Genome-scale metabolic model of the diatom Thalassiosira pseudonana highlights the importance of nitrogen and sulfur metabolism in redox balance

Abstract

Diatoms are unicellular photosynthetic algae known to secrete organic matter that fuels secondary production in the ocean, though our knowledge of how their physiology impacts the composition of dissolved organic matter remains limited. Like all photosynthetic organisms, their use of light for energy and reducing power creates the challenge of avoiding cellular damage. To better understand the interplay between redox balance and organic matter secretion, we reconstructed a genome-scale metabolic model of Thalassiosira pseudonana strain CCMP 1335, a model for diatom molecular biology and physiology, with a 60-year history of studies. The model simulates the metabolic activities of 1,432 genes via a network of 2,792 metabolites produced through 6,079 reactions distributed across six subcellular compartments. Growth was simulated under different steady-state light conditions (5-200 μmol photons m-2 s-1) and in a batch culture progressing from exponential growth to nitrate-limitation and nitrogen-starvation. We used the model to examine the dissipation of reductants generated through light-dependent processes and found that when available, nitrate assimilation is an important means of dissipating reductants in the plastid; under nitrate-limiting conditions, sulfate assimilation plays a similar role. The use of either nitrate or sulfate uptake to balance redox reactions leads to the secretion of distinct organic nitrogen and sulfur compounds. Such compounds can be accessed by bacteria in the surface ocean. The model of the diatom Thalassiosira pseudonana provides a mechanistic explanation for the production of ecologically and climatologically relevant compounds that may serve as the basis for intricate, cross-kingdom microbial networks. Diatom metabolism has an important influence on global biogeochemistry; metabolic models of marine microorganisms link genes to ecosystems and may be key to integrating molecular data with models of ocean biogeochemistry.

Fig 1. Diagram illustrating the principle reactions involved in generating ATP and balancing the ATP/NADPH ratio in diatom plastids.

Black shapes: enzymatic complexes, gray bars: plastidial or mitochondrial membranes, dashed lines: grouping of plastidial and mitochondrial reactions, green: reactions that produce NAD(P)H equivalents, red: reactions that consume NAD(P)H equivalents, blue: ATP producing and consuming reactions. Abbreviations: L-glutamate (glu__L), L-glutamine (gln__L), 2-phosphoglycolate (2pglyc), glycolate (glyclt), glyoxylate (glx), glycine (gly), L-serine (ser__L), L-malate (mal__L), fumarate (fum), succinate (succ), dehydroascborbate (dhascb), L-ascorbate (ascb__L), oxidized glutathione disulfide (gthox), reduced glutathione (gthrd), reduced ferredoxin (Fd), reduced thioredoxin (Trx), plastoquinone (PQ), plastoquinol (PQH₂), NADH ubiquinone oxidoreductase (NADHOR), ATP synthase (ATPS). Number symbols indicate reaction references in the text including: cyclic electron flow (CEF, ❶), nitrate assimilation (❷), sulfate assimilation (❸), ribulose-1,5-bisphosphate oxygenase (RUBISO, ❹), the Mehler reaction (❺), energetic coupling between the plastid and the mitochondria (❻), plastid terminal oxidase (PTOX, ➐), alternative oxidase (AOX, ❽), cytochrome c oxidase (CYOO, ❾).

Fig 2. Biomass composition and detailed pigment composition of Thalassiosira pseudonana acclimated to three different light levels (5, 60, 200 μmol photons m^-2 s^-1).

The radius of each circle is scaled to the total cellular dry weight. To demonstrate the effect of pigment composition (highlighted in green) on photon absorption at different wavelengths, photon absorption was plotted for each acclimated cell during illumination at 100 μmol photons m^-2 s^-1 under a cool white fluorescent bulb. The contribution of each pigment is integrated over each 20 nm of the light spectrum.

Fig 3. Parsimonious Flux Balance Analysis (pFBA) predictions (in mmol e^- gDW^-1 h^-1) for the contribution of different reactions to the dissipation of reductants generated by photosynthesis and organic carbon utilization in iTps1432 across three chemostats maintained at 5, 60, and 200 μmol photons m^-2 s^-1.

Differences in biomass composition and biosynthesis of metabolites impact the contribution of C, N, and S assimilation reactions (grey, blue, yellow, respectively). Respiration reactions are distributed across several different organelles including the plastid, peroxisome, and mitochondria (green, purple, red, respectively). Flux predictions with (a) baseline constraints on respiration, (b) photorespiration included, (c) photorespiration and energetic coupling included.

Fig 4. Target biomass composition and comparison between simulated and experimental biomass elemental ratios for the transition of iTps1432 from exponential to stationary phase under nitrogen-limited conditions.

(a) Biomass composition was measured 0, 1, 3, 7, and 10 days after transfer to fresh media with low nitrate and used to calculate the biomass objective function at each time point. (b) C:N, N:P, and C:P ratios were measured 0, 1, 2, 3, 5, 7, and 10 days after transfer and compared with simulated biomass elemental composition. See S4 Fig for plots of error rates between predictions and experimental measurements.

Fig 5. Comparison between simulated and experimental (solid, dashed lines) media nutrient concentration and biomass concentration for the transition of iTps1432 from exponential to stationary phase under nitrogen-limited conditions.

(a) NO₃ (+ NO₂), NH₄, Si(OH)₄, and PO₄ were measured 0, 1, 3, 7, and 10 days after transfer to fresh media with low nitrate and compared to simulated media nutrient concentrations (b) Cell concentration (cells mL^-1) was measured each day and converted to biomass concentration (g L^-1) where measurements of cell dry weight were available and compared with the simulated biomass concentration. 12: 12 h light: dark cycles are depicted with white and grey stripes.

Fig 6. Simulated respiratory flux predictions (in mmol O₂ gDW^-1 h^-1) and the constraints affecting respiration for the transition of iTps1432 from exponential to stationary phase under nitrogen-limited conditions.

(a) Flux through the PSII reaction is driven by photon absorption and is constrained by measurements taken 0, 1, 3, 7, and 10 days after transfer to fresh media with low nitrate (green, mmol O₂ mg chl a^-1 h^-1, [25], S3 Dataset). The 95% confidence intervals were converted from mmol O₂ mg chl a^-1 h^-1 to mmol O₂ gDW^-1 h^-1 using simulated mg chl a / gDW and set as lower and upper bounds of the reaction. Steady-state N-replete ± 95% confidence interval (black) measurement of net oxygen production [24] was plotted at the 0 h time point. Simulation results are plotted here for each 3 h increment in mmol O₂ gDW^-1 h^-1 (black). (b) Plot of simulated new biomass O: C molar ratio during the light (black) and dark (grey) period. (c) The photosynthetic quotient was calculated from the simulated new biomass elemental composition during the light period (black) and also contributes to the respiratory flux results. Oxygen and inorganic carbon assimilation were left unconstrained during the dark period (grey). (d) Respiratory flux predictions and contribution of different respiration reactions (plastid: green, mitochondria: red, peroxisome: purple). 12: 12 h light: dark cycles are depicted with white and grey stripes.

Fig 7. Simulated metabolite excretion by iTps1432 during the transition from exponential to stationary phase under nitrogen-limited conditions across a range of possible photosynthetic quotient constraints.

Note that y-axes are scaled to each metabolite and metabolites are listed in order of highest to lowest maximum concentration. Metabolites composed of organic carbon are labeled red, organic nitrogen are blue, organic sulfur are yellow, and inorganic are black.

Fig 8. Simulated N & S (blue, yellow) metabolic flux predictions (in mmol e^- gDW^-1 h^-1) during the transition of iTps1432 from exponential to stationary phase under nitrogen-limited conditions with a 102% photosynthetic quotient constraint.