing UAE efficiency [14,16]. This association between ripening physiology and extraction method enables ripe fruit to maximise all three responses at the same time (Table 5): 120.71 mg C3G/L (+162% vs. turning), 672.46 mg GAE/100 g (+46.7% vs. turning), 5857.55 µmol TE/100 g (+63.6% vs. turning).

The findings demonstrate that C. tomentosum ripening has a progressive bioactive compound accumulation model, which was promoted via vacuolar acidification with a transient metabolic valley in turning. This pattern, along with favorable UAE conditions (57% methanol, 70% amplitude, 15 min) turns ripe fruit into raw materials for natural colorants, nutraceuticals and functional foods from Amazonian berries.

4. Conclusions

Box-Behnken optimization established UAE conditions (57% methanol, 70% amplitude, 15 min) that maximized anthocyanin yield (120.71 mg C3G/L), total phenols (672.46 mg GAE/100 g), and antioxidant capacity (5857.55 µmol TE/100 g) from ripe C. tomentosum fruits with high predictive accuracy (R² > 0.95, 2.1% deviation). Ripening from green to ripe stage triggered a 162% increase in anthocyanins and 46.7% increase in total phenols, driven by vacuolar acidification (pH 4.27) that stabilizes anthocyanin structure and enhances UAE extractability. The transitory metabolic valley at turning stage (−29.2% phenols, −34.8% antioxidant capacity) reflects biosynthetic reorganization during cell expansion, establishing an ontogenic pattern where acidification predominates over sugar accumulation. Proximate composition (90.45% moisture, 2.04% ash, 2.49% fiber, 1.70% protein) positions C. tomentosum as a hydrophilic matrix comparable to Andean berries, favoring hydroalcoholic extraction of water-soluble bioactives. These findings establish ripe C. tomentosum fruits as a promising raw material for natural colorants and functional food ingredients derived from underexploited Amazonian biodiversity.