Effects and mechanisms of glyphosate as phosphorus nutrient on element stoichiometry and metabolism in the diatom Phaeodactylum tricornutum

Abstract

The ability to utilize dissolved organic phosphorus (DOP) gives phytoplankton competitive advantages in P-limited environments. Our previous research indicates that the diatom Phaeodactylum tricornutum could grow on glyphosate, a DOP with carbon-phosphorus (C-P) bond and an herbicide, as sole P source. However, direct evidence and mechanism of glyphosate utilization are still lacking. In this study, using physiological and isotopic analysis, combined with transcriptomic profiling, we demonstrated the uptake of glyphosate by P. tricornutum and revealed the candidate responsible genes. Our data showed a low efficiency of glyphosate utilization by P. tricornutum, suggesting that glyphosate utilization costs energy and that the alga possessed an herbicide-resistant type of 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase. Compared to the P-limited cultures, the glyphosate-grown P. tricornutum cells up-regulated genes involved in DNA replication, cell growth, transcription, translation, carbon metabolism, and many genes encoding antioxidants. Additionally, cellular C and silicon (Si) increased remarkably while cellular nitrogen (N) declined in the glyphosate-grown P. tricornutum, leading to higher Si:C and Si:N ratios, which corresponded to the up-regulation of genes involved in the C metabolism and Si uptake and the down-regulation of those encoding N uptake. This has the potential to enhance C and Si export to the deep sea when P is limited but phosphonate is available. In sum, our study documented how P. tricornutum could utilize the herbicide glyphosate as P nutrient and how glyphosate utilization may affect the element content and stoichiometry in this diatom, which have important ecological implications in the future ocean.IMPORTANCEGlyphosate is the most widely used herbicide in the world and could be utilized as phosphorus (P) source by some bacteria. Our study first revealed that glyphosate could be transported into Phaeodactylum tricornutum cells for utilization and identified putative genes responsible for glyphosate uptake. This uncovers an alternative strategy of phytoplankton to cope with P deficiency considering phosphonate accounts for about 25% of the total dissolved organic phosphorus (DOP) in the ocean. Additionally, accumulation of carbon (C) and silicon (Si), as well as elevation of Si:C ratio in P. tricornutum cells when grown on glyphosate indicates glyphosate as the source of P nutrient has the potential to result in more C and Si export into the deep ocean. This, along with the differential ability to utilize glyphosate among different species, glyphosate supply in dissolved inorganic phosphorus (DIP)-depleted ecosystems may cause changes in phytoplankton community structure. These insights have implications in evaluating the effects of human activities (use of Roundup) and climate change (potentially reducing DIP supply in sunlit layer) on phytoplankton in the future ocean.

Keywords: Phaeodactylum tricornutum; element stoicheometry; glyphostae; molecular mechanism; phosphorus nutrient.

Fig 1

Responses of P. tricornutum to different P nutrient environments. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. (A) Growth curves. Arrow indicates the time for cell harvesting for RNA isolation and RNA-seq; (B) ¹³C signal and cell concentrations in P. tricornutum cells in ¹³C-Glyphosate cultures compared to that in P- cultures; (C) AP activity. * indicates P < 0.05 in t-test between Gly and P- cultures.

Fig 2

Elemental contents and stoichiometry of P. tricornutum in response to different P nutrient environments. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. (A) Cellular POC content and cell concentrations. (B) Cellular PON content and cell concentrations. (C) Cellular PP content and cell concentrations. (D) Cellular Si content and cell concentrations. (E) Cellular Si:C ratio and cell concentrations. (F) Cellular Si:N ratio and cell concentrations. * indicates P < 0.05 while ** indicates P < 0.001 in t-test in comparison of P- vs. f/2 or Gly vs. f/2.

Fig 3

Transcriptomic profiles of P. tricornutum under different P nutrient conditions. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. (A) Principal component analysis using gene expression levels (normalized as FPKM) in each sample. (B) Heatmap with FPKM (normalized with z-score) of all the screened DEGs identified at a cutoff of |log₂ (fold change)| ≥1 and P adjust (Q-value) ≤0.001 from the three comparisons (P- vs. f/2, Gly vs. f/2, and Gly vs. P-). To facilitate proper presentation of the comparative information across all conditions, FPKM were normalized with z-score by row, and the positive z-score in the color bar indicates the up-regulation of the corresponding gene while negative z-score in the color bar indicates the down-regulation of that. The heatmap color strength in each row represents the FPKM (normalized with z-score), from blue (lowest), white to red (highest). (C) Venn diagram showing the number of shared and unique DEGs in different comparisons. (D) Histogram showing the number of up- and down-regulated DEGs in each comparison group.

Fig 4

Heatmap showing relative expression levels of DEGs in major metabolic pathways under different P nutrient conditions. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. DEGs in this heatmap were classified according to the functional annotation with KEGG database. Membrane transport here includes all identified ABC transporters. Data shown are FPKM normalized with z-score by row, and the positive z-score in the color bar indicates the up-regulation of the corresponding gene while negative z-score in the color bar indicates the down-regulation of that. The heatmap color strength in each row represents the FPKM (normalized with z-score), from yellow (lowest), red to black (highest).

Fig 5

Expression and classification of genes specifically responding to glyphosate. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. DEGs in these heatmaps were classified according to the annotation with major functional databases (KEGG, GO, and NR), and if this step did not yield a match, based on InterPro protein domain or family prediction, these DEGs were then mapped to the KEGG pathway database to further identify the metabolic pathway involved. Data shown are FPKM normalized with z-score by row, and the positive z-score in the color bar indicates the up-regulation of the corresponding gene while negative z-score in the color bar indicates the down-regulation of that. The heatmap color strength in each row represents the FPKM (normalized with z-score), from blue (lowest), yellow to red (highest).

Fig 6

Alignment and the conserved amino acid residues among phnC homologs in P. tricornutum, E. huxleyi, and Nostoc sp. The consensus sequences are indicated under each conserved domain. The sequences of Walker A motif (G-GGKST) is nucleotide-binding fold. Both ATP-binding cassette transporter (ABC transporter) signature motif (LSGGQ) and Walker B motif (hhhhDE: four aliphatic residues followed by two negatively charged residues) are thought to be responsible for ATP hydrolysis in the phnC superfamily.

Fig 7

Expression dynamics of metabolic pathways and some specific genes in P. tricornutum under different P nutrient conditions. f/2, control (f/2 + Si, P-replete); P-, P-limited; Gly, glyphosate as the sole P source replacing Pi in f/2. Text in rectangles indicates metabolic pathways or proteins, and heatmaps under them indicate expression levels of the involved genes (sum of FPKM of all genes identified in the corresponding pathway or protein) in the three groups (f/2, P-, and Gly). Genes identified in different metabolic pathways and proteins were performed via KEGG, GO, and NR databases. FPKM was normalized with z-score for each metabolic pathway or protein, and the positive z-score in the color bar indicates the up-regulation of the corresponding pathway or protein while negative z-score in the color bar indicates the down-regulation of those. The heatmap color strength in each row represents the FPKM (normalized with z-score), from blue (lowest), white to red (highest).