Identification of Polyunsaturated Fatty Acids Synthesis Pathways in the Toxic Dinophyte Alexandrium minutum Using 13C-Labelling

The synthetic pathways responsible for the production of the polyunsaturated fatty acids 22:6n-3 and 20:5n-3 were studied in the Dinophyte Alexandrium minutum. The purpose of this work was to follow the progressive incorporation of an isotopic label (13CO2) into 11 fatty acids to better understand the fatty acid synthesis pathways in A. minutum. The Dinophyte growth was monitored for 54 h using high-frequency sampling. A. minutum presented a growth in two phases. A lag phase was observed during the first 30 h of development and had been associated with the probable temporary encystment of Dinophyte cells. An exponential growth phase was then observed after t30. A. minutum rapidly incorporated 13C into 22:6n-3, which ended up being the most 13C-enriched polyunsaturated fatty acid (PUFA) in this experiment, with a higher 13C atomic enrichment than 18:4n-3, 18:5n-3, 20:5n-3, and 22:5n-3. Overall, the 13C atomic enrichment (AE) was inversely proportional to number of carbons in n-3 PUFA. C18 PUFAs, 18:4n-3, and 18:5n-3, were indeed among the least 13C-enriched FAs during this experiment. They were assumed to be produced by the n-3 PUFA pathway. However, they could not be further elongated or desaturated to produce n-3 C20-C22 PUFA, because the AEs of the n-3 C18 PUFAs were lower than those of the n-3 C20-C22 PUFAs. Thus, the especially high atomic enrichment of 22:6n-3 (55.8% and 54.9% in neutral lipids (NLs) and polar lipids (PLs), respectively) led us to hypothesize that this major PUFA was synthesized by an O2-independent Polyketide Synthase (PKS) pathway. Another parallel PKS, independent of the one leading to 22:6n-3, was also supposed to produce 20:5n-3. The inverse order of the 13C atomic enrichment for n-3 PUFAs was also suspected to be related to the possible β-oxidation of long-chain n-3 PUFAs occurring during A. minutum encystment.


Introduction
Phytoplankton, at the basis of food webs, are the main producers of long-chain polyunsaturated fatty acids (LC-PUFAs), 20:5n-3 and 22:6n-3 [1][2][3]. n-3 PUFAs are essential compounds for marine organisms because they are not always able to synthesize them in sufficient quantities and then have Attempts to understand the synthesis pathway of microalgae FA mainly focused on identifying gene coding for the different enzymes (desaturase or elongase) involved in PUFA formation. For transcriptomic studies, they are challenging targets to enhance the production of 20:5n-3 and 22:6n-3 rich single-cell oil. This could help in counteracting the expected lack of these two healthy compounds in the future. Additionally, new isotopic techniques have been developed in recent years and present an important progress in understanding metabolic fluxes. The principle is to monitor the incorporation of an isotopic label, introduced as a substrate, into targeted organic macromolecules such as fatty acids [49][50][51][52] to quantify lipid production and identify synthesis pathways. This had already been applied in E. coli [53], yeast [52], and microalgae [33,[54][55][56]. In compound-specific isotopic analysis (CSIA), these techniques couple gas chromatography with mass spectrometry (GC-c-IRMS) and allow the direct resolution of the isotopic composition of organic macromolecules, including fatty acids [57][58][59][60].
The present study aims to investigate the synthesis pathways of n-3 PUFAs by the toxic Dinophyte Alexandrium minutum using a stable isotope ( 13 C) labelling experiment. This species is responsible for a harmful bloom (red tide) that can cause the severe mortality of marine organisms and has been extensively investigated. Three culture balloons followed; two replicates received 13 CO 2 , while the last one was used as a control and was bubbled with unlabeled CO 2 of petrochemical origin (depleted in 13 C). The incorporation of the 13 C was monitored in eleven FAs for 54 h at a high to moderate temporal resolution. The progressive incorporation of the 13 C-labelled CO 2 into FA (from precursors to n-3 PUFA) and quantification of main fatty acid production allowed us to constrain FAS, elongase/desaturase, as well as the PKS pathway and their involvement in PUFA production by A. minutum. In parallel with the 13 C-monitoring, growth and other cellular parameters (morphology, viability, and lipid content) were monitored by flow cytometry analysis to assess the microalgae physiological status.

Algal Culture and Isotopic Labelling
This experiment was conducted following the experimental design described by Remize et al. [33]. Briefly, the toxic Dinophyte A. minutum (strain AM89BM, isolated from a bloom in the Bay of Morlay, France) were cultured in batch (total volume of 2 L) under continuous light (24 h light cycle, 100 µmoles photons m −2 ·s −1 at 20 • C). The culture was prepared with 1750 mL of sterile sea water, 250 mL of algal inoculum (pre-cultured for five days to reach the exponential growth phase), and 2 mL of nutrient medium (L1 medium) [61]. The culture was kept sterile during the entire experiment. Two replicate balloons received the 13 C-label (named, later on, Alm1 for balloon 1 and Alm2 for balloon 2). A third balloon was used to check the absence of a negative impact of 13 C-labelled CO 2 . This control balloon was prepared with petrochemical CO 2 (non-13 C labelled) but has not been analyzed further for fatty acid composition.
The two replicates were subjected to a pre-culturing phase of four days before the start of the monitoring. The isotopic labelling was performed using pure 13 C-CO 2 gas (Sigma-Aldrich, Saint Louis, MO, USA, <3%atom 18O, 99.0%atom 13 C). The CO 2 introduction in the culture was started just before the first sampling time (t 0 ) and was maintained for 54 h (t 54 ). The growth was then controlled by 13 C-CO 2 addition using a pH-stat system, which supplied the culture when the pH was superior to 8.00 ± 0.05 ( Figure 1).

Figure 1.
Experimental design of the enrichment experiment. The 13 CO 2 is supplied to the culture depending on its pH. To sample the algae, pliers are used to close/open the tubes/ways needed to first put the balloon under pressure and then allow sampling and finally rinse the tubes after sampling.

Sample Collection
Sampling was performed regularly for 54 h at the following time points: 0, 2.5, 5, 10, 20, 24, 30, 48, 54 h (nine samples per balloons). The experimental design was created to collect the culture medium without opening the balloon and to avoid bacterial contamination and the introduction of atmospheric CO 2 . At each sampling time, a total volume comprised of between 82 and 142 mL was collected for (i) the flow cytometry (FCM) analysis of cellular parameters, (ii) the isotopic analysis of particulate organic carbon ( 13 C-POC) and dissolved inorganic carbon ( 13 C-DIC) by EA-IRMS, (iii) fatty acid (FA) analysis in neutral lipids (NLs) and polar lipids (PL) by GC-FID, and (iv) the compound-specific isotope analysis (CSIA) of FA ( 13 C-FA) by GC-c-IRMS, as described in the following paragraphs.

Flow Cytometry Analysis
Algae cellular variables were measured using an Easy-Cyte Plus 6HT flow cytometer (Guava Merck Millipore ® , Darmstadt, Germany) equipped with a 488 nm blue laser; detectors of forward (FSC) and side (SSC) light scatters; and three fluorescence detectors: green (525/30 nm), yellow (583/26 nm), and red (680/30 nm). Cell variables-i.e., forward scatter (forward scatter, FSC), side scatter (side scatter, SSC), and red fluorescence (FL3, red emission filter long pass, 670 nm, a proxy of chlorophyll content)-were used to identify and select the A. minutum cell population. FSC and SSC give, respectively, information on the relative size and complexity of cells [62][63][64]. The flow cytometry measurements were performed on fresh (live) samples.
Two fluorescent probes were used to assess the viability and lipid content according to Lelong et al. (2011) and Seoane et al. [65]. SYTOX (Molecular Probes, Invitrogen, Eugene, OR, USA at a final concentration of 0.05 µM) was used to estimate the percentage of dead cells in the Dinophyte samples [66]. The BODIPY probe (BODIPY 505/515 FL; Molecular Probes, Invitrogen, Eugene, OR, USA, final concentration of 10 mM) was used as a proxy of the lipid reserves [65].
The concentration of bacteria was also monitored during the experiment according to Seoane et al. [65] using SYBR ® Green staining (Molecular Probes, Invitrogen, Eugene, OR, USA, #S7563). The results are expressed as the concentration of bacteria per mL.

POC Concentration and Bulk Carbon Isotopic Composition
Samples (40-70 mL) for Particulate Organic Carbon (POC) concentration and stable isotopic composition were processed as described in Remize et al. [33]. The POC concentrations of all samples were measured using a CE Elantech NC2100 (ThermoScientific, Lakewood, NJ, USA) according to the protocol of the United States Environmental Protection Agency [67] with acetanilide (99.9% purity, C 8 H 19 NO CASRN 103-84-4) as a standard. The POC concentrations are given in mmol L −1 . The bulk carbon isotopic composition ( 13 C-POC) was analyzed by continuous flow on an Elemental Analyzer (EA, Flash 2000; Thermo Scientific, Bremen, Germany) coupled with a Delta V+ isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany). Calibration was performed with international standards and the in-house standards described in Table 1.

DIC Concentration and Bulk Carbon Isotopic Composition
Samples for the Dissolved Inorganic Carbon (DIC) concentration and stable isotopic composition were collected from the filtrate of POC samples and processed as described in Remize et al. [33] and following the protocol of Assayag et al. [68]. The DIC was measured using a gas bench coupled with a Delta Plus mass spectrometer from Thermo Scientific, Bremen, Germany (GB-IRMS).

Isotopic Data Processing
We used the atomic abundance of 13 C in percent (%atom of 13 C) to express the results. Conversion between delta notation and %atom 13 C notation can be performed according to Larsson et al. [69]: where ( 13 C/ 12 C) PDB = 0.0112372, the ratio of 13 C to 12 C in the international reference VPDB standard. The atomic enrichment (AE) was then calculated from the atom % 13 C-POC and corrected with the values of the control balloon (POC control = 1.08%) and atom % 13 C DIC corrected from the control balloon (DIC control = 1.12%) according to the following equations:

Lipid Extraction
After filtration on pre-combusted (450 • C, 6 h) 47 mm GF/F filters (porosity 0.7 µm) with 40-70 mL of culture and the addition of boiling water to stop the lipase activity, the lipids were extracted by diving the filter into a 6 mL solvent mixture (chloroform:methanol; 2:1 v/v). The lipid extracts were flushed with nitrogen and stored at −20 • C until analysis.

Separation of Neutral and Polar Lipids
The lipid extracts were separated into neutral and polar fractions following the method of Le Grand et al. [70] and as described previously in Remize et al. [33]. In brief, 1 mL of extract was evaporated to dryness with nitrogen, recovered with three washes of 0.5 mL of chloroform:methanol (98:2 v:v; final volume 1.5 mL), and spotted at the top of a silica gel column (40 mm × 4 mm, silica gel 60A 63-200 µm 70-230 mesh rehydrated with 6% H 2 O, Sigma-Aldrich, Darmstadt, Germany). The neutral lipid fraction (NL) was eluted using chloroform:methanol (98:2 v:v; 10 mL) and the polar lipid fraction (PL) with methanol (20 mL). Both were collected in glass vials containing an internal standard (C23:0 as free fatty acid, 2.3 µg). The lipid fractions were then stored at −20 • C until further analysis.

Fatty Acid Analysis by Gas Chromatography (GC-FID and GC-c-IRMS)
Transesterification and fatty acids methyl esters (FAME) analysis with gas chromatography (GC-FID and GC-c-IRMS) were conducted according to the protocol described by Mathieu-Resuge et al. [71]. The analysis of the FAME was performed on a Varian CP8400 gas chromatograph (Agilent, Santa Clara, CA, USA) using simultaneously two parallel separations on polar (ZBWAX: 30 mm × 0.25 mm ID × 0.2 µm, Phenomenex, Torrance, CA, USA) and apolar columns (ZB5HT: 30 m × 0.25 mm ID × 0.2 µm, Phenomenex, Torrance, CA, USA). The FAME were identified by the comparison of their retention times with those of commercial and in-house standards mixtures, as shown for the two columns used in the Supplementary Data ( Figure S1). The individual fatty acid concentrations obtained in µg·L −1 by GC-FID were also expressed in µmolC·L −1 (µg·L −1 /molecular weight of individual fatty acid × carbon number of individual fatty acid) to ease the comparison with the POC concentrations expressed in µmolC L −1 as well. The total fatty acid concentrations in neutral lipids, polar lipids, and the sum of both fractions (the latter named hereafter TFA) were expressed in µmolC·L −1 .
The compound-specific isotope analyses (CSIA) of FAME were performed on a Thermo Fisher Scientific GC ISOLINK TRACE ULTRA (Bremen, Germany) using the same apolar column as for the FAME analysis. Only the 11 fatty acids with the highest concentrations (> 1 µmolC·L −1 ) as measured by the GC-FID analyses were considered for CSIA, as they presented a signal amplitude on the GC-c-IRMS superior to the threshold value of 800 mV defined by Mathieu-Resuge et al. [71]. The minor FA presenting a too-low signal amplitude (<800 mV) on the GC-c-IRMS did not allow precise isotope ratio analysis. 18:1n-9 and 18:3n-3 were co-eluted for GC-c-IRMS on the apolar column so they were considered together. Examples of GC-c-IRMS chromatograms for C 18 PUFAs 18:4n-3 and 18:5n-3 as well as the confirmation of their identification by GC-MS are available in the Supplementary Data ( Figure S2).
The FA atomic enrichment (AE FA ) was estimated with the same method as for AE POC and AE DIC , with the %atom 13 C for each FA given by GC-c-IRMS analysis (AE FAcontrol = AE FAnat = 1.08%).

Compound Specific Isotope Data Processing
To evidence the FA conversion of fatty acid A into fatty acid B in A. minutum, we calculated the AE FA ratio (R) of product B over hypothesized precursor A. If the calculated ratio was close to one with a confidence interval calculated at α = 0.1 (defined arbitrarily), the fatty acids A and B were supposed at the equilibrium in terms of the label incorporated and can be assumed to be synthesized simultaneously or very closely. If the ratio was below one, the transformation of A into B was possible but slower. On the contrary, if the ratio was above one, A could not be considered as a major precursor of B. R was defined as follows.
where A the fatty acid is supposed to be a precursor of fatty acid B and AE FA(A) and AE FA(B) are their respective atomic enrichments at each sampling time.

Statistical Analysis
To assess the potential effect of the time and difference between balloons during algae development and 13 CO 2 incorporation, a PERMANOVA analysis was used on the FA percentage separately in NL and PL. The Spearman test was conducted on the fatty acid abundance in both PL and NL to explore the correlations between individual fatty acids.

Bulk POC and DIC and Their 13 C-Labelling
The Dissolved Inorganic Carbon (DIC) and POC atomic enrichments with time are shown in Figure 4. The DIC atomic enrichment increased quickly between t 0 and t 10 (from 0 to 50%), remained stable between t 10 and t 30 (around 50%), and increased to around 60% at t 48 and t 54 . The two balloons had a similar DIC enrichment. On the contrary, the POC atomic enrichment was continuous but slower from t 0 to t 54 . The Alm1 had a slightly higher POC enrichment after t 30 . The final POC enrichments for Alm1 and Alm2 were 58% and 45%, respectively.

Bulk POC and DIC and Their 13 C-Labelling
Overall, there was more PL (61 ± 11%) than NL (39 ± 11%) (Figure 5b), but their respective proportions varied greatly during the culture; the percentage of NL over TL decreased from 56% to 25% between t 0 and t 30 and then increased back to about 60% at t 54 (Figure 5c).
For A. minutum, 36 FAs were identified but 24 were present in a trace amount (<1% of TFA). The 12 main FAs were the ones presented in Figure 5a. Table S1 of the Supplementary Data gives the concentration in µg L −1 and µmolC L −1 for all the fatty acids identified in this study. The PERMANOVA conducted on PL and NL FA showed a significant difference between the sampling times for the fatty acid percentages (p-value < 0.05), but no significant difference was observed between balloons. The replicates could then be considered together. The SFAs and PUFAs were homogenously represented in the NL (40% each, 20% for MUFA), and while in the PL PUFA was the main category of FA (63% against 32% for SFA and only 5% for MUFA). 16:0, 18:1n-9, 14:0, 18:2n-6, and 18:3n-3 were the main FAs of the NL fraction (respectively, 27 ± 2%, 19 ± 5%, 10 ± 0.004%, 10 ± 1%, and 9 ± 1%). For the PL fraction, 16:0 (23 ± 8%) and 22:6n-3 (20 ± 3%) were the most abundant FA, followed by 18:4n-3 (15 ± 4%) (Figure 5a). At t 48 , in the NL fraction the PUFA and SFA proportions were augmented (39% to 43% for both), while in PL the PUFA decreased (68% between t 0 and t 30 and 50% after t 30 ) (data not shown). The MUFA decreased in NL (23% between t 0 and t 30 and 11% after t 30 ) and remained relatively stable in PL. The n-3 PUFAs were always more important than the n-6 PUFAs in both NLs (28% vs. 11%, respectively) and in PL (51% vs. 11%). The bacterial fatty acids (including branched, C 15 , C 17 and C 21 FA) of the A. minutum-associated bacteria communities (non-axenic culture) remained below 0.8% in the NL and PL fractions and were not further followed.

Fatty Acids 13 C-Enrichment
The atomic enrichments (AE) with time of the 11 FAs analyzed by GC-c-IRMS are presented in Figure 6. All the studied FAs appeared to have a similar continuous enrichment to that of POC. The AE of 18:1n-9+18:3n-3 and 18:2n-6 in the NL fraction was saturated at t 54 with an enrichment level superior to AE DIC (56%). All the FAs presented an increasing 13 C level from t 0 to t 54 . The enrichment dynamics were slightly different between t 0 to t 30 (phase I) and t 48 and t 54 (phase II).

Discussion
This experiment investigated the 13 CO 2 incorporation into fatty acids to understand how n-3 long-chain polyunsaturated fatty acids were synthesized by the Dinophyte Alexandrium minutum. Both DIC and POC were enriched rapidly after only five hours of incubation, and this was before the algae reached its exponential growth phase at t 30 . The algae physiology (cell size, cell complexity, and viability) was not negatively impacted by the 13 CO 2 bubbling and remained stable during the experiment. However, the neutral lipid and chlorophyll contents appeared to undergo some modification during the 54 h of monitoring and will be discussed in the next sections.
The O 2 -independent PKS pathway was identified in different microalgae taxa [18,30]. It relies on the same four basic reactions as the FAS pathway (condensation, reduction, dehydration, reduction) and is responsible for long-chain polyunsaturated fatty acids synthesis, such as 20:5n-3 and 22:6n-3 [19,75]. Contrarily to the FAS pathway, the PKS pathway is less energy-consuming because it requires fewer reduction and dehydration steps [18]. Identifying the FA precursors of this pathway is difficult, as these intermediates might not be released from the enzyme protein into the fatty acids pool and would therefore be present in very small concentrations [19]. Several PKS genes have been identified in microalgae, including Dinophytes, in recent years [25,26,30]. Regarding the rapidity of 22:6n-3 synthesis and the absence of more enriched precursors during the monitoring, we suggested that 22:6n-3 was produced directly from the PKS pathway in A. minutum (Figure 7). Moreover, because the enrichment was similar in both the NL and PL fractions, we proposed that the PKS pathway in this Dinophyte might exist to produce 22:6n-3 in both membrane lipids (PL) and reserve lipids (NL). , white corresponds to neutral lipids and black to polar lipids. The PKS intermediates written in grey are not detected during the study, as the enzyme proteins might not release them until the final product is synthesized [19]. The circles represent the enzymes involved in these pathways: KS: 3-ketoacyl synthase; KR: 3-ketoacyl-ACP-reductase; DH: dehydrase; 2.2I/2.3I: 2-trans, 2-cis, or 2-trans, 3-cis isomerase; ER: enoyl-reductase. The numbers at the top of each pathways represent the suspected order of 20:5n-3 and 22:6n-3 synthesis according to the 13 C-enrichment dynamics observed in this study.
During our monitoring, A. minutum exponential growth started only after 30 h and resulted in an eight-fold increase in cell abundance. Before this growing phase, despite presenting a stable viability the algae underwent modifications of their cellular content with a decrease in the chlorophyll content (FL3) and neutral lipid content (BODIPY). The diminution of the chlorophyll content might be an indicator of reduced photosynthesis following a stress [76,77], while the diminution of neutral lipid content would reflect a remobilization of the stored lipids. This could be associated with encystment [42]. Indeed, Dinophyte species are able to form temporary cysts in response to unfavorable environments. Temporary cysts enable functioning over shorter time scales to answer the perturbation of environmental conditions [42]. The formation of temporary cysts can be associated with the repression of photosynthesis activity (reduction in chlorophyll concentration) [42]. During temporary encystment, the cysts rely on their reserves, which can also explain why we observed a decrease in the NL concentration before t 30 and a slight decrease in the TFA content after t 30 in A. minutum.
In parallel with the loss of neutral lipids and decrease in TFA concentrations, most of the FAs were nevertheless enriched in priority in the NL fraction, especially 22:6n-3 and other n-3 PUFAs (18:4n-3, 18:5n-3, 20:5n-3, 22:5n-3). However, the succession of FA atomic enrichment into fatty acids and the atomic enrichment ratios of these n-3 PUFAs did not support the existence of an active n-3 pathway to produce C 20 and C 22 PUFAs from their potential C 18 PUFAs precursors. Indeed, 18:4n-3 and 18:5n-3 were enriched after C 20 and C 22 PUFAs. Furthermore, the 22:5n-3/20:5n-3 and 22:6n-3/22:5n-3 ratios were >1, implying that 20:5n-3 and 22:5n-3 cannot be the respective precursors of 22:5n-3 and 22:6n-3. Thus, it seemed improbable that these C 20 and C 22 PUFAs were synthesized by the known "conventional" n-3 pathway. Two explanations can be proposed. First, as for 22:6n-3, 20:5n-3 might have been produced from a different and independent PKS pathway ( Figure 7). This implies that PKS enzymes do not release their intermediates, but only the final PUFA end product, 20:5n-3 or 22:6n-3. The absence of 16:4n-3, a mainstream PKS pathway intermediate [19], had already been reported in Dinophytes by Leblond and Lasiter [78]. Furthermore, Jang et al. [37] have demonstrated that PKS genes were upregulated during the formation of resting cysts in Dinophytes. This could also be the case in temporary cysts, and this then would support the use of a PKS pathway to produce LC-PUFAs without O 2 during encystment. Based on the enrichment dynamics, 22:6n-3 synthesis by PKS would be faster than the hypothesized parallel PKS pathway for 20:5n-3. The existence of such a faster route to produce 22:6n-3 in Dinophytes might also explain why, in some species such as Crypthecodinium cohnii, putative n-3 PUFAs intermediates were only identified in trace amounts [79]. C. cohnii, an exclusive heterotroph Dinophyte, was previously speculated to produce 22:6n-3 primarily by the PKS pathway [80]. We can then assume that the 22:6n-3 PKS pathway could be active in both heterotroph and mixotroph Dinophytes. However, it might not be the case for the 20:5n-3 PKS routes, as C. cohnii did not produce it in significant quantities [79].
The degradation of PUFA by β-oxidation coupled with fatty acid remodeling by lipase could also explain the decreasing order of n-3 PUFA enrichment, from a longer to a shorter chain length, as observed in A. minutum. As suspected earlier, it appeared that quiescent cells repressed their photosynthetic activities during temporary encystment [42]. However, energy metabolism, glycolysis, and even β-oxidation remain active [44]. β-oxidation is the principal pathway implicated in FA degradation and occurs in the peroxisome of plants and microalgae [81]. It relies on four enzymatic reactions (oxidation, hydration, dehydrogenation, and thiolytic cleavage) to form the end product acetyl-CoA [81]. The LC PUFA 22:6n-3 is the most abundant fatty acid found in the triacylglycerol (TAG) of Dinophytes [82] and is found in large proportion in phospholipids [74]. Enzymes responsible for the cleavage of reserve lipids (NL) or membrane lipids (PL) could release 22:6n-3 from the glycerol backbone and then join the free fatty acids (FFA) pool to be degraded by β-oxidation in the peroxisomes. PUFAs such as 20:5n-3 and 18:5n-3 are the intermediates of 22:6n-3 β-oxidation. They can be further degraded into acetyl-CoA, or serve as an acyl-donor for the de novo synthesis of TAG. This hypothesis would explain why all the n-3 PUFAs were primarily enriched in the NL fraction. The remodeling of membrane lipids during encystment had already been observed by Lichtlé and Dubacq [83]. Moreover, Xu et al. [84] reported that the 20:5n-3 issued from 22:6n-3 β-oxidation was recycled into TAG in yeast after the expression of type-2 diacylglycerol acyltransferase (DGAT2) from Thalassiosira pseudonana. Their study provided evidence that DGAT2 could be involved in both the de novo synthesis of TAG as well as the TAG remodeling with β-oxidation intermediates. Dinophytes were reported to have type 2 DGAT [85]. However, the enzymes (lipases) involved in the cleavage of TAG and membrane lipids as well as β-oxidation processes remain to be fully investigated in Dinophytes.
According to the atomic enrichment, 18:4n-3 was synthesized after all the long-chain n-3 PUFAs. It was likely synthesized from 18:3n-3 by the conventional n-3 pathway for both the PL and NL, fractions even if we were not able to prove it with ratio calculation as 18:3n-3 and 18:1n-9 coeluted. ∆6 desaturase involved in the desaturation of 18:3n-3 into 18:4n-3 has been identified in the Dinophyte Lingulodinium polyedrum, allowing the production of 18:4n-3-rich galactolipids [86]. The ratio 18:5n-3/18:4n-3 was below one in NL (<0.55), indicating a feasible transformation of 18:4n-3 into 18:5n-3, although this ratio was above one in PL (>1.67). The 18:5n-3 could be formed from 18:4n-3 in NL by the action of ∆3 desaturase, as suggested by Joseph [87] to explain the presence of this unusual FA in Dinophytes. The 18:5n-3 could also be synthesized by the PKS pathway in the PL fraction, since it had been previously identified as a mainstream PKS fatty acid [19] (Figure 8). C 18 PUFAs, especially 18:4n-3 and 18:5n-3, seemed to be final products in A. minutum and not be further desaturated, as they presented some of the lowest enrichments in our experiment. 18:4n-3 and 18:5n-3 have been reported to be the dominant fatty acids of Dinophyte glycolipids, especially MGDG and DGDG [78,88]. In A. minutum, 18:4n-3 and 18:5n-3 might have a similar location. 18:5n-3 and 18:4n-3 have been also described to be present in free fatty acids, revealing a higher toxicity compared to that of other PUFAs in Dinophytes [47,48]. They could then be involved in A. minutum toxicity. Figure 8. Hypothesized pathways to produce C 18 FA in A. minutum. Numbers in boxes correspond to the final AE value (white for neutral lipids and black for polar lipids). The triangles symbolize the desaturases (front-end in yellow and methyl-end in purple), the circles the enzymes involved in the PKS pathway (KS: 3-ketoacyl synthase; KR: 3-ketoacyl-ACP-reductase; DH: dehydrase; 2.2I: 2-trans, 2-cis isomerase; 2.3I: 2-trans, 3-cis isomerase; ER: enoyl reductase). The ways with dashed arrows appear to be unlikely or cannot be proven with the enrichment dynamics. NL and PL written on the routes allowing the synthesis of 18:5n-3 indicate the fraction considered for each pathway.
18:2n-6 was always more enriched than the other n-6 PUFA, 22:5n-6, in NL. Their enrichments were even similar in the PL. 18:2n-6 and 22:5n-6 are both linked in the classical n-6 pathway [19]. Based on the atomic enrichments, 18:2n-6 could be initially produced in the NL fraction and then transported to the membranes (PL).Řezanka et al. [82] showed than 18:2n-6 and 16:0 were mainly found in phospholipids (PCs) in Dinophytes. Even though the enrichment of the known precursors of 22:5n-6 cannot be quantified due to their low concentrations, 18:2n-6 could be introduced into PC to be furtherly desaturated into 22:5n-6. Finally, 22:5n-6 could be integrated into TAG. Whether PC is involved in acyl-editing mechanisms during TAG synthesis remains to be described in a Dinophyte. Another hypothesis regarding 22:5n-6 synthesis could be the existence of a n-6 PKS pathway.
After t 30 , the algae entered into a vegetative phase and grew exponentially. It was associated with an increase in the SFA and PUFA proportion in the NL and a decrease in the PUFA proportion in the PL. TAGs are useful to exit quiescence and facilitate the initiation of the vegetative phase in Dinophytes [89]. 18:1n-9+18:3n-3 and 18:2n-6 became more enriched than 22:6n-3 in the NL after t 30 . The 18:1n-9 produced by the FAS pathway can be used in the de novo TAG synthesis pathway, called the Kennedy pathway, and be acylated in the positions sn-1 and sn-2 of DAG [90,91]. We assumed that de novo TAG synthesis has probably been reinitiated in A. minutum, and that it was directly driven to reserve lipid synthesis to compensate for those lost during excystment (increase in NL proportion and concentration after t 30 ).

Conclusions
The fatty acid synthesis in the Dinophyte Alexandrium minutum went through different routes. The PKS pathway appeared to be a particularly fast synthetic process, responsible for the high enrichment and production of the polyunsaturated fatty acid 22:6n-3. 22:6n-3 seemed to have a central role in maintaining a good physiological state, including during encystment. In our study, we assumed that the lag phase observed reflected the possibility of temporary encystment in A. minutum. It was characterized by a significant decrease in the proportion of neutral lipids, corresponding to the consumption of the reserve lipids. The 22:6n-3 might be degraded during this process, and its degradation products might be involved in the re-synthesis of triacylglycerol during algae excystment. The enrichment dynamics of C18 PUFAs revealed that they are unlikely to be involved in the further desaturation and elongation steps of n-3 C 20 -C 22 PUFAs. They appeared to be the final products of the classical n-3 pathway. The 18:5n-3 atomic enrichment makes possible its origin from the desaturation of 18:4n-3, the degradation of longer fatty acids such as 20:5n-3, or the PKS pathway. These C 18 PUFAs have been assumed to play some role in A. minutum, such as in its toxicity. Further studies are needed to better constrain the PUFA synthesis pathway in A. minutum, and especially to further demonstrate the involvement of the PKS pathway following 22:6n-3 and 20:5n-3 synthesis.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.