The life cycle of fatty acids in oleaginous stramenopiles: from synthesis to maturation and beyond

Abstract

Stramenopiles, including diatoms, eustigmatophytes and thraustochytrids, are major contributors to global primary productivity and promising platforms for lipid accumulation and very-long-chain polyunsaturated fatty acid biosynthesis. Their evolutionary history, shaped by secondary endosymbiosis, has produced highly compartmentalised cellular architectures that strongly influence fatty acid metabolism. Yet how fatty acid synthesis, trafficking, remodelling and degradation are coordinated across cellular compartments remains incompletely understood. Here, we review the life cycle of fatty acids in marine oleaginous stramenopiles, based on studies in a set of model species, from their de novo synthesis in plastids to their cellular fate. We summarise plastidial pathways producing saturated and unsaturated fatty acids and examine fatty acid export, activation into acyl-CoAs and incorporation into extraplastidial glycerolipids. Particular attention is given to elongation and desaturation pathways in the endoplasmic reticulum that generate essential very-long-chain polyunsaturated fatty acids such as eicosapentaenoic and docosahexaenoic acids. We further discuss the cytosol-to-plastid trafficking of eicosapentaenoic acid (referred to as the ‘omega pathway’, by analogy with the ‘prokaryotic’ and ‘eukaryotic pathways’ in angiosperms), the catabolic routes of fatty acid degradation, and the enzymatic as well as non-enzymatic oxidation of fatty acids to prduce oxylipins and isoprostanoids. Together, recent genetic, biochemical and lipidomic studies reveal remarkable metabolic plasticity in stramenopiles and highlight key unresolved questions relevant to cell biology and microalgal biotechnology.

Introduction

The Stramenopiles (alternatively known as Heterokonta) are mostly microbial eukaryotes [1] and constitute a superphylum within the kingdom Chromista [2]. They stem from a complex series of endosymbiotic events, primarily involving the engulfment of a red alga (Sect. 2). The monophyletic group Stramenopila was established through a combination of morphological and molecular investigations, initially focusing on SSU-rDNA genes [3, 4]. The term stramenopile is derived from the Latin stramen (‘straw’) and pilos (‘hair’), referring to the distinctive, hair-like tripartite tubular protrusions known as mastigonemes, which are found on the flagella of many stramenopiles. Mastigonemes have long been recognised as a defining characteristic of stramenopile cells [5, 6], although they have been secondarily lost in certain clades, such as diatoms [7]. Stramenopiles comprise two well-supported clades: Bigyra and Gyrista [8–11] and can be found in a wide range of habitats, including freshwater and soil environments [12, 13], oceans, Arctic waters [12, 14, 15], the deep sea, and deserts [16–19]. They include photosynthetic unicellular organisms, classified within the monophyletic group known as Ochrophytes, which is estimated to have emerged over 500 million years ago [20–22], as well as non-photosynthetic ones, like Labyrinthulomycetes [23]. Ochrophytes constitute a highly diverse clade of eukaryotic autotrophs [24], ranging from photosynthetic nano- and picoalgae to large kelps.

Stramenopiles, like all living organisms, possess membranes that function as barriers, separating internal and external environments through semi-selective permeability. The plasma membrane delineates the intracellular space from the extracellular environment, while in eukaryotes organellar membranes compartmentalise cellular structures such as mitochondria, the nucleus, and lipid droplets [25–33]. Membranes are predominantly composed of lipids, which include fatty acids (FAs) as key structural components. FAs are composed of a hydrocarbon chain with a terminal carboxyl group (-COOH). The carbon of the carboxyl group is designated C1, and subsequent carbons are numbered sequentially from this position. Alternatively, carbons may be identified using Greek-letter nomenclature, where the α-carbon denotes the carbon adjacent to the carboxyl group and the terminal carbon is referred to as the ω (omega) carbon (Fig. 1). FAs are classified based on the length of their acyl chains and the degree of unsaturation. In unsaturated FAs, the position of double bonds can be specified using two nomenclatures: the Δ system counts from the carboxyl end considering the carbon atom of the carboxylic group as C1 and the n- (or ω) system counts from the opposite end. For example, the same 16-carbon monounsaturated palmitoleic acid can be denoted as 16:1 ω7 (or n-7) or 16:1Δ⁹ (Fig. 1A).

Fig. 1.

Numbering and nomenclature of fatty acid carbon atoms. A Schematic representation of a C16 fatty acid illustrating the two commonly used carbon numbering systems. Carbon atoms can be numbered starting either from the carboxyl end (right), where the carboxyl carbon is designated as C1 and the adjacent carbon as the α-carbon, or from the methyl end (left), referred to as the ω (or n) end. In the ω/n nomenclature, carbons are counted from the terminal methyl group, such that the first double bond position defines the ω (or n) family (e.g. ω7 or n-7). Double bond positions can also be indicated using the Δ notation, which refers to the position relative to the carboxyl carbon (e.g. Δ9). This dual nomenclature is widely used in lipid biology to describe fatty acid structure, unsaturation patterns and functional properties. B Representation of a linear and skewed FA. Created with biorender (https://www.biorender.com)

Saturated FAs are linear molecules that exhibit greater stability compared to mono- or polyunsaturated FAs. In contrast, the presence of double covalent bonds in the acyl chain introduces a kink (Fig. 1B), altering the steric interactions of FAs within glycerolipids. For a comprehensive overview of FA chemistry and structure, see Scrimgeour et al., (2020) and Rustan and Drevon, (2005) [34, 35].

In the present review, we cover recent advances concerning the pathways of FA metabolism in model marine oleaginous stramenopiles, from their synthesis, to their metabolic and physiological utilisation, degradation and then give an outlook to the oxidation and transport. Over the last two decades, marine oleaginous stramenopiles have attracted significant interest due to their exceptional production of high quantities of lipids. This remarkable feature has earned them the designation ‘oleaginous’, positioning these organisms as promising candidates for both fundamental research and biotechnological applications, including their potential as a novel feedstock in bioindustry.

Nomenclaure and shorthand notation

For clarity, throughout this review we adopt the shorthand lipid nomenclature derived from mass spectrometry-based conventions for FA and glycerolipids [36, 37]. In this system, a FA is designated by the number of carbon atoms in the acyl chain followed by the number of double bonds, separated by a colon. For example, palmitoleic acid is denoted 16:1, indicating a 16-carbon chain with one double bond. For triacylglycerols, we use the abbreviation TAG (rather than TG) for consistency with our previous publications. Within each lipid class, molecular species are indicated by the class name followed by the total number of carbon atoms and double bonds in the acyl chains (e.g., PC 34:2). Regiospecific information is indicated according to accepted conventions. When an underscore separates the two FAs (e.g., 16:0_18:1), the positional distribution between the sn-1 and sn-2 positions is unknown. In contrast, a solidus (e.g., 16:0/18:1) indicates that the FAs have been assigned to the sn-1 and sn-2 positions, respectively. In the oleaginous model species discussed in this review, the regiospecificity of most lipid molecular species has been experimentally validated [38–40]; therefore, positional notation is used where supported by experimental evidence.

Origin of stramenopiles from complex endosymbiotic events

Secondary endosymbiosis is the engulfment of a photosynthetic eukaryote by a heterotrophic host cell. This process has occurred on multiple occasions throughout evolutionary history, resulting in the formation of intricate secondary plastids [41–44]. Two instances of green algal engulfment and an indeterminate number involving red algal endosymbionts have been proposed [45].

Secondary endosymbiotic events involving chlorophyte endosymbionts occurred independently in the ancestors of euglenophytes and chlorarachniophytes [46, 47]. In chlorarachniophytes, the plastid is enclosed by a four-membrane envelope and retains a remnant of the endosymbiont’s nucleus, known as a nucleomorph [48]. In contrast, the plastid of euglenophytes is limited by only three membranes, and lacks any rudimentary structures derived from the endosymbiont’s cytosol [49].

Rhodophyte-derived organisms such as the Chromalveolata, include cryptophytes, haptophytes, dinoflagellates, apicomplexans, and stramenopiles [50] and the endosymbiotic events that produced them remain a topic of ongoing debate. The plastids of the chromalveolates, now sometimes referred to as “CASH” (cryptophytes, alveolates, stramenopiles, and haptophytes), all contain chlorophyll c and are closely related based on plastid phylogenies [51, 52].

The ‘chromalveolate hypothesis’ proposes that the Chromalveolata supergroup originated from a single secondary endosymbiotic event in which a red alga was engulfed by an unknown heterotrophic eukaryote [53]. This hypothesis assumes that heterotrophic lineages within Chromalveolata independently lost their plastids. However, it is now established that once an endosymbiotic organelle is acquired, its complete loss is very rare [54]. In the alveolates and stramenopiles alone, the minimum number of plastid losses required to support the chromalveolate hypothesis ranges from five to seven [55]. Given the rarity of plastid loss, the chromalveolate hypothesis has become difficult to uphold.

Alternatively, the plastids of chromalveolate may have arisen through multiple independent secondary endosymbiotic events involving closely related red algae. However, the secondary plastids of cryptophytes, stramenopiles, and haptophytes share a specialised protein sorting system, phylogenetically related and derived from the Endoplasmic Reticulum-Associated Degradation (ERAD) machinery of red algae [56–58]. This system, known as SELMA (Symbiont-specific ERAD-Like Machinery), facilitates the transport of proteins across the second outermost plastid membrane [59, 60]. These findings strongly suggest a common endosymbiotic origin for the plastids in CASH lineages. This raises the question of how a shared origin is possible if multiple independent endosymbioses are required.

The third hypothesis, known as the ‘rhodoplex hypothesis’, suggests that these plastids originated from a combination of secondary and higher-order endosymbiotic events (tertiary, quaternary, etc.). Instead of multiple independent secondary endosymbiosis events, the hypothesis proposes that a single secondary endosymbiotic event occurred, after which other red algal-derived plastids were acquired through tertiary or even quaternary endosymbiosis [52, 61]. Current evidence indicates that the cryptophyte lineage is the most likely source of the original secondary endosymbiosis [22, 62, 63]. These findings suggest that the plastids in stramenopiles, alveolates, and haptophytes may have been acquired via tertiary endosymbiosis from cryptophytes, or possibly even more complex quaternary endosymbiosis [22].

Large-scale comparative genomic analyses have also suggested that a cryptic green-algal endosymbiont in chromalveolates may have preceded the well-known red algal symbiosis. Approximately 1,700 green algal genes, comprising around 16% of the diatom proteome, were identified in their genome [64]. Furthermore, in silico predictions and experimental data on plastid-targeted proteins in photosynthetic stramenopiles indicate that roughly 25% of the ancestral diatom plastid proteome consisted of nucleus-encoded proteins from green algae [65].

Diatoms, eustigmatophytes, phaeophytes, and thraustochytrids: key members of the stramenopile subkingdom

Stramenopiles are broadly divided into two major groups, Gyrista and Bigyra. Most photosynthetic stramenopiles within Gyrista belong to the Ochrophyta lineage [8, 10, 11, 66, 67]. Ochrophyta can be further subdivided into three clades: SI, SII, and SIII.

The SI clade includes, among others, the Phaeophyceae (brown algae), which are multicellular ochrophytes. Ectocarpus siliculosus, a filamentous brown alga with uniseriate branched filaments, serves as a model organism for molecular and developmental studies [68]. Brown algae are predominantly marine and include ecologically significant kelp species that form extensive underwater forests [69]. Some of these, such as the giant kelp Macrocystis pyrifera, one of the most widely distributed macroalgal species, can reach heights of up to 20 m. In addition to their ecological significance, several brown algal species are of considerable nutritional and economic value. For example, Undaria pinnatifida is a widely consumed edible seaweed, traditionally used in both cuisine and medicine across many Asian countries, including China, Korea, and Japan [70].

The SII lineage includes the Eustigmatophyceae, a class of unicellular, coccoid protists that occur either as solitary cells or in loosely associated colonies. Approximately 30 species have been formally described, inhabiting freshwater, soil, and marine environments. Some freshwater species, such as Monodus subterraneus, and marine genera like Nannochloropsis and Microchloropsis, have gained considerable biotechnological interest due to their high capacity for neutral lipid accumulation (triacylglycerol, TAG), making them promising candidates for biofuel and nutraceutical production [71].

The SIII lineage comprises the diatoms (Bacillariophyta), the most species-rich and extensively studied group of ochrophytes. This clade includes model pennate species such as Phaeodactylum tricornutum and the high-lipid-producing Fistulifera solaris, as well as centric diatoms like Thalassiosira pseudonana and Cyclotella cryptica. These organisms serve as key models for molecular, cellular, and metabolic studies. Ecologically, diatoms are major contributors to global primary productivity, accounting for approximately 40% of the Earth’s oxygen production and about 20% of global photosynthesis, despite representing only ~ 1% of the total photosynthetic biomass [72]. Diatoms dominate phytoplankton communities in marine, freshwater, and brackish environments, and they are distributed across a broad range of climatic zones, from tropical to polar regions [73].

Diatoms, eustigmatophytes, and phaeophytes all possess complex plastids surrounded by four membranes (Fig. 2). The plastid architecture has been well-characterised in the diatom P. tricornutum [74] and the eustigmatophyte Nannochloropsis oculata [75]. The outermost membrane, known as the epiplastidial membrane (EpM), is continuous with the outer membrane of the host nuclear envelope. Below, lies the periplastidial membrane (PPM), followed by the two membranes of the plastid envelope—the outer envelope membrane (oEM) and the inner envelope membrane (iEM). A distinctive vesicular compartment, termed the “blob,” is located between the PPM and oEM, though its function remains unresolved (Fig. 2) [74]. The lipid composition of these membranes has yet to be characterised.

Fig. 2.

Schematic representation of a diatom cell structure. The photosynthetic organelle containing thylakoids (Thyl) is not a chloroplast as classically encountered in red or green algae or in plants, but a complex plastid with a distinctive architecture linked to the nuclear envelope and the endomembrane system. De novo fatty acid synthesis occurs in the stroma of this complex plastid. The inner and outer envelope membranes (iEM and oEM, respectively) derive from the chloroplast envelope of the red alga engulfed during secondary endosymbiosis. A periplastidial membrane (PPM), thought to originate from the endomembrane system of the red algal symbiont, surrounds these membranes. The PPM is connected to a vesicular network forming a structure known as the blob [74, 215]. The identification of glycerolipid-synthesizing enzymes within the blob highlights its potential role as a central hub in the fatty acid life cycle of photosynthetic stramenopiles [133, 144]. A fourth membrane, the epiplastid membrane (EpM), delineates the outer boundary of the organelle and is continuous with the outer nuclear envelope (oNE). As such, the EpM connects the plastid to the endomembrane system, which comprises the nuclear membranes (N) and the endoplasmic reticulum (ER). A hinge region is present where the PPM forms a membrane contact site with the inner nuclear envelope (iNE), potentially harbouring molecular machinery for the trafficking of small molecules and information between the nucleus and the plastid. The mitochondrion (M) and the peroxisome (Px), both contain enzymes involved in β-oxidation, complete the fatty acid life cycle. Scheme based on information from [32, 74, 144, 216, 217]

In contrast, thraustochytrids, members of the Bigyra clade within the phylum Labyrinthulomycetes, are obligate heterotrophic, unicellular marine protists. These organisms are ecologically important in both coastal and pelagic ecosystem [76, 77] and are found across a wide range of marine habitats, from tropical to Antarctic waters, and at depths up to 2,000 m [30]. Thraustochytrids are particularly abundant in detritus-rich environments such as mangrove forests, estuaries, and surface sediments. Genera such as Schizochytrium and Aurantiochytrium are noted for their exceptional ability to biosynthesise high levels of very-long-chain polyunsaturated fatty acids (VLC-PUFAs), particularly ω3-docosahexaenoic acid (DHA, 22:6) [30].

Glycerolipids and fatty acids in stramenopiles

The membranes of living organisms are composed of lipids, proteins and carbohydrates [78]. Membrane glycerolipids form a bilayer and contain a polar head group oriented towards the aqueous environment and a hydrophobic tail situated within the bilayer core. Glycerolipids are composed of a glycerol backbone with FAs esterified at the sn-1 and sn-2 positions and a ‘polar head’ group bonded to the sn-3 position, thereby determining the lipid class.

Stramenopiles typically contain nine classes of membrane glycerolipids (excluding lysolipids) (Fig. 3); namely, five phospholipids, three glycolipids, and either one, two or three betaine lipids (depending on the species [79]). Furthermore, they contain a non-structural neutral lipid, triacylglycerol (TAG), contained in specialised organelles called the lipid droplets (LDs) [33] and composed of three FAs esterified to a glycerol backbone.

Fig. 3.

Schematic representation of membrane lipid classes in photosynthetic stramenopiles. Created with biorender (https://www.biorender.com)

Phosphoglycerolipids (PLs) and betaine lipids (BLs) are typically found in the endomembrane systems of simple eukaryotes. The PL classes include three anionic lipids (phosphatidic acid (PA), phosphatidylglycerol (PG), and phosphatidylinositol (PI)) and two zwitterionic ones (phosphatidylcholine (PC) and phosphatidylethanolamine (PE)). Three distinct BLs have been identified in stramenopiles: 1(3),2-diacylglyceryl-3(1)-O-4’-(N, N,N-trimethyl)-homoserine (DGTS), 1(3),2-diacylglyceryl-3(1)-O-2’-(N, N,N-trimethyl)-β-alanine (DGTA), and 1,2-diacylglyceryl-3-(O-carboxyhydroxymethylcholine (DGCC). BLs frequently occur in organisms lacking PC or both PC and PE [80, 81], however, in stramenopiles, they co-exists with both PC and PE [79]. Additionally, PA and PG are present in both plastids and endomembrane systems (Fig. 3).

Glycoglycerolipids (GLs) are a conserved feature of all photosynthetic organisms [82, 83], serving as integral components of plastidial membranes. The GL classes comprise two galactolipids, monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG), along with one anionic sulfolipid, sulfoquinovosyldiacylglycerol (SQDG).

In all these glycerolipid classes, FA composition varies with different lengths and degrees of unsaturation, which modify their biochemical properties [84–86].

The major FAs present in stramenopiles include 14:0, various degrees of unsaturation of C16 and C18 FAs, 20:4, 20:5 and 22:6 [30, 39, 87–94]. In osmotrophic, saprophytic thraustochytrids, 15:0 and 17:0 can be also found (see next section) (Table 1).

Table 1.

FAs reported in stramenopiles. Information about the melting point were retrieved from PhysProp in PubChem if not differently indicated (* = Hazardous Substances Data Bank (HSDB) [100]), ** = Human Metabolome Database (HMDB))

IUPAC	Common name	PubChem CID	Abbreviation	Melting point [°C]
Tetradecanoic acid	Myristic acid	11,005	14:0	53.9
Pentadecanoic acid			15:0	52.3**
Hexadecanoic acid	Palmitic acid	985	16:0	61.8
(9Z)-Hexadec-9-enoic acid	Palmitoleic acid	445,638	16:1	-0.1
(3E)-3-hexadecenoic acid		97,748	16:1 trans	NA
(9Z,12Z)-hexadeca-9,12-dienoic acid	Palmitolinoleic acid	5,312,427	16:2	NA
(6Z,9Z,12Z)-hexadeca-6,9,12-trienoic acid		52,921,838	16:3	NA
(6Z,9Z,12Z)-6,9,12,15-Hexadecatetraenoic acid		11,957,734	16:4	NA
Heptadecanoic acid		10,465	17:0	61.3**
Octadecanoic acid	Stearic acid	5281	18:0	68.8
cis-9-octadecenoic acid	Oleic acid	445,639	18:1	13.4
(9Z,12Z)-octadeca-9,12-dienoic acid	Linoleic acid	5,280,450	18:2	-6.9*
(9Z,12Z,15Z)-octadeca-9,12,15-trienoic acid	α-linolenic acid	5,280,934	18:3 n-3	-16.5
(6Z,9Z,12Z)-octadeca-6,9,12-trienoic acid	γ-linolenic acid	5,280,933	18:3 n-6	NA
(6Z,9Z,12Z,15Z)-octadeca-6,9,12,15-tetraenoic acid	Stearidonic acid	5,312,508	18:4	NA
(11Z,14Z,17Z)-icosa-11,14,17-trienoic acid	Dihomolinolenic acid	5,312,529	20:3 n-3	NA
(8Z,11Z,14Z)-icosa-8,11,14-trienoic acid	Dihomo-γ-linolenic acid	5,280,581	20:3 n-6	NA
(5Z,8Z,11Z,14Z)-icosa-5,8,11,14-tetraenoic acid	Arachidonic Acid	444,899	20:4	-49.5
(5Z,8Z,11Z,14Z,17Z)-icosa-5,8,11,14,17-pentaenoic acid	Eicosapentaenoic acid (EPA), timnodonic acid	446,284	20:5	NA
(7Z,10Z,13Z,16Z,19Z)-Docosa-7,10,13,16,19-pentaenoic acid	Docosapentaenoic acid, clupanodonic acid	5,497,182	22:5	NA
(4Z,7Z,10Z,13Z,16Z, 19Z)-docosa-4,7,10,13,16,19-hexaenoic acid	Docosahexaenoic acid (DHA), cervonic acid	445,580	22:6	NA
Tetracosanoic acid	Lignoceric acid	11,197	24:0	84.2**

De novo fatty acid synthesis in stramenopiles

De novo FA synthesis in ochrophytes like in all photosynthetic organisms occurs predominantly within the plastids [95] This process begins with the carboxylation of acetyl-CoA (a two-carbon molecule derived from several pathways [96–98]) by acetyl-CoA carboxylase (ACCase), an ATP-dependent reaction that incorporates a bicarbonate-derived carbon to form malonyl-CoA [99] (a three-carbon molecule) (Fig. 4). ACCase can occur in a homomeric and heteromeric isoforms but photosynthetic stramenopiles seem to have conserved mainly the homomeric ones [100]. The homomeric isoform is composed of a single large polypeptide that includes all catalytic domains arranged in tandem: the biotin carboxylase domain (BC), which contains the ATP-binding motif and catalyses CO₂ fixation [101]; the biotin carboxyl carrier protein (BCCP), which serves as the attachment site for the biotin prosthetic group [102]; the carboxyltransferase domain (CT), composed of structurally distinct α- and β-subunits (α-CT and β-CT). The α-CT domain binds acetyl-CoA, which undergoes carboxylation, while the β-CT subunit binds carboxybiotin [102].

Fig. 4.

FAS and PUFA synthase. Created with biorender (https://www.biorender.com)

In P. tricornutum two homomeric nuclear-encoded ACCase, PtACC1 (Phatr3_EG01955) and PtACC2 (Phatr3_J55209) have been identified [100], with PtACC1 being the focus of several studies in different diatoms [102–107]. In the eustigmatophyte Nannochloropsis oceanica, however, the available data are inconsistent among sequenced strains. Two multifunctional monomeric ACCase encoding genes were identified in the genome of strain IMET1 [108] whereas only the multimeric form was identified in strain CCMP1779 of the same species [109]. A TBLASTN and a BLASTP search using the PtACC1 and PtACC2 amino acid sequences in the Microchloropsis gaditana (formerly known as Nannochloropsis gaditana [110]) genomes present in EnsemblProtist [111] did not produce single hits leading to the conclusion that only the heteromeric isoform exists. Thraustochytrids, the only non-photosynthetic oleaginous stramenopile included in this review, possess monomeric ACCase that has been genetically manipulated to enhance FA and lipid production [30, 77, 112–116].

Subsequently, malonyl-CoA is converted to malonyl-ACP by malonyl-CoA: ACP transacylase (MCMT). In N. oceanica, P. tricornutum, and the thraustochytrid Schizochytrium, overexpression of the endogenous MCMT resulted in an increase in VLC-PUFAs [117–119].

De novo FAs are ultimately synthesised by a multi-subunit bacterial type II FA synthase (FASII) complex whose evolutionary history was investigated by Kohli and colleagues [120] in 213 strains (representing 152 genera) of marine eukaryotes using the Marine Microbial Eukaryote Sequencing Project (MMETSP) database [121]. The complex is comprised of four key enzymes: ketoacyl-ACP synthase (KAS), ketoacyl-ACP reductase (KAR), hydroxyacyl-ACP dehydratase (HAD), and enoyl-ACP reductase (ENR). The KAS enzyme catalyses the condensation of malonyl-ACP with an acetyl-CoA molecule, yielding 3-ketoacyl-ACP while releasing CO₂. The resulting 4-carbon 3-ketoacyl-ACP then undergoes a series of modifications, including reduction by KAR, dehydration by HD, and a second reduction by ENR (Fig. 4). This ultimately yields a 6-carbon acyl-ACP. In the stramenopile kingdom, the FASII cycle typically repeats six or seven times to produce C14-ACP or C16-ACP, respectively.

In addition to FASII, M. gaditana also possesses a cytosolic type I FAS, typically involved in the elongation of FAs beyond C18, or supporting FASII to cope with a high demand for FA. In M. gaditana, an increase in the transcription of FASI was observed under high light conditions accompanied by a decrease in FASII transcription, indicating a shift in FA synthesis activities from the chloroplast to the cytoplasm [122].

In plants, 16:0 FA can undergo elongation to 18:0 via a plastidial β-ketoacyl-ACP synthase II (KASII) [123]. However, the question of whether plastids extend FA synthesis to 18 carbons or limit it to 16 carbons remains unresolved in stramenopiles [39, 124, 125]. The low levels of 18-carbon FAs in plastidial lipids, coupled with the identification of a Δ0-elongase localised in the extraplastidial environment of eustigmatophytes and diatoms that catalyses the elongation of 16:0-CoA to 18:0-CoA, strongly suggest that the synthesis of 18-carbon FAs in stramenopile plastids is very likely minimal or absent [124, 126].

Thraustochytrids possess two biosynthetic pathways to FAs, one based on FASI, which is in most species incomplete, and a PUFA-synthase pathway, the so-called polyketide synthase (PKS) pathway [30, 112, 127]. The two pathways share the same entry building block (acetyl-CoA) carboxylated into malonyl-CoA by ACCase, in the committed step of even-numbered FA synthesis (Fig. 4). The in vivo overexpression of ACCase (in concert with the overexpression of ATP-citrate lyase to increase the cytosolic acetyl-CoA pool) in Schizochytrium sp. induces a lipid accumulation and an over-production of 22:6 [115]. This is where the two pathways diverge, FASI produces C14:0 and C16:0, whereas PUFA synthase (Fig. 4) leads to the synthesis of 22:5 and 22:6 [30, 112]. The PUFA synthases contain three dehydratase domains, some of these domains also exhibiting an additional isomerase activity. These combined activities not only introduce a double bond but can also convert a trans double bond into a cis configuration [128, 129]. The regulation and coordination of the PUFA synthase domains remain an area of ongoing research. In thraustochytrids, both the FAS and PUFA synthase pathways are present, and neither appears to be redundant or dispensable [130].

Noteworthy, thraustochytrid FA arsenal includes odd-numbered carbon chains as well. When the culture medium contains branched-chained amino-acids, propionyl-CoA (a 3 C molecule) is produced and it enters the FA synthesis metabolism together with acetyl-CoA [131], as experimentally demonstrated by feeding Aurantiochytrium limacinum with acetate and propionate [30].

Metabolic journey of a plastidial fatty acid: from 16:0 to 16:4 in plastidial glycerolipids

16:0-ACP-to 16:1-ACP desaturation and esterification to plastidial lipids

De novo FA synthesis results in the production of 14- and 16-carbon saturated acyls, which are linked to ACP in the plastid. No desaturations occur on 14:0. Nevertheless, up to four unsaturations may be introduced on the initial 16:0. These reactions occur within the plastid. In P. tricornutum, the desaturation of 16:0 to 16:1^Δ9 occurs on palmitoyl-ACP and is catalysed by a soluble acyl-ACP Δ9-desaturase, designated as palmitoyl-ACP desaturase (PtPAD, Phatr3_J9316), in the stroma [126]. The orthologue of PtPAD in N. oceanica has also been identified and experimentally localised to the plastid [132].

Further desaturations of 16:1^Δ9 are conducted by plastid-localised [133] membrane-bound desaturases on acyl chains esterified at the sn-1 and sn-2 positions of the glycerol backbone in lipids [134]. In the plastid, acylation of glycerol-3-phosphate (G3P) is catalysed by plastidial glycerol-3-phosphate acyltransferase (ATS1) and lysophosphatidic acid acyltransferase (ATS2), which incorporate FAs at the sn-1 and sn-2 positions, respectively. Noteworthy, once 16:0 has been esterified to the glycerol backbone of glycerolipids, it remains in its saturated form and cannot undergo further desaturation because PtPAD is specific to the acyl carrier protein-bound form of the 16:0-ACP and does not act on glycerolipid-bound 16:0 [126]. Subsequent desaturations in the plastid occur on esterified 16:1 [134].

From the FA composition at the sn-1 position of plastidial GLs (MGDG, DGDG, SQDG) and PG, it can be inferred that ATS1 displays a high degree of versatility in FA utilisation. In P. tricornutum, these FAs include 14:0, 16:0, 16:1, and potentially 20:5 [39]. In the case of M. gaditana, 20:5 is also present in the sn-2 position of MGDG and DGDG (Fig. 5) [38].

Fig. 5.

Schematic representation of plastid FA and lipid biosynthesis and of the journey of a FA from the plastid to the cytosol in the model microalgae Phaeodactylum tricornutum and Microchloropsis gaditana. FAs are de novo synthesised in the plastid and subsequently incorporated into membrane lipids through a conserved pathway. These reactions lead to the production of major plastid membrane components, including phosphatidylglycerol and galactolipids, which are essential for photosynthetic membranes. The diagram highlights the main metabolic steps and cellular compartments, while indicating typical fatty-acid chain lengths and degrees of unsaturation. FA indicated in brown are typical of P. tricornutum whereas the ones indicated in blue are found in M. gaditana. FAs in black are shared between the two models. Regiospecificity was determined with the methods described by Jouhet et al., 2017 [218]. Created with biorender (https://www.biorender.com)

Significant differences are observed between diatoms and eustigmatophytes in the FA composition at the sn-2 position of plastidial glycerolipids. In diatoms, 20:5 and 14:0 are absent from the sn-2 position [39], whereas in eustigmatophytes MGDG 20:5/14:0 and 20:5/20:5 have been reported [38] (Fig. 5). These compositional differences indicate that, in diatoms, ATS2 utilises 16:0, 16:1, and 24:0, but not 14:0 or 20:5. In eustigmatophytes, however, the presence of 20:5 at the sn-2 position suggests that ATS2 has the potential to utilise FAs such as 14:0, 16:0, 16:1, and 20:5.

In P. tricornutum, the enzyme encoded by gene Phatr3_J3262, designated as PtATS1, is predicted to be targeted to the plastid. Although the overexpression of PtATS1 has been demonstrated to induce an increase of 2.9-fold in TAGs, with alterations in their FA composition [135], insight into the role of this enzyme in the synthesis of plastidial glycerolipids has yet to be demonstrated.

Two PtATS2 genes have been characterised [136]. PtATS2a (Phatr3_J11916) localises within the iEM, whereas PtATS2b (Phatr3_J43099) is located at the thylakoids. The observed alterations in glycerolipid profiles are consistent with the proposed function of both PtATS2 enzymes in catalysing the acylation of sn-1-acyl-glycerol-3-phosphate substrates containing 20:5 at the sn-1 position. In M. gaditana, a single gene, designated MgATS1, is predicted to encode an ATS1 protein with plastidial localisation. Additionally, three distinct ATS2 are predicted.

Li and colleagues [137] observed that microalgal plastidial GPATs (ATS1) display a preference for acyl-ACP when tested in comparison with acyl-CoAs. This preference is associated with the presence of specific residues, namely K204, R212 and R266, which act as interaction sites with the substrates in the trebouxiophycean green alga Lobosphaera incisa MiGPAT1 (previously identified as Myrmecia incisa [138] and cited under this designation in [137]). Although the aforementioned results originate from the investigation of green microalgae, analogous residues that mimic analogous positions on protein sequences have been identified in PtATS1. The conservation of these residues suggests their potential role as critical determinants for the recognition of plastidial GPATs in microalgae, positioning them as promising targets for investigating enzyme specificity.

16:1-to16:4 desaturation on platidial lipids

In P. tricornutum, the desaturation of 16:1 to 16:4 esterified on plastidial glycerolipids is mediated by membrane-bound desaturases such as PtFAD6 (Phatr3_J48432) localised in the oEM [133]. PtFAD6 catalyses the conversion of 16:1^Δ9 to 16:2^Δ9,12 in acyl chains esterified at either the sn-1 or sn-2 positions of MGDG, DGDG and SQDG [125]. PtFAD6 expression increases during stationary phase [139] but is downregulated under high-light conditions [140]. PtPlastidΔ6FAD (Phatr3_EG02619) is predicted to desaturate 16:2^Δ9,12 likely at the sn-2 position of MGDG to produce 16:3^Δ6,9,12. Despite previous suggestions that the enzyme does not act on the sn-1 position, the detection of MGDG 32:6 (most likely MGDG 16:3/16:3) suggests potential enzymatic activity at both positions, unless acyl-editing processes is involved in the synthesis of this lipid species. Furthermore, the Plastidω3FAD/FAD7 (Phatr3_J41570) enzyme is predicted to catalyse the terminal desaturation of 16:3^Δ6,9,12 at the sn-2 position of MGDG, resulting in the formation of 16:4^Δ6,9,12,15 [134].

The 16:3^Δ6,9,12 FA is predominantly associated with both galactolipids, whereas 16:4^Δ6,9,12,15 is largely restricted to MGDG. Only trace amounts of 16:4 are detected in DGDG, and its abundance increases exclusively under low-temperature conditions [141]. These observations raise important questions regarding the lipid environment in which desaturation reactions occur. Both 16:3 and 16:4 can be present in 20:5-containing MGDG and DGDG; however, within the 16 C–16 C molecular species, they are strictly confined to MGDG and are absent from DGDG. Consistently, 16:2 represents the highest degree of unsaturation detected in DGDG 16 C–16 C species. This distribution supports the hypothesis that the progressive desaturation of 16:1 to 16:4 occurs primarily on MGDG. In particular, the conversion of 16:2 to 16:3 and the subsequent desaturation step from 16:3 to 16:4 appear to be restricted to MGDG 16 C–16 C, as these highly unsaturated species are not detected in DGDG 16 C–16 C. By contrast, 20:5-containing galactolipids may provide an alternative platform for desaturation reactions, allowing the conversion of 16:2 to 16:3 and of 16:3 to 16:4 through the action of the oEM-localised desaturases PtPlastidΔ6FAD and Plastidω3FAD/FAD7, respectively. An open question is whether PtPlastidΔ6FAD is unable to use DGDG 32:3 (mainly composed of DGDG 16:1/16:2) as a substrate but can act on 36:7 DGDG (mainly composed of DGDG 20:5/16:2), or whether DGDG species are excluded altogether from its substrate spectrum, with desaturation occurring exclusively on 16:2-containing MGDG.

Plastidial phosphatidylglycerol (PG) desaturation at the sn-2 position by FAD4: specific (16:0 → 16:1^Δ3 (trans)) desaturation

It is widely accepted that PG is the sole phosphoglycerolipid synthesised within the chloroplast envelope. A hallmark feature of chloroplast PG is the presence of a 16:1^Δ3 (trans) FA at the sn-2 position, which is crucial for its structural and functional interactions with components of the photosynthetic apparatus [142]. In Arabidopsis thaliana, the desaturation of 16:0 to 16:1^Δ3 (trans) occurs at the sn-2 position of PG and is catalysed by the membrane-bound desaturase FAD4 [143].

A homolog of A. thaliana FAD4 in P. tricornutum (Phatr3_J5271) has been experimentally localised to the stroma [133]. The stromal localisation of PtFAD4 in diatoms is distinct from the membrane-bound FAD4 in plants [143], indicating the potential for a different mode of operation for these two homologous enzymes.

In P. tricornutum, PG predominantly exists in two molecular species, 36:6 (highly likely 20:5/16:1) and 32:1 (highly likely 16:1/16:0). Prior research established, on the basis of phospholipase A₂ hydrolysis, that the 16:1 at the sn-1 position corresponds to 16:1^Δ3 (cis), while 16:1 at the sn-2 position corresponds to 16:1^Δ3 (trans) [40]. Molecular characterisation of PtPAD, responsible for 16:1^Δ9 synthesis, shows that knocking PtPAD out results in the complete loss of PG 16:1/16:0, while overexpression lines display a statistically significant increase in these species [126], thereby confirming the positional assignment described above [40]. In PtPAD knockout lines the dramatic reduction of 16:1^Δ9 is accompanied by a concomitant increase in 16:0 [126]. Such an excess of 16:0 is exported out the plastid, where its desaturation is presumed to be impossible [134] and supports the synthesis of 20:5 (which is enriched in PtPAD KO lines [126]) via elongation to 18:0 and ensuing desaturation and elongations. The decrease in 16:1^Δ9 affects the MGDG synthesis which is reduced almost one third in PtPAD KO lines [126] whereas PG markedly increase, especially due to PG 36:6 [126]. This evidence, suggests that the 16:1 at the sn-2 position in this species may be desaturated directly on the PG by the action of PtFAD4. Although no molecular characterisation of PtFAD4 is still available, it is tempting to hypothesise that the product of PtFAD4is indeed the enigmatic 16:1^Δ3 (trans).

Given that 16:1^Δ3 (trans) is exclusively found in plastidial PG, it is likely that the PG 36:6 species is plastid-localised. In contrast, the PG 32:1 species, which shares compositional similarity with extraplastidial phosphatidylinositol (PI), likely represents an extraplastidial PG species [39]. PGPS (Phatr3_J8663), the enzyme converting CDP-DAG to PG, has been identified in the plastid proteome of P. tricornutum [133], but its exact localisation remains experimentally unverified. In contrast, PGPP (Phatr3_J6288), the enzyme responsible for converting phosphatidylglycerolphosphate (PGP) to PG, has been experimentally localised to the blob. If PGPP is indeed the sole enzyme responsible for PG synthesis in diatoms, as suggested by the absence of additional hits when the PtPGPP protein was blasted against the P. tricornutum genome, the blob could represent a key platform for PG biosynthesis.

The blob seems to play a crucial role in lipid production and dispatching within and beyond the plastidial membranes as suggested by Guéguen and co-workers [144] as well. Once produced in the blob, PG can either be reintroduced in the inner compartments of the plastid to reach thylakoids or be exported. This distinction is critical due to the uncertainty surrounding the mode of action of FAD4 in diatoms. If PtFAD4 functions as in plants, acting on the 16:0 esterified on PG, then the PG 20:5/16:0 could be synthesised in the blob, imported back into the plastid, and subsequently desaturated by PtFAD4 into PG 20:5/16:1 (trans). Despite these advancements, many aspects of PG biosynthesis in P. tricornutum remain unclear, necessitating further investigation into the localisation and functionality of enzymes such as PGPS, PGPP, and FAD4, as well as the trafficking of PG between plastidial and extraplastidial compartments.

Following a fatty acid beyond the plastid

Thioesterase-mediated release of fatty acids from ACP: a crucial step for export from the plastid

Prior to the export of FAs from the plastid, acyl-ACP thioesterases catalyse the hydrolysis of the thioester bond between the acyl chain and the ACP, resulting in the formation of free FAs (FFAs) and ACPs. Through this step, the quantity and type of FAs exported are determined. A plastid-localised Hotdog-fold thioesterase, designated PtTES1 (Phatr3_J33198), has been characterised in P. tricornutum (Fig. 5). However, it has been proposed that it functions more as an acyl-CoA thioesterase and thus may not be involved in acyl-ACP hydrolysis prior to export [145, 146]. It is possible that other thioesterases are responsible for acyl-ACP hydrolysis. One such enzyme is the palmitoyl-protein thioesterase (PtPPT, Phatr3_J10454), which is predicted to be localised in the plastid. Nevertheless, further molecular characterisation of this enzyme is necessary.

The overexpression of PtPAD resulted in an increase in the production of 16:1 in the plastid, which subsequently led to an increase in 16:1-containing extraplastidial lipids [126]. This indicates that the 16:1 FAs produced by PtPAD can be exported for the synthesis of extraplastidial lipids. It can therefore be assumed that both 16:0-ACP and 16:1-ACP can be hydrolysed by thioesterases prior to their export.

Fatty acid export from the plastid

The process by which newly synthesised FAs are exported from the plastid stroma and cross its four membranes remains largely unknown. In angiosperms, this export may occur via facilitated diffusion, potentially involving systems similar to FA Export 1 (FAX1) localised at the inner envelope membrane (iEM) of the plastid [147]. Once FAs reach the oEM, long‑chain acyl‑CoA synthetases (LACS) catalyse the activation of FFAs into acyl‑CoA, a necessary step for their subsequent incorporation into glycerolipid [148]. Once formed, acyl‑CoA may be stabilised and trafficked within the cell by acyl‑CoA‑binding proteins (ACBPs), which bind medium‑ and long‑chain acyl‑CoA with high affinity and can facilitate their intracellular exchange between membranes and compartments [149–151].

Another major mechanism influencing how FAs are incorporated into extraplastidial lipid pools is acyl‑editing, in which lyso‑phospholipids such as lysophosphatidylcholine (lysoPC) are reacylated into PC by lysophosphatidylcholine acyltransferases (LPCATs). LPCATs transfer an acyl group from acyl‑CoA to lysoPC, thereby modifying the acyl composition of membrane lipids and participating in the exchange of fatty acids between acyl‑CoA and phospholipid pools [152]. LysoPC itself can move between membranes more readily than PC, and its reacylation at ER-plastid contact sites contributes to the incorporation of exported FA into ER phospholipid metabolism without implying that lysoPC per se is the direct transporter of plastid‑produced FAs [153]. In P. tricornutum, the enzyme PtLPCAT1 (Phatr3_J20460), an acyl‑CoA:lysophosphatidylcholine acyltransferase, has been characterised as a key acyl‑editing enzyme involved in the synthesis of triacylglycerol and galactoglycerolipids and in remodelling phospholipids, including the incorporation of VLC-PUFAs into PC and downstream lipid pools [154, 155]. PtLPCAT1 localises to the plastid–ER membrane interface, consistent with a role in mediating acyl‑CoA incorporation into PC at sites of lipid trafficking between these compartments [155].

In addition to carrier proteins and acyl‑editing enzymes, the movement of acyl chains may also involve ATP‑binding cassette (ABC) transporters, such as ABCA9 in Arabidopsis, which has been implicated in lipid homeostasis and could participate in acyl‑CoA or FA transport [156]. Taken together, these mechanisms suggest that FAs exported from the plastid may travel in multiple biochemical forms, as unesterified FFAs, as acyl‑CoA bound to ACBPs, or as acyl groups exchanged via lysoPC/PC acyl‑editing cycles, with distinct enzymatic and transporter systems participating at different steps. In stramenopiles, however, the mechanisms for FA export across the four plastid membranes remain entirely uncharacterised.

Activation of 16:0 and 16:1 by Long Acyl-CoA Synthetase (LACS)

In order for FFAs to be esterified on a glycerol backbone or elongated by elongases, they have to be activated through the esterification with coenzyme A (CoA) to form acyl-CoA. Acyl-CoA synthetases and long-chain acyl-CoA synthetases (LACS) catalyse this reaction. P. tricornutum and M. gaditana contain five and six putative LACS genes, respectively.

In P. tricornutum, the activation of newly exported 16:0 and 16:1 FAs may be mediated by PtACSL2 and PtACSL4, both localised in the cytoplasm. Hao and colleagues [157] demonstrated that PtACSL2 has a high affinity for 16:0. The affinity of PtACSL2 and PtACSL4 for 16:1 has yet to be investigated. The orthologue of PtACSL4 (Phatr3_J45510) in M. gaditana, a bubblegum acyl-CoA synthetase (MgACSBG), is involved in the production of 16:1-CoA [38], which suggests that PtACSL4 may play a similar role in P. tricornutum.

Elongation of 16:0-CoA to 18:0-CoA by Δ0-ELO elongase

The 16:0-CoA can then be elongated to 18:0-CoA via an ER-localised saturated FA elongase (Δ0-ELO). M. gaditana genome encodes for seven saturated FA elongases, among which NgΔ0‐ELO1. NgΔ0-ELO1 has been functionally validated to catalyse the elongation of 16:0-CoA to 18:0-CoA [124]. The KO of NgΔ0‐ELO1 in M. gaditana resulted in a reduction in eicosapentaenoic acid (EPA, 20:5) levels and a specific decrease in MGDG. These functional genomic analyses further presented evidence indicating that a portion of the EPA utilised in MGDG synthesis is synthesised through a directed process initiated during the elongation of palmitic acid by NgΔ0‐ELO1. Despite being localised in the endoplasmic reticulum, this enzyme serves therefore as a catalytic channel, initiating the production of 20:5 specifically incorporated into galactolipids. Very recent evidence from the related eustigmatophyte N. oceanica corroborates the importance of the C16-to-C18 elongation step in determining the flux toward EPA [158]. A minor elongase isoform, NoELO2, has been identified as a key control point for directing FA precursors into the ω3 biosynthetic pathway. Heterologous expression confirmed that NoELO2 elongates 16:0 to 18:0. NoELO2 was reported to localise to the endoplasmic reticulum closely associated with the plastid [158], however, inspection of the subcellular fluorescence pattern suggests a localisation in the EpM, as the signal extends around the nucleus. Given that in stramenopiles the EpM is continuous with the oNE [74, 75], this distribution is consistent with an EpM localisation, placing the enzyme at a strategic interface between plastidial FA synthesis and extraplastidial lipid metabolism. Overexpression of NoELO2 in alga markedly enhanced EPA levels, whereas KO mutants showed a reduced EPA content [158]. Taken together, the findings from M. gaditana and N. oceanica show a conserved metabolic role for Δ0-elongases in controlling the entry of 16:0 fatty acids into the biosynthetic pathway to EPA. In both species, Δ0-elongases seem to function as metabolic gatekeepers that regulate the availability of C18 intermediates required for long-chain PUFA biosynthesis.

Two homologues of NgΔ0-ELO1 have been identified in P. tricornutum, PtΔ0-Elo1 (Phatr3_J16376) and PtΔ0-Elo2 (Phatr3_J49867). The malonyl-CoA required for the elongation reaction could be supplied by extraplastidial acetyl-CoA carboxylase (PtACC2).

The discovery of Δ0-ELOs highlights an innovation in stramenopiles, allowing them to utilise a distinct pathway for the production of 18:0-CoA, which differs from that deciphered in angiosperms. Indeed, it can be inferred that stramenopiles employ an extraplastidial elongation of 16:0-CoA to produce 18:0-CoA, independent of a plastidial β-ketoacyl-ACP synthase II (KASII) pathway. Nevertheless, a KASII orthologue has been identified in P. tricornutum (Phatr3_J37367): its functional characterisation needs therefore to be elucidated.

Desaturation of 16:0-CoA to 16:1-CoA and 18:0-CoA to 18:1-CoA by acyl-lipid desaturase ADS1

In the cytosol of P. tricornutum, 18:0-CoA is likely desaturated by a putative Δ9-oleyl desaturase, PtADS1 (Phatr3_J29797), localised in the EpM [133]. This yields 18:1^Δ9-CoA. It is important to note that 18:1^Δ9-CoA cannot be derived from the elongation of 16:1^Δ9-CoA, as the elongation of 16:1^Δ9-CoA would instead yield 18:1^Δ11-CoA. Additionally, PtADS1 may desaturate 16:0-CoA to 16:1^Δ9-CoA, as evidenced by the presence of residual 16:1^Δ9-CoA in PtPAD KO mutants [126].

Extraplastidial PA and DAG generation

Newly synthesised and exported FAs can be channelled either into de novo glycerolipid assembly or into acyl editing pathways [152, 159–161]. In angiosperms, freshly exported FAs are preferentially incorporated into phosphatidylcholine (PC) by lysoPC acyltransferase (LPCAT), which acylates lysoPC and regenerates PC. This incorporation constitutes the entry point of acyl chains into the so-called Lands’ cycle. Once esterified at the sn-2 position of PC, acyl chains can undergo desaturation and other modifications [159–161]. Subsequently, phospholipase-mediated cleavage at the sn-2 position releases a FFA and regenerates lysoPC, which can again be reacylated by LPCAT, thereby completing the cycle. Through this dynamic process, plants efficiently remodel acyl chains on PC and prevent the accumulation of highly saturated FAs in membrane lipids. The overall contribution and mechanistic relevance of the Lands’ cycle in stramenopiles, however, remain poorly understood and warrant further investigation.

If FAs from the cytosolic pool are not directed to acyl editing, they are instead channelled into glycerolipid biosynthesis. When esterified onto extraplastidial glycerolipids such as phospholipids and BLs, these FAs can undergo further desaturation. Extraplastidial glycerolipid production occurs similarly to plastidial glycerolipid synthesis: FAs are esterified onto a glycerol backbone at the sn-1 and sn-2 positions through the actions of GPAT (Phatr3_J54709) and lysophosphatidic acid acyltransferase (LPAAT, Phatr3_J45551) (Fig. 5). It is worth noting that an LPAAT activity was attributed to the gene Phatr3_J20460 [162] based on the reported acylation of LysoPC to produce PC [154]. However, by definition, LPAAT (lysophosphatidic acid acyltransferase) catalyses the acylation of lysophosphatidic acid (LPA), not LysoPC; therefore, the reported enzymatic activity is more consistent with that of a lysophosphatidylcholine acyltransferase (LPCAT) rather than an LPAAT. Extraplastidial GPAT and LPAAT enzymes are capable of utilising a broad range of FAs, including 16:0, 16:1, 18:0 and 18:1. Given the paucity of 16:0 and 18:0 in extraplastidial lipids, GPAT and LPAAT have possibly a preference for monounsaturated FAs, which serve as substrates for subsequent desaturation. Homologous and heterologous overexpression studies have demonstrated that elevated GPAT expression is associated with increased TAG content [135, 163–166], indicating a pivotal function for GPAT in this biosynthetic pathway. The in vitro and in vivo substrate specificity and role in membrane glycerolipid homeostasis of GPATs in diatoms and eustigmatophytes has yet to be fully characterised.

Four LPAATs have been characterised in N. oceanica [167]. LPAT1 and LPAT2 are localised to the ER, while LPAT3 and LPAT4 are unexpectedly localised to the periphery of LDs. LPAT1 is the most involved in the formation of membrane lipids and exhibits a high affinity for 16:1. LPAT2 and LPAT3 may also play a role in membrane lipid formation, albeit to a lesser extent, by adding 18:1 at the sn-2 position of a LPA. LPAT2, LPAT3 and LPAT4 appear to play a significant role in the formation of TAGs [167].

EPA synthesis combines ω6 and ω3 pathways in Ochrophyta

Pulse-chase experiments in P. tricornutum have demonstrated that EPA can be synthesised via multiple pathways, with the predominant route involving Δ6-desaturation of 18:2^Δ9,12 to 18:3^Δ6,9,12, utilising intermediates of both the ω6 and ω3 pathways [168]. The most active pathway is as follows: 18:0 → 18:1^Δ9 → 18:2^Δ9,12 → 18:3^Δ6,9,12 → 18:4^Δ6,9,12,15 → 20:4^Δ8,11,14,17 → 20:5^{Δ5,8,11,14,17}.

In eustigmatophytes, the ω6 pathway prevails [169, 170], with the most active pathway being: 18:0 → 18:1^Δ9 → 18:2^Δ9,12 → 18:3^Δ6,9,12 → 20:3^Δ8,11,14 → 20:4^Δ5,8,11,14 → 20:5^{Δ5,8,11,14,17}.

It is difficult to ascertain the precise glycerolipid platforms involved in FA desaturation in P. tricornutum because they are present across all extraplastidial lipid classes. In contrast, in eustigmatophytes such as M. gaditana or Monodus subterraneus, the FAs 18:1, 18:2 and 18:3 are exclusively found in PC, whereas 20:3 and 20:4 are specifically associated with DGTS and PE. It can therefore be inferred that the desaturation of 18:1 to 18:3 occurs on PC, while the desaturation processes from 20:3 to 20:5 are likely to be confined to DGTS and/or PE in M. gaditana and potentially eustigmatophytes [38, 171].

Membrane-bound desaturase-mediated desaturation from 16:1 to 16:3 and 18:1 to 18:3 or 18:4

In P. tricornutum, the membrane-bound desaturases involved in the desaturation from 18:1 to 18:4 have all been experimentally localised in the EpM [133], which is continuous with the ER membrane [74] (Fig. 2). The characterised PtFAD2 (Phatr3_J25769) enzyme has been demonstrated to catalyse the conversion of 18:1^Δ9 to 18:2^Δ9,12 [125]. The most active pathway involves the conversion of 18:2^Δ9,12 to 18:3^Δ6,9,12 via the action of a Δ6 Cytb-fused front end desaturase initially named PtD6 [172] then subsequently renamed PtERΔ6FAD (Phatr3_J29488) [134]. 18:3^Δ6,9,12 is supposed to be desaturated again by an undefined ω3FAD to 18:4^Δ6,9,12,15 which still needs to be molecularly characterised [134]. Nevertheless, Moreno and co-workers [173] measured comparable quantities of 18:3^Δ6,9,12 and 18:3^Δ,9,12,15, the latter being likely produced from 18:2^Δ9,12 by the action of a ω3FAD. In conclusion, it is hypothesised that the pathway to 18:4^Δ6,9,12,15 may be the resultant of the action of two enzymes sharing the same substrate, 18:2^Δ9,12, and producing each the substrate for the other. It is worth noting that only PtERΔ6FAD was formally characterised [172].

In M. gaditana, it has been proposed that the desaturation of 18:1 to 18:3 occurs on PC, as these FAs are exclusive to PC [38]. The production of 18:4 is not observed in M. gaditana.

In P. tricornutum, traces of 16:2 and 16:3 have been identified in PC, DGTA, and PE. This indicates that the PtFAD2 and ERΔ6FAD enzymes may also catalyse the conversion of 16:1 into 16:2 and 16:3, respectively.

Elongation of 18:3 to 20:3-CoA and 18:4 to 20:4-CoA: acyl release, activation, and subsequent elongation

The elongation of FAs occurs exclusively on acyl-CoA, rather than on glycerolipid-bound acyls. Therefore, for newly produced 18:3^Δ6,9,12 or 18:4^Δ6,9,12,15 esterified onto extraplastidial lipids to undergo elongation, they must first be liberated as FFAs. The liberation process can be achieved by lipases or by the potentially reversible action of LPAATs, which produces FFAs and lysophoslipids. Three distinct classes of phospholipases are capable of releasing FAs from phospholipids, phospholipases A1 (PLA1), phospholipases A2 (PLA2), and phospholipases B (PLB). PLA1 and PLA2 catalyse the hydrolysis of the sn-1 and sn-2 acyl ester bonds of phospholipids, respectively, resulting in the generation of FFAs and 2-acyl-1-lysophospholipids and 1-acyl-2-lysophospholipids, respectively. PLB exhibits both PLA and lysoPLA activities, and is capable of sequentially removing two FAs from phospholipids. In P. tricornutum, specific phospholipases, such as the EpM-localised enzymes PtPLA2.2 [133] and/or PtLPCAT1 [154], may release FAs from phospholipids. Other putative phospholipases (like Phatr3_J44005 and Phatr3_J44066 both harbouring a PLA1 domain as per InterProScan [174]) and/or LPAATs may play a role.

Once released, FFAs are activated to acyl-CoA by LACS. In the case of M. gaditana, 18:3-CoA may be produced by MgACSBG [38]. In P. tricornutum, 18:4-CoA may be synthesised by PtACSL4, which is a homolog of MgACSBG. Nevertheless, the involvement of other LACS enzymes in this process cannot be ruled out.

Subsequently, 18:3^Δ6,9,12-CoA and 18:4^Δ6,9,12,15-CoA can be elongated to 20:3^Δ8,11,14-CoA and 20:4^Δ8,11,14,17-CoA, respectively, by the action of a Δ6-elongase (Phatr3_J22274 and Phatr3_J20508).

Re-esterification of 20:3-CoA and 20:4-CoA by LPAAT in extraplastidial lipids and final desaturation for EPA production

In order to proceed with further desaturation, it is first necessary to re-esterify 20:3-CoA and 20:4-CoA onto lysoPC or lysoBL. The esterification of acyl groups onto lysolipids is catalysed by LPAATs. P. tricornutum PtLPCAT1 has been suggested to be able to incorporate 20:4^Δ8,11,14,17-CoA into PC [154], thereby enabling its subsequent desaturation to 20:5^{Δ5,8,11,14,17}. This desaturation is mediated by the characterised PtERΔ5FAD.1 (Phatr3_J46830) [172] and/or the predicted PtERΔ5FAD.2 (Phatr3_J22459) [134].

In M. gaditana, 20:3-CoA is exclusively re-esterified onto DGTS or PE by specific LPAATs. The re-esterified 20:3 is then desaturated to 20:4 by ERΔ5FAD and subsequently to 20:5 by MgERω3FAD [134].

The elongation/desaturation process from 20:5 to 22:6

A lipid-esterified 20:5 can be released by lipases or LPAATs and subsequently activated into 20:5-CoA by an ACS/LACS. PtACSL1 (Phatr3_J20143), which has a strong affinity for 20:5 and is localised in the EpM, may be responsible for this activation [157]. The 20:5-CoA is then elongated by a Δ5-elongase (ptELO5a, Phatr3_J9255) [175], resulting in the formation of 22:5^{Δ7,10,13,16,19}. Subsequently, 22:5-CoA is reintegrated into PC, or BLs by a LPAAT for final desaturation to 22:6^{Δ4,7,10,13,16,19}, catalysed by PtERΔ4FAD (Fig. 6) [134].

Fig. 6.

Proposed pathway for fatty acid lifecycle of Phaeodactylum tricornutum. Created with biorender (https://www.biorender.com)

Insights into the omega pathway: the three hypotheses

20:5 is notably abundant in plastid glycerolipids, particularly in MGDG and DGDG. Despite being a VLC-PUFA synthesised in the endoplasmic reticulum or the EpM, 20:5 has to be transported back to the plastid for incorporation into plastid glycerolipids. The current understanding of the mechanism by which 20:5 is transported back to the plastid is limited. This process is referred to as the ‘omega pathway’ [83]. The precise manner in which 20:5 completes its return journey to the plastid remains to be elucidated.

Three hypotheses have been proposed regarding the import of 20:5 into the plastid (Fig. 7):

Fig. 7.

Possible scenarios for the “Omega pathway” in Phaeodactylum tricornutum. The “Omega pathway” refers to the process by which 20:5 acyl groups are transferred from their site of synthesis to plastid glycerolipids. The precise route(s) and molecular determinants of this fatty acid trafficking remain unknown. Created with biorender (https://www.biorender.com)

• The DAG or PA import hypothesis posits that: 20:5 synthesised in extra-plastidial lipids and esterified onto PC, PE, or BLs has the potential to enter the plastid in the form of DAG or PA. The conversion of phospholipids or BLs into DAG or PA requires the hydrolysis of polar head groups by lipases. The formation of PA can be catalysed by phospholipase D (PLD) [176], such as Phatr3_J12431 in P. tricornutum, which cleaves the terminal phosphodiester bond of phospholipids, releasing PA and water-soluble head groups. PA may then be transported into the plastid, where it is dephosphorylated by phosphatidic acid phosphatase (PAP) to produce DAG, subsequently utilised in plastidial lipid synthesis [177]. Alternatively, DAG can be generated by a phospholipase C (PLC), which hydrolyses the glycerophosphate ester bond of phospholipids to yield DAG and phosphorylated head groups [178]. The DAG could potentially be transported into the plastid and directly incorporated into lipid biosynthesis. However, no PLC homolog has been identified in the P. tricornutum and M. gaditana genomes. Importantly, for this alternative scenario, PA and DAG molecules must possess FA profiles compatible with plastidial lipids in order to have this hypothesis validated. In P. tricornutum, this suggests that molecules with 20:5 at the sn-2 position or containing 18-carbon FAs are unlikely to be imported into the plastid.

• The LPA import hypothesis posits that: phospholipids are first converted to PA by PLD, which is subsequently hydrolysed by PLA2s or LPAATs to produce LPA. The LPA is then transported into the plastid, where ATS2 catalyses the esterification of a FA to the sn-2 position. In P. tricornutum, the FAs added are 16:0 or 16:1, while in M. gaditana, the additions include 16:0, 16:1, or 20:5.

• The FFA import hypothesis posits that: 20:5 is imported as a FFA. Enzymes such as PLA1, PLA2, and PLB could enable the release of FFAs from the glycerol backbone of phospholipids or BLs [178]. The FFAs may then be converted into acyl-CoA by LACS [38, 157], thereby promoting their import into the plastid. Indeed, in P. tricornutum, the KO of the LACS enzyme PtACSL1 has been demonstrated to result in a reduction in the levels of 20:5-containing MGDG [157].

Once within the plastid, thioesterases may release the acyl group from CoA [179], allowing the FA to be converted into 20:5-ACP by an acyl-ACP synthetase (AAS). The activated form of 20:5 could then be directly employed by the acyltransferase ATS1 for incorporation into plastid lipid biosynthesis pathways. Nevertheless, if this pathway is indeed correct, it suggests that P. tricornutum’s PtATS2 may be unable to utilise 20:5, given that this FA is not typically found in plastidial lipids at the sn-2 position.

The potential utilisation of 20:5 on a recycled diacyl moiety derived from extraplastidial lipids, as observed in plants, has yet to be evidenced in P. tricornutum. In contrast to Arabidopsis thaliana, where TGD proteins are involved directly or indirectly in the return of diacyl (DAG or PA) backbones to the plastid [180], no homologues of these proteins have been identified in P. tricornutum [124]. Furthermore, the diacyl composition of plastidial lipids in P. tricornutum does not reflect the one found in phospholipids and BLs. The LPA and FFA import hypotheses are equally probable, as evidenced by the reduced levels of 20:5-containing MGDG and DGDG observed in the PtATS2 KO [136], consistent with the fact that PtATS2 esterifies a 16 C FA at the sn-2 position of a 20:5-containing LPA for the production of 20:5-containing MGDG and DGDG.

Degradation of FAs in mitochondria and peroxisomes

The final stage in a FA lifecycle is its breakdown, which releases carbon and energy, vital for cellular metabolism and essential functions. This catabolic process is initiated through the hydrolysis of FAs esterified in glycerolipids, particularly in triacylglycerols (TAGs) [95]. The liberated FAs may then be repurposed for membrane lipid biosynthesis or directed into β-oxidation to produce acetyl-CoA units and reducing equivalents (NADH and FADH₂), both critical for cellular energy and biosynthesis [181].

In diatoms [182, 183], eustigmatophytes [184] and different thraustochytrids like A. limacinum [112, 116, 185, 186] and Hondaea fermentalgiana [113] TAG lipases have been identified and characterised. One P. tricornutum and two N. oceanica orthologue of SUGAR-DEPENDENT1 (SDP1), the primary TAG lipase in Arabidopsis thaliana [187], were identified as PtTGL1 [182] and NoTGL1 and NoTGL2 [184]. Knockdown of the stroma-localised PtTGL1 leads to TAG accumulation, especially during the stationary growth phase in P. tricornutum, whereas single and double KO of NoTGL1 and NoTGL2 did not induce an increase of TAGs in N. oceanica mutants. Nevertheless, both enzymes proved involved in TAG metabolism [184]. Another lipase in the P. tricornutum genome, OmTGL, was identified by [181] and is localised to the oEM. OmTGL knockdown lines exhibit TAGs enriched in 20:5, suggesting that OmTGL may participate in the hydrolysis of TAGs containing VLC-PUFAs.

Another multifunctional lipase, with acyltransferase, phospholipase, and lipase activities, was identified and initially characterised in Thalassiosira pseudonana. Named Comparative Gene Identification (CGI-58), after the homologous human protein [188], a homologous gene was also identified in P. tricornutum. CGI-58 knockdown mutants in T. pseudonana show increased TAG levels and larger LDs across various culture conditions. In P. tricornutum, recent studies localised CGI-58 to the mitochondria [189], while in animals and plants, it is typically found in LDs or peroxisomes.

In plants, β-oxidation (also referred to as the Lynen spiral) occurs exclusively in peroxisomes, possibly similarly to eustigmatophytes [190]. However, in diatoms and thraustochytrids, β-oxidation can take place in both the mitochondrial matrix and peroxisomes [185, 191–193].

In Aurantiochytrium limacinum β-oxidation plays a pivotal role in providing energy to the motile zoospores stage of its life cycle [30, 112, 185, 186]. During the non-motile stage, A. limacinum cells accumulate TAGs using carbon derived from the decomposition of mangrove leaves. Upon receiving a specific stimulus, vegetative cells differentiate into zoosporangia that maturate and release numerous reniform, flagellated zoospores that inherit lipid droplets from the vegetative cells [186]. The transition from vegetative to motile stage is accompanied by a reprograming of up to 50% of their genome with peroxisomal β-oxidation highly up-regulated [185]. This suggests that in thraustochytrids, peroxisomes degrade preferentially VLC-PUFAs [192], particularly abundant in their TAGs [112].

The β-oxidation process initiates with the release of FAs from glycerolipids by the action of different TAG lipases. After release from glycerolipids, FAs are first activated to acyl-CoA by LACS enzymes before entering β-oxidation. In P. tricornutum, PtACSL3 and PtACSL5 are proposed to catalyse FA activation for subsequent β-oxidation in the mitochondria and peroxisomes, respectively [157]. The activated acyl-CoAs are transported into the mitochondria or peroxisomes by type I or II carnitine acyltransferases (CATs).

The core steps of β-oxidation are highly similar between mitochondria and peroxisomes, differing primarily in the enzyme that catalyses the first step (Fig. 8). In this initial step, a double bond forms between carbons C2 (the β-carbon, hence the term β-oxidation) and C3, catalysed by an acyl-CoA dehydrogenase (ADE) in mitochondria or an acyl-CoA oxidase (AOX) in peroxisomes, yielding trans-enoyl-CoA. During the second step, a hydroxyl group is added at the C3 position by enoyl-CoA hydratase (EHY), forming hydroxyacyl-CoA. The third step involves dehydrogenation by 3-hydroxyacyl-CoA dehydrogenase (HADH), converting hydroxyacyl-CoA into keto-acyl-CoA. Steps two and three can occur simultaneously through a bifunctional EHY/HADH enzyme. Finally, 3-ketoacyl-CoA thiolase (KAT) catalyses the release of an acetyl-CoA molecule via thiolysis, thereby shortening the acyl-CoA FA chain by two carbons, allowing the shortened chain to re-enter the β-oxidation cycle [194]. How FA catabolism is distributed between peroxisomes and mitochondria, and how this partitioning is dynamically regulated in response to developmental, physiological, and environmental cues, remain major open questions in stramenopiles.

Fig. 8.

Schematic representation of the pathways for Acyl-CoA degradation via peroxisomal or mitochondrial β-oxidation in diatoms. CATs: carnitine acyltransferases I and II; ADE: acyl-CoA dehydrogenase; AOX: acyl-CoA oxidase; EHY: enoyl-CoA hydratase; HADH: 3-hydroxyacyl-CoA dehydrogenase; KAT: 3-ketoacyl-CoA thiolase. Adapted from Tanaka et al., 2022 [219]. Created with biorender (https://www.biorender.com)

Where could the journey lead next?

Oxidation of FA

One of the possible paths in the metabolic journey of FAs in stramenopiles is enzymatic or non-enzymatic oxidation into oxylipins and isoprostanoids, respectively [195–201]. Oxylipins comprise a structurally diverse class of oxygenated FAs derivatives generated from PUFAs through either enzymatic or non-enzymatic oxidation reactions. Enzymatic pathways typically involve dioxygenases such as lipoxygenases (LOX) or cyclooxygenases (COX). Non-enzymatic oxidation, driven by reactive oxygen species, also contributes substantially to oxylipin diversity [202]. Functionally, oxylipins act as signalling molecules regulating development, stress responses, defence mechanisms and inter-organismal communication.

Prostaglandins represent a specific subgroup of oxylipins belonging to the eicosanoid family. They are characterised by a cyclopentane ring structure and are typically derived from C20 PUFAs and they are synthesised through the COX pathway [203]. Although first described in mammals, prostaglandins are now identified in a wide range of marine organisms, including macroalgae [204].

Closely related to prostaglandins are isoprostanoids, produced from free radical-mediated peroxidation of PUFAs rather than COX activity. Their formation is independent of enzymatic control and they are widely used as robust biomarkers of oxidative stress. In addition to their diagnostic value, accumulating evidence suggests that some isoprostanoids may exert biological activity, thereby contributing to redox-dependent signalling networks [202].

FA and lipid transport in stramenopile

FA and lipid trafficking in stramenopiles remains comparatively understudied, possibly reflecting both the unicellular lifestyle of many representatives and their complex evolutionary history. Photosynthetic stramenopiles, including diatoms and eustigmatophytes, originate from secondary endosymbiosis [21, 28, 43, 47] and harbour mosaic genomes shaped by contributions from multiple ancestral lineages [48, 65, 205]. It is therefore conceivable that lipid transport systems in these organisms have evolved through the integration, modification, or loss of components inherited from distinct genomic sources. Searches for homologues of known FA and lipid transport proteins in the genomes of P. tricornutum and M. gaditana have yielded limited results. Notably, both species lack clear homologues of the plastidial Fatty Acid Exporter (FAX) proteins that mediate FA export from chloroplasts to the cytosol in Archaeplastida, including green algae [206]. In plants, FAX proteins localised at the inner plastid envelope facilitate the movement of newly synthesised FAs toward extraplastidial glycerolipid biosynthesis [147, 207]. The absence of canonical FAX proteins in P. tricornutum and M. gaditana raises the question of whether alternative, yet unidentified, mechanisms ensure FA transfer across plastid membranes in stramenopiles. In this context, the peculiar plastid architecture of photosynthetic stramenopiles may play a crucial role. Their plastids are surrounded by four membranes [74, 75], with the third membrane often tightly associated with the inner nuclear envelope and the outermost membrane continuous with the outer nuclear membrane and the endoplasmic reticulum [26]. Such structural continuity may facilitate metabolite exchange and lipid fluxes between compartments, potentially reducing the need for dedicated exporters or enabling alternative trafficking routes at membrane contact sites.

LACS are present in both P. tricornutum [157] and M. gaditana [208] genomes and have been involved in intracellular FA trafficking as well as in the metabolism of eicosapentaenoic acid (20:5), particularly under nutrient deprivation. By converting free FAs into acyl-CoA thioesters, LACS enzymes likely play a pivotal role in directing acyl chains toward specific metabolic fates and compartments [157, 208].

Acyl-CoA-binding proteins (ACBPs), which in plants participate in acyl-CoA trafficking and homeostasis [209], appear to be only partially conserved in stramenopiles. A functional ACBP has been identified in P. tricornutum [210]. This PtACBP localises to the endomembrane system, and its KO impairs cell division, lipid droplet degradation, and EPA metabolism, revealing its importance in lipid homeostasis. However, obvious homologues have not been identified in M. gaditana, suggesting lineage-specific diversification or functional replacement.

Beyond vesicular trafficking [211], non-vesicular lipid transport [212] represents an additional, largely unexplored dimension. In higher plants, large lipid transfer proteins such as VPS13 mediate bulk lipid movement at membrane contact sites by forming hydrophobic channels bridging two organelles [212, 213]. Members of the VPS13 family have diversified across Archaeplastida, reflecting a complex evolutionary history [214]. BLAST searches reveal putative VPS13 homologues in P. tricornutum and M. gaditana, yet their size and structural complexity make functional characterisation challenging. Whether these proteins participate in lipid exchange between plastids, ER, LDs, or other organelles in stramenopiles remains an open question.

Altogether, FA and lipid transport in stramenopiles emerges as a promising but still fragmented field. The combination of atypical plastid architecture, mosaic evolutionary origin, and partial conservation of canonical transport factors suggests that these organisms may employ unique or hybrid mechanisms for intracellular lipid trafficking. Dissecting these pathways will not only refine our understanding of lipid metabolism in complex plastid-bearing lineages but may also uncover novel targets for metabolic engineering in oleaginous microalgae.

Conclusions

The life cycle of FAs in stramenopiles, shaped by the massive and highly specific demand for acyl groups in membrane and storage glycerolipids, is intricately organised. It encompasses de novo synthesis in plastids; elongation and desaturation in the cytoplasm and endoplasmic reticulum; entry into specific pathways for incorporation into distinct lipid classes; trafficking to all membrane compartments; remodelling in response to environmental contexts; and ultimately degradation via β-oxidation in mitochondria and peroxisomes. Stramenopiles display remarkable versatility in their lipid metabolism, as evidenced by their ability to produce VLC-PUFAs such as EPA and DHA, critical for both cellular function and biotechnological applications.

Key recent findings include the identification of distinct enzymatic pathways to FA synthesis, transport, and degradation. These processes reveal the evolutionary sophistication of stramenopiles, likely arising from complex endosymbiotic events that shaped their plastidial and extraplastidial lipid metabolism. The presence of specialised enzymes demonstrates a high degree of compartmentalisation and regulation, allowing precise control over FA composition and distribution. Membrane polar lipid classes appear as important platforms for FA modifications (in particular PC, PE and BL), and sometimes as cargoes for transfers within the cell (e.g., MGDG). Still the release of FFA forms in specific subcellular compartments appears as essential in allowing intracellular relocations, including across plastid membranes.

Furthermore, the integration of multiple pathways, including the ω6 and ω3 desaturation routes, illustrates the metabolic flexibility of species like P. tricornutum and M. gaditana. These mechanisms offer promising avenues for enhancing the production of high-value lipids through genetic engineering and optimising strains for industrial purposes.

Recent studies have allowed the identification and characterisation of enzymes involved in glycerolipid metabolic pathways. However, gaps remain in understanding FA export across plastid membranes, the precise mechanisms of lipid trafficking, and the role of acyl-CoA synthetases in coordinating extraplastidial lipid metabolism. Addressing these gaps through further research will refine our knowledge of lipid biology in stramenopiles and expand their potential for sustainable biotechnological applications.

Abbreviations

ABC

ATP Binding Cassette

ACBP

Acyl CoA-binding proteins

ACCase

Acetyl CoA Carboxylase

ACP

Acyl Carrier Protein

ACSL

Acyl-CoA Synthetase Long-chain

ADS1

Acyl-lipid Desaturase 1

AOX

Acyl-CoA Oxidase

ATS1

Acyltransferase of the sn-1 position

ATS2

Acyltransferase of the sn-2 position

BCCP

Biotin Carboxyl Carrier Protein

BC

Biotin Carboxylase

BL(s)

Betaine Lipid(s)

CASH

Cryptophytes, Alveolates, Stramenopiles, Haptophytes

CGI-58

Comparative Gene Identification-58

COX

Cyclooxygenases

CT

Carboxyltransferase

DAG

Diacylglycerol

DGCC

Diacylglyceryl-carboxyhydroxymethylcholine

DGDG

Digalactosyldiacylglycerol

DGTA

Diacylglyceryl-trimethyl-β-alanine

DGTS

Diacylglyceryl-trimethylhomoserine

DHA

Docosahexaenoic Acid

EHY

Enoyl-CoA Hydratase

ENR

Enoyl-ACP Reductase

EpM

Epiplastidial Membrane

EPA

Eicosapentaenoic Acid

ER

Endoplasmic Reticulum

ERAD

Endoplasmic Reticulum-Associated Degradation

FA(s)

Fatty Acid(s)

FAD

Fatty Acid Desaturase

FASI

Fatty Acid Synthase type I

FASII

Fatty Acid Synthase type II

FAX

Fatty Acid Exporter

FFA(s)

Free Fatty Acid(s)

G3P

Glycerol-3-Phosphate

GPAT

Glycerol-3-Phosphate Acyltransferase

GL

Galactoglycerolipids

HAD

Hydroxyacyl-ACP Dehydratase

HADH

3-Hydroxyacyl-CoA Dehydrogenase

iEM

Inner Envelope Membrane

KAR

Ketoacyl-ACP Reductase

KAS

Ketoacyl-ACP Synthase

KAT

3-Ketoacyl-CoA Thiolase

KO

Knockout

LACS

Long-chain Acyl-CoA Synthetase

LD(s)

Lipid Droplet(s)

LOX

Lipoxygenase

LPA

Lysophosphatidic Acid

LPAAT / LPAT

Lysophosphatidic Acid Acyltransferase

LPCAT

Lysophosphatidylcholine Acyltransferase

lysoPC

Lysophosphatidylcholine

MCMT

(FabD) Malonyl-CoA:ACP Transacylase

MGDG

Monogalactosyldiacylglycerol

MMETSP

Marine Microbial Eukaryote Sequencing Project

oEM

Outer Envelope Membrane

PA

Phosphatidic Acid

PC

Phosphatidylcholine

PE

Phosphatidylethanolamine

PG

Phosphatidylglycerol

PI

Phosphatidylinositol

PKS

Polyketide Synthase

PL

Phosphoglycerolipids

PLA1

Phospholipase A1

PLA2

Phospholipase A2

PLB

Phospholipase B

PLC

Phospholipase C

PLD

Phospholipase D

SELMA

Symbiont-specific ERAD-Like Machinery

SQDG

Sulfoquinovosyldiacylglycerol

TAG(s)

Triacylglycerol(s)

TBLASTN

Translated BLAST (protein vs nucleotide)

VLC-PUFA(s)

Very-Long-Chain Polyunsaturated Fatty Acid(s)

VPS13

Vacuolar Protein Sorting 13

Authors’ contributions

A.A. and Y.S. conceived the structure of the review, performed the literature search, and wrote the first draft of the manuscript. C.S.-V. redesigned and prepared all figures. E.M. contributed to the critical discussion of the literature and to manuscript revision. A.A. conceived and supervised the study, coordinated the work, critically revised the manuscript and wrote the final version of the manuscript. All authors read and approved the final manuscript.

Funding

Authors are supported by the French National Research Agency (GRAL Labex ANR-10-LABEX-04, EUR CBS ANR-17-EURE-0003, ANR AlpAlga ANR-20-CE02-0020, ANR DIM ANR-21-CE02-0021, PEPR Algadvance A-22-PEBB-0002, Glyco@Alps Cross-Disciplinary Program; Grant ANR-15-IDEX-02) and the Human Frontier Science Program (HFSP 2024 Trapped In ice). YS was supported by a PhD grant from CEA. CSV was supported by a PhD grant from INRAE (BAP program). Lipid analyses were performed at the LIPANG platform supported by the Rhône-Alpes Region, the FEDER funds, and French National Research Agency (GRAL Labex ANR-10-LABEX-04, EUR CBS ANR-17-EURE-0003).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.