Fernanda Anselmo Moreira¹, Bruno Ruiz Brandão da Costa², Patricia Menezes Ferreira Rodrigues¹, Eduardo Luís Martins Catharino¹, Agnès Borbon³, Adalgiza Fornaro⁴, Silvia Ribeiro de Souza¹, Cláudia Maria Furlan²

Climate change is reshaping environments worldwide. Two key stressors driven by climatic shifts are air pollution and drought, common in urban areas where they compromise ecosystem services provided by urban forests. Under such conditions, plants may undergo major metabolic adjustments. This study investigated the metabolomic responses of two native Atlantic Forest species – Alchornea sidifolia (AS) and Guarea macrophylla (GM) – sampled in the Morro Grande Forest Reserve (RMG, less urbanized) and the Instituto de Biociências Forest Reserve (Matão-IAG, more urbanized), both located in the Metropolitan Area of São Paulo (MASP), Brazil. Sampling occurred during both the dry (Aug–Sep 2023) and rainy (Jan–Mar 2024) seasons. Leaf samples from six individuals per species, per area and season, were extracted with chloroform/methanol/water (12:5:1). The polar phase, after derivatization, was analyzed by GC-MS. Spectral data were processed using GNPS, followed by multivariate (PCA, PERMANOVA, Random Forest) and univariate (Mann-Whitney and effect size) analyses using MetaboAnalyst and R. Annotation was based on analytical standards (level 1), retention index and mass spectra (level 2), and molecular networking with Cytoscape (level 3). After filtering (blank ratio and balance score >65%), 238 and 243 features were retained for AS and GM, respectively. Of these, 98 (94% of total relative abundance) and 99 (82%) compounds were annotated. Data were imputed using LOD/5, sum-normalized, log10-transformed, and auto-scaled. Compounds were categorized into nine chemical classes. Fingerprint analysis showed significant group differences for AS (F = 3.75; p = 0.005), but not for GM (F = 2.04; p = 0.088). AS exhibited more pronounced metabolic adjustments in Matão-IAG, with significant seasonal separation (F = 2.92; p = 0.003), unlike RMG (F = 1.22; p = 0.158). In Matão-IAG, the dry season showed elevated levels of amino acids (proline, alanine, threonine) and organic acids (phosphoric, succinic, isocitric, among others), which displayed high effect sizes and ranked among the most important features in the Random Forest analysis. Conversely, the rainy season was marked by higher disaccharide levels. In RMG, only a few compounds, mainly polyols, showed significant seasonal variation. In contrast, GM showed a distinct pattern: a clear separation between seasons was found in RMG (F = 2.25; p = 0.008), but not in Matão-IAG (F = 0.96; p = 0.434). In RMG, GM showed elevated levels of organic acids (shikimic, ribonic, oxalic) and phenolic compounds (quinic, p-coumaric) during the rainy season, with strong effect sizes and higher importance in Random Forest compared to the dry season. These species- and site-specific responses likely reflect combination of functional traits and local environmental factors. AS, a fast-growing pioneer species, appears more sensitive to environmental stressors such as drought and pollution, which may explain its stronger responses in the more urbanized Matão-IAG, particularly under combined stress during the dry season. In contrast, GM, a non-pioneer species, exhibited greater shifts in RMG, potentially due to factors such as microclimatic variability, resource availability, or biotic interactions within this larger and less urbanized fragment. Our findings indicate that both seasonal and spatial metabolomic plasticity are species-dependent and modulated by the interplay between urbanization, climate, and plant traits.

Agradecimentos: FAPESP grant numbers: 2020/07141-2, 2022/07326-8, 2022/13213-1.