Nitrogen isotope compositions (δ15N) in sedimentary rocks are extensively used to investigate the biogeochemical nitrogen cycle through geological times. This use relies on the observation that, in modern continental platforms and anoxic basins, surface sediments faithfully record the δ15N of primary producers, assuming that it was similar in the past. Over Earth’s history, however, surface environments experienced profound changes, including the transition of ammonium-dominated to nitrate-dominated waters and the transition from exclusively microbial ecosystems to ecosystems including multicellularity, which make modern environments significantly different compared to earlier ones, potentially invalidating the fundamental assumption that surface sediments faithfully record the δ15N of primary producers. In order to improve our understanding of the nitrogen isotopic information contained in the early Earth’s sedimentary rock record, we investigate here the nitrogen isotope systematics in a microbial, nitrate free and ammonium-rich modern system, the Dziani Dzaha Lake. In this modern system, the δ15N of the reduced dissolved inorganic nitrogen (i.e., NH4+ and NH3) in the water column is close to ∼7‰ . δ15N of suspended particulate matter (SPM) show a similar average value in surface waters (i.e., where SPM is massively composed of active primary producers), but increases up to 14‰ in the deeper part of the water column during periods when it is enriched in dissolved reduced species (i.e., CH4, H2S/HS− and NH4+/NH3). Surface sediments δ15N, with values comprised between 10 and 14 ‰, seem to preferentially record these positive isotopic signatures, rather than those of active primary producers. We propose here that the observed isotopic pattern is mainly linked to the assimilation of ammonium strongly enriched in 15N by isotope exchange with ammonia under basic conditions. Although ammonium assimilation seems here to be responsible for a significant isotopic enrichment due to the basic conditions, in neutral anoxic environments inhabited by similar microbial ecosystems, this process may also significantly impact the δ15N of primary producers towards more negative values. This would have strong implications for our interpretation of the Precambrian sedimentary record as this finding challenges one the fundamental hypotheses underlying the use of sedimentary δ15N in paleo-oceanographic reconstructions, i.e. that surface sediments faithfully record the δ15N of active primary producers in the photic zone.