Experimental strains S. costatum and P. globosa were both isolated from the coastal water in the SCS and cultured in Erdschreiber’s medium (modified from the original Plymouth seawater recipe, Rosowski and Parker, 1971; Supplementary Table S1) at (24 ± 1)°C and 50 μmol photons m⁻² s⁻¹ under a 24 h light: 0 h dark cycle. Phaeocystis. globosa was kept in the solitary form throughout the experimental period, and colony formation was not observed. This could be explained by no environmental stresses, such as grazing, in the laboratory. Na₂SiO₃·9H₂O with a final concentration of 106 μM was added to the culture of S. costatum. Erdschreiber’s medium lacking Na₂HPO₄ was used as a P-free medium in which Pi-containing compounds were substituted with NaCl. Algal cells in the logarithmic growth phase under P-repleted condition were collected by centrifugation, washed three times with P-free Erdschreiber’s medium, and resuspended into 500 mL Erlenmeyer flasks containing 250 mL Erdschreiber’s medium (PO₄³⁻ concentration: 67.0 μM) as the control group (+P) and 250 mL P-free medium (PO₄³⁻ concentration: 0.0 μM) as the treatment group (−P). The initial biomass of experimental algae was set to an OD₇₃₀ of 0.15 corresponding to Chl a content of ~0.6 μg/mL in S. costatum and ~ 0.4 μg/mL in P. globosa. All cultures were shaken three times every day during the 7-day cultivation period. All treatments were performed in triplicates.

To test the effect of intracellular P on the growth of S. costatum and P. globosa, the contents of polyphosphate body (PolyP) in cells were quantified according to Martin and Van Mooy (2013). Cells were harvested by centrifugation (Eppendorf 5810R centrifuge, 4,000 × g, 3 min). 10 μL proteinase K (20 mg/mL) was added to the pellets and incubated at 37°C for 30 min. Then centrifuged for 1 min at 16100 × g and the supernatant was removed. PolyP contents of samples were measured after the addition of DAPI (final concentration: 100 μM) and made on a fluorescence spectrofluorometer (F-2700, Hitachi, Japan) at an excitation wavelength of 415 nm and emission wavelength of 550 nm, with an integration time of 0.5 s.

Various chlorophyll fluorescence tools combined with protein measurements were used to examine the different effects of P limitation on PSII in S. costatum and P. globosa. The photosynthetic acclimation mechanisms of S. costatum and P. globosa were explored in terms of induction of NPQ and D1 turnover.

Growth characteristics of cells were studied by monitoring the optical density at 730 nm (OD₇₃₀) using a spectrophotometer. F_v/F_m was obtained from the JIP-test after dark-adapted for 15 min at room temperature according to Strasser et al. (2004) with a dual modulation kinetic fluorometer FL3500 (PSI, Brno, Czechia). These two values were obtained from cultures every day.

The rapid light response curve included steps of actinic irradiance: 2, 7, 14, 28, 45, 64, 92, 131, 181, 245, 321, 412, 540, 684, 869, 1,109, and 1,410 μmol·m⁻² s⁻¹, with a 30-s interval between steps using a Dual-PAM 100 (Walz, Effeltrich, Germany) under the illumination with the measuring light at 460 nm and the red actinic light at 635 nm. The built-in fitting function was used to fit the rapid light response curves with the EP model and derive the maximum electron transport rate (ETR_max), light-harvesting efficiency (alpha), and the point of light saturation (Ik).

Samples were dark-adapted for 15 min at room temperature before fluorescence measurement using a dual modulation kinetic fluorometer FL3500 (PSI, Brno, Czechia). Under conditions of continuous red actinic light (630 nm) at a high intensity of 2000 μmol·m⁻² s⁻¹, the Chl a fluorescence transient was recorded up to 1 s on a logarithmic timescale, with data acquisition every 10 μs for the first 2 ms and for every 1 ms thereafter. The curves were normalized between F_o and F_m and plotted as V_t changes on a logarithmic scale [V_t = (F_t − F_o)/(F_m − F_o)], where F_t is the fluorescence in respective time, F_o is the minimal fluorescence and F_m is the maximum fluorescence. The quantum yield of electron transport (φE₀), the reduction in end acceptors on the PSI electron acceptor side (RE₀/RC), and the amount of active PSII RCs per CS (RC/CS₀) values were obtained from the JIP-test.

Samples of 3 mL were adjusted to an OD₇₃₀ of 0.7 and dark-adapted for 10 min before measurement. The Q_A⁻ reoxidation kinetics after a single turnover flash was performed with a double-modulation fluorometer FL6600 (Photon Systems Instruments, Brno, Czechia; Eisenstadt et al., 2008). Both measuring flashes (4 μs) and actinic flashes (50 μs) were provided by computer-controlled light-emitting diodes. Q_A⁻ reoxidation kinetic curves were normalized and plotted as [F_(t)-F₍₀₎] changes on a logarithmic scale, where F_(t) is the fluorescence in respective time and F₍₀₎ is the initial fluorescence.

The relaxation of the flash-induced increase in Chl a fluorescence yield reflects the reoxidation of Q_A⁻ via forward electron transport to Q_B and reverse reactions with the S₂ state of the oxygen-evolving complex (OEC; Cao and Govindjee, 1990; Dau, 1994). The fast phase is attributable to the reoxidation of Q_A⁻ by Q_B. The middle phase arises from Q_A⁻ reoxidation in the PSII reaction center that has an empty Q_B site at the time of the flash and has to bind to a plastoquinone (PQ) molecule from the PQ pool. The slow phase reflects Q_A⁻ reoxidation with the S₂ state of the OEC, thus causing backward electron transport through the equilibrium of Q_A⁻Q_B and Q_AQ_B⁻ (Zhang et al., 2017). When Q_A⁻ reoxidation kinetics were determined in the presence of 20 μM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), the fluorescence decay reflects Q_A⁻ reoxidation via charge recombination with the S₂ state of the OEC (Vass et al., 1999).

According to Beauchemkin et al. (2007), the Q_A⁻ reoxidation kinetics curves were fitted by the three-exponential component decay equation:

F_(t) = A₁ exp. (−t/T₁) + A₂ exp. (−t/T₂) + A₃ exp. (−t/T₃) + A₀

Where F(t) is the variable fluorescence yield, A₀ to A₃ are the amplitudes, and T₁ to T₃ are the time constants from which the half-life values can be calculated as t_1/2 = ln 2 T.

To further determine the proportion of inactive PSII centers (PSII_X), S-state tests were performed with FL6600 after concentrated and dark-adapted samples as described in section 2.5. The population of the PSII_X center was measured by the difference between F_o and the fluorescence level 100 ms after the fourth flash (F₄ = F_400ms/F_o − 1) because the fluorescence decay after the fourth flash is controlled almost entirely by inactive centers (Lavergne and Leci, 1993). Given the decreased relative variable fluorescence, a revised equation was proposed according to Pan et al. (2008) as follows: PSII_X(%) = F₄ × 100/(F_300ms/F_o − 1).

Samples were dark-adapted for 30 min before measurement using a Dual-PAM 100. The slow induction kinetic curves at a specific process (Delay: 30 s, Measure light: 46 μmol photons m⁻² s⁻¹, Saturable Pulse: 20000 μmol photons m⁻² s⁻¹, Width: 300 ms, Clock: 1 min). NPQ was calculated as Bilger and Bjorkman (1990): NPQ = (Fm − Fm′)/Fm′, where Fm′ is the maximum fluorescence measured after the samples are exposed to continuous red actinic light (635 nm) of 321 μmol photons m⁻² s⁻¹.

Total cellular proteins were extracted as previously described by Thangaraj et al. (2021) with minor modifications. Cells were harvested and resuspended in lysis buffer (40 mM Tris–HCl, pH 8.0) supplemented with 1 mM phenylmethanesulfonyl fluoride (PMSF) as the protease inhibitor. The cells were lysed using a noise isolating chamber (3 s on, 3 s off, Scientz, China) for 20 min on ice with a whole-cell lysate and then centrifuged (Eppendorf 5810R centrifuge) at 1,800 × g for 10 min at 4°C to remove cell debris. Protein concentrations were determined using the BCA assay (Beyotime, China).

Protein samples were subjected to 12.5% SDS-PAGE where each gel lane was loaded with equal amounts of 10 μg, stained with Coomassie brilliant blue R250, or transferred to polyvinylidene fluoride membranes. Subsequently, each membrane was blocked for 1 h in 5% skimmed milk and probed using rabbit primary anti-D1 (1:4000, PhytoAB, United States) and anti-PsbO (1:2000, PhytoAB, United States) antibodies. Immunodetection was performed using a goat-anti-rabbit secondary antibody conjugated to horseradish peroxidase (1:5000). Proteins were visualized on the basis of the intensities of immunoreactions using an ImageQuant LAS 4000 Mini system (GE Healthcare) and semi-quantitated by ImageJ 1.52a (Wayne Rasband, United States).

The clearance of the D1 protein untreated and treated with lincomycin (block counteracting repair processes, Wu et al., 2012) was plotted and compared. Samples in triplicates were supplemented with 1,000 μg/mL lincomycin and incubated in the dark for 10 min, to allow the antibiotic to penetrate the cells and inhibit ribosome function. All samples were then shifted to 50 μmol photons m⁻² s⁻¹ and collected on days 0, 1 and 4 for later protein immunodetection. Newly synthesized D1 protein in PSII repair was calculated as D1_newly = D1_active − D1_blocked, where D1_active is the change in the active D1 level in the absence of lincomycin (PSII repair active), and D1_blocked is the change in the active D1 level in the presence of lincomycin (PSII repair blocked).

All experiments were performed in triplicates, and the results are presented as means ± standard deviations (SD). Statistical significance of differences between treatments was compared using two-way ANOVA with the least-significant difference (LSD) by SPSS 18.0 (IBM, United States).

To test the endogenous phosphorus in initial experimental cells, intracellular PolyP in S. costatum and P. globosa in +P and − P groups on days 0, 3 and 7 were measured (Supplementary Figure S1). The contents of PolyP in S. costatum and P. globosa cells were both significantly lower in the +P group compared with the −P group on day 0 (p < 0.01, ANOVA), indicating that the above two strains have the ability to store phosphate and form the PolyP when P is sufficient. The contents of PolyP in S. costatum and P. globosa in −P groups were 0.005 ± 0.0016 and 0.003 ± 0.0003 nmol per OD₇₃₀ on day 0, and 0.003 ± 0.0002 and 0.002 ± 0.0008 nmol per OD₇₃₀ on day 7, respectively (Supplementary Figure S1), which indicated that the cells in −P groups were P-limited for the entire experiment duration.

The cell propagation of S. costatum was completely inhibited after exposure to P limitation. The OD₇₃₀ of cells cultured in −P group was significantly lower after 4-day cultivation (Figure 1A, p < 0.01, ANOVA), implying that S. costatum experienced growth inhibition under the P-limited condition. In comparison, P. globosa cell growth in the absence of P showed OD₇₃₀ similar to that of the control (p > 0.05), which could still maintain a logarithmic period (Figure 1B).

To investigate if the decreased PSII activity under P limitation was associated with the decrease in the PSII protein level, the effects of P limitation on PSII subunits D1 and PsbO were further examined. The decrease in D1 proteins in S. costatum and P. globosa in the +P group (Figures 8A,C) could be explained by the gradual consumption of nutrient elements in batch culture experiments (Keren et al., 1997). P limitation results in 100% loss of the D1 protein in S. costatum in 4 days (Figures 8A,C, p < 0.01, ANOVA). However, a 60% decrease was found in the D1 level of P. globosa under P limitation for 6 days (Figures 8A,C, p < 0.05, ANOVA). The PsbO level was stable at 25% of the original value in the later stage of growth under P limitation in P. globosa. However, a 100% decrease in PsbO protein level was observed in S. costatum after 2 days of P limitation (Figures 8B,D, p < 0.01, ANOVA). Therefore, P limitation resulted in a larger and faster decrease in D1 and PsbO in S. costatum than in P. globosa.

Cells can repair damaged PSII by replacing photo-damaged proteins in PSII with newly synthesized subunits (Nixon et al., 2010; Komenda et al., 2012). Wu et al. (2012) blocked the counteracting repair processes by adding lincomycin to estimate the newly synthesized D1 protein in PSII repair. To further investigate if stable PSII activity under P limitation in P. globosa was associated with PSII repair, the newly synthesized D1 was plotted and compared. In both S. costatum and P. globosa, treatment with lincomycin elicited a greater drop in the D1 protein level in the absence of P than in the presence of P (Figures 9A–C). The presence of lincomycin resulted in a greater loss of the D1 protein in S. costatum than in P. globosa under both P-repleted and P-limited conditions (Figures 9A–C). Without lincomycin, the loss of D1 protein content is still higher in S. costatum than in P. globosa under both two conditions (Figures 9A–C). These results indicated that the PSII was impaired more seriously in S. costatum than in P. globosa under both conditions. When comparing the newly synthesized D1 protein between P-limited and P-repleted conditions, we found that the newly synthesized D1 protein decreased in S. costatum (Figure 9D, p < 0.05, ANOVA) while increased in P. globosa after P limitation (Figure 9E, p < 0.01, ANOVA), These results suggested that P limitation inhibited the PSII repair in S. costatum but accelerated it in P. globosa.

To explore whether S. costatum and P. globosa enhance their NPQ activity under P limitation to dissipate excess light as heat for protection from potential photo-oxidative damage, the NPQ values were obtained from the slow induction kinetic curves of S. costatum and P. globosa in the +P and − P groups for various periods of time (Supplementary Figure S4). The NPQ value of the −P group was significantly higher than that of the +P group, which was expressed in P. globosa (p < 0.05, ANOVA, Figure 10B), indicating that P limitation induced higher NPQ. The NPQ values of S. costatum were maintained at a low level for various periods, and the increase was not significant in the −P group compared with the +P group (p > 0.05, ANOVA, Figure 10A). Notably, P. globosa had >5-fold higher NPQ relative to the S. costatum at the same actinic light of 321 μmol photons m⁻² s⁻¹ red light (Figure 10), suggesting that the NPQ capacity of P. globosa cell was higher than that of S. costatum. Considering that the NPQ prevents photodamage of the D1 protein in PSII (Lepetit et al., 2012), we suggested that the above more active D1 turnover (Figure 9E) and stable PSII activity (Figure 3B) were due to the higher NPQ capacity P. globosa under P limitation.

Phaeocystis globosa blooms usually occur after diatom blooms in winter to early spring in the SCS and similar trends were also observed in the coastal waters of high latitude North Sea (Lancelot et al., 2005). Karasiewicz et al. (2018) suggested both abiotic and biotic interactions favored Phaeocystis blooms with great contribution from the preceding diatoms. The silicate exhaustion has historically been thought to be a major explanation for the appearance of P. globosa (Lancelot et al., 1987; Reid et al., 1990). The role of P in the transition of phytoplankton communities was reviewed by Lin et al. (2016).

P is an essential nutrient for the growth and proliferation of phytoplankton. P limitation inhibited the growth of S. costatum remarkably but had no obvious effect on the growth of P. globosa over a period of 7 days (Figure 1), which strongly suggested that P. globosa has more tolerance to P limitation compared with S. costatum. Moreover, the PSII activity of S. costatum was inhibited by P limitation, whereas P. globosa maintained photosynthetic activity after P limitation based on the estimation of F_v/F_m and other chlorophyll fluorescence parameters (Figure 2). The obvious growth and photosynthetic activity inhibition of P limitation in S. costatum are in accordance with those reported previously (Zhang et al., 2016).

In this study, the pre-incubation in P-free medium before this experiment was not done since S. costatum cannot keep photosynthetic activity after the pre-incubation. P pools associated with cells were identified previously as intracellular P pool and cell surface-adsorbed P pool (Jin et al., 2021). Both S. costatum and P. globosa can take up P in excess of its immediate cellular demand under P-sufficient conditions and store it in the form of PolyP, so-called luxury uptake, and break down this PolyP store upon P stress (Schoemann et al., 2001; Solovchenko et al., 2019). As shown in Supplementary Figure S1, both of them have the similar ability to store PolyP. However, the maintenance of growth and photosynthetic activity in P. globosa under P limitation might be caused by its higher tolerance to the P limitation as it still kept a stable photosynthetic activity after 25 days of P limitation (Supplementary Figure S2).

Photosynthesis can be divided into the light reactions, in which light energy is stored as ATP and NADPH, and the dark reactions, in which the products of the light reactions are used to reduce inorganic C (Green and Durnford, 1996). Phosphorus is an essential element in compounds such as ATP, NADPH, nucleic acids, phospholipids and sugar phosphates, all of which play important roles in photosynthesis. Numerous groups have demonstrated that P limitation causes a decline in photosynthesis in a number of different ways (Heraud et al., 2005; Shelly et al., 2005). It was proposed that a decreased supply of Pi reduced the ATP synthesis/levels, thereby, reducing both phosphorylation and CO₂ assimilation (Lewis et al., 1994; Campbell and Sage, 2006). Other studies also reported that the photosynthetic machinery composition or electron transport activity was also affected under P-limited condition (Van Rensen and Vredenberg, 2009; Carstensen et al., 2018). Wykoff et al. (1998) reported that P deprivation seriously inhibited the PSII activity but not that of PSI in C. reinhardtii. In this study, we mainly focused on the performance of PSII.

P limitation on their photosystem II, a multiprotein complex present in the thylakoid membranes of oxygenic photosynthetic organisms, is the center of photosynthesis that uses light to drive water oxidation and PQ reduction (Iwata and Barber, 2004). PSII is artificially divided into the donor sides (OEC), reaction centers, and acceptor sides. The donor sides split water and evolve oxygen (Wilson and Jain, 2018). The reaction center of PSII carries out photochemical reactions, including the primary charge separation and subsequent electron transfer from water to PQ (Semenov et al., 2011). On the electron acceptor side of PSII, Q_A (a one-electron acceptor PQ) takes up one electron from Phe_D1⁻ (a pheophytin molecule) and transfers it to Q_B (a two-electron acceptor PQ; Zhang et al., 2017). In this study, various chlorophyll fluorescence tools were used to examine the physiological response to P limitation in and around PSII.

P limitation inhibited electron transfer from Q_A⁻ to Q_B in S. costatum and P. globosa, which was suggested by the increase in the fluorescence intensity of phase J (Figures 4A,B), the increase in the decay half-life of the fast phase obtained from the Q_A⁻ reoxidation kinetics test (Table 1) and the increase in the proportion of PSII_X centers (Figures 7A,B). The increase in PSII_X centers, also called Q_B-nonreducing PSII centers, under P limitation was also observed in Chlamydomonas reinhardtii, Chlorella, and Scenedesmus (Wykoff et al., 1998; Dao and Beardall, 2016; Markou et al., 2017). These results provide circumstantial evidence that both S. costatum and P. globosa accumulate inactive PSII unable to reduce PQ under P limitation, which might be a way to acclimate to abiotic stress including P starvation. Previous studies have found that these centers are resistant to photoinhibition and capable of facilitating excess excitation energy dissipation for photosynthetic organisms under abiotic stress conditions (Neale and Melis, 1990; Hill and Ralph, 2006). A higher concentration of Q_B-nonreducing PSII centers was also shown in the triazine-resistant plant Chenopodium album (Van Rensen and Vredenberg, 2009).

The impairment of the donor side in PSII by P limitation has been reported in wheat, rice, citrus, and other higher plants (Lin et al., 2009; Veronica et al., 2017; Meng et al., 2021). From the Q_A⁻ reoxidation kinetics test in the presence of DCMU (Table 1), the donor side of PSII in S. costatum was damaged after exposure to P limitation for 2 days. These results are in accordance with the observed positive K-step in the OJIP transients (Figure 4C) and a sharp decrease in RC/CS₀ (Figure 5A) of S. costatum under P limitation. However, because the changes in these parameters resulting from P limitation were small or even negligible in P. globosa, the donor side of PSII in P. globosa might be damaged by P limitation but not seriously. The fact that P limitation caused more serious damage to OEC in S. costatum than in P. globosa was further confirmed by the different PsbO expression patterns in the two species under P limitation. PsbO is an extrinsic protein that plays a crucial role in the structure and function of the OEC of PSII (Popelkova and Yocum, 2011). Jain et al. (2005) reported that P limitation suppresses the expression of psbO (PsbO protein encoding gene) in Arabidopsis thaliana, thus impairing the structural integrity of OEC and consequently its function. In this study, a 100% decrease of PsbO in S. costatum was observed after exposure to P limitation for 2 days, which was in accordance with the physiological result that P limitation resulted in a serious impairment of the donor side in PSII in S. costatum. However, PsbO in P. globosa showed a slow decrease and remained at a certain level in the later stage of culture under P limitation (Figures 8B,D), reflecting smaller and slighter damage of OEC caused by P limitation in P. globosa.

Impairment of PSII finally led to the inhibition of electron transport to PSI. P limitation inhibiting electron transport to PSI has been demonstrated in barley (Carstensen et al., 2018) and dinoflagellate Peridinium bipes revealed by the decrease in φE₀ and RE₀/RC (Yang et al., 2020). P limitation resulted in the gradual and significant decline in φE₀ and RE₀/RC values in S. costatum (Figure 5A) rather than in P. globosa. These results confirmed that the P limitation only leads to the inhibition of electron transport after Q_A⁻ to the PSI electron acceptor side of S. costatum. However, the electron transport to PSI in P. globosa was not affected by P limitation (Figure 5B).

Thus, P limitation caused serious PSII damage with obvious impairment of both donor and acceptor sides of PSII in S. costatum; P. globosa under P limitation showed minor impairment of OEC of PSII compared with S. costatum and no obvious inhibition of electron transport to PSI. D1 constitutes the core of the PSII reaction center. PSII inactivation in C. reinhardtii subjected to P deprivation is the result of enhanced photoinhibition shown by enhanced D1 degradation (Wykoff et al., 1998; Malnoe et al., 2014). In this study, P limitation resulted in a much faster and larger decrease in D1 in S. costatum than in P. globosa, which is associated with the worse performance of PSII in the former than in the latter.

PSII repair involves partial disassembly of the damaged complex, selective proteolytic degradation, and replacement of the damaged subunit (predominantly the D1 reaction center subunit) by de novo synthesized copy and reassembly. Turnover of the D1 protein is required for PSII repair and restoration of PSII photochemical activity after photoinhibition (Nixon et al., 2010). P limitation causes photodamage of PSII in S. costatum by not only directly accelerating photoinactivation but also inhibiting D1 restoration. Inhibited D1 restoration under P starvation has also been reported in D. tertiolecta (Heraud et al., 2005). This may be because protein synthesis (protein turnover) is energetically costly, requiring 10.8 ATP per peptide bond, and P limitation often causes a decline in nucleotides such as ATP and GTP (Murata and Nishiyama, 2018). In contrast, P. globosa under P limitation showed increased D1 restoration obtaining a smaller decrease in the D1 level compared with S. costatum. It remains to be explained how P. globosa synthesizes a high amount of D1 under P limitation. P. globosa still maintains active D1 synthesis while ceasing synthesis of most other proteins, including the photosynthetic apparatus subunits in P limitation (Feng et al., 2015). Under unfavorable P-limited conditions, P. globosa invests much more energy in optimization and maintaining of light harvesting (such as de novo D1 synthesis), but not under P-replete conditions. This flexibility of P. globosa provides an advantage in the fluctuating environment. Similarly, the acceleration of D1 synthesis has been reported in Emiliania huxleyi under short-term P depletion (Loebl et al., 2010) and in the cyanobacterium Synechocystis sp. PCC 6803 under oxidative stress (Yu et al., 2014). Loebl et al. (2010) also found that the superior ability of PSII repair in E. huxleyi helps maintain significant PSII function over at least 38 days under nitrogen depletion. Undoubtedly, P. globosa, but not S. costatum, could counteract photoinactivation by the acceleration of PSII repair, thus maintaining stable photosynthetic activity under P limitation.

The photosynthetic apparatus showed higher resistance in P. globosa, which might also be caused by more effective energy dissipative mechanisms. NPQ allows for the safe dissipation of excess excitation energy as heat in light-harvesting antenna complexes (LHCs) of PSII and prevents over-reduction in the electron transport chain that may lead to reactive oxygen species (ROS) production (Muller et al., 2001). It thereby prevents photodamage of the D1 protein in PSII, which is the primary target of ROS-generated oxidative stress (Wu et al., 2011; Lepetit et al., 2012). The main part of NPQ in microalgae is called qE, which relies on the interconversion of specific pigments by the xanthophyll cycle. The xanthophyll cycle in green and brown algae is composed of violaxanthin, antheraxanthin, and zeaxanthin (VAZ cycle), while in dinoflagellates, diatoms, and haptophytes, it consists of diadinoxanthin and diatoxanthin (Dd-Dt cycle; Eisenstadt et al., 2008; Goss and Jakob, 2010).

Both P. globosa and S. costatum possess the Dd-Dt cycle and can promote NPQ (Lavaud et al., 2002; Harris et al., 2005). From the induction kinetics of NPQ (Figure 10B), P. globosa cells could enhance their NPQ activity as an adaptive strategy under P limitation. This can be explained as follows: P limitation reduces the orthophosphate concentration in the chloroplast stroma and causes the inhibition of ATP synthase activity. Consequently, protons accumulate in the thylakoids and cause a high trans-thylakoidal proton gradient activating the NPQ mechanism (Lavaud et al., 2004; Lavaud and Lepetit, 2013). Enhancement of NPQ is also an adaptive strategy under P limitation in the dinoflagellate Karlodinium veneficum (Cui et al., 2017), diatom Thalassiosira pseudonana (Li et al., 2021), and chlorophytes D. tertiolecta (Petrou et al., 2008) and C. reinhardtii (Wykoff et al., 1998). Indeed, the induction of LHCSR (Light Harvesting Complex Stress Related), which controls the recruitment of NPQ in C. reinhardtii, has also been reported under P deprivation (Moseley et al., 2006). Thus, NPQ, as a photosynthetic mechanism to cope with P limitation, plays an important role in maintaining the balance of the energy budget in P-limited P. globosa cells.

Su et al. (2012) and Lavaud et al. (2016) pointed out that S. costatum had limited capacity for induction of NPQ. Previous studies suggested two molecular explanations for the low NPQ in S. costatum (Lavaud and Lepetit, 2013): (1) a lower amount of Dt molecules involved in NPQ, and (2) lower capacity to form large functionally disconnected oligomeric fucoxanthin–chlorophyll protein complexes caused by the different organization and composition of LHC of PSII. Ultimately, the inability of S. costatum to promote a similarly strong NPQ as in P. globosa leads to a higher susceptibility of S. costatum to photoinhibition during exposure to P limitation.

P limitation was considered one of the factors favoring the population shift from diatoms or other early succession species to late succession species such as dinoflagellates, haptophytes, and pelagophytes, which are better adapted to this new set of environmental conditions (Lin et al., 2016). Considering that photosynthesis is the most important driver of cell growth and other physiological processes in phytoplankton, these different photosynthetic responses in algal bloom species may also contribute to their collapse or well-being under P limitation, thus affecting the transition of phytoplankton communities from early succession species to late succession species.

Liu et al. (2013) suggested that the photosynthetic capacity of S. costatum was limited by P limitation, and F_v/F_m decreased in S. costatum under P limitation (Li and Sun, 2016). This study demonstrated that P limitation seriously impaired PSII in S. costatum, and this species did not have a strong ability to induce high NPQ and accelerate the PSII repair cycle to decrease photosynthetic impairment resulting from P limitation. From the results of the present study, P. globosa could maintain relatively stable photosynthetic activity under P limitation by inducing high NPQ and accelerating D1 restoration (Figure 11). The photosynthesis activity might affect the P limitation tolerance because the energy sources for Pi uptake and the synthesis of APase in microalgae are mostly derived from photosynthesis (Pandey, 2006; Gorbunov and Falkowski, 2022). The superior photosynthetic performance of P. globosa under P limitation was undoubtedly conducive to its well-being, which is consistent with previous ecological investigations (Wang K. et al., 2021; Wang X. D. et al., 2021).

The succession of other species under P limitation has also been reported to be closely correlated with photosynthetic responses. An apparent bloom-forming species succession with the shifting from diatom blooms (mainly S. costatum) in the early spring to long-lasting and large-scale dinoflagellate blooms dominated by Prorocentrum donghaiense was observed in the coastal waters of the East China Sea (Lu et al., 2022), which is also reported to be in accordance with the different response of S. costatum and P. donghaiense to the phosphate exhaustion (Li et al., 2011; Ou et al., 2020). Although the study about the comparison of photosynthesis response to P limitation of these two strains is still limited, it is clear that P. donghaiense had higher photosynthetic activity and potential than S. costatum under eutrophic but relatively P-limited conditions according to Li and Sun (2016). P limitation is also one of the critical factors allowing the coccolithophorid E. huxleyi to bloom after diatom blooms in temperate coastal and oceanic areas (Tyrrell and Merico, 2004; Lessard et al., 2005; Oguz and Merico, 2006), although these conditions are not mandatory for E. huxleyi blooms (Lessard et al., 2005). Zhao et al. (2015) reported that P limitation resulted in greater damage to the photosynthetic apparatus in Phaeodactylum tricornutum, while E. huxleyi under P limitation showed increased xanthophyll cycle pigment accumulation and more active transformation from diadinoxanthin to diatoxanthin than P. tricornutum. Hence, the different photosynthetic responses to P limitation in some algal bloom species might be associated with their survival, dominance, and succession under P limitation.