ULTRASONIC PRETREATMENT FOR PROTEIN RECOVERY FROM Amaranthus viridis FLOUR

Amaranthus viridis is an underused edible plant with potential as an alternative protein source (Jin et al., 2025). Obtaining protein-rich fractions from green biomass requires extraction strategies that improve protein recovery while limiting the co-extraction of nonprotein constituents (Pérez-Vila et al., 2022). This study evaluated ultrasonic pretreatment before alkaline extraction to obtain protein concentrates from A. viridis flour, focusing on nominal power and sonication time. The plant material was dried at 65 °C for approximately 20 h and ground. Preliminary precipitation tests covered pH 1.5–5.0; pH 2.5 was selected because residual soluble protein did not differ significantly from that at pH 2.0, allowing less severe acidification. Ultrasonic pretreatment was performed using a 20 kHz probe sonicator at nominal powers of 250, 500, and 750 W for 10, 20, or 30 min, in pulsed mode (3 s on/3 s off). The nine conditions were compared with conventional extraction without sonication pretreatmet. Alkaline extraction was conducted at 40 °C for 2 h, followed by centrifugation, acid precipitation at pH 2.5, and concentrate recovery and drying. Experiments were conducted in triplicate. Concentrate protein content was determined by Kjeldahl, while soluble protein after alkaline extraction was assessed by Bradford (Kruger, 2009). Global extraction yield represented concentrate mass relative to initial flour mass, whereas protein recovery represented the proportion of initial protein recovered in the concentrate. Data were analyzed by ANOVA and Tukey’s test at a 5% significance level. Ultrasonic effects depended on the processing conditions. Pretreatment at 750 W for 30 min produced the highest protein recovery (27.3 ± 2.5%) and concentrate protein content (81.58 ± 1.10%), compared with 7.57 ± 2.50% and 33.2 ± 5.2%, respectively, for conventional extraction. This was the only pretreatment that significantly increased protein recovery relative to the control (p < 0.05). Global extraction yield increased from 6.09 ± 0.37% to 7.37 ± 0.58%. These results highlight the importance of assessing both concentrate composition and mass recovery when selecting extraction conditions. In contrast, Bradford-assessed soluble protein did not differ significantly among treatments (p > 0.05). Although ultrasound can modify protein structure and functionality (Rahman and Lamsal, 2021), this divergence warrants further investigation of precipitation behavior and analytical responses before assigning a mechanism. Ultrasonic pretreatment did not consistently improve protein recovery across all tested conditions. Among those evaluated, 750 W for 30 min was the most promising for obtaining a protein-rich concentrate from A. viridis. Further assessment of energy requirements and concentrate functionality is needed to evaluate its practical application. The authors acknowledge FAPESP (2025/17032-0), CNPq (420526/2023-4), and the University of São Paulo Unified Scholarship Program (PUB).