Significant N₂ fixation by heterotrophs, photoheterotrophs and heterocystous cyanobacteria in two temperate estuaries

Abstract

Nitrogen (N) fixation is fueling planktonic production in a multitude of aquatic environments. In meso- and poly-haline estuaries, however, the contribution of N by pelagic N₂ fixation is believed to be insignificant due to the high input of N from land and the presumed absence of active N₂-fixing organisms. Here we report N₂ fixation rates, nifH gene composition and nifH gene transcript abundance for key diazotrophic groups over 1 year in two contrasting, temperate, estuarine systems: Roskilde Fjord (RF) and the Great Belt (GB) strait. Annual pelagic N₂ fixation rates averaged 17 and 61 mmol N m(-2) per year at the two sites, respectively. In RF, N₂ fixation was mainly accompanied by transcripts related to heterotrophic (for example, Pseudomonas sp.) and photoheterotrophic bacteria (for example, unicellular diazotrophic cyanobacteria group A). In the GB, the first of two N₂ fixation peaks coincided with a similar nifH-expressing community as in RF, whereas the second peak was synchronous with increased nifH expression by an array of diazotrophs, including heterotrophic organisms as well as the heterocystous cyanobacterium Anabaena. Thus, we show for the first time that significant planktonic N₂ fixation takes place in mesohaline, temperate estuaries and that the importance of heterotrophic, photoheterotrophic and photosynthetic diazotrophs is clearly variable in space and time.

Figure 1

Map depicting the sampling locations. One sampling site was located in the shallow RF estuary and the other in the GB strait.

Figure 2

Seasonal changes in environmental data collected in RF (left panels) and the GB (right panels). Concentrations of inorganic nutrients and DOC for RF (a) and GB (b). (c) and (d) show concentrations of Chl a, bacterial production and bacterial abundance in RF and in GB, respectively. Note the difference in bacterial abundance units between stations. Error bars represent standard deviations.

Figure 3

N₂ fixation measured throughout the year in RF (a) and the GB (b). Total N₂ fixation in diurnal light cycle treatments is depicted as open circles, whereas closed circles represent the equivalent fixation in bottles incubated in complete darkness. N₂ fixation data from the two different size fractions (>10 μm or <10 μm) are depicted as bars where available. Error bars indicate standard deviations.

Figure 4

Neighbor-joining tree of representative sequences from the 84 most abundant OTUs (each representing ⩾250 sequences) resulting from the 97% nucleotide sequence similarity clustering. Each of the OTUs are designated a number chosen by Mothur. Bootstrap values ⩾50% are represented by size-proportional gray circles in the tree. The heatmap indicates the relative abundance of sequences associated with a given OTU in each of the following types of samples: GB, RNA, <10 μm (A), GB, RNA, >10 μm (B), GB, DNA, <10 μm (C), GB, DNA, >10 μm (D), RF, RNA, <10 μm (E), RF, RNA, >10 μm (F), RF, DNA, <10 μm (G), RF, DNA, >10 μm (H). The affiliations of the OTUs with the canonical nifH clusters (Chien and Zinder, 1996) is indicated by roman numerals.

Figure 5

The relative abundances of OTUs identified to be part of the core nifH-expressing microbiome over time in RF (a) and in the GB (b). The OTU-names are given to the left of the heatmap and the closest cultivated relative is indicated to the right of the heatmap. For reference, the volumetric N₂ fixation rates showed in Figure 3 are given above the heatmaps. These rates are split into size fractions where available.

Figure 6

Results from the reverse-transcription quantitative PCR (RT-qPCR) and quantitative PCR (qPCR) assays performed on RNA and DNA samples taken from February to November in RF and from March to November in the GB. RT-qPCR results from RF (a) and the GB (b). qPCR results from RF (c) and the GB (d). The quantification limit (QL) is defined as the number of gene copies and transcripts needed in situ to match the lowest standard; for example, 10 target transcripts in 4 ng of RNA. QL was calculated for all samples and is depicted as a solid black line. OTUs that were detectable, but not quantifiable, are depicted as bars without error bars at the QL line. Error bars indicate standard deviations.