Effect of iron limitation on the isotopic composition of cellular and released fixed nitrogen in Azotobacter vinelandii

Most biological nitrogen transformations have characteristic kinetic isotope effects used to track these processes in modern and past environments. The isotopic fractionation associated with nitrogen fixation, the only biological source of fixed nitrogen (N), provides a particularly important constraint for studies of nitrogen cycling. Nitrogen fixation using the ‘canonical’ Mo-nitrogenase produces biomass with a δ¹⁵N value of ca. −1‰ (vs. atmospheric N₂). If the ‘alternative’ V- and Fe-only nitrogenases are used, biomass δ¹⁵N can be between −6‰ and −7‰. These biomass values are assumed to be relatively invariant and to reflect the cellular level expressed isotope effect of nitrogen fixation. However, field and laboratory studies report wide ranges of diazotrophic biomass δ¹⁵N (from −3.6‰ to +0.5‰ for Mo-based nitrogen fixation). This variation could be partly explained by the release of dissolved organic N (DON) that is isotopically distinct from biomass. The model nitrogen fixer Azotobacter vinelandii secretes siderophores, small molecules that aid in Fe uptake and can comprise >30% of fixed nitrogen. To test whether siderophores (and other released N) can decouple biomass δ¹⁵N from the isotope effect of nitrogen fixation we measured the isotopic composition of biomass and released N in Fe-limited A. vinelandii cultures fixing nitrogen with Mo- and V-nitrogenases. We report that biomass δ¹⁵N was elevated under Fe limitation with a maximum value of +1.2‰ for Mo-based nitrogen fixation. Regardless of the nitrogenase isozyme used, released nitrogen δ¹⁵N was also 2–3‰ lower than biomass δ¹⁵N. Siderophore nitrogen was found to have a slightly higher δ¹⁵N than the rest of the DON pool but was still produced in large enough concentrations to account for increases in biomass δ¹⁵N. The low δ¹⁵N of siderophores (relative to biomass) is consistent with what is known from compound specific isotope studies of the amino acids used in siderophore biosynthesis, and indicates that other amino-acid derived siderophores should also have a low δ¹⁵N. The implications for studies of nitrogen fixation are discussed.