1 INTRODUCTION

Dinoflagellates are the major agents responsible for harmful algal blooms (HABs) in the global ocean, which cause serious ecological, environmental, and health problems (Anderson et al., 2002; Wells et al., 2015, 2020). Studies have shown that dinoflagellates usually form blooms in the low-nutrient conditions or environments (Zhou et al., 2017; Xiao et al., 2018), but the underlying molecular mechanisms remain poorly understood. As a limiting nutrient in the ocean, phosphorus (P) is recognized as one of the key factors regulating cell growth and bloom formation of marine dinoflagellates (Do Rosário Gomes et al., 2014; Fricke et al., 2015; Shi et al., 2017). Dissolved inorganic P (DIP) and dissolved organic P (DOP) are the two major available sources of P for marine dinoflagellates (Dyhrman, 2016). However, DIP concentrations are usually low, at less than 0.5 μmol/L, which cannot fulfill the needs of dinoflagellates' growth, let alone their bloom formation (Benitez-Nelson, 2000; Cañellas et al., 2000; Ou, 2006; Ou et al., 2015; Dyhrman, 2016). DOP comprises a significant portion of the total P in both oceanic and coastal waters and it can be utilized by some dinoflagellate species (Ou et al., 2015; Dyhrman, 2016; Shi et al., 2017; Zhang et al., 2019a, b). It has been postulated that the capacity of dinoflagellates to utilize DOP in DIP-deficient ambient conditions is essential to their success and bloom formation in the ocean (Zhang et al., 2014, 2018, 2019a, b; Ou et al., 2015; Shi et al., 2017). Yet a comprehensive understanding of the dinoflagellates' response to changing ambient DIP or DOP is still limited, especially at the molecular level.

Prorocentrum donghaiense is a typical dinoflagellate species that causes extensive HABs in the coastal East China Sea, which seriously threatens the marine ecosystem, mariculture, and environmental health (Li et al., 2014; Zhou et al., 2017; Yu et al., 2018). The strong ability of P. donghaiense to utilize DOP in a low DIP environment is reportedly an important reason for why it can gain a competitive advantage and form blooms (Ou, 2006; Ou et al., 2015; Zhang et al., 2018, 2019b). However, our understanding of P. donghaiense in response to changes in either ambient DIP or DOP is surprisingly limited. Our previous study showed that the physiological responses of the P-deficient P. donghaiense to DIP and DOP resupply are similar (Zhang et al., 2019b). Building on that finding, this study compared global protein expression profiles of P. donghaiense under P-replete, P-deficient, and DIP-and DOP-resupplied conditions by using a Tandem Mass Tag (TMT)-based quantitative proteomic approach that combined the corresponding transcriptomic database and characterized the differentially expressed proteins. Our results reveal a significantly different proteomic response of P. donghaiense to DIP versus DOP resupply, and that DOP could provide both P and carbon sources for sustained cell growth which triggers bloom formation of P. donghaiense under the low DIP environment.

2 MATERIAL AND METHOD 2.1 Marine organism and culture conditions

The Prorocentrum donghaiense strain was isolated from the East China Sea, in May 2014, and routinely maintained in K-medium at 20 ℃ under a 14-h: 10-h light: dark cycle (Keller et al., 1987), with a light intensity of approximately 100 μmol/(m²·s) provided by fluorescent lamps. Before the experiment, the culture was treated with an antibiotic cocktail to minimize bacterial contamination and then inoculated into fresh K-medium after rinsing it with autoclaved seawater (Zhang et al., 2016).

2.2 Experimental design

To investigate the response of P. donghaiense to ambient P-deficiency and the utilization mechanism of DOP in a low DIP environment, four treatments were set up as described in our previous study (Zhang et al., 2019b), which consisted of the P-replete, P-deficient, DIP-resupplied, and DOP-resupplied groups with different ecological relevance (Supplementary Table S1). Each treatment had triplicate biological repeats and the initial cell density was 8.0×103 cells/mL. At the beginning of the experiment, cultures with 10.0-μmol/L and 0.2-μmol/L Na₂HPO ₄ were respectively defined as the P-replete and P-deficient groups. According to the results of physiological response (Zhang et al., 2019b), the P-replete cells entered the exponential phase on day 4, while the P-deficient cells exhibited P limitation on day 6 according to the result of alkaline phosphatase activity (APA) (Supplementary Fig.S1). At this point, a final concentration of 10.0-μmol/L Na₂HPO₄ or Glucose-6-phosphate (G-6-P) was added to the three P-deficient cultures, respectively, for recovery, constituting the DIP- and DOP-resupplied groups.

2.3 Protein preparation

According to the physiological responses of P.donghaiensepreviously reported on (Supplementary Fig.S1) (Zhang et al., 2019b), there is no significant change in cell density on day 4 and 6 in the P-deficient group, but the result of APA shows that the cells are limited by P on day 6 but not on day 4 (Supplementary Fig.S1). Therefore, in order to reflected their real P status and the difference between the P-replete group and the P-deficient group, samples of the P-replete cells on day 4, the P-deficient cells on day 6, and the DIP- and DOP-resupplied-28 h cells were collected for quantitative proteomic analysis, using three replicates.

Protein extraction was conducted using the lysis-buffer extraction method (Zhang et al., 2015). Briefly, the samples were first ground by liquid nitrogen and then transferred to a 1.5-mL centrifuge tube containing a lysis buffer (8-mol/L urea, 1% Triton-100, 65-mmol/L dithiothreitol, and 0.1% protease inhibitor cocktail). Total protein was obtained after sonication, centrifugation, acetone precipitation, and air-drying. Finally, protein content was quantified using a 2D Quant Kit (GE Healthcare, USA).

2.4 Peptide labeling and fractionation

In this study, 100 μg of protein was taken from each sample, and 100-mmol/L borane-triethylamine complex (TEAB) was added to dilute the urea concentration to less than 2 mol/L. Trypsin was added in a mass ratio of trypsin: protein=1:50, and this solution was further enzymatically hydrolyzed at 37 ℃; then trypsin was added once more (trypsin: protein=1:100), and the enzymatic hydrolysis continued for 4 h at 37 ℃. After trypsin digestion, the peptide was desalted by a Strata X C18 (Phenomenex, USA) and vacuum-dried. Peptides were redissolved in 0.5-mol/L TEAB and then labeled with the 6-plex TMT kit according to its manual: Tag₁₂₆, P-replete; Tag₁₂₇, P-deficient; Tag₁₂₈, DIP-resupplied; Tag₁₂₉, DOP-resupplied. The sample was subjected to liquid phase separation by high pH reversed-phase high-performance liquid chromatography (RP-HPLC), using an Agilent 300 liquid phase system (C18 separation column, 5-μm particle size, 4.6-mm inner diameter, 250 mm in length). Peptides were first separated into 80 fractions, by mixing with 2%-60% acetonitrile in a 10-mmol/L ammonium bicarbonate solution of pH 10 for 80 min, and then the peptides were combined into 20 fractions and vacuum-dried.

2.5 LC-MS/MS analysis

Each fraction was dissolved in 0.1% formic acid (FA) and then loaded onto a reversed-phase pre-column (Acclaim PepMap 100, Thermo Scientific, USA), and the peptide was separated via a reversed-phase analytical column (Acclaim PepMap RSLC, Thermo Scientific). The elution procedure went as follows: each sample was first eluted with 0.1% FA and 98% acetonitrile for 45 min, then the linear gradient was increased from 6% to 23%, and increased again, from 23% to 36% within 15 min, after which it rose to 85% within 5 min. The elution was carried out on an EASY-n LC1000 ultra-performance liquid chromatography (UPLC) system (Thermo Scientific), at a constant flow rate of 280 mL/min, and the ensuing peptides were analyzed by a Q Exactive TM Plus hybrid quadrupole-Orbitrap mass spectrometer (Thermo Scientific, USA). Peptide fragments were separated by the UPLC system and ionized by implantation into the nanospray ionization (NSI) ion source, after which they were coupled to the online UPLC system using tandem mass spectrometry (MS/MS) in a Q Exactive TM Plus.

2.6 Bioinformatics analysis

Raw data for mass spectrometry was converted to the materials and geometry format (MGF) format, using Proteome Discoverer software, and the protein identification conducted with the Mascot search engine (v.2.3.0, Matrix Science, UK). In this study, the corresponding transcriptome of each of the four treatments was also analyzed. After de novo assembly with the Trinity software (Release-201302251), overall 220 440, 222 259, 229 711, and 225 876 unigenes were obtained from the P-replete, P-deficient, DIP-resupplied and DOP-resupplied cells, respectively (Zhang et al., 2019b). All datasets were deposited in the SRA database (BioProject ID: PRJNA522720). The putative amino acid sequences translated from the coding sequence (CDS) of unigenes served as the protein database. The following search parameters were designated: the enzyme was set to Trypsin, the maximum allowable number of missed sites was set to 2, the fixed modifier was set to carbamidomethyl (C), the variable modifier was set to oxidation (M), TMT 6-plex (N-term), and TMT 6-plex (K); the mass error setting peptide precursor ion mass tolerance was 10×10-6, with a fragment mass tolerance of 0.02 Da; the P-value was set to < 0.05 and the peptide cutoff (Score cutoff) was set to > 20; the false discovery rate (FDR) of the peptide was controlled (to be less than 1%). According to a previous study (Zhang et al., 2015), peptides at the 95% confidence interval were used for protein identification, for which every confident designation involved at least one unique peptide. The quantification of protein was performed using Mascot and normalized by the median ratio in that software. Differentially expressed proteins (DEPs) were distinguished by pairwise comparisons, with P-values ≤0.05 and a fold change ≥1.3 (up-regulated) or ≤0.77 (down-regulated) used as the threshold criteria.

To further understand the function and characteristics of the identified proteins, each was annotated accordingly into four broad categories: Gene Ontology (GO), protein domain, Kyoto Encyclopedia of Genes and Genomes (KEGG), and subcellular location. GO annotation was derived from the UniProt-GOA database (http://www.ebi.ac.uk/GOA/). Firstly, converting identified protein ID to UniProt ID and then mapping to GO IDs by protein ID. If some identified proteins were not annotated by UniProt-GOA database, the InterProScan soft was used to annotate GO function of proteins based on protein sequence alignment method. Domain functional description of identified proteins was annotated by InterProScan (a sequence analysis application) based on protein sequence alignment method, and the InterPro domain Database (http://www.ebi.ac.uk/interpro/) was used. KEGG database was used to annotate protein pathway. Firstly, KEGG online service tool KAAS was used to annotate proteins based on KEGG database descriptions. Then the annotation results were mapped to the KEGG pathway database using KEGG online service tool KEGG mapper. The subcellular localization was predicted using subcellular localization predication soft wolfpsort (version: PSORT/PSORT Ⅱ) (Feng et al., 2017).

GO and KEGG pathway functional enrichment analysis was performed by the Fisher's exact test, and a corrected P value < 0.05 was considered as significant (Feng et al., 2017). Briefly, the numbers of DEPs and all quantitative proteins in each GO term or pathway were calculated, and then the total numbers of DEPs annotated by GO terms or pathway and all quantitative proteins annotated by GO terms or pathway were calculated, finally, the four numbers were used to calculate Fisher' exact test P values and enrichment folds (Feng et al., 2017). For further hierarchical clustering analysis based on DEPs functional classification, we first collated all the categories obtained after enrichment along with their P values, and then filtered for those categories which were at least enriched in one of the clusters with P value < 0.05. This filtered P value matrix was transformed by the function x=-log10 (P value).

3 RESULT 3.1 Proteome overview

A total of 351 739 mass spectra were generated, of which 43 882 spectra were matched to 19 903 peptides with a spectrum-utilizing rate of about 12.5%. The mass error distribution of all peptides identified was near zero and most of them were less than 0.02 Da, which meant the mass accuracy of the MS data met the requirements for robust analysis (Fig. 1a), and the length of most peptides was distributed between 8 and 16 amino acids, being consistent with the known properties of tryptic peptides (Fig. 1b).

Overall, 8 348 proteins were identified from 15 553 unique peptides in P. donghaiense. To ascertain the repeatability of the experiment, Pearson's correlation coefficient and the relative standard deviation (RSD) were calculated to evaluate the repeatability of each protein's relative quantification (Fig. 1). There was a significant positive correlation between the replicates, with negative or no correlations found between the different treatments, thus indicating the quantitative repeatability was sound (Fig. 1c). The average RSD of the biological replicates was < 0.05, and all the quantiles were < 1, further indicating that the quantitative repeatability of the experimental samples was very good (Fig. 1d).

3.2 Differentially expressed proteins

According to the screening criteria for DEPs, the data of three replicates were combined according to the protein ID, to calculate three mean values: that of each protein's relative expression among the different comparison groups, the RSD, and the differentially expressed t-test statistic. In all, 1 182 proteins were significantly altered among the four treatments. Compared with the P-replete cells, 138 and 213 proteins were significantly up-regulated and down-regulated in the P-deficient cells, respectively (Supplementary Fig.S2). After the recovery of DIP and DOP, 141 and 125 proteins were up-regulated and 175 and 101 proteins were down-regulated, respectively, relative to the P-deficient cells (Supplementary Fig.S2). Despite no significant difference in the dinoflagellate's physiological response between the DIP- and DOP-resupplemented cultures (Zhang et al., 2019b), its proteomic responses nonetheless varied remarkably, including 129 up-regulated proteins and 114 down-regulated proteins (Supplementary Fig.S2).

In further exploring the biological processes associated with DEPs, the up-regulated DEPs were mainly associated with lysosome, flavone and flavonol biosynthesis, taurine and hypotaurine metabolism, and oxidative phosphorylation in the P-deficient cells when compared with the P-replete cells, while the down-regulated DEPs were found mainly involved in porphyrin and chlorophyll metabolism, ribosome, terpenoid backbone biosynthesis, chloroalkane and chloroalkene degradation, and steroid biosynthesis (Fig. 2). After the DIP resupply, processes such as porphyrin and chlorophyll metabolism, nitrogen metabolism, photosynthesis, and ribosome were up-regulated compared with the P-deficient cells, whereas taurine and hypotaurine metabolism, lysosome, and flavonoid biosynthesis were down-regulated (Fig. 2). Yet some processes, such as ribosome, nitrogen metabolism, vitamin B6 metabolism and alanine, aspartate and glutamate metabolism, were up-regulated, while lysosome, starch and sucrose metabolism, and taurine and hypotaurine metabolism were down-regulated, following the DOP resupply when compared with the P-deficient cells (Fig. 2).

Different biological responses were also discerned between the DIP- and DOP-resupplied cells. These processes were mainly related to vitamin B6 metabolism, biosynthesis of stilbenoid, diarylheptanoid, and gingerol, as well as flavonoid biosynthesis and photosynthesis (Fig. 2). This pointed to different utilization mechanisms of DIP and DOP by P. donghaiense.

3.3 P metabolism-related proteins

A total of 20 P metabolism-related proteins involved in phosphate transport, P reallocation, organic P utilization, and non-P lipid utilization were identified. Two synaptogenesis protein 1/phosphate system positive regulatory protein 81/xenotropic and polytropic retrovirus receptor 1 (SPX), N-terminal proteins related to intracellular P reallocation were up-regulated, by about 2.2-fold, in the P-deficient cells compared with the P-replete cells, but they were down-regulated by 0.7-fold after the both DIP- and DOP-resupply that lasted for 28 h (Table 1). Organic P utilization enzymes, glycerophosphodiester phosphodiesterase, and a cyclic phosphodiesterase-like enzyme were all up-regulated about 1.4-fold in the P-deficient cells; however, no significant change was detected after the P resupplementation. Additionally, two arylsulfatases involved in non-P lipid utilization were significantly up-regulated in the P-deficient cells but down-regulated significantly after the DIP resupply (Table 1).

4 DISCUSSION 4.1 Responses of P. donghaiense to ambient P deficiency

Phosphate is generally considered to be the only P form that can be directly utilized by phytoplankton (Cembella et al., 1982, 1984; Currie et al., 1986). Under P-deficient conditions, many phytoplankton species, such as the diatoms Skeletonema costatum and Thalassiosira pseudonana, enhance phosphate transport to capture more P from the surrounding environment and/or reallocate their intracellular P (Dyhrman et al., 2012; Zhang et al., 2015; Dyhrman, 2016; Lin et al., 2016). Two inorganic phosphate transporters were identified in P. donghaiense, but they varied insignificantly under the P-deficient condition (Fig. 3, Table 1), while DIP transport was not enhanced in the P-deficient cells, similar to the dinoflagellate Karenia mikimotoi (Lei and Lü, 2011). Some phytoplankton species, such as Alexandrium catenella, S. costatum, and T. pseudonana, can luxuriously absorb P from the environment and store it in vacuoles in the form of polyphosphate when the ambient P is not limiting, so that later this P can be released and utilized under P-deficient conditions (Jauzein et al., 2010; Dyhrman et al., 2012; Secco et al., 2012; Fu et al., 2013; Zhang et al., 2016). SPX-domain-containing proteins play an important role in the maintenance of phosphate homeostasis in plants (Secco et al., 2012). In our study, the SPX proteins were significantly up-regulated in the P-deficient cells (Fig. 3, Table 1), indicating that intracellular P reallocation was enhanced and intracellular P was an important P source for P-deficient cells, similar to the diatoms S. costatum and T. pseudonana (Dyhrman et al., 2012; Zhang et al., 2016). Organic P, another vital P source in the ocean, can also be used by most phytoplankton species (Dyhrman et al., 2012; Lin et al., 2016; Zhang et al., 2016, 2018). Under P deficiency, many phytoplankton species can use intracellular and extracellular organic P (Dyhrman et al., 2012; Zhang et al., 2016, 2018; Gong et al., 2017; Luo et al., 2017). However, the utilization mechanisms of organic P among different phytoplankton species are poorly understood. Phospholipids, including glycerophospholipids and sphingolipids, are crucial components of cell membranes (Paolo and Suzanne, 2009). In our study, glycerophosphodiester phosphodiesterase was up-regulated in the P-deficient cells (Fig. 3, Table 1). The most prominent function of phospholipase is the degradation of membrane lipids; yet phospholipase is also involved in vesicle trafficking and signal transduction, and it may act as a link between cell membranes and proteins (Munnik and Musgrave, 2001; Bargmann and Munnik, 2006). Our results suggested that membrane phospholipids might be a critical intracellular organic P source for P-deficient cells (Fig. 3), similar to S. costatum (Zhang et al., 2016). Some phytoplankton species, such as A. catenella, S. costatum, and T. pseudonana, can use non-P lipids instead of phospholipids to reduce the P requirement of their cells under P-deficient conditions; for example, some diatoms and cyanobacteria can utilize sulfolipid and betaine lipids (Yu et al., 2002; Van Mooy et al., 2009; Jauzein et al., 2010; Dyhrman et al., 2012; Fu et al., 2013; Zhang et al., 2016). P deficiency significantly increases the sulfolipid metabolism of P. donghaiense at both the transcriptional (Shi et al., 2017) and protein level (Zhang et al., 2019b), which implies that P. donghaiense could reduce its P requirement by utilizing sulfolipid in response to an ambient P deficiency. Furthermore, our results indicated that the cells' P requirement under the ambient P-deficiency condition was mainly derived from intracellular polyphosphates, phospholipids, and organic P, with organic P becoming the main P source once intracellular P has been exhausted (Fig. 3). In this study, however, we did not detect significant changes in endocytosis at the protein level (Fig. 3) when compared with the transcriptional level (Zhang et al., 2019b).

Energy fuels the the life activities of all organisms. In this study, ATP synthase subunits (such as β and δ) were down-regulated in the P-deficient P. donghaiense cells compared with the P-replete cells (Supplementary Table S2), similar to A. catenella (Zhang et al., 2014, 2019a). Glycolysis can release a large amount of ATP, which is the oldest and most primitive way for organisms to obtain energy (Harris, 2013). Phosphofructokinase (PFK), phosphoglycerate kinase (PGK), and pyruvate kinase (PK) are all involved in glycolysis and they were down-regulated under P deficiency (Supplementary Table S2). PFK is a rate-limiting enzyme for glycolysis (TeSlaa and Teitell, 2014), while PGK and PK catalyze two reactions to produce ATP (Plaxton, 1996; Harris, 2013). However, we found that the Ⅴ-type H+ ATPase was significantly up-regulated (P < 0.05) under P deficiency (Supplementary Table S2). In plants, the Ⅴ-type H+ ATPase obtains energy by hydrolyzing ATP, transferring H+ across the membrane to form electrochemical gradients inside and outside the membrane, which provides electrochemical potential energy for the membrane transport system of other substances (Beyenbach and Wieczorek, 2006). Our results show that P deficiency inhibited the production of ATP in P. donghaiense, which might have led to insufficient energy for protein, lipid, and nucleic acid synthesis, and cell growth (Zhang et al., 2018, 2019b). In addition, enzymes localized in the lysosome, such as deoxyribonuclease Ⅱ and iduronate 2-sulfatase, were significantly up-regulated in the P-deficient cells (Supplementary Table S2). Under nutrient-deficient conditions, plant autophagy is induced, and macromolecular substances such as intracellular proteins are degraded in the lysosome, while intracellular nitrogen and carbon sources are reused to sustain biosynthesis, energy metabolism, and the reactivation of nutrients (Doelling et al., 2002; Bassham, 2009). Our results indicated that P. donghaiense engaged in multiple strategies to adapt to the ambient DIP deficiency and mainly used organic P as its P source under a P-deficient condition.

4.2 Different response of P. donghaiense to DIP and DOP resupply

Our previous study revealed that the physiological responses of P-deficient P. donghaiense to DIP and DOP resupply were not significantly different, but their transcriptome responses did differ (Zhang et al., 2019b). In the current study, significant differences were also observed at the proteome level (Fig. 2).

4.2.1 Photosynthesis

The paramount feature of phytoplankton is their ability to synthesize organic matter using light energy via photosynthesis (Falkowski and Raven, 2013). After the DIP or DOP resupply, cytochrome b6 and f in P. donghaiense were up-regulated only in the DIP-resupplied cells (Fig. 4, Supplementary Tables S3- S6), a result opposite to that found for A. catenella (Zhang et al., 2015). However, there was no significant difference in porphyrin and chlorophyll metabolism between cells in the DIP and DOP resupply groups (Fig. 4, Supplementary Tables S3-S6). The cytochrome b6f complex is a membrane protein that is essential for the light reaction in plants (Cramer et al., 2013). Light energy harvested by chlorophyll and other pigments during the light reaction drives the transfer of photosynthetic electrons through photosystems Ⅰ and Ⅱ (Gooch, 2011). Once the electrons are finally transferred to nicotinamide adenine dinucleotide phosphate (NADPH) and ATP, this completes conversion of light energy to chemical energy and this energy is now available for the carbon fixation reaction (Gooch, 2011; Walker et al., 2014). Both NADPH and ATP synthesis require P, and DIP is more easily utilized by phytoplankton than DOP, so P. donghaiense expressed more cytochrome b6 and f to sustain the electron transfer. Our results showed that DIP is more conducive than DOP for recovering photosynthetic energy production in P. donghaiense.

4.2.2 Secondary metabolism

Plant secondary metabolism is a metabolic process derived from primary metabolism, which figures prominently in plants' environmental adaptation, especially for their interaction with and defense response against bacterial pathogens (Erb and Kliebenstein, 2020). Compared with the P-deficient P. donghaiense cells, the secondary metabolites, such as flavonoid, stilbenoid, diarylheptanoid and gingerol, which are related to plant defense response and scavenging free radicals, were significantly down-regulated after the DIP resupply (Fig. 2, Supplementary Table S3). However, no significant changes were observed in these processes after the DOP resupply. Nitrogen and P deficiency affect the content of flavonoids in plants, and P deficiency increases the content of flavonoid in tomato at the early stage of fruit ripening (Stewart et al., 2001). In our study, DIP was evidently more conducive than DOP for reducing the content of flavonoids and other secondary metabolites in P-deficient P. donghaiense cells. Since DIP was more easily utilized by P. donghaiense than DOP and also restored by cellular environmental defenses, more energy could be saved.

4.2.3 Vitamin B6 metabolism

Vitamin B6 is a cofactor for many metabolic enzymes—especially amino acid metabolism—and a strong biological antioxidant, which can effectively quench singlet oxygen and superoxide, and exert anti-abiotic stress effects in plants (Szydlowski et al., 2013). Pyridoxine biosynthesis protein (PDX), being a rate-limiting enzyme in the synthesis of vitamin B6, is mainly active in the plasma membrane and inner membrane system, where it modulates the functioning of the plant cell membrane (Chen and Xiong, 2005). After the DOP resupply, we found that PDX was down-regulated compared with the P-deficient cells (Fig. 4, Supplementary Table S5), whereas the DIP-resupplied cells went unchanged. When plants are exposed to strong light, low temperature, or other abiotic stresses, PDX coding genes are usually up-regulated (Herrero and Daub, 2007). Overexpression of PDX in Arabidopsis can increase the content of vitamin B6, prolong the growth period of cells, increase cell volume, and augment resistance to oxidative stress and other environmental stresses (Raschke et al., 2011). Therefore, DOP increased the P-deficient P. donghaiense cells' resistance more than DIP did, by enhancing the synthesis of vitamin B6.

4.2.4 Amino acid metabolism

Protein is the main carrier of biological function, and amino acids are the basic units of all proteins (Morot-Gaudry et al., 2001; Hildebrandt et al., 2015). In our experiment the alanine, aspartate, glutamate, cysteine and methionine metabolisms were each up-regulated in the DOP-resupplied cells, yet no significant change (P > 0.05) was evident after the DIP resupply (Fig. 4, Supplementary Tables S4-S6). In comparison with the P-deficient cells, S-adenosylmethionine (SAM) synthetase and methionine S-adenosyl transferase—two key enzymes of SAM synthesis—and glutamate synthase (GS) were all significantly up-regulated (P≤0.05) following the DOP resupply (Fig. 4, Supplementary Table S5). Because SAM can improve cell metabolism and ensure normal mitochondrial function and membrane fluidity, its accumulation plays a key role in regulating cell membrane stability and reactive oxygen metabolism balance, and for shoring up resistance in plants to P stress (Lu, 2000; Moffatt and Weretilnyk, 2001). GS, a key enzyme involved in ammonia assimilation, catalyzes ammonia and glutamic acid to form glutamine, which is the first step of ammonia assimilation metabolism (Temple et al., 1998; Lea and Miflin, 2003). Glutamine is a nitrogen donor for the biosynthesis of amino acids, nucleotides, and chlorophyll (Kan et al., 2015), but nitrogen metabolism was up-regulated in both DIP- and DOP-resupplied cells (Fig. 4, Supplementary Tables S4-S7). This result indicated that the actual form of P did not affect nitrogen metabolism in spite of (or because of) the importance of nitrogen to cells.

Overall, our study shows that those metabolic processes directly requiring P, such as photosynthesis and cellular environmental defense, respond rapidly to DIP resupply while they varied insignificantly in the DOP-resupplied cells (Supplementary Tables S4- S7). The metabolic processes that responded to DOP resupply were mainly those dependent on C, such as amino acid metabolism (Fig. 4). Collectively, these results indicate that intracellular DOP assimilation provides not only P but also C for the biosynthesis of various substances essential for cell growth, and the coupled utilization of P and C triggered by DOP might be an important reason resulting in the occurrence of P. donghaiense blooms in a P-deficient environment.

5 CONCLUSION

Understanding the adaptation and response of marine dinoflagellates to change in ambient P is always an ecological endeavor. In this quantitative proteomic study, we found that P. donghaiense initiated multiple strategies to adjust to an ambient P-deficiency, such as intracellular P reallocation and organic P and non-P lipid utilization, but, interestingly, it did not enhance its inorganic phosphate transport. This suggests P. donghaiense preferred to use DOP rather than DIP under ambient P-deficient conditions. Although no significant difference in their physiological response was observed between the DIP- and DOP-resupplied cells (Zhang et al., 2019b), their proteomic response differed remarkably: photosynthesis was up-regulated and secondary metabolism was down-regulated only in the DIP-resupplied cells, while alanine, aspartate and glutamate metabolism, cysteine and methionine metabolism, and vitamin B6 metabolism were all up-regulated in the DOP-resupplied cells, indicating different utilization mechanisms of DOP and DIP by the dinoflagellate cells. When compared with DIP, a more efficient utilization of DOP might be an important reason why bloom occurrences of P. donghaiense occur under P-deficient conditions. Further in-situ proteomic research is necessary, especially in combination with the analysis of DOP composition and concentration in natural marine environments during the bloom period of P. donghaiense, to help fully unveil the role of DOP in the bloom formation of P. donghaiense.

6 DATA AVAILABILITY STATEMENT

The datasets that support the findings of this study are available from the corresponding author on reasonable request.

Electronic supplementary material

Supplementary material (Supplementary Tables S1-S7 and Figs.S1-S2) is available in the online version of this article at https://doi.org/10.1007/s00343-021-1030-0.