Regulation of inorganic carbon acquisition in a red tide alga (Skeletonema costatum): the importance of phosphorus availability

Skeletonema costatum is a common bloom-forming diatom and encounters eutrophication and severe carbon dioxide (CO₂) limitation during red tides. However, little is known regarding the role of phosphorus (P) in modulating inorganic carbon acquisition in S. costatum, particularly under CO₂ limitation conditions. We cultured S. costatum under five phosphate levels (0.05, 0.25, 1, 4, 10 µmol L⁻¹) and then treated it with two CO₂ conditions (2.8 and 12.6 µmol L⁻¹) for 2 h. The lower CO₂ reduced net photosynthetic rate at lower phosphate levels (< 4 µmol L⁻¹) but did not affect it at higher phosphate levels (4 and 10 µmol L⁻¹). In contrast, the lower CO₂ induced a higher dark respiration rate at lower phosphate levels (0.05 and 0.25 µmol L⁻¹) and did not affect it at higher phosphate levels (> 1 µmol L⁻¹). The lower CO₂ did not change relative electron transport rate (rETR) at lower phosphate levels (0.05 and 0.25 µmol L⁻¹) and increased it at higher phosphate levels (> 1 µmol L⁻¹). Photosynthetic CO₂ affinity (1/K_0.5) increased with phosphate levels. The lower CO₂ did not affect photosynthetic CO₂ affinity at 0.05 µmol L⁻¹ phosphate but enhanced it at the other phosphate levels. Activity of extracellular carbonic anhydrase was dramatically induced by the lower CO₂ in phosphate-replete conditions (> 0.25 µmol L⁻¹) and the same pattern also occurred for redox activity of the plasma membrane. Direct bicarbonate (${\mathrm{HCO}}_{\mathrm{3}}^{-}$) use was induced when phosphate concentration was more than 1 µmol L⁻¹. These findings indicate P enrichment could enhance inorganic carbon acquisition and thus maintain the photosynthesis rate in S. costatum grown under CO₂-limiting conditions via increasing activity of extracellular carbonic anhydrase and facilitating direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ use. This study sheds light on how bloom-forming algae cope with carbon limitation during the development of red tides.

1 Introduction

Diatoms are unicellular photosynthetic microalgae that can be found worldwide in freshwater and oceans. Marine diatoms account for 75 % of the primary productivity for coastal and other nutrient-rich zones and approximately 20 % of global primary production (Field et al., 1998; Falkowski, 2012), hence playing a vital role in the marine biological carbon pump as well as the biogeochemical cycling of important nutrients, such as nitrogen and silicon (Nelson et al., 1995; Moore et al., 2013; Young and Morel, 2015). Diatoms usually dominate the phytoplankton communities and form large-scale blooms in nutrient-rich zones and upwelling regions (Bruland et al., 2001; Anderson et al., 2008; Barton et al., 2016). Nutrient enrichment is considered to be a key factor that triggers algal blooms, albeit the occurrence of diatom blooms may be modulated by other environmental factors, such as temperature, light intensity, and salinity (Smetacek and Zingone, 2013; Jeong et al., 2015). When inorganic nitrogen and phosphorus are replete, diatoms can outcompete chrysophytes, raphidophytes, and dinoflagellates (Berg et al., 1997; Jeong et al., 2015; Barton et al., 2016) and dominate algal blooms due to their quicker nutrient uptake and growth rate.

In normal natural seawater (pH 8.1, salinity 35), ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ is the majority (∼90 %) of total dissolved inorganic carbon (DIC, 2.0–2.2 mM). CO₂ (1 %, 10–15 µM), which is the only direct carbon source that can be assimilated by all photosynthetic organisms, only accounts for 1 % of total DIC. Diatoms' ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), catalyzing the primary chemical reaction by which CO₂ is transformed into organic carbon, has a relatively low affinity for CO₂ and is commonly less than half saturated under current CO₂ levels in seawater (Hopkinson and Morel, 2011), suggesting that CO₂ is limiting for marine diatoms' carbon fixation. To cope with the CO₂ limitation in seawater and maintain a high carbon fixation rate under the low-CO₂ conditions, diatoms have evolved various inorganic carbon acquisition pathways and CO₂ concentrating mechanisms (CCMs), for instance, active transport of ${\mathrm{HCO}}_{\mathrm{3}}^{-}$, the passive influx of CO₂, multiple carbonic anhydrase (including both common (α, β, γ, found in all algae) and unusual (δ, ζ, found only in diatoms) families that carry out the fast interconversion of CO₂ and ${\mathrm{HCO}}_{\mathrm{3}}^{-}$), and assumed C4-type pathway (using phosphoenolpyruvate to capture more CO₂ in the periplastidal compartment), to increase the concentration at the location of Rubisco and thus the carbon fixation (Hopkinson and Morel, 2011; Hopkinson et al., 2016). Skeletonema costatum is a worldwide diatom species that can be found from equatorial to polar waters. It usually dominates large-scale algal blooms in eutrophic seawaters (Wang, 2002; Li et al., 2011). When blooms occur, seawater pH increases and CO₂ decreases because the dissolution rate of CO₂ from the atmosphere cannot catch up with its removal rate caused by intensive photosynthesis of algae. For instance, pH level in the surface waters of the eutrophic Mariager Fjord, Denmark, could be up to 9.75 during algal blooms (Hansen, 2002). Consequently, S. costatum experiences very severe CO₂ limitation when blooms occur. To deal with it, S. costatum has developed multiple CCMs (Nimer et al., 1998; Rost et al., 2003). However, contrasting findings were reported. Nimer et al. (1998) documented that extracellular carbonic anhydrase activity in S. costatum was only induced when CO₂ concentration was less than 5 µmol L⁻¹ while Rost et al. (2003) reported that activity of extracellular carbonic anhydrase could be detected even when CO₂ concentration was 27 µmol L⁻¹. Chen and Gao (2004) showed that S. costatum had little capacity in direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ utilization. Conversely, Rost et al. (2003) demonstrated that this species could take up CO₂ and ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ simultaneously.

Phosphorus (P) is an indispensable element for all living organisms, serving as an integral component of lipids, nucleic acids, adenosine triphosphate (ATP), and a diverse range of other metabolites. Levels of bioavailable phosphorus are very low in many ocean environments and phosphorus enrichment can commonly increase algal growth and marine primary productivity in the worldwide oceans (Davies and Sleep, 1989; Müller and Mitrovic, 2015; Lin et al., 2016). Due to the essential role of phosphorus, extensive studies have been conducted to investigate the effect of phosphorus on photosynthetic performances (Geider et al., 1998; Liu et al., 2012; Beamud et al., 2016) and growth (Jiang et al., 2016; Reed et al., 2016; Mccall et al., 2017) as well as phosphorus acquisition, utilization, and storage (Lin et al., 2016; Gao et al., 2018a). Some studies show the essential role of phosphorus in regulating inorganic carbon acquisition in green algae (Beardall et al., 2005; Hu and Zhou, 2010). In terms of S. costatum, studies regarding the inorganic carbon acquisition in S. costatum focus on its response to variation in CO₂ availability. The role of phosphorus in S. costatum's CCMs remains unknown. Based on the connection between phosphorus and carbon metabolism in diatoms (Brembu et al., 2017), we hypothesize that phosphorus enrichment could enhance inorganic carbon utilization and hence maintain high rates of photosynthesis and growth in S. costatum under CO₂ limitation conditions. In the present study, we aimed to test this hypothesis by investigating the variation in CCMs (including active transport of ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ and carbonic anhydrase activity) and photosynthetic rate under five levels of phosphate and two levels of CO₂ conditions. We also measured redox activity of the plasma membrane as it is deemed to be critical to activate carbonic anhydrase (Nimer et al., 1998). Our study provides helpful insights into how bloom-forming diatoms overcome CO₂ limitation to maintain a quick growth rate during red tides.

2 Materials and methods

2.1 Culture conditions

Skeletonema costatum (Grev.) Cleve from Jinan University, China, was cultured in f/2 artificial seawater with five phosphate levels (0.05, 0.25, 1, 4, 10 µmol L⁻¹) by adding different amounts of NaH₂PO₄⋅2H₂O. The cultures were carried out semicontinuously at 20 ^∘C for 7 days. The light irradiance was set as 200 µmol photons m⁻² s⁻¹, with a light and dark period of 12 : 12. The cultures were aerated with ambient air (0.3 L min⁻¹) to maintain the pH at around 8.2. The cells during the exponential phase were collected and rinsed twice with DIC-free seawater that was made according to Xu et al. (2017). Afterwards, cells were resuspended in fresh media with two levels of pH (8.20 and 8.70, respectively corresponding to ambient CO₂ (12.6 µmol L⁻¹) and low CO₂ (2.8 µmol L⁻¹) under corresponding phosphate levels for 2 h before the following measurements, with a cell density of 1.0×10⁶ mL⁻¹. The concentrations of DIC were 2109±36 and 1802±38 µmol (kg seawater)⁻¹, respectively. Cell density was determined by direct counting with an improved Neubauer hemocytometer (XB-K-25, Qiu Jing, Shanghai, China). This transfer aimed to investigate the effects of phosphate on DIC acquisition under a CO₂ limitation condition. The pH of 8.70 was chosen considering that it is commonly used as a CO₂ limitation condition (Nimer et al., 1998; Chen and Gao, 2004) and also occurs during algal blooms (Hansen, 2002). A total of 2 h should be enough to activate CCMs in S. costatum (Nimer et al., 1998). The cell density did not vary during the 2 h of pH treatment. All experiments were conducted in triplicates.

2.2 Manipulation of seawater carbonate system

The two levels of pH (8.20 and 8.70) were obtained by aerating the ambient air and pure nitrogen (99.999 %) until the target value and were then maintained with a buffer of 50 mM tris (hydroxymethyl) aminomethane HCl. The cultures were open to the ambient atmosphere and the rise of culture pH was due to algal photosynthesis below 0.02 units (corresponding to the decrease in CO₂ less than 0.7 and 0.2 µmol L⁻¹ for pH 8.20 and 8.70 treatments, respectively) during the 2 h of pH treatment. CO₂ level in seawater was calculated via CO2SYS (Pierrot et al., 2006) based on measured pH and total alkalinity (TAlk), using the equilibrium constants of K1 and K2 for carbonic acid dissociation (Roy et al., 1993) and the ${\mathrm{KSO}}_{\mathrm{4}}^{-}$ dissociation constant from Dickson (1990).

The pH_NBS was measured using a pH meter (pH 700, Eutech Instruments, Singapore) that was equipped with an Orion^® 8102BN Ross combination electrode (Thermo Electron, USA) and calibrated with standard National Bureau of Standards (NBS) buffers (pH = 4.01, 7.00, and 10.01 at 25.0 ^∘C; Thermo Fisher Scientific Inc., USA). TAlk was determined at 25.0 ^∘C using Gran acidimetric titration on a 25 mL sample with a TAlk analyzer (AS-ALK1, Apollo SciTech, USA), using the precision pH meter and an Orion^® 8102BN Ross electrode for detection. To ensure the accuracy of TAlk, the TAlk analyzer was regularly calibrated with certified reference materials from Andrew G. Dickson's laboratory (Scripps Institute of Oceanography, USA) at a precision of ±2 µmol kg⁻¹.

2.3 Chlorophyll fluorescence measurement

Chlorophyll fluorescence was measured with a pulse modulation fluorometer (PAM-2100, Walz, Germany) to assess electron transport in photosystem II (the first protein complex in the light-dependent reactions of photosynthesis) and the possible connection between electron transport and redox activity of the plasma membrane. The measured light and actinic light were 0.01 and 200 µmol photons m⁻² s⁻¹, respectively. The saturating pulse was set to 4000 µmol photons m⁻² s⁻¹ (0.8 s). Electron transport in photosystem II (electron transport rate, ETR, µmol e⁻ (mg Chl a)⁻¹ s${}^{-\mathrm{1}})=$ 0.5 × E ×Φ_PSII × ${\overline{a}}^{\ast }$ (Dimier et al., 2009; Alderkamp et al., 2012), where E (µmol photons m⁻² s⁻¹) is the ambient light density, Φ_PSII (dimensionless) is the PSII photochemical efficiency and ${\overline{a}}^{\ast }$ is Chl a-specific absorption coefficient (m⁻² (mg Chl a)⁻¹). Since ${\overline{a}}^{\ast }$ is light dependent, we used the value of 0.0138 based on the study of Lefebvre et al. (2007) in which the light density is very close to ours.

2.4 Estimation of photosynthetic oxygen evolution and respiration

The net photosynthetic and respiration rates of S. costatum were measured using a Clark-type oxygen electrode (YSI model 5300, USA) that was held in a circulating water bath (cooling circulator; Cole-Parmer, Chicago, IL, USA) to keep the setting temperature (20 ^∘C). A total of 5 mL of samples was transferred to the oxygen electrode cuvette and was stirred during measurement. The light intensity and temperature were maintained the same as that in the growth condition. The illumination was provided by a halogen lamp. The increase in oxygen content in seawater within 5 min was defined as the net photosynthetic rate. To measure dark respiration rate, the samples were placed in darkness and the decrease in oxygen content within 10 min was defined as the dark respiration rate given the slower oxygen variation rate for dark respiration. Net photosynthetic rate and dark respiration rate were presented as µmol O₂ (10⁹ cells)⁻¹ h⁻¹.

To obtain the curve of net photosynthetic rate versus DIC, seven levels of DIC (0, 0.1, 0.2, 0.5, 1, 2, and 4 mM) were made by adding different amounts of NaHCO₃ to the Tris buffered DIC-free seawater (pH 8.20). The algal samples were washed twice with DIC-free seawater before transferring to the various DIC solutions. Photosynthetic rates at different DIC levels were measured under a saturating irradiance of 400 µmol photons m⁻² s⁻¹ and growth temperature. The algal samples were allowed to equilibrate for 2–3 min at each DIC level, during which a linear change in oxygen concentration was obtained and recorded. The parameter, photosynthetic half saturation constant (K_0.5, i.e., the DIC concentration required to obtain half of the DIC-saturated maximum rate of photosynthetic O₂ evolution), was calculated from the Michaelis–Menten kinetics equation (Caemmerer and Farquhar, 1981): $V=V_{\max }\times$ [S] ∕ (K_0.5+ [S]), where V is the real-time photosynthetic rate, V_max is the maximum photosynthetic rate, and [S] is the DIC concentration. The value of 1/K_0.5 represents photosynthetic DIC affinity. K_0.5 for CO₂ was calculated via CO2SYS (Pierrot et al., 2006) based on pH and TAlk, using the equilibrium constants of K1 and K2 for carbonic acid dissociation (Roy et al., 1993) and the ${\mathrm{KSO}}_{\mathrm{4}}^{-}$ dissociation constant from Dickson (1990).

2.5 Measurement of photosynthetic pigment

To determine the photosynthetic pigment (Chl a) content, 50 mL of culture was filtered on a Whatman GF∕ F filter, extracted in 5 mL of 90 % acetone for 12 h at 4 ^∘C, and centrifuged (3000 g, 5 min). The optical density of the supernatant was scanned from 200 to 700 nm with a UV–VIS spectrophotometer (Shimadzu UV-1800, Kyoto, Japan). The concentration of Chl a was calculated based on the optical density at 630 and 664 nm: Chl $a=\mathrm{11.47}\times$ OD${}_{\mathrm{664}}-\mathrm{0.40}\times$ OD₆₃₀ (Gao et al., 2018b) and was normalized to picograms per cell.

Figure 1 Net photosynthetic rate (a), dark respiration rate (b), and ratio of respiration rate to net photosynthetic rate (c) in S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

2.6 Measurement of extracellular carbonic anhydrase activity

Carbonic anhydrase activity was assessed using the electrometric method (Gao et al., 2009). Cells were harvested by centrifugation at 4000 g for 5 min at 20 ^∘C, washed once, and resuspended in 8 mL Na barbital buffer (20 mM, pH 8.2). A total of 5 mL of CO₂-saturated icy distilled water was injected into the cell suspension, and the time required for a pH decrease from 8.2 to 7.2 at 4 ^∘C was recorded. Extracellular carbonic anhydrase (CA_ext) activity was measured using intact cells. CA activity (E.U.) was calculated using the following formula: E.U. = 10 $\times (T_{\mathrm{0}}/T-\mathrm{1}$), where T₀ and T represent the time required for the pH change in the absence or presence of the cells, respectively.

2.7 Measurement of redox activity in the plasma membrane

The redox activity of plasma membrane was assayed by monitoring the change in K₃FeCN₆ concentration that accompanied reduction of the ferricyanide to ferrocyanide. The ferricyanide K₃FeCN₆ cannot penetrate intact cells and has been used as an external electron acceptor (Nimer et al., 1998; Gao et al., 2018b). Stock solutions of K₃FeCN₆ were freshly prepared before use. A total of 5 mL of samples was taken after 2 h of incubation with 500 μmol K₃FeCN₆ and centrifuged at 4000 g for 10 min (20 ^∘C). The concentration of K₃FeCN₆ in the supernatant was measured spectrophotometrically at 420 nm (Shimadzu UV-1800, Kyoto, Japan). The decrease in K₃FeCN₆ during the 2 h of incubation was used to assess the rate of extracellular ferricyanide reduction that was presented as µmol (10⁶ cells)⁻¹ h⁻¹ (Nimer et al., 1998).

Table 1 Two-way analysis of variance for the effects of CO₂ and phosphate on net photosynthetic rate, dark respiration rate, and ratio of respiration to photosynthesis of S. costatum. CO₂*phosphate means the interactive effect of CO₂ and phosphate, df means degree of freedom, F means the value of the F statistic, and Sig. means the p value.

2.8 Cell-driving pH drift experiment

To obtain the pH compensation point, the cells were transferred to sealed glass vials containing a fresh medium (pH 8.2) with corresponding phosphate levels. The cell concentration for all treatments was 5.0×10⁵ mL⁻¹. The pH drift of the suspension was monitored at 20 ^∘C and a 200 µmol photons m⁻² s⁻¹ light level. The pH compensation point was obtained when there was no further increase in pH.

2.9 Statistical analysis

Results were expressed as means of replicates ± standard deviation and data were analyzed using the software SPSS v.21. The data from each treatment conformed to a normal distribution (Shapiro–Wilk, P > 0.05) and the variances could be considered equal (Levene's test, P > 0.05). Two-way ANOVAs were conducted to assess the effects of CO₂ and phosphate on net photosynthetic rate, dark respiration rate, ratio of net photosynthetic rate to dark respiration rate, rETR, Chl a, K_0.5, CA_ext, reduction rate of ferricyanide, and pH compensation point. Least significant difference (LSD) was conducted for post hoc investigation. Repeated measurements of ANOVA were conducted to analyze the effects of DIC on net photosynthetic rate and the effect of incubation time on media pH in a closed system. Bonferroni was conducted for post hoc investigation as it is the best reliable post hoc test for repeated measurements of ANOVA (Ennos, 2007). The threshold value for determining statistical significance was P < 0.05.

3 Results

3.1 Effects of CO₂ and phosphate on photosynthetic and respiratory performances

The net photosynthetic rate and dark respiration rate in S. costatum grown at various CO₂ and phosphate concentrations were first investigated (Fig. 1). CO₂ interacted with phosphate on net photosynthetic rate, with each factor having a main effect (Table 1 and Fig. 1a). Post hoc LSD comparison (P= 0.05) showed that 2.8 µmol of CO₂ reduced net photosynthetic rate when the phosphate level was below 4 µmol L⁻¹ but did not affect it at the higher phosphate levels. Under the condition of 12.6 µmol of CO₂, net photosynthetic rate increased with phosphate level and reached the plateau (100.51±9.59 µmol O₂ (10⁹ cells)⁻¹ h⁻¹) at 1 µmol L⁻¹ phosphate. Under the condition of 2.8 µmol of CO₂, net photosynthetic rate also increased with phosphate level but did not hit the peak (101.46±9.19 µmol O₂ (10⁹ cells)⁻¹ h⁻¹) until 4 µmol L⁻¹ of phosphate. In terms of dark respiration rate (Fig. 1b), phosphate had a main effect on it and it interacted with CO₂ (Table 1). Specifically, 2.8 µmol of CO₂ increased dark respiration rate at 0.05 and 0.25 µmol L⁻¹ phosphate levels, but did not affect it when the phosphate level was above 1 µmol L⁻¹ (LSD, P < 0.05). Regardless of CO₂ level, respiration rate increased with phosphate availability and stopped at 1 µmol L⁻¹.

Figure 2 Relative electron transport rate (rETR) in S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

The ratio of respiration to photosynthesis ranged from 0.23 to 0.40 (Fig. 1c). Both CO₂ and phosphate had a main effect, and they interacted on the ratio of respiration to photosynthesis (Table 1). The level of 2.8 µmol of CO₂ increased the ratio when phosphate was lower than 4 µmol L⁻¹ but did not affect it when phosphate levels were 4 or 10 µmol L⁻¹.

Both CO₂ and phosphate affected ETR and they also showed an interactive effect (Fig. 2 and Table 2). For instance, post hoc LSD comparison showed that 2.8 µmol of CO₂ did not affect ETR at lower phosphate levels (0.05 and 0.25 µmol L⁻¹) but increased it at higher phosphate levels (1–10 µmol L⁻¹). Regardless of CO₂ treatment, ETR increased with phosphate level (0.05–4 µmol L⁻¹) but the highest phosphate concentration did not result in a further increase in ETR (LSD, P > 0.05).

Table 2 Two-way analysis of variance for the effects of CO₂ and phosphate on relative electron transport rate (rETR), Chl a, and CO₂ level required to obtain half of the DIC-saturated maximum rate of photosynthetic O₂ evolution (K_0.5) of S. costatum. CO₂*phosphate means the interactive effect of CO₂ and phosphate, df means degree of freedom, F means the value of the F statistic, and Sig. means the p value.

The content of Chl a was measured to investigate the effects of CO₂ and phosphate on photosynthetic pigment in S. costatum (Fig. 3). Both CO₂ and phosphate affected the synthesis of Chl a and they had an interactive effect (Table 2). Post hoc LSD comparison (P= 0.05) showed that 2.8 µmol of CO₂ did not affect Chl a at 0.05 or 0.25 µmol L⁻¹ of phosphate but stimulated Chl a synthesis at higher phosphate levels (1–10 µmol L⁻¹). Irrespective of CO₂ treatment, Chl a content increased with phosphate level and reached the plateau (0.19±0.01 pg cell⁻¹ for 12.6 µmol of CO₂ and 0.23±0.01 pg cell⁻¹ for 2.8 µmol of CO₂) at 4 µmol L⁻¹ of phosphate.

Figure 3 Photosynthetic Chl a content in S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

To assess the effects of CO₂ and phosphate on photosynthetic CO₂ affinity in S. costatum, the net photosynthetic rates of cell exposure to seven levels of DIC were measured (Fig. 4). After curve fitting, the values of K_0.5 for CO₂ were calculated (Fig. 5). CO₂ and phosphate interplayed on K_0.5 and each had a main effect (Table 2). The level of 2.8 µmol of CO₂ did not affect K_0.5 at the lowest phosphate level but reduced it at the other phosphate levels. Under the condition of 12.6 µmol of CO₂, higher phosphate levels (0.25–4 µmol L⁻¹) reduced K_0.5 and the highest phosphate level led to a further decrease to 2.59±0.29 µmol kg⁻¹ of seawater compared to the value of 4.00±0.30 µmol kg⁻¹ of seawater at 0.05 µmol L⁻¹ of phosphate. The pattern with phosphate at 2.8 µmol of CO₂ was the same as 12.6 µmol of CO₂.

Figure 4 Net photosynthetic rate as a function of DIC for S. costatum grown at various phosphate concentrations after 12.6 (a) and 2.8 µmol CO₂ (b) treatments. The error bars indicate the standard deviations (n=3).

3.2 The effects CO₂ and phosphate on inorganic carbon acquisition

To investigate the potential mechanisms that allow cells to overcome CO₂ limitation during algal blooms, the activity of CA_ext, a CCM-related enzyme, was estimated under various CO₂ and phosphate conditions (Fig. 6a). Both CO₂ and phosphate had a main effect and they interacted on CA_ext activity (Table 3). Post hoc LSD comparison (P= 0.05) showed that 2.8 µmol of CO₂ induced more CA_ext activity under all phosphate conditions except for 0.05 µmol L⁻¹ levels, compared to 12.6 µmol of CO₂. Under the condition of 12.6 µmol of CO₂, CA_ext activity increased (0.04–0.10 EU (10⁶ cells)⁻¹) with phosphate level and stopped increasing at 1 µmol L⁻¹ of phosphate. Under the condition of 2.8 µmol of CO₂, CA_ext activity also increased (0.04–0.35 EU (10⁶ cells)⁻¹) with phosphate level but reached the peak at 4 µmol L⁻¹ of phosphate. The redox activity of the plasma membrane was also assessed to investigate the factors that modulate CA_ext activity (Fig. 6b). The pattern of redox activity of the plasma membrane under various CO₂ and phosphate conditions was the same as that of CA_ext activity. That is, CO₂ and phosphate had an interactive effect on redox activity of the plasma membrane, each having a main effect (Table 3).

Figure 5 Half saturation constant (K_0.5) for CO₂ in S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

To test cells' tolerance to high pH and obtain pH compensation points in S. costatum grown under various CO₂ and phosphate levels, changes of media pH in a closed system were monitored (Fig. 7). The media pH under all phosphate conditions increased with incubation time (Table 4). Specifically speaking, there was a steep increase in pH during the first 3 h; afterwards the increase became slower and it reached a plateau at 6 h (Bonferroni, P < 0.05). Phosphate had an interactive effect with incubation time (Table 4). For instance, there was no significant difference in media pH among phosphate levels during the first 2 h of incubation but then divergence occurred and they stopped at different points. Two-way ANOVA showed that CO₂ treatment did not affect pH compensation point but phosphate had a main effect (Table 3). Under each CO₂ treatment, pH compensation point increased with phosphate level, with the lowest of 9.03±0.03 at 0.05 µmol L⁻¹ and highest of 9.36±0.04 at 10 µmol L⁻¹ of phosphate.

Table 3 Two-way analysis of variance for the effects of CO₂ and phosphate on CA_ext activity, redox activity of plasma membrane, and pH compensation point of S. costatum. CO₂*phosphate means the interactive effect of CO₂ and phosphate, df means degree of freedom, F means the value of the F statistic, and Sig. means the p value.

4 Discussion

4.1 Photosynthetic performances under various CO₂ and phosphate conditions

The lower CO₂ availability reduced the net photosynthetic rate of S. costatum grown at the lower phosphate levels in the present study. However, Nimer et al. (1998) demonstrated that the increase in pH (8.3–9.5) did not reduce photosynthetic CO₂ fixation of S. costatum and Chen and Gao (2004) reported that a higher pH (8.7) even stimulated the photosynthetic rate of S. costatum compared to the control (pH 8.2). The divergence between our and the previous studies may be due to different nutrient supply. Both Nimer et al. (1998) and Chen and Gao (2004) used f/2 media to grow algae. The phosphate concentration in f/2 media is ∼36 µmol L⁻¹, which is replete for physiological activities in S. costatum. Skeletonema costatum grown at higher phosphate levels (4 and 10 µmol L⁻¹) also showed similar photosynthetic rates for the lower and higher CO₂ treatments. Our finding combined with the previous studies indicates phosphorus plays an important role in dealing with low CO₂ availability for photosynthesis in S. costatum.

Figure 6 CA_ext activity (a) and reduction rate of ferricyanide (b) in S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

Different from net photosynthetic rate, 2.8 µmol of CO₂ did not affect rETR at lower phosphate levels (0.05 and 0.25 µmol L⁻¹) and stimulated it at higher phosphate levels (1–10 µmol L⁻¹). This interactive effect of CO₂ and phosphate may be due to their effects on Chl a. The level of 2.8 µmol CO₂ induced more synthesis of Chl a at higher phosphate levels (1–10 µmol L⁻¹). This induction of lower CO₂ on photosynthetic pigment is also reported in green algae (Gao et al., 2016). More energy is required under lower CO₂ to address the more severe CO₂ limitation and thus more Chl a is synthesized to capture more light energy, particularly when phosphate was replete. Although P is not an integral component for chlorophyll, it plays an important role in cell energetics through high-energy phosphate bonds, i.e., ATP, which could support chlorophyll synthesis. The stimulating effect of P enrichment on photosynthetic pigment is also found in the green alga Dunaliella tertiolecta (Geider et al., 1998) and the brown alga Sargassum muticum (Xu et al., 2017). The increased photosynthetic pigment in S. costatum could partially explain the increased rETR and photosynthetic rate under the higher P conditions.

Figure 7 Changes of pH in a closed system caused by photosynthesis of S. costatum grown at various phosphate concentrations after 12.6 and 2.8 µmol CO₂ treatments. The error bars indicate the standard deviations (n=3).

4.2 Ratio of respiration to photosynthesis

The ratio of respiration to photosynthesis in algae indicates carbon balance in cells and carbon flux in marine ecosystems as well (Zou and Gao, 2013). The level of 2.8 µmol of CO₂ increased this ratio in S. costatum grown at the lower P conditions but did not affect it under the higher P conditions, indicating that P enrichment can offset the carbon loss caused by carbon limitation. To cope with CO₂ limitation, cells might have to obtain energy from dark respiration under lower P conditions as it seems infeasible to acquire energy from the low ETR, which led to the increased dark respiration. However, 2.8 µmol of CO₂ induced higher ETR under P-replete conditions and energy used for inorganic carbon (CO₂ and ${\mathrm{HCO}}_{\mathrm{3}}^{-}$) acquisition could be from the increased ETR. Therefore, additional dark respiration was not triggered, avoiding carbon loss. Most studies regarding the effect of CO₂ on ratio of respiration to photosynthesis focus on higher plants (Gifford, 1995; Ziska and Bunce, 1998; Cheng et al., 2010; Smith and Dukes, 2013); little is known about phytoplankton. Our study suggests that CO₂ limitation may lead to carbon loss in phytoplankton but P enrichment could alter this trend, regulating carbon balance in phytoplankton and thus their capacity in carbon sequestration.

Table 4 Repeated measurements of analysis of variance for the effects of CO₂ and phosphate on pH change during 10 h of incubation. Time*CO₂ means the interactive effect of incubation time and CO₂. Time*phosphate means the interactive effect of incubation time and phosphate. Time*CO₂*phosphate means the interactive effect of incubation time, CO₂, and phosphate. df means degree of freedom, F means the value of the F statistic, and Sig. means the p value.

4.3 Inorganic carbon acquisition under CO₂ limitation and phosphate enrichment

Decreased CO₂ can usually induce higher inorganic carbon affinity in algae (Raven et al., 2012; Wu et al., 2012; Raven et al., 2017; Xu et al., 2017). In the present study, the lower CO₂ did increase inorganic carbon affinity when the P level was higher than 0.25 µmol L⁻¹ but did not affect it when P was 0.05 µmol L⁻¹, indicating the important role of P in regulating cells' CCMs in response to environmental CO₂ changes. The level of 2.8 µmol of CO₂ induced larger CA activity when P was above 0.25 µmol L⁻¹ but did not increase it at 0.05 µmol L⁻¹ of P, which could explain the interactive effect of P and CO₂ on inorganic carbon affinity as CA can accelerate the equilibrium between ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ and CO₂ and increase inorganic carbon affinity. Regardless of CO₂, P enrichment alone increased CA activity and inorganic carbon affinity. P enrichment may stimulate the synthesis of CA by supplying required ATP. In addition, P enrichment increased the redox activity of the plasma membrane in this study. It has been proposed that redox activity of the plasma membrane could induce extracellular CA activity via protonation extrusion of its active center (Nimer et al., 1998). Our result that the pattern of CA is exactly the same as that of redox activity of the plasma membrane shows a compelling correlation between CA and redox activity of the plasma membrane. The stimulating effect of P on redox activity of plasma membrane may be due to its effect on ETR. The increased ETR could generate excess reducing equivalents, particularly under CO₂-limiting conditions. These excess reducing equivalents would be transported from the chloroplast into the cytosol (Heber, 1974), supporting the redox chain in the plasma membrane (Rubinstein and Luster, 1993; Nimer et al., 1999) and triggering CA activity.

4.4 Direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ utilization due to phosphate enrichment

A pH compensation point over 9.2 has been considered a sign of direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ use for algae (Axelsson and Uusitalo, 1988) as the CO₂ concentration is nearly zero at a pH above 9.2. This criterion has been justified based on experiments for both micro and macro-algae. For instance, the marine diatom Phaeodactylum tricornutum, with a strong capacity for direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ utilization, has a higher pH compensation point of 10.3 (Chen et al., 2006). In contrast, the red macroalgae, Lomentaria articulata and Phycodrys rubens, that cannot utilize ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ directly, and whose photosynthesis only depends on CO₂ diffusion, have pH compensation points of less than 9.2 (Maberly, 1990). In terms of S. costatum, it has been reported to have a pH compensation point of 9.12, indicating a very weak capacity for direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ utilization (Chen and Gao, 2004). Our study demonstrates that the pH compensation point of S. costatum varies with the availability of P. It is lower than 9.2 under P-limiting conditions but higher than 9.2 under P-replete conditions, suggesting that the capacity of direct ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ utilization is regulated by P availability. Contrary to CO₂ passive diffusion, the direct use of ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ depends on positive transport that requires energy (Hopkinson and Morel, 2011). P enrichment increased ETR in the present study and the ATP produced during the process of electron transport could be used to support ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ positive transport. In addition, the increased respiration at higher P levels can also generate ATP to help ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ positive transport. Our study indicates that P enrichment could trigger ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ direct utilization and hence increase inorganic acquisition capacity of S. costatum to cope with CO₂ limitation.

4.5 CCMs and red tides

In the development of red tides, the pH in seawater can be very high along with extremely low CO₂ availability due to intensive photosynthesis (Hansen, 2002; Hinga, 2002). For instance, the pH level in the surface waters of the eutrophic Mariager Fjord, Denmark, is often above 9 during dinoflagellate blooms (Hansen, 2002). Diatoms are the causative species for red tides and S. costatum could outcompete other bloom algae (dinoflagellates Prorocentrum minimum and Alexandrium tamarense) under nutrient-replete conditions (Hu et al., 2011). However, the potential mechanisms are poorly understood. Our study demonstrates S. costatum has multiple CCMs to cope with CO₂ limitation and the operation of CCMs is regulated by P availability. The CCMs of S. costatum are hampered under P-limiting conditions and only function when P is replete. This finding may explain why diatoms could overcome carbon limitation and dominate red tides when P is replete as well as the shift from diatoms to dinoflagellates when P is limiting (Mackey et al., 2012).

5 Conclusions

The present study investigated the role of P in regulating inorganic carbon acquisition and CO₂ concentrating mechanisms in diatoms for the first time. The intensive photosynthesis and quick growth during algal blooms usually result in a noticeable increase in pH and decrease in CO₂. Our study demonstrates that P enrichment could induce activity of extracellular carbonic anhydrase and direct utilization of ${\mathrm{HCO}}_{\mathrm{3}}^{-}$ in S. costatum to help overcome CO₂ limitation, as well as increasing photosynthetic pigment content and rETR to provide required energy. This study provides important insight into the connection of phosphorus and carbon acquisition in diatoms and the mechanisms that help S. costatum dominate algal blooms.