Amino acids (AA) represent the most abundant identifiable biomolecule class in marine dissolved organic nitrogen (DON) and provide powerful proxies for DON degradation state. In particular, the D-enantiomers of AA (D-AA) are known to be derived mainly from bacteria, making them ideal tracers for bacterially derived N. However, despite the widespread use of D-AA tracers, it remains unclear if the accumulation of different D-AA species in the ocean indicates that most DON arises from direct bacterial sources or from continual bacterial alteration of eukaryotic algal material. This difference has major implications for our understanding of sources and cycling mechanisms of DON in the ocean. Here, we present the most extensive D-AA suite ever reported in younger, high molecular weight (HMW) DON contrasted with older, low molecular weight (LMW) solid phase extracted (SPE) DON from the central Atlantic and Pacific Oceans. We evaluate D-AA in these two contrasted MW fractions in the context of multiple common AA-based proxies and bulk DOM radiocarbon (Δ14C) data. Specifically, we assess if D-AA in HMW and LMW SPE-DON are most consistent with 1) preformed bacterial source signals, 2) progressive bacterial degradation/alteration of eukaryotic algal sources, or 3) gradual, continued resynthesis/addition of new bacterial biomolecules during ocean circulation. Our results suggest that AA-containing molecules in HMW and LMW SPE-DON fractions are almost entirely distinct, with independent bacterial sources and degradation mechanisms. In HMW DON, all measured indices support a surface-produced, semi-labile component which is progressively altered in the mesopelagic with increasing radiocarbon age. In contrast, for LMW SPE-DON, AA-based proxies showed conflicting results. Some proxies (D/L ratios of most D-AA, non-protein AA, mol% Gly, and %C-AA) indicated LMW SPE-DON was less labile and more degraded than HMW DON. However other proxies (D/L-Ala, the N isotope based ΣV parameter, and the commonly used DI index) indicated equivalent or even less degradation and resynthesis in the isolated LMW material compared to HMW material, suggesting a disconnect in the underlying mechanisms reflected by different proxies. Finally, AA composition and degradation state in the subsurface samples of both HMW and LMW DON varied little with increasing radiocarbon age, suggesting that HMW and LMW pools may cycle independently. Together, our results suggest that AA DON sources, while almost entirely bacterial, may be more diverse than previously believed, and that much of the hydrolysable AA pool in the ocean may not be derived from proteinaceous material. Overall, these observations support the microbial nitrogen pump idea, but with compositionally unique refractory components in both HMW and LMW material which resist degradation over millennial timescales.
Read full abstract