Transcriptional responses of Trichodesmium to natural inverse gradients of Fe and P availability

Abstract

The filamentous diazotrophic cyanobacterium Trichodesmium is responsible for a significant fraction of marine di-nitrogen (N₂) fixation. Growth and distribution of Trichodesmium and other diazotrophs in the vast oligotrophic subtropical gyres is influenced by iron (Fe) and phosphorus (P) availability, while reciprocally influencing the biogeochemistry of these nutrients. Here we use observations across natural inverse gradients in Fe and P in the North Atlantic subtropical gyre (NASG) to demonstrate how Trichodesmium acclimates in situ to resource availability. Transcriptomic analysis identified progressive upregulation of known iron-stress biomarker genes with decreasing Fe availability, and progressive upregulation of genes involved in the acquisition of diverse P sources with decreasing P availability, while genes involved in N₂ fixation were upregulated at the intersection under moderate Fe and P availability. Enhanced N₂ fixation within the Fe and P co-stressed transition region was also associated with a distinct, consistent metabolic profile, including the expression of alternative photosynthetic pathways that potentially facilitate ATP generation required for N₂ fixation with reduced net oxygen production. The observed response of Trichodesmium to availability of both Fe and P supports suggestions that these biogeochemically significant organisms employ unique molecular, and thus physiological responses as adaptations to specifically exploit the Fe and P co-limited niche they construct.

Fig. 1. Transect and surface nutrient profiles.

a Sampling stations 1–7 for the RRS James Cook-JC150 (GEOTRACES GApr08) cruise across the subtropical North Atlantic Ocean (June–August 2017 at ~22 ^oN) from Guadeloupe (French Caribbean) to Tenerife (Canary Islands). Background colour shows the satellite-derived chlorophyll-a concentration (mg/m³) from July 2017 [100]. Surface nutrient concentrations (nM) for (b) dFe, (c) DIP and (d) DOP (in grey) including measurements from the seven sampling stations (in black). dFe is a sub-set of data from Kunde et al. [38].

Fig. 2. Diazotroph abundances and measured bulk N₂ fixation rates.

a Relative abundance of the nifH amplicon for Trichodesmium sp. (T. thiebautii, T. spiralis), UCYN-A, γ-Proteobacterium 24774A11 and other non-identified nifH sequences. b In situ bulk measured N₂ fixation rates (nM per day). c Trichodesmium nifH gene abundance (genes L⁻¹) quantified by qPCR.

Fig. 3. Trichodesmium metatranscriptome.

a Principal component analysis (PCA) of all the 4044 Trichodesmium OGs across stations from all 17 samples calculated from the variance stabilising transformation (VST) using DESEq2 [68]. Colour gradient reflects measured surface dFe concentrations (0.3–1 nM). Numbers represent the sample stations. b Transcript expression patterns of the 1057 most differentially expressed OGs (DEOGs) across stations. Columns are individual samples clustered based on basis of Euclidean distance annotated with station number. Colour scale indicates high (red) to low (blue) DESEq2 VST OG normalised abundance scaled per row. Four clusters of expression patterns (labelled I–IV) were obtained from k-mean optimal clustering by Euclidean distance. Corresponding nutrient concentrations are shown above heatmap (nM).

Fig. 4. Trichodesmium nutrient stress.

Correlation matrix of normalised counts from selected gene OG associated with Fe (blue) and P (red) metabolism: fructose-bisphosphate aldolase FBA class-I, isiA, Heme oxygenase, idiA, ferritin, fur, isiB, pstS/sphX, pstAC, phoX, phnD, phnE, phnI, surE, ptxB, ptxC; against measured dFe, DOP and DIP concentrations (nM). Pearson correlation coefficient r value indicated inside boxes (n = 17). Colour scale indicates significant positive correlation (violet), negative correlation (green) or non-significant (white) (Pearson parametric correlation test, p value < 0.05).

Fig. 5. Co-regulation of nitrogen fixation metabolism.

a Pearson correlation matrix for the normalised counts of the 17 samples for the nitrogenase genes against selected genes involved in fixed nitrogen uptake (nitrite/nitrate assimilation and transport), PSI, PSII, electron flow pathways, and vitamin B₁₂ metabolism. Colour scale indicates r value with crosses on non-significant correlations (Pearson test, p value > 0.05). b Averaged DESEq2 normalised counts for nifH across stations. Ratios of the normalised counts per station for (c) nifH:PSI (psaA), and (d) isiA:PSI (psaA).

Fig. 6. Transcription patterns related to resource availability in situ.

a Transcription patterns (Fig. 3a) mapped into resource space. Dashed lines are ±2 standard deviations above the average nutrient concentration of dFe at stations 6–7 and DOP at stations 1–2. Colour scale indicates averaged normalised z-score across stations from the OGs for each of clusters I–IV (see Fig. 3). Symbol sizes indicate the averaged expression for the structural genes of nitrogenase nifHDK. b Gene examples found on each expression pattern plotted on (a).