Amino acid enantiomers in old and young dissolved organic matter: Implications for a microbial nitrogen pump

Dissolved organic nitrogen (DON) represents the largest reservoir of fixed N in the surface ocean and a significant portion accumulates in the deep sea, where it can persist for millennial time scales. However, like the dissolved organic carbon (DOC) pool, the origin and composition of long-lived, refractory DON remains largely unknown. In recent years, the “microbial carbon pump” hypothesis has emerged from abundant evidence showing that microbial processes are primarily responsible for refractory DOC accumulation. However, a similar mechanism for DON has rarely been investigated. In the study of DON, spectroscopic evidence has indicated a primarily amide composition, implying a dominant contribution from peptides. Therefore, if an analogous “microbial nitrogen pump” controls refractory DON accumulation, the amino acid component should bear increasing signatures of microbial origin with increasing age. Here we investigate the microbial sequestration of N via the production of refractory DON, for the first time considering together DOM Δ14C with amino acid (AA) molar abundance (Mol%) and D/L ratio (as a tracer for prokaryotic input). Measurements were made on a unique set of high and low molecular weight (HMW, LMW) DOM isolates with 14C ages and chemical compositions generally consistent with semi-labile and refractory DOM respectively. The samples were collected in the North Pacific Subtropical Gyre where deep waters contain some of the oldest DOC in the world ocean. We observe higher D/L ratios in older, LMW DOM isolates for almost all analyzed AAs. Using mass spectral data, we also quantify three D-AAs in all samples (D-valine, D-phenylalanine, and D-leucine), which have not previously been confirmed in ocean DOM. These newly identified D-AAs are concentrated in the LMW refractory DOM fraction and have oceanographically consistent depth profiles. Our results suggest that several novel D-AA subgroupings may be unique tracers for different prokaryotic source processes. D-alanine appears to have largely independent cycling from the other D-AAs with a connection to the production of HMW DON, which we hypothesize is linked to water column peptidoglycan. In contrast, D-leucine, D-valine, and D-phenylalanine appear to be most strongly related to the production of LMW DON. Trends in both the HMW and LMW fractions suggest a linkage to sinking particles and local microbial transformations, implying that LMW DON has a direct biological source rather than originati