We used nitrogen stable isotopes to study the regulation of nitrogen fixation by filamentous cyanobacteria. Nitrogen fixation was found to be almost insensitive to combined nitrogen, along a gradient from the Himmerfjarden sewage treatment plant discharge to the open sea. We found similarly low cyanobacterial (mostly Aphanizomenon sp.) delta N-15-values at all stations, despite significant differences along the bay in both total nitrogen concentrations in water and delta N-15 in seston (particles <10 mu m), the latter used as a proxy for algae growing on combined nitrogen alone. Only late in the productive season, when cyanobacterial biomass was declining or already low, did elevated delta N-15 suggest uptake of combined nitrogen. However, this coincided with an increase in the contribution of Dolichospermum spp. to overall diazotrophic biomass and may indicate uptake of combined nitrogen by this species. These results indicate that almost all nitrogen used for growth by nitrogen-fixing cyanobacteria in the study area comes from nitrogen fixation, and very little from uptake of dissolved combined nitrogen. This study was part of a whole ecosystem experiment analyzing the effects of nitrogen removal in a sewage treatment plant discharging to the Himmerfjarden Bay, northern Baltic Sea Proper.

Heterocyst frequencies in Baltic Sea Aphanizomenon sp. were similar along a strong nutrient gradient from the discharge point of a sewage treatment plant at the head of the Himmerfjarden bay to the open sea. Filaments lacked heterocysts in winter and for over a month after the spring bloom had depleted combined nitrogen in the surface layer. Heterocyst-free filaments in spring contained granulate structures that decreased in abundance simultaneously as colony nitrogen content decreased, but delta N-15 remained unchanged, indicative of storage of fixed nitrogen in over-wintering Aphanizomenon sp. filaments. Heterocyst formation was initiated when water temperature was sufficient to form a shallow seasonal pycnocline that allowed filaments to be exposed to enough light to initiate growth and a subsequent intracellular shortage of nitrogen. During the growth season, heterocyst frequency varied significantly with maximum values in early summer (May), lower values in mid-August that coincided with maximum temperatures and an increase in late summer. Heterocyst frequencies decreased with increased temperatures, suggesting a more efficiently functioning nitrogenase enzyme. Based on data from three seasons, filament C: P ratios did not correlate with heterocyst frequencies, neither did reduced heterocyst frequencies coincide with high dissolved inorganic nitrogen concentrations. Increased heterocyst frequencies, however, resulted in decreased C:N ratios, probably as more heterocysts likely increase nitrogen fixation.