Trichodesmium metaproteomes reflect the differential influence of resource availability across ocean regions

Abstract

The diazotroph Trichodesmium is an important contributor to marine dinitrogen fixation, supplying nitrogen to phytoplankton in typically nitrogen-limited ocean regions. Identifying how iron and phosphorus influence Trichodesmium activity and biogeography is an ongoing area of study, where predicting patterns of resource stress is complicated by the uncertain bioavailability of organically complexed iron and phosphorus. Here, a comparison of 26 metaproteomes from picked Trichodesmium colonies identified significantly different patterns between three ocean regions: the western tropical South Pacific, the western North Atlantic, and the North Pacific Subtropical Gyre. Trichodesmium KEGG submodule signals differed significantly across regions, and vector fitting showed that dissolved iron, dissolved inorganic phosphorus, and temperature significantly correlated with regional metaproteome patterns. Patterns of iron and phosphorus stress marker proteins previously validated in culture studies showed significant enrichment of a phosphorus stress signal in the western North Atlantic and an iron stress signal in the North Pacific. Populations in the western tropical South Pacific appeared to modulate their proteomes in response to both dissolved iron and dissolved inorganic phosphorus bioavailability, with significant enrichment of iron and phosphorus stress marker proteins, concomitant proteome restructuring, and significant decreases in the relative abundance of the dinitrogen fixation protein, NifH. These signals recapitulate established regional patterns of resource stress on phytoplankton communities released from nitrogen stress. Evaluating community stress patterns may therefore predict resource controls on diazotroph biogeography. These data highlight how Trichodesmium modulates its metabolism in the field and provide an opportunity to more accurately constrain controls on Trichodesmium biogeography and dinitrogen fixation.

Keywords: Trichodesmium; co-limitation; diazotrophy; metaproteomics.

Figure 1

Spatial distribution of Trichodesmium metaproteome sampling locations and unique proteins by region. (A) Map of Trichodesmium samples with shape indicating the ocean regions (western North Atlantic, western tropical South Pacific, and North Pacific subtropical gyre). (B) Venn diagram of 1729 Trichodesmium proteins that are either shared or uniquely detected among regions based on UniProt ID.

Figure 2

Differences in Trichodesmium protein spectral counts across regions. (A) Correspondence analysis of normalized spectral counts of all Trichodesmium proteins from all samples (PERMANOVA: P = 1e-4, F = 16.9). The explanatory variables included were determined to be significant based on vector analysis. DIP and dFe values are co-located with metaproteome samples except for western North Atlantic dFe values, which were calculated as an average of past cruise data (Table S1; see Rouco et al. [42]). (B) Heatmap of the percentage of KEGG-identified spectral counts by KEGG submodule (Table S7). Percentages are log-transformed. Samples were averaged across regions and KEGG submodule allocations were taken as a percentage of the total KEGG-identified protein spectral counts. White boxes signify that proteins from this submodule were not detected in this region. KEGG submodules in bold are examples that vary by region and are discussed in the text.

Figure 3

Stress signal enrichment for (A) P and (B) Fe stress signals displayed at the regional level. The P stress marker proteins included PhnM, PhnD, PhoX, SphX, SphA, and SqdB and the Fe stress proteins included IsiA, IdiA, Fld1, and FbaA (Table S4). Asterisks indicate significant differences between regions (P < .05) based on Kolmogorov–Smirnov testing.

Figure 4

Plots of average Fe stress marker signal versus average P stress marker signal by sample (Table S6), where gray-scale overlay indicates the spectral counts of the alkaline phosphatase proteins (A) PhoX and (B) PhoA and (C) the Fe-containing subunit of nitrogenase, NifH. Overlaid spectral counts were normalized by average spectral count across all samples, then divided by the maximum value of that protein across all samples. Sample region is differentiated by shape. Horizontal and vertical lines represent minimum values of the P and Fe stress marker signals in regions experiencing single stress of P or Fe, respectively.

Figure 5

Nutrient stress patterns for Trichodesmium alone and the whole chlorophyll-containing community across ocean regions. Trichodesmium metaproteome stress patterns are displayed (squares) and compared to community stress responses [1] at locations closest to the regions sampled here (circles). Nitrogen (N) stress is indicated in blue, phosphorus (P) stress in black, and iron (Fe) stress in red. Shapes with multiple colors indicate co-stress. The outer ring indicates a serial limiting response, where adding the serially limiting nutrient enhanced growth more than adding the primary limiting nutrient alone [1]. The color scale is surface ocean nitrate (μmol kg⁻¹) from the World Ocean Atlas, NOAA [118, 119].