Effects of phosphorus-induced changes on the growth, nitrogen uptake, and biochemical composition of Pavlova pinguis and Hemiselmis cf. andersenii

The understanding of the phosphorus-induced changes in the biochemical composition of microalgae is of great importance for achieving efficiency in high-value lipid production. To study the chemoplasticity of Pavlova pinguis (Haptophyceae) and Hemiselmis cf. andersenii (Cryptophyceae), their growth, carotenoid and chlorophyll a content, and their monosaccharide and lipid profiles were analyzed against several phosphorus (P) regimes: low (LP), medium (MP), and high (HP). For both microalgal cultures, increasing initial P concentrations showed a positive effect on biomass productivities. Carbon-rich pools presented significant differences (p< 0.05) for P. pinguis against P treatments, in contrast to H. cf. andersenii. Differential responses to P-induced changes in microalgae monosaccharide and lipid profile were observed. Hemiselmis cf. andersenii increased its proportion in galactose (up to 3 times) from LP to HP conditions, whereas P. pinguis decreased (up to 20%) its glucose proportion from LP to HP conditions. For P. pinguis, the lowest amount (13.12 mg g−1 dw) of sterols was observed at LP conditions, in contrast to its carotenoid content (4.32 mg g−1 dw). P-replete conditions were the most effective in inducing high-value lipid accumulation. Non-targeted lipid analysis revealed which samples would need to be processed to fully exploit its high-value lipids, namely H. cf andersenii under MP and HP conditions. This study demonstrated that P played an important role in carbon allocation, nitrogen uptake, and lipid regulation on P. pinguis and H. cf. andersenii, and that P-replete conditions could be useful for optimizing high-value lipids with potential for nutraceutical and pharmaceutical fields.


Introduction
Increasing concerns regarding consumer safety, environmental sustainability, and regulatory issues over synthetic materials have turned consumers' attention to natural products (Vieira et al. 2020). Microalgae are often presented as promising sustainable cell factories for their ability to convert atmospheric carbon dioxide, water, inorganic nutrients, and sunlight to high-value compounds (Fu et al. 2016). These organisms are primary producers of essential nutrients that perform vital functions in the human organism and whose dietary intake is mandatory (Khozin-Goldberg et al. 2011). Improvement of microalgae production often involves the application of several stress-inducement strategies from which nutrient stress is the most employed due to its low cost and easy applicability at both lab and large-scale cultivation (Singh et al. 2016).
Phosphorus (P) plays a significant role in algal growth and cell division due to its role as an essential component of nucleic acids, phospholipids, and phosphorylated sugars (Roopnarain et al. 2014). Besides algal growth, P is involved in metabolic processes like signal transduction, photosynthesis, and energy transfer (in the form of adenosine triphosphate (ATP) and nicotinamide adenine dinucleotide phosphate (NADP + ) (Roopnarain et al. 2014;Yaakob et al. 2021). Due to its fast recycling, P limitation has been observed in microalgae natural environments (Alipanah et al. 2018). Microalgal P requirements are known to vary considerably between species ranging from 0.001 to 0.179 g L −1 (Roopnarain et al. 2014). This macronutrient can be uptaken by microalgae in the form of orthophosphate (most commonly used; PO 3− 4 ) and polyphosphate (Alipanah et al. 2018;Yaakob et al. 2021). Moreover, in P-replete conditions, algae can store excess P in polyphosphate bodies (luxury uptake) (Grobbelaar 2013).
To cope with P limitation, microalgae modulate several metabolic pathways affecting their biomass composition (Alipanah et al. 2018). From the main macromolecular pools, lipids and carbohydrates are especially affected by internal and external P supplies (Grobbelaar 2013). According to Markou et al. (2012), carbohydrate accumulation due to P-starvation is reported to occur in diatoms, green algae, and cyanobacteria. Nevertheless, a different response was reported to occur in the Haptophyta Isochrysis galbana in which P-starvation stimulated lipid accumulation by up to 50% (Yaakob et al. 2021).
Phytosterols, polyunsaturated fatty acids (PUFA), and lipid-soluble compounds, like carotenoids and tocopherols, are important for a variety of nutraceutical and pharmaceutical purposes, and their market is continuously growing (Vieira et al. 2020). These compounds are wellknown for their health benefits in humans including neuroprotective, anti-inflammatory, anti-cancer, and antioxidant activities as well as their action as chemoprotective agents in cardiovascular diseases (Luo et al. 2015;Mudimu et al. 2017;Harwood 2019). Most studies evaluating the effects of P stress on microalgae have focused on the changes of specific classes of lipid families such as fatty acids and sterols (Piepho et al. 2010(Piepho et al. , 2012. Nonetheless, a non-targeted lipophilic characterization was applied by Fernandes et al. (2021) focusing on nitrogen-induced changes instead of phosphorus.
Although its wide taxonomic diversity and metabolic differences, studies on the effect of P on microalgal lipid profile have been limited to few species most of them belonging to the Chlorophyta phylum (Cañavate et al. 2017). In previous studies, the haptophyte Pavlova pinguis ) and the cryptophyte H. cf. andersenii Fernandes et al. 2021) have demonstrated to be promising candidates for highvalue lipid production. However, the mechanisms that led to carbon allocation and lipid remodeling in response to environmental-induced perturbations are poorly explored for these two microalgae. Therefore, to increase knowledge on the chemoplasticity (ability of microalgae to change its chemical composition in response to extrinsic stimuli by reorganizing its metabolome) of P. pinguis and H. cf. andersenii, their carotenoid and chlorophyll a content, and monosaccharide and lipid profiles were analyzed against several phosphorus regimes.

Growth and culture conditions
The marine haptophyte Pavlova pinguis (RCC 1539) was obtained from the Roscoff Culture Collection (RCC), whereas the marine cryptophyte Hemiselmis cf. andersenii (BEA 0118B) was obtained from the Spanish Bank of Algae (BEA). The microalgal cultures were started with a 1:10 inoculum from exponentially growing cultures into 1 L of media as in Sun et al. (2018), with initial cell concentrations of 0.54 × 10 6 cells mL −1 for H. cf. andersenii, and 1.29 × 10 6 cells mL −1 for P. pinguis. Microalgal cell cultures were grown on 1 L of sterile medium with pH adjusted to 7.0 under 70 μmol photons m −2 s −1 , light intensity with 16:8 h (light:dark cycles) at 25 °C as in . For the phosphorus-induced changes, f/2-Si medium (Guillard and Ryther 1962;Guillard 1975) was used as the basal formula (Table 1), and compressed air (100 mL min −1 ) was used for aeration of cell cultures and as a carbon source. The different phosphorus regimes were performed by adjusting sodium dihydrogen phosphate monohydrate (NaH 2 PO 4 .H 2 O) concentrations to 3.62 µM (low phosphorus; LP), 36.20 µM (medium phosphorus; MP), and 72.40 µM (high phosphorus; HP). The N:P molar ratios for LP, MP, and HP treatments were 244:1, 24:1, and 12:1, respectively. The LP concentration was chosen to trigger a reduced phosphorus supply close to phosphorus depletion without compromising biomass Table 1 Culture medium composition of f/2-Si medium (Guillard and Ryther 1962;Guillard 1975 production. The HP level was chosen to study the effects of excess phosphorus in a tolerable supplementation in microalgae. At the end of the early stationary phase, microalgae were harvested by centrifugation for 7 min. at 3720 × g and the pellets washed twice with a 0.09 g L −1 NaCl solution. As in Fernandes et al. (2021), the harvest days were determined as those on which the cell concentration was similar for 2 to 3 consecutive counting days, supported by non-significant differences (p > 0.05). Algal growth was monitored daily by measuring the optical density of the microalgal culture at 550 nm with an ultraviolet/ visible spectrophotometer (UV-6300PC, VWR, China). A calibration curve (R 2 = 0.999, P. pinguis; R 2 = 0.993, H. cf. andersenii) plotting cell concentration (cells mL −1 ) against absorbance was used to determine cell growth. The calibration curve was formed under normal conditions, MP conditions (equivalent to the initial P concentrations in f/2-Si growth medium). The cell concentration was estimated by counting cells with a 0.1 mm deep improved Neubauer hemocytometer (Marienfield Superior) in a light microscope (Olympus BX41). This method can bring some inaccuracies specifically when cell morphology, shape, size, and content of scattering substances change (Becker 1994;Lee et al. 2013). To minimize these sources of error, the wavelength of 550 nm is recommended by Becker (1994), namely, to avoid the range of absorption of photosynthetic pigments (Samiee-Zafarghandi et al. 2018). The logistic model of Xin et al. (2010) was used to describe the microalgal growth (Eq. (1)): where K (cells mL −1 ) is the carrying capacity, X (cells mL −1 ) is the cell concentration in time t (day −1 ), a is a constant that refers to the position of the origin, and r (day −1 ) is the specific growth rate. Fitting of the model to both microalgae raw data was performed with Solver GRG (generalized reduced gradient) nonlinear solving method in Microsoft Excel selecting a, r, and K as decision variables.

-N ) determination
Nitrogen (N) was determined as nitrate ( NO − 3 ) by the ultraviolet spectrophotometric screening method as reported by Wan et al. (2013). A calibration curve (R 2 = 0.998) was performed using 220 nm optical density against nitrogen concentrations (0-4.07 mg L −1 ) from NaNO 3 solutions. Before nitrogen determinations, liquid samples were filtered using a 0.45 μm filter, diluted with deionized water, and acidified with 1 M HCl. Absorbances were read with an ultraviolet/ visible spectrophotometer (UV-6300PC, VWR, China). The N removal (%) was calculated as follows: (1) where N i is the initial nitrogen concentration (mg L −1 ), N f is the final nitrogen concentration (mg L −1 ).

Pigment determination
Carotenoids and chlorophyll a extraction and determination were performed according to Fernandes et al. (2021). For extraction, 10 mg of microalgae dried biomass was used and 7 mL of a cold 80% acetone solution was added. After homogenization, the mixture was put on ultrasounds for 90 min. Then, samples were centrifuged, supernatant transferred to pre-chilled tubes, pellets continuously washed with extraction solution, and supernatant was passed through 0.45 µm filters. For chlorophyll a (Chl a) and carotenoid (Car) determination through an ultraviolet-visible spectrophotometer, the following equations used by Chen and Vaidyanathan (2013) were employed:

Lipid extraction
Lipid extraction was performed at the end of batch cultures. Microalgal biomass lipid extraction was performed accordingly with a modified Bligh and Dyer (1959) described by Fernandes et al. (2016). In brief, to 50 mg of dried microalgal biomass, additions of 3 mL of a methanol:chloroform mixture (2:1; v:v), 400 μL of a saturated solution of potassium chloride (KCl), 2 mL of chloroform, and 2 mL of distilled water were performed. Then, the organic phase was removed and transferred to pre-weighted tubes, and lipids were determined by gravimetric quantification.

Alkaline hydrolysis
Alkaline hydrolysis was performed according to Santos et al. (2015). Thus, to two lipid aliquots, 0.5 M of NaOH in aqueous methanol was added and left to react at 100 °C for 1 h in a nitrogen atmosphere. Acidification to pH 2 was performed with 1 M HCl, and the hydrolysis products were further extracted with dichloromethane. (2)

Trimethylsilyl derivatization and GC-MS analysis
Before gas chromatography-mass spectrometry (GC-MS) analysis, lipid aliquots before and after alkaline hydrolysis were derivatized to trimethylsilyl (TMS) ethers and/or esters according to Santos et al. (2015). Following the addition of tetracosane (internal standard) to the lipid extracts before and after alkaline hydrolysis, additions of 250 μL of pyridine, 250 μL of N,O-bis(trimethylsilyl)trifluoroacetamide, and 50 μL of trimethylchlorosilane were performed. Then, samples were incubated at 70 °C for 30 min. GC-MS analysis was performed using a gas chromatographer (6890 N, Agilent Technologies, China) equipped with a mass selective detector (5973Network, Agilent Technologies, USA) and a ValcoBond capillary column from VICI Valco (30 m × 0.25 mm inner diameter, 0.25 μm film thickness) using the conditions described by . The TMS derivatives were identified by comparing the retention times and mass spectra fragmentation with those obtained through injection of the standards and literature. Semi-quantitative analysis was made through the calculated response factor of each standard towards the internal standard (tetracosane). The standards used were mannose, trans-ferulic acid, nonadecan-1-ol, eicosan-1-ol, 5α-chlolestane, cholesterol, stigmasterol, and hexadecenoic and nonadecanoic acids. Four replicates were performed for each GC-MS analysis with the results being presented as the mean value ± standard deviation (SD) expressed in milligrams per gram of dry biomass weight (dw).

Monosaccharide reduction and derivatization
For monosaccharide analysis, two-stage acidic hydrolysis followed by monosaccharide reduction and acetylation was performed according to modified Blakeney et al. (1983) and described by Fernandes et al. (2017). Briefly, to 10 mg of sample, 400 µL of 72% sulfuric acid was added and left for 3 h at 20 °C. Next, water was added, and mixtures placed at 100 °C for 2.5 h. At the end of acidic hydrolysis, the internal standard (myo-inositol) was added to the hydrolysate. The reduction of monosaccharides to alditols was performed to 1 mL of hydrolysate by addition of 200 µL of a solution of 25% ammoniac (NH 3 ) solution and 100 µL of 3 M NH 3 containing 150 mg mL −1 of sodium borohydride. The mixtures were incubated at 30 °C for 1 h, and 50 µL of glacial acetic acid was added twice to the samples with homogenization in between. Alditol acetylation was performed by adding 450 µL of 1-methylimidazole and 3 mL of acetic anhydride to 300 µL of the previous mixture. Then, the solution was incubated at 30 °C for 30 min. Following water addition (3.75 mL), alditol acetates were extracted with dichloromethane (2.5 mL) and further washed three times with water.

GC-MS analysis
Monosaccharides were analyzed as alditol acetates by gas chromatography (Agilent HP 6890) mass spectrometry (Agilent 5973Network) and a capillary column (DB-225 J & W; 30 m × 0.25 mm inner diameter, 0.15-μm film thickness) from Agilent. After optimization, the GC-MS conditions were as follows: the inlet temperature was 220 °C and initial column temperature was 210 °C for 3 min. Then, the column temperature increased at a rate of 1 °C min −1 until reaching 220 °C and 10 °C min −1 until 230 °C which was kept for 2 min. The transfer line temperature was 240 °C, the split ratio was 50:1, and helium was used as the carrier gas with a ramped flow rate of 1.0 mL min −1 for 6 min increasing to 1.2 mL min −1 and remaining constant until the end of the run. The derivatized monosaccharides were identified by comparing the retention times and mass spectra fragmentation with those obtained through injection of the standards. The monosaccharide quantification was made through the calculated response factor of each standard towards the internal standard (myo-inositol). The standards used were myo-inositol, L( +)arabinose, D( +)xylose, D( +)galactose, D( +)glucose, D( +)mannose, monohydrate D( +)rhamnose, and D( +)fucose purchased at Sigma-Aldrich (USA). Four replicates were performed for each GC-MS analysis with the results being presented as the mean value ± SD of monosaccharides expressed in mg g −1 dw.

Statistical analysis
Analysis of variance (ANOVA) followed by Tukey post hoc analysis was performed using IBM SPSS Statistics 26 software, p values < 0.05 were considered statistically significant, whereas for principal component analysis (PCA), Pearson correlation, and hierarchical clustering heat maps, the Metaboanalyst 5.0 webserver (Pang et al. 2021) was used. Before PCA, Pearson correlation, and hierarchical clustering heat maps, features were median normalized, log 2 transformed, and auto scaled. For hierarchical clustering heat maps, Euclidean clustering distance and ward clustering method were used.

Microalgal growth and chemical composition
Phosphorus is an essential nutrient for the growth and division of microalgal cells (Roopnarain et al. 2014). Growth dynamics of Hemiselmis cf. andersenii (Fig. 1a) and Pavlova pinguis ( Fig. 1b) were analyzed under different initial phosphorus regimes. In Fig. 1, it is possible to observe that microalgal growth varied with the different initial phosphorus supplies, with microalgae at LP conditions presenting lower cell harvest density (Table 2) and reaching early stationary phase more quickly than microalgae grown at MP and HP conditions. Biomass productivity determines the prospective of microalgae for efficient high-value compounds. According to Patel et al. (2012), higher P loadings are often associated with higher biomass accumulation and productivity. For H. cf. andersenii and P. pinguis cultures, significant differences (p < 0.05) were observed for biomass productivity across P treatments, with increasing phosphorus concentrations exhibiting a positive effect on biomass productivities, with the greatest values being achieved at HP conditions 30.05 and 43.91 mg L −1 day −1 respectively. These values are greater than those reported by Slocombe et al. (2015) of 13 mg L −1 day −1 for another cryptophyte (Rhodomonas reticulata).
In H. cf. andersenii, nitrogen uptake was affected by initial phosphorus concentrations in growth media. Figure S1a shows that at day 2 the nitrogen uptake by H. cf. andersenii under LP conditions ceased despite only 41% of nitrogen had been removed. This indicates that a nutrient other than nitrogen has become limiting, influencing not only nitrogen acquisition but also the beginning of the stationary phase verified at day 3 of cultivation ( Fig. 1a). Since this occurs in H. cf. andersenii cultures with a reduced initial phosphorus concentration, these results suggest that phosphorus was the limiting nutrient at these conditions, instead of nitrogen. Values are expressed as mean ± standard deviation, n = 3. The determination coefficients (r 2 ) for the growth models were higher than 0.90. * Values are not significantly different (p > 0.05) among cultivation days, within each treatment Different results were obtained for H. cf. andersenii cultures under higher phosphorus levels where no variations (p > 0.05) on nitrogen removal were verified for the treatments MP and HP (Table 2).
For P. pinguis, no differences were found for the nitrogen removal efficiency (Table 2; Fig. S1b) among treatments. These results suggest that increasing initial phosphorus concentrations, which in turn decreases N:P ratios, has no direct effect on nitrogen uptake by P. pinguis. These observations reinforce the assumption that the nutrient uptake ability of microalgae depends not only on its availability, environmental conditions, and nutrient ratios but also on biological factors such as species (Grobbelaar 2013).
Carotenoids (photoprotective pigments) to chlorophyll a (light-harvesting pigment) ratio (Car/Chl a) can serve as an indicator of carotenogenesis, and as a marker of physiological stress in microalga under poor environmental conditions (Chen et al. 2017;Alipanah et al. 2018;Jo et al. 2020). For this ratio, significant variations (p < 0.05) under different initial phosphorus supplies for both microalgae studied were observed (Fig. 2). For H. cf. andersenii, LP conditions (light green; Fig. S2a) presented low carotenoids (0.61 mg g −1 dw), Chl a (1.72 mg g −1 dw), and Car/Chl a (0.36) ratio values, whereas for P. pinguis, a carotenoid accumulation was observed through an increase of 54% from MN conditions (2.81 mg g −1 dw) to LP conditions (4.32 mg g −1 dw) and with the highest Car/Chl a (0.77) ratio being observed at these conditions. The low pigments (chlorophyll a and carotenoids) accumulation verified for H. cf. andersenii under LP conditions might be derived from the combined effect of phosphorus limitation and reduced nitrogen uptake (Table 2; Fig. S1a), affecting not only cell division but also the photosynthetic efficiency of this microalga.
Color differences were verified between the microalgae exposed to distinct initial phosphorus concentrations (Fig. S2 a-b). In the last days of cultivation, H. cf. andersenii LP cultures (Fig. S2a) presented a light green, whereas the same microalga at MP and HP conditions displayed a dark green. These visual changes reflect the lowest pigments amount verified at LP conditions, in contrast to H. cf. andersenii under MP and HP conditions. For P. pinguis (Fig. S2b), LP cultures presented a yellowish color, whereas MP and HP cultures presented a brown color. A previous study (Gorgônio et al. 2013) associated the yellowish color of a haptophyte with higher levels of carotenoids, this observation being consistent with the results obtained in this study.
Microalgae-derived carotenoids have a high market potential due to their several biological activities such as antioxidant activity, hypolipidemic, anti-inflammatory, and hypotensive (Le Goff et al. 2019). Therefore, the results obtained for carotenoids and Car/Chl a ratio indicate that phosphorus limitation constitutes a cost-efficient strategy for carotenoid production by P. pinguis for high-value market. Similar behavior to P. pinguis was observed for other haptophytes (Isochrysis galbana) in which P-starvation initiated chlorophyll reduction and carotenoid accumulation (Roopnarain et al. 2014). The accumulation of carotenoids observed for P. pinguis may have been a protective mechanism against photo-oxidative stress prompted by the LP conditions (Roopnarain et al. 2014). Microalgae fixed carbon can be distributed among three main macromolecular pools: protein, carbohydrate, and lipid . In general, microalgae use carbohydrate as their primary carbon storage pool switching their carbon partitioning towards energy-rich storage compounds like lipids under poor environmental conditions ). However, this partition among carbon-rich pools (carbohydrate and lipid) can be species-specific (Fernandes et al. 2016). For H. cf. andersenii, the lipid/monosaccharide ratio remained stable across treatments with non-significant differences (p > 0.05) being observed for this parameter (Table 2). In this microalga, the lipid/monosaccharide ratio lower than one seems to indicate that carbohydrates are the main carbon-rich pool for this microalga. For P. pinguis, the lipid/monosaccharide ratio was significantly (p < 0.05) affected by initial phosphorus supplies reaching its lowest value (1.26) at LP conditions, suggesting that at these conditions P. pinguis displaced its carbon partitioning towards carbohydrate synthesis (Table 2). Moreover, an opposite trend between the Car/Chl a ratio and the lipid/monosaccharide ratio suggests that the Car/Chl a ratio could be a simple indicator of the carbon partitioning among lipid and carbohydrate.
A convertible relationship between carbohydrate and lipid has been reported to occur in other marine microalgae such as the haptophyte Isochrysis sp., the cryptophyte Rhodomonas marina, and the ochrophyte Nannochloropsis gaditana, as a response to environmental fluctuations (Fernandes et al. 2016(Fernandes et al. , 2017. In the study performed by Fernandes et al. (2016), Isochrysis sp., which belongs to the same phylum of P. pinguis, also accumulated carbohydrates instead of lipid towards low nutrient conditions. However, this study targeted the change of overall nutrient availabilities, by adjusting the volume of a commercial nutrient solution, against carbon supply, which remained constant. Improvements of carbohydrate production though phosphorus limitation was reported for the cyanophyte Arthrospira platensis (Markou et al. 2012).
As shown in Table 2, P. pinguis displayed the more complex mixture of monosaccharides presenting six identified compounds: rhamnose, arabinose, xylose, mannose, galactose, and glucose. For both H. cf. andersenii and P. pinguis, glucose and mannose were the main building blocks of carbohydrate accounting for 97-98% of total monosaccharides and 93-96% of total monosaccharides, respectively. Looking at the relative abundance of monosaccharides, it is possible to observe that H. cf. andersenii and P. pinguis remodeled differently their monosaccharide composition in response to initial phosphorus inputs. Thus, H. cf. andersenii increased its proportions in galactose (threefold) from LP conditions towards HP conditions (Fig. 3a), whereas P. pinguis decreased by 20% its glucose relative abundance from LP conditions towards HP conditions (Fig. 3b). The main storage product of Haptophyta is chrysolaminaran which is a β-1,3glucan ( Barsanti and Gualtieri 2006). Thus, the verified lowest lipid/monosaccharide ratio along with the highest monosaccharide content and glucose proportion indicates that P. pinguis responded to LP conditions by synthesizing this storage product. P-starvation is often referred as an effective strategy to enhance lipid accumulation by microalgae (Satpati et al. 2016;Yang et al. 2018;Yaakob et al. 2021). Table 2 shows that lipid content was significantly (p < 0.05) affected by the different phosphorus regimes applied in both microalgae studied. For H. cf. andersenii, lipid levels ranged between 12.58 and 20.48% dw, and for P. pinguis, lipid amount varied between 21.87 and 25.11% dw. In both microalgae, the greatest lipid values were observed at MP conditions, while the lowest values were found at LP conditions for H. cf. andersenii and at HP conditions for P. pinguis. According to Saha et al. (2013), most eukaryotic microalgae proportionally accumulate lipids and carotenoids as a response to environmental stresses. In the present study, H. cf. andersenii showed a similar trend for both lipid and carotenoids as a response to phosphorus-induced changes.
In contrast with previous studies (Khozin-Goldberg and Cohen 2006; Satpati et al. 2016;Alipanah et al. 2018), LP conditions did not cause an increase in the lipid content in both microalgae. Additionally, H. cf. andersenii responded to LP conditions by decreasing its lipid content. The same observation was made by Reitan et al. (1994) where two (Nannochloris atomus and Tetraselmis sp.) from the seven microalgae studied responded to phosphorus limitation by decreasing its lipid content. The way by which microalgae biochemical composition is modulated in response to an external stimulus is dependent on several factors, namely species-specific differences, composition of the media, physical parameters (e.g., temperature, light intensity, CO 2 aeration), and methods of cultivation (Akgül et al. 2021).

Lipid remodeling
Several studies (Cañavate et al. 2017;Alipanah et al. 2018;Murakami et al. 2018;Kokabi et al. 2019) have reported the reorganization of microalgal cellular lipids under phosphorus-induced changes. Thus, to study the impact of phosphate-induced changes on H. cf. andersenii and P. pinguis lipid composition, an untargeted approach analyzing its esterified and unesterified composition was applied. Through Tables 3 and 4, it is possible to observe that H. cf. andersenii and P. pinguis displayed different patterns of lipid remodeling as a response to phosphorus-induced changes. For all lipid sets analyzed (fatty acids, aliphatic alcohols, sterols, monoglycerides, and other compounds), significant differences (p < 0.05) were observed across treatments in both microalgae studied.
As building blocks of lipids, fatty acids were the most affected by alkaline hydrolysis, increasing up to 43% in H. cf. andersenii at HP conditions and up to 41% in P. pinguis at LP conditions. Within fatty acids, polyunsaturated fatty acids (PUFA) presented the highest increase after hydrolysis for H. cf. andersenii samples (up to 81% at HP conditions), whereas for P. pinguis, saturated fatty acids (SFA) presented the greatest increase (up to 46% at LP conditions). Since PUFA are often attributed to polar lipids, whereas SFA are mainly enriched in triacyclglycerols (Shen et al. 2016), this may indicate that P. pinguis has a richer composition in neutral lipids (triacylglycerols) than H. cf. andersenii.
The production of enhanced amounts of nutritionally important fatty acids by microalgae would have added value as an ingredient for various functional food, cosmetics, and nutraceutical products (Saha et al. 2013). Figure 4 shows the effect of phosphorus-induced changes on H. cf. andersenii (Fig. 4a) and P. pinguis (Fig. 4b) fatty acid dietary ratios before and after alkaline hydrolysis. H. cf. andersenii presented more variations in fatty acid dietary ratios than P. pinguis which seemed to be more stable. For H. cf. andersenii, the ratios more affected by alkaline hydrolysis were the EPA to 4,7,10,13,16,19-docosahexaenoic acid (DHA; C22ω6) and PUFA/SFA ratios, which increased up to 45 and 69% after saponification respectively, indicating that the saponifiable fraction of this microalga is richer in EPA and PUFA. Moreover, the lowest value of PUFA/SFA ratio (BH: 0.48; AH: 0.81) observed at LP conditions shows that H. cf. andersenii tend to produce SFA in low phosphorus environments. This observation is consistent with a previous study performed for microalga Chlorophyta Scenedesmus, in which it was assumed that during phosphorus limitation this microalga tends to produce SFA . H. cf. andersenii also displayed the highest ∑ω3/∑ω6 ratio (BH: 6.50; AH: 6.78) showing its potential as a rich source of ω3 fatty acids.
Decreasing dietary SFA intake and replacing it with PUFA have been mentioned as the main dietary strategy for humans to benefit from PUFA preventive and protective activities against chronic inflammatory, cancer, diabetes, atherogenesis, and Alzheimer's disease (Shang et al. 2017;López et al. 2019). Since PUFA differ in their physiological functions, the balance between the ω3 and ω6 fatty acids, along with EPA and DHA, should be considered simultaneously (Shang et al. 2017). Thus, the ability of H. cf. andersenii to adjust its PUFA/SFA, ∑ω3/∑ω6, and EPA/DHA constitutes a great advantage for modulating its composition according to the supplementation purposes. Moreover, the study of the nutritional ratios before and after alkaline hydrolysis in both microalgae allows seeing which strategies can be better employed to fully exploit microalgae potential for PUFA production.
Sterols, carotenoids, and phytol synthesis, initiate with the biosynthesis of the isoprenoids precursors: isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) (Sasso et al. 2012). For P. pinguis, the lowest amount (13.12 mg g −1 dw) of sterols was observed at LP conditions, in contrast to its carotenoids (4.32 mg g −1 dw). Since phosphorus is an elemental constituent of isoprenoid precursors, the previous observation may indicate that at LP conditions P. pinguis promote the accumulation of carotenoids (photoprotective pigments) at the expense of sterols (membrane components). For H. cf. andersenii, the lowest amount (1.83 mg g −1 dw) of sterols was also verified at LP conditions. As with the present study, the decrease in sterol content as a response to low phosphorus environments has been reported to occur in other microalgae (Chlorophyta Scenedesmus quadricauda, and Chlamydomonas globosa, Cryptophyta Cryptomonas ovata, Bacillariophyta Cyclotella meneghiniana) (Piepho et al. 2010(Piepho et al. , 2012. Marine microalgae have developed efficient strategies to cope with poor phosphorus environments by remodeling membrane lipids in phosphorus-free lipids (such as betaine lipid and triacylglycerols) instead of phospholipids (Cañavate et al. 2017;Murakami et al. 2018;Kokabi et al. 2019). Two metabolic pathways have been described for triacylglycerols biosynthesis: an acyl-CoA-dependent pathway 2.12 ± 0.17 a 2.88 ± 0.12 b 2.48 ± 0.14 c 2.24 ± 0.14 A 2.42 ± 0.18 A 2.60 ± 0.24 A (often called Kennedy pathway) and an acyl-CoA-independent pathway (Cagliari et al. 2011). In the Kennedy pathway, the intermediates phosphatidic acid and diacylglycerol are also used for the synthesis of phospholipids ).
For both microalgae, the greatest amounts of monoglycerides were verified for the treatments with low phosphorus loadings (LP and MP treatments). For P. pinguis, the greatest increase in fatty acid levels after alkaline hydrolysis Values (means ± SD of four replicates) in the same row, not sharing a common superscript, are significantly different (p < 0.05). Differences among treatments assessed by one-way ANOVA followed by Tukey post-hoc analysis are represented by superscript lowercase letters for H. cf. andersenii, and by superscript uppercase letters for P. pinguis. All the compounds containing hydroxyl and/or carboxyl groups are identified as the correspondent trimethylsilyl (TMS) derivatives. 1 Contains cis and trans isomers. n.d., non detected. * identified as mono-TMS ether  (Table 3), and the greatest amounts were found at LP conditions, where the highest monosaccharide levels (180.27 mg g −1 dw) were observed. As with phospholipids, glycosyl sterols are amphiphilic molecules which can be used by microalgal cells to maintain the organization and fluidity of membranes (Fernandes et al. 2021). Therefore, the previous observations seem to indicate that both microalgae responded to LP conditions by synthesizing non-phosphorus-containing lipids to offset the degradation of phospholipids. As with nucleic acids, phospholipids are the major phosphorus reservoirs within microalgal cell (Mooy et al. 2009). Previous studies have reported that phosphorus scarcity induced lipid remodeling including the replacement of membrane phospholipids by non-phosphorus-containing lipids
Cholesterol-lowering and anti-aging activities have been described for brassicasterol derived from cryptophytes (Abidizadegan et al. 2021). To produce this sterol by H. cf. andersenii, P-replete conditions such as those in MP and HP conditions should be used. ALA and 9,12-octadecadienoic acid (LA; C18:2ω6) are two essential fatty acids that are needed for normal growth and development of animals; however, they lack the ability to synthesize these precursors of PUFA (DHA and EPA), which, in turn, makes its dietary intake mandatory. Thus, the rich composition of H. cf. andersenii saponifiable lipids in essential fatty acids (ALA and LA), and long-chain PUFA (EPA and DHA) at MP and HP cultures, reinforces that P-replete conditions are good to boost H. cf. andersenii composition in health-promoting lipids.
Within lipid-soluble compounds, tocopherols are often referred to as an underexploited resource from microalgae (Carballo-Cárdenas et al. 2003;Mudimu et al. 2017). α-Tocopherol is especially known for its antioxidant activity and health properties such as prevention of light-induced pathologies, degenerative disorders, cardiovascular diseases, and cancer (Carballo-Cárdenas et al. 2003). In the industrial field, this compound is used in the preservation of food, in cosmetics, and as an additive in animal field. Thus, for α-tocopherol production, the best conditions in the present study were MP and HP conditions before hydrolysis for H. cf. andersenii, and LP conditions for P. pinguis. Mudimu et al. (2017) performed a screen of microalgae strains for α-tocopherol content and studied the influence of nitrate reduction on α-tocopherol production. In this study, α-tocopherol content in microalgae ranged between 20.52 and 1445.66 µg g −1 dw, and nutrient reduction leads to an increase in α-tocopherol content. In this study, P. pinguis presented the highest amount of this molecule accounting 170 µg g −1 dw at LP conditions.
For H. cf. andersenii, hierarchical cluster analysis (Fig. 6a) showed the same trends displayed by principal component analysis, with LP treatments showing the deepest lipid remodeling when compared to MP and HP conditions before and after alkaline hydrolysis. In P. pinguis (Fig. 6b) after hydrolysis, LP and MP conditions are closely related according to hierarchical cluster analysis. This trend is different from that displayed by PCA. This may be derived from P. pinguis cultured at LP and MP conditions which presented the lowest content of unsaponifiable material, with the greatest increase of total identified compounds with alkaline hydrolysis being verified at LP conditions (31%).

Conclusions
This work provided novel insights on the response of marine microalgae to phosphorus-induced changes and their potential as a tool for optimizing high-value lipid production. Increasing initial phosphorus concentrations enhanced Pavlova pinguis and Hemiselmis cf. andersenii biomass productivities, making these conditions suitable for efficient high-value compounds production. Monosaccharide data, along with lipid/monosaccharide ratio, indicates that P. pinguis responded to low phosphorus conditions by synthesizing a β-1,3-glucan, while H. cf. andersenii seemed to divert its carbon to lipid production regardless of the treatment applied. In P. pinguis, the Car/Chl a could be a simple indicator for determining carbon allocation among carbon-rich pools. In H. cf. andersenii, the co-production of carotenoids and lipids against phosphorus-induced changes constitutes a major advantage for microalgae-based industries. Phosphorus-induced changes showed to be an effective tool for inducing lipid remodeling in P. pinguis and H. cf. andersenii, with P-replete conditions being the most effective to induce the accumulation of high-value lipids in both microalgae. andersenii and P. pinguis grown under low, medium, and high phos-phorus supplementations, respectively, and analyzed before hydrolysis. LPH, MPH, and HPH stand for H. cf. andersenii and P. pinguis grown under low, medium, and high phosphorus supplementations, respectively, and analyzed after hydrolysis. Loading's descriptions can be seen in Tables 1 and 2 Fig. 6 Hierarchical clustering analysis heat maps based on Euclidean clustering distance and the ward clustering method, n = 4 replicates, for Hemiselmis cf. andersenii (a) and Pavlova pinguis (b) grown under different phosphorus supplementations. Feature descriptions can be seen in Tables 3 and 4. The heat maps reflect the relative levels of metabolites among different treatment groups, the color scheme is associated with the elevation and reduction in metabolite level through treatments: dark blue, lowest; dark red, highest. HP, MP, and HP stand for H. cf. andersenii and P. pinguis grown under low, medium, and high phosphorus supplementations, respectively, and analyzed before hydrolysis. LPH, MPH, and HPH stands for H. cf. andersenii and P. pinguis grown under low, medium, and high phosphorus supplementations, respectively, and analyzed after hydrolysis