Feeding in mixoplankton enhances phototrophy increasing bloom-induced pH changes with ocean acidification

Abstract Plankton phototrophy consumes CO2, increasing seawater pH, while heterotrophy does the converse. Elevation of pH (>8.5) during coastal blooms becomes increasingly deleterious for plankton. Mixoplankton, which can be important bloom-formers, engage in both photoautotrophy and phagoheterotrophy; in theory, this activity could create a relatively stable pH environment for plankton growth. Using a systems biology modelling approach, we explored whether different mixoplankton functional groups could modulate the environmental pH compared to the extreme activities of phototrophic phytoplankton and heterotrophic zooplankton. Activities by most mixoplankton groups do not stabilize seawater pH. Through access to additional nutrient streams from internal recycling with phagotrophy, mixoplankton phototrophy is enhanced, elevating pH; this is especially so for constitutive and plastidic specialist non-constitutive mixoplankton. Mixoplankton blooms can exceed the size of phytoplankton blooms; the synergisms of mixoplankton physiology, accessing nutrition via phagotrophy as well as from inorganic sources, enhance or augment primary production rather than depressing it. Ocean acidification will thus enable larger coastal mixoplankton blooms to form before basification becomes detrimental. The dynamics of such bloom developments will depend on whether the mixoplankton are consuming heterotrophs and/or phototrophs and how the plankton community succession evolves.

. Diagrammatic representation of the model, configured for a constitutive mixoplankton. Interconnectivity is shown between the state variables for structural C ( C C), metabolite C ( M C), N and P biomass, C-biomass allocated to the photosystem (chloroplast; PM C), and chlorophyll (Chl). This is shown specifically for a constitutive mixoplankton (CM). Forms of dissolved organic matter (DOM) are shown as dissolved organic C (DOC) and dissolved free amino acids (DFAA). Inorganic nutrients include dissolved inorganic P (DIP), ammonium and nitrate. Total cell-C, T C, = C C+ M C+ PM C. The C-status is related to M C: T C, Nstatus to N: T C and P-status to P: T C. Pool M C equates to the 'metabolic pool' in Fig. 2.
Depending on the application, additional state variables are included: • protCells (cells m -3 ): cells, required for a dynamic description of cell-size with nutrient status, diel light cycle and temperature • protSi (mgSi m -3 ); organism-Si, required for diatoms • protANA (mgNA m -3 ); acquired nucleic acid material from phototrophic prey, required to support acquired phototrophy in plastidic specialist non-constitutive mixoplankton (pSNCM) NOTE: In the functional equations below, for brevity, state variable names in the equations do not include the prefix 'prot'.

Growth and nutrient status
Ultimately growth is a function of the nutritional status of the organism (in terms of elements C, N and P) and the maximum growth rate potential. The latter varies with temperature, T, around the value of µmax at a reference temperature, µmaxRT. growth = f{C-status, N-status, P-status, µmax, losses} µmax = f{µmaxRT, T} The nutrient status defines the health of the organism in terms of C ( M C: T C), N (N: T C) and P (P: T C), and is a function of various inputs and outputs. Inputs are associated with the use of dissolved organic substrates via osmotrophy, prey via phagotrophy, and also the use of inorganics via phototrophy. Losses occur through respiration and regeneration, and also through the leakage of metabolites as dissolved organic matter (DOM), some of which may be recovered via osmotrophy.

Losses = f{C-respiration, N-regeneration, P-regeneration, DOM-leak}
Growth is associated with catabolic (including basal) and anabolic respiration, part of which is associated with specific dynamic action (SDA) during prey digestion and assimilation. Anabolic respiration is affected by the flows of resources via the different trophic mechanisms. Nitrate assimilation incurs an additional cost for reduction of nitrate to nitrite to ammonium. There are also losses of C, N, P required to preserve organism stoichiometry within the bounds of acceptable C:N:P.
Cell division occurs when the cell reaches a critical size (which varies with nutrient status and temperature affecting the growth rate), and typically occurs in phototrophs within a specific part of the diel light:dark (LD) cycle. division = f{size, critical size, LD} critical size = f{T, C-status, N-status, P-status, growth} The size of the organism affects predation for phagotrophy, and whether it itself is likely to encounter its own predator.

Osmotrophy
Osmotrophy depends on the concentration of the substrate, [DOM], the C:N:P status of that material, and the uptake kinetics parameters of the maximum uptake rate ( DOM Vmax) and the substrate affinity (i.e., DOM Vmax/KDOM). The uptake kinetics depend on the nutrient status of the organism; cells that are nutrientstressed have a higher uptake potential and a high affinity. Against the gains from osmotrophy there are losses with the leakage of DOM. At especially high growth rates, which require a high nutrient status and hence a cell that is replete with the internal metabolite pool containing mM concentrations, DOM inevitably leaks. Osmotrophy may recover some of that leakage. The net leakage of N-containing DOM (as amino acids) is most significant during N-replete growth conditions, while leakage of DOC (sugars) occurs especially with high rates of phototrophy, including when N becomes exhausted and the cell has yet to down-regulate photosynthesis.

Phagotrophy and voiding of waste
Phagotrophy brings in resources from the assimilation of prey biomass; note the plural in prey-assimilations in the equations. Prey need to be encountered (which depends on the sizes of the predator organism and of the prey, their respective motilities and turbulence), captured (which, like predator motility, varies with satiation, and also with the 'taste' of the prey as affected by its stoichiometric quality), and then ingested. These processes are prey-species specific; the collective biomass from many ingestions, perhaps of different prey organisms, is then digested. During digestion a fraction of the ingested prey is subjected to voiding (depending on the assimilation efficiency, AE, predator satiation and the food quality), and another fraction is lost associated with specific dynamic action (SDA) as the prey biomass is subjected to catabolic and then anabolic processes. The internal recycling of regenerated inorganic nutrients is a critical step in mixoplankton ( Fig. 2

Phototrophy
Photosynthesis depends on light, the availability of dissolved organic C (DIC, especially as CO2 and HCO3 -), photopigment content (Chl:C), the value of alpha governing the initial slope of the light-photosynthesis curve, and the maximum rate of C-fixation (Pmax). The value of Pmax is set by the size of PM C: T C. For organisms with a constitutive ability to photosynthesise, both Chl:C and Pmax are modulated by the demand for C and energy, reflected by the organisms' nutritional status and growth rate potential. For non-constitutive mixotrophs (NCM), phototrophy is acquired from captured phototrophic prey. Light is a function of the photon flux density at the water surface and of attenuation within the water (which varies with the biomass of the pigmented organisms). Light also varies over the diel light:dark cycle; this imparts a diel cycle on phototrophy that then feeds through to affect osmotrophy and phagotrophy via feedback processes.

Inorganic nutrient assimilations
Inorganic nutrients are sourced both internally, as regenerative products of prey assimilation, and externally; use of the former takes priority and will be affected by prey C:N:P. The uptake of DIN is affected also by the P-status of the organism. The uptake kinetics for ammonium (NH4) provide for development of an enhanced uptake capability over that for nitrate (NO3), with that development also commencing at a higher N-status. The latter results in ammonium being taken up 'in preference' to nitrate. There is no 'inhibition' term controlling NO3-assimilation by [NH4]; if the supply of ammonium from internal recycling plus external sources cannot meet the demand, then the ability to use nitrate is derepressed.

Controlling physiological processes
The model uses the normalised value of the nutrient quotas (i.e., 0 at minimum quota, 1 at optimal quota) for C, N and P (see Fig. S2) to control physiological processes. In essence these controls replicate the biochemical events of (de)repression and, consistent with allosteric controls, exploit sigmoidal curve functions. For example, the N:C quota value is used to controls the use of ammonium and nitrate (see Flynn 2001).
The value of RelMC (defined as M C : T C; see Fig. S1) was used to modulate the (de)repression of phototrophy and phagotrophy in the work described here.
The curve form for CCu is: For the curve shown in Fig. S3, CCuH=4, CCuK=0.2. The model is not sensitive to the value of these parameters.
The generic curve form for the resource acquisition controls is: For the PS and Pred curves shown in Fig.S2, H=8 and K=0.4. For the more repressed curves ('phag>phot', or 'phot>phag'), H=8 and K=0.8.

Fig. S2.
Control of growth, resources acquisition via phototrophy and phagotrophy through reference to the value of RelMC. The relative availability of C for growth is defined by CCu, increasing as the relative size of the metabolite pool (RelMC) increases. Curves describing the (de)repression of phototrophy and phagotrophy are shown for when they are similar (PS and Pred respectively), and also for when either of these processes is de-repressed when the cell is of a lower C-status (i.e. RelMC is lower). Shifting these curves provides a mechanism to alter the balance of phagotrophy vs phototrophy ('phag>phot', or 'phot>phag'). Note that the cross-over point between CCu and the resource acquisition trophic control curves affects the potential maximum growth rate attainable when solely exploiting that trophic mode. For example, when phagotrophy provides a backup mechanism for when phototrophy cannot supply sufficient C to maintain a high enough RelMC ('Pred; photo>phag'), this aligns with a growth rate maximum exploiting phagotrophy of ca. 50% (Response ≈0.5).
For phagotrophy (which brings in a complete package of C,N,P) the above described control dominates regulation of the activity. However, phototrophy is also controlled by the acquisition of N (as ammonium and nitrate) and P (as phosphate); those interactions are modulated by N:C and P:C respectively. Phototrophy is further affected by synthesis and operation of the photosystems, which are modulated via RelMC and N:C.  Modelling the carbonate chemistry system, alkalinity and pH The following describes the ocean acidification sub-model, including the carbonate chemistry system and changes in pH and alkalinity with biological activity. This is the sub-model used by Flynn et al. (2012Flynn et al. ( , 2015Flynn et al. ( , 2016. The model was operated with an Euler integration routine (step size 0.015625 d). As described here, it was configured to enable an exploration of the effects of changing different abiotic parameters and also of the balance of phototrophic vs heterotrophic events.
The model includes traps to halt the simulation if certain conditions are met, notably if a state variable goes negative, or if the operator has requested that the model pauses at regular times to enable changes in operational parameters. Changes in nutrients affect alkalinity and the size of the carbonate system which provides the greater proportion of seawater pH buffering. Here those biotic-linked changes are controlled by the constant deltaC and C:N:P ratios etc.
Calculations of the sea water chemistry operate using units of µmol kg -1 , requiring various transforms and corrections with salinity and temperature from the units used to describe the biological functionality. Gas exchange of CO2 (if enabled) moves DIC in and out of the system at the water surface as functions of wind speed, atmospheric pCO2, salinity and temperature. Calculation of the different components of the carbonate system responsible for most of the seawater buffering are also related to salinity and temperature as well as the current acidity. Calculation of omega (dissolution of) calcite, are included here for completeness.
Changes in total alkalinity occur with physiology. There are decreases (i.e. TA_out is +ve) with ammonium use and with calcification, and increases (i.e. TA_out is -ve) with nitrate and phosphate use. TA_out is effectively TA uptake into organisms. For this implementation, rates of biological action are just set using the constant deltaC. The calculation of acidity and total alkalinity is an iterative process as changes in each affect the other via dissociations of the key ionic compounds.   (k1*k2/(H^2+H*k1+k1*k2))*DICm µmol kg -1 current CO3 -concentration dcf (999.842594 + 67.939520e-3*T-9.095290e-3*T^2 + 100.168500e-6*T^3 -1.120083e-6*T^4 + 6.536332e-9*T^5 + a_1*S + 'b -2'*S^1.5 + c*S^2)/1.0e3 correction factor for water density at given T and S; to convert µmol/Kg to µmol/L multiply by dcf    The consequences of phototrophy vs heterotrophy on seawater acidity starting at equilibria with pCO2 = 300 atm or 600 atm, are demonstration in Fig. S3. This shows the consequence of phototrophy or heterotrophy for seawater acidity (i.e., for changes in [H + ]) and provides an insight of the changes in acidity that develop during protist plankton growth especially as the different protist functional groups release or take up different inorganic nutrients that affect alkalinity and thence pH.
Positive values of δ (i.e., an increase in δ C for biomass in Fig. S3) shows the impact of phototrophy with Cfixation removing CO2 from the seawater with or without concurrent assimilation of dissolved inorganic P (DIP) and dissolved inorganic N (DIN, as ammonium or as nitrate). Differences between these plots show how the nutrient assimilations affect alkalinity and thus further modify the impact of the C-fixation upon acidity. Thus, the use of ammonium (the removal of which decreases alkalinity) increases acidity and partially counters the basification caused by CO2 removal with C-fixation. Phototrophy using nitrate increases basification. Heterotrophic processes, as occur with phagotrophy, are associated with respiration (i.e., a negative δ C for biomass in Fig. S3); this adds CO2 to the seawater causing acidification. With that respiration there is also a regeneration of DIP and DIN (as ammonium only). The addition of ammonium raises alkalinity, partly compensating for the acidification caused by CO2 release.
Changes in [H + ] in the 600 atm pCO2 scenario, from an initial value almost double that in the 300 atm pCO2 scenario, are greater because of the lower pH buffering capacity at the initial pCO2 equilibrium state.

Fig. S3 Changes in [H + ] with phototrophy and heterotrophy. Decreases in [H +
] occur during Cfixation, associated with the production (δC >0) of organics and the concurrent inorganic nutrient assimilation. Increases in [H + ] occur during heterotrophy, associated with the loss (δ gC <0) of organics, and hence regeneration of CO2 and inorganic nutrients. The equilibrium state (δ gC =0), in water of salinity 35 and at 15°C, was with pCO2 of 300 atm (p300, blue lines) or 600 atm (p600, red lines), but thereafter excludes air-sea gas exchange. The different line types show changes due to just the assimilation or regeneration of CO2 (solid lines), or additionally coupled with Redfield C:N:P assimilation or regenerations of phosphate and dissolved inorganic N. The N was either considered as ammonium ('NH4'; dotted lines), or for assimilation only as nitrate ('NO3'; dashed lines). For reference, the pH values are indicated at the equilibrium states and the minimum and maximum [H + ] values on the y-axis. Note the y-axis is a log scale.  (Fig. 3) increased during the early part of the simulation but was then taken up by the GNCM bringing DOC down to near zero. See Figure  legends for further details.  Simulations were run in closed systems originally equilibrated to pCO2 of 300 (blue) or 600 atm (red). During each day, the first 0.7d was illuminated, hence the sinusoidal form of many of the proximal values. The two parts of the split y-axis for acidity are of equal value (i.e., 1 nmol kg -1 ). Acidity increases with net heterotrophy, and decreases with net phototrophy. The greater difference between proximal and bulk acidities for the p600 series vs the p300 series is because of the lower buffering capacity in the seawater for the former scenario; the biological events were unaffected by acidity in these simulations (see also Fig. S3).

Balance of Phototrophic vs Phagotrophic CO2 Dynamics
The following describes an estimation of the balance point between phagotrophy and phototrophy to achieve a net zero CO2 exchange in a mixoplankton.

PhotC = Umax-Ass
This C input from phototrophy is associated with the assimilation of ammonium described by PhotC*NC (gN gC -1 d -1 ), which costs the amount of C described by AR. The anabolic respiration rate with phototrophy (PhotR; gC gC -1 d -1 ) is: PhotR = PhotC*NC*AR The total C fixed and retained by the cell is PhotC+PhotR. However, we then need to add the C that is fixed but leaks out as DOC. The total C-fixation rate (GrossPS; gC gC -1 d -1 ) is thus: The net CO2 fixation with phototrophy (NetPS) is: The balance of zero net CO2 exchange is given when respiration from phagotrophy balances the net photosynthetic rate, i.e., when PhagR = NetPS.
With the above tabulated values, this balance is achieved with the input and output values given in the From this, the ratio of total C inputs from phagotrophy to phototrophy is (Ing:GrossPS = 0.9:0.3 = ) 3.
Using a value of DOCfrac=0, this ratio is 2.91