Nitrogen Addition Reshapes Soil Carbon Molecular Composition via Nitrate–Enzyme Interactions in Soybean–Maize Intercropping

Nitrogen (N) fertilization is a fundamental agronomic practice that governs crop productivity, yet its effects on the molecular composition and chemical stability of soil organic carbon (SOC) remain poorly understood, especially in cereal–legume intercropping systems. Traditional studies have focused on total SOC stocks rather than molecular-level changes, and the mechanistic pathway linking N addition to SOC functional group transformation remains unclear. This study addressed these critical gaps by investigating how graded N addition (0, 180, 270, and 360 kg N ha−1) reshapes SOC chemistry in a subtropical soybean–maize intercropping system. Soil physicochemical properties, inorganic N pools, N-transformation enzyme activities (urease, nitrate reductase, and glutaminase), microbial biomass indices, labile organic carbon fractions (particulate, mineral-associated, and dissolved organic carbon), and SOC functional groups characterized by Fourier transform infrared (FTIR) spectroscopy were quantified across a two-year field experiment (2024–2025). Results showed that increasing N rates significantly elevated nitrate nitrogen (NO3−-N) accumulation while depressing soil pH. Nitrogen-transformation enzymes, especially nitrate reductase and glutaminase, responded strongly and positively to the N gradient. Microbial biomass carbon (MBC) and nitrogen (MBN) increased with moderate N input but exhibited saturation or decline at 360 kg N ha−1, accompanied by reduced microbial carbon use efficiency (CUE) and a lower MBC/MBN ratio. Among labile carbon fractions, dissolved organic carbon (DOC) was the most responsive pool, increasing markedly with N addition and correlating strongly with NO3−-N. FTIR analysis revealed that N addition shifted SOC functional group composition toward chemically recalcitrant structures: the relative abundances of aromatic C=C and carbonyl C=O groups increased significantly, whereas labile C–O groups declined. Random forest modelling identified C=C, NO3−-N, and DOC as the three most influential predictors of SOC chemical composition. Structural equation modelling (SEM) demonstrated a sequential mechanistic pathway: N fertilization increased NO3−-N, which stimulated glutaminase activity and enhanced DOC, ultimately promoting C=C/C=O stabilization and explaining 91.3% of the variance in SOC aromaticity. These findings reveal that N addition does not merely augment SOC quantity but fundamentally transforms its molecular architecture toward greater chemical stability through a nitrate-mediated, enzyme–labile carbon coupling mechanism. This study provides a novel spectroscopic–mechanistic framework for understanding carbon–nitrogen interactions in intercropping agroecosystems and informs precision N management strategies aimed at simultaneous crop production and long-term soil carbon sequestration.