The intrinsic relevance of nitrogen removal pathway to varying nitrate loading rate in a polycaprolactone-supported denitrification system.

Up to date, the intrinsic association of nitrate loading rate (NLR) with treatment performance of solid-phase denitrification (SPD) systems is still ambiguous. To address this issue, three continuous up-flow bioreactors were configured. They were packed with polycaprolactone (PCL) under a filling ratio of 30%, 60% or 90% and were operated under a varying NLR of 0.34 ± 0.01-3.99 ± 0.12 gN/(L·d). Results showed that the denitrification efficiency was high (RE > 96%) and stable except the case with the highest NLR, which was mainly attributed to the lack of available carbon sources. At the phylum or genus level, most of the detected dominant bacterial taxa were either associated with organics degradation or nitrogen metabolism. The difference in bacterial community structure among the three stages was mainly caused by NLR rather than the filling ratio. Moreover, as the NLR got higher, the Bray-Curtis distance between samples from the same stage became shorter. By the results of gene or enzyme prediction performed in PICRUSt2, the main nitrogen metabolism pathways in these reactors were denitrification, dissimilatory nitrate reduction to ammonium (DNRA), assimilatory nitrate reduction to ammonium (ANRA) and nitrogen fixation. Moreover, aerobic and anaerobic nitrate dissimilation coexisted in the systems with the latter playing a dominant role. Finally, denitrification and DNRA occurred under both high and low NLR conditions while nitrogen fixation and ANRA preferred to occur under low NLR environments. These findings might help guide practical applications.