Genotype × water deficit interactions drive contrasting phenotypic responses and production outcomes in Brazilian and Andean maize
Phenotypic plasticity, water deficit, genotype × environment interaction, maize breeding, drought tolerance, multivariate analysis, production efficiency, harvest index
Water deficit represents a major constraint to maize production worldwide, yet genotypes from different breeding backgrounds may exhibit distinct response patterns with contrasting implications for productivity. This study investigated genotype × water deficit interactions in two contrasting maize genotypes during pre-flowering: INIA 618 (high-altitude Andean breeding program) and B2782PWU (Brazilian tropical breeding program). Plants were subjected to well-watered (90% WHC) and water deficit (50% WHC) treatments for 20 days during the V14 stage. We employed an integrative approach combining physiological, biochemical, anatomical, and biometric assessments with advanced multivariate analyses, including the Multivariate Plasticity Index (MVPi) to quantify coordinated phenotypic responses. Results revealed fundamentally different response strategies with distinct production implications: INIA 618 exhibited a resource conservation strategy characterized by exceptional physiological plasticity (MVPi = 148.7), emphasizing osmolyte accumulation (proline increased 3.2-fold, soluble sugars 2.8-fold) and anatomical modifications (reduced stomatal density, increased mesophyll thickness). In contrast, B2782PWU employed a metabolic maintenance strategy with moderate physiological plasticity (MVPi = 17.6) but higher biometric plasticity (MVPi = 6.2), sustaining growth processes while adjusting biomass allocation patterns. Critically, physiological plasticity showed strong negative correlations with both grain yield and harvest index (r = -0.43 to -0.53, p < 0.001), revealing fundamental trade-offs between stress tolerance mechanisms and productive performance. Principal component regression outperformed simple correlations and multiple linear regression for predicting production outcomes, with harvest index showing stronger relationships with phenotypic traits than grain yield (R² up to 0.54 vs 0.034). Principal component analysis revealed that the first two components explained 49.7% of phenotypic variation, with clear separation between genotypes and treatments validating the genotype × environment interaction framework. These findings demonstrate that breeding background significantly influences drought response patterns and provide quantitative evidence for optimizing genotype selection and management strategies in water-limited environments. The contrasting strategies identified offer complementary approaches for developing drought-resilient cultivars: incorporating INIA 618-type tolerance mechanisms for severe stress environments while utilizing B2782PWU-type efficiency traits for moderate stress conditions with recovery potential.