Phosphorus fertilizer input level regulates soil organic carbon physical fraction sequestration by influencing the microbial community

Microbe-driven soil organic carbon (SOC) turnover has received worldwide attention because of its ability to improve soil fertility, increase crop productivity, and achieve C neutrality. The fertilization regime is the main factor regulating this process. To date, most related studies have focused on the effects of urea or nitrogen (N) fertilizer levels on SOC accumulation. However, knowledge is lacking concerning the relationships among phosphorus (P) fertilizer levels, soil microbial communities, and turnover of SOC fractions. Herein, a continuous 4-year in situ field experiment was conducted after straw retention with the following treatments combined with regular N and potassium (K) fertilization: (i) regular P fertilizer (P + NK); (ii) 25 % reduction in P fertilizer (0.75P + NK); (iii) 50 % reduction in P fertilizer (0.5P + NK); and (iv) no P fertilizer (NK). Maize yield, SOC fractions and microbial communities responded distinctly to different P fertilizer levels. Regular fertilization resulted in the highest maize yield, macroaggregate proportion, and aggregate mean weight diameter. A significant decrease in particulate organic carbon (POC) was observed under NK. Moreover, significant decreases in mineral-associated organic carbon (MaOC) were observed under 0.5P + NK and NK compared with those under regular fertilization. Moreover, turnover of SOC fractions was strongly associated with microbial clusters and keystone taxa. Linear regressions indicated close associations between communities in clusters 2 and 3 and POC and MaOC. Random forest models further predicted that keystone taxa in the co-occurrence network may significantly explain SOC turnover. Overall, there were significant correlations between the bacterial richness of Chitinophagaceae and Saprospiraceae (within cluster 3) and those of POC and MaOC. Specifically, the fungal richness of Lasiosphaeriaceae (within cluster 2) was significantly positively correlated with only MaOC. Overall, fungi, rather than bacteria, drove the function of specific microbial clusters and thus affected SOC fraction turnover. The Lasiosphaeriaceae-driven cluster 2 community facilitated MaOC sequestration, whereas the Chitinophagaceae- and Mortierellaceae-driven cluster 3 communities facilitated both POC and MaOC accumulation. Our findings strengthen our understanding of the relationships among P fertilizer reduction, microbial communities and SOC fractions. Furthermore, we optimized the fertilization regime for sustained crop yield. Specifically, reducing P fertilization by 25 % is a win–win strategy for optimizing fertilization and promoting soil fertility.