Fractional Factorial Design-Guided Optimization of Flavonoid Content and Antioxidant Capacity in Justicia gendarussa Leaf Extracts and UHPLC-HRMS-Based Metabolite Characterization

Background/Objectives: Extraction conditions can influence the chemical profile and antioxidant properties of medicinal plant extracts. This study aimed to optimize ethanol-based extraction of Justicia gendarussa leaves and characterize the optimized extract through UHPLC-HRMS metabolite profiling, with application to cultivated plant materials. Methods: A two-level half-fractional factorial design was used to assess the effects of five extraction factors—liquid-to-solid ratio, temperature, ethanol concentration, extraction time, and pH—on three response variables: total flavonoid content (TFC), DPPH radical-scavenging activity, and ferric reducing antioxidant power (FRAP). The optimized condition was validated experimentally, characterized by UHPLC-Q Exactive Plus Orbitrap HRMS with data-dependent MS/MS, and applied to cultivated leaf and stem materials under six fertilization treatments, analyzed using two-way ANOVA (organ × treatment). Results: The liquid-to-solid ratio × temperature interaction was the dominant determinant of TFC, and the liquid-to-solid ratio alone dominated both DPPH and FRAP. Desirability-based optimization identified a condition of 100 mL g−1 liquid-to-solid ratio, 30 °C, 20% ethanol, 30 min, and pH 2. Confirmatory extraction resulted in 16.43 ± 0.30 mg QE g−1 DW TFC, 12.30 ± 0.02 µmol TE g−1 DW DPPH, and 190.56 ± 2.58 µmol TE g−1 DW FRAP, in close agreement with model predictions (0.10–1.54% error). UHPLC-HRMS/MS detected 52 putatively annotated metabolites, including a candidate gendarusin A/B (apigenin di-C-arabinoside)-type C-glycosylflavone. Leaves showed significantly higher TFC, CUPRAC, and FRAP than stems regardless of fertilization treatment (organ effect p < 0.0001), whereas the leaf–stem difference in DPPH capacity varied significantly with treatment (interaction p = 0.0010). Conclusions: These findings provide a validated extraction condition for the measured responses and support further targeted metabolite evaluation of J. gendarussa.