Pathways of lipid metabolism in marine algae, co-expression network, bottlenecks and candidate genes for enhanced production of EPA and DHA in species of Chromista

Abstract

The importance of n-3 long chain polyunsaturated fatty acids (LC-PUFAs) for human health has received more focus the last decades, and the global consumption of n-3 LC-PUFA has increased. Seafood, the natural n-3 LC-PUFA source, is harvested beyond a sustainable capacity, and it is therefore imperative to develop alternative n-3 LC-PUFA sources for both eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3). Genera of algae such as Nannochloropsis, Schizochytrium, Isochrysis and Phaedactylum within the kingdom Chromista have received attention due to their ability to produce n-3 LC-PUFAs. Knowledge of LC-PUFA synthesis and its regulation in algae at the molecular level is fragmentary and represents a bottleneck for attempts to enhance the n-3 LC-PUFA levels for industrial production. In the present review, Phaeodactylum tricornutum has been used to exemplify the synthesis and compartmentalization of n-3 LC-PUFAs. Based on recent transcriptome data a co-expression network of 106 genes involved in lipid metabolism has been created. Together with recent molecular biological and metabolic studies, a model pathway for n-3 LC-PUFA synthesis in P. tricornutum has been proposed, and is compared to industrialized species of Chromista. Limitations of the n-3 LC-PUFA synthesis by enzymes such as thioesterases, elongases, acyl-CoA synthetases and acyltransferases are discussed and metabolic bottlenecks are hypothesized such as the supply of the acetyl-CoA and NADPH. A future industrialization will depend on optimization of chemical compositions and increased biomass production, which can be achieved by exploitation of the physiological potential, by selective breeding and by genetic engineering.

Figure 1

n-6 fatty acids (%) versus n-3 fatty acids (%) in different food sources and microalgae [51,52,53,54,55,56].

Figure 2

Content of LC-PUFAs of different lipid classes in selected Chromistans under different environmental conditions. Due to the experimental setups of the referred papers, not all lipid classes of the organisms could be assembled. Monodus subterraneus [67], N. oculata [68], Pavlova lutheri [69], P. tricornutum [70]. Blue: ARA (20:4n-6); red: EPA (20:5n-3); green: DHA (22:6n-3). MGDG: monogalactosyldiacylglycerol; NL: neutral lipids; DGCC: diacylglycerylcarboxyhydroxymethylcholine; DGDG: digalactosyldiacylglycerol; DGGA: diacylglyceryl glucuronide; DGTS: diacylglyceryltrimethylhomoserine; DGTA: diacylglycerylhydroxymethyl-trimethylalanine; SQDG: sulfoquinovosyldiacylglycerol; PC: phosphatidylcholine; PE: phosphatidylethanolamine; PG: phosphatidylglycerol.

Figure 3

Co-expression network of 106 genes associated to the FA metabolism in P. tricornutum. The co-expression network can visually be divided into two subclusters. Subcluster 1 (blue, left square) contains mainly genes of the mitochondrial TCA cycle and β-oxidation. Subcluster 2 (red, right square) includes genes of the plastidial-located de novo FA synthesis and the endoplasmatic n-3 LC-PUFA biosynthesis. Color code: TAG biosynthesis (light purple); TCA cycle (red); ACCase (acetyl-CoA carboxylase); de novo FA and HTA (16:3n-4) synthesis (light blue); Predicted elongases and desaturases (dark blue); Predicted EPA pathway (turquoise); Acetyl-CoA precursors and transporter (light red); Acyl-CoA synthetases, ATpase4 (gray); Mitochondrial or peroxisomal located β-oxidation and FA elongation (yellow); Kennedy pathway, phospholipid-, glycerolipids, sphingolipid and sterol biosynthesis (green); Ca²⁺-dependent lipid-binding protein, amid hydrolase, DHHC palmitoyltransferase, serine incorporator, ATP-binding protein (ABC) transporter (purple). Shapes in the cluster indicate the localization of enzymes encoded by the gene: Triangle, mitochondria; Square, chloroplast; Diamond, peroxisome; Circle, no prediction. Transcription data of five microarray datasets from P. tricornutum submitted to GEO, NCBI (GSE12015, GSE17237, GSE31131, GSE42039 and GSE42514; [93,94,95,96,97]) were used to construct based on log2 expression ratios from the experiments, an unweighted co-expression network by using Cytoscape (version 2.8.3) and the force directed drawing algorithm [98]. The network represents 106 genes related to lipid metabolism with similar transcriptional profiles and includes 311 calculated gene-pairs with Pearson correlation values r > 0.85.

Figure 4

Simplified overview of the compartments, the main pathways and the metabolites in most Chromista; calvin cycle, fatty acid synthesis, tricarboxylic acid cycle, polyunsaturated FA pathway, β-oxidation and lipid synthesis shown in black arrows. Involved enzymes are shown in red: ACCase, acetyl-CoA carboxylase; ACS, Acyl-CoA synthetase, ACP, acyl carrier protein; CoA, coenzyme A; ATP:CL, ATP-citrat lyase; ENR, enoyl-ACP reductase; FAT, fatty acyl-ACP thioesterase; HD, 3-hydroxyacyl-ACP dehydratase; KAR, 3-ketoacyl-ACP reductase; KAS, 3-ketoacyl-ACP synthase; LACS, long chain acyl CoA synthetase; MAT, malonyl-CoA:ACP transacylase; ME, malic enzyme; PDC, pyruvate dehydrogenase complex; PUFA, polyunsaturated fatty acid; TAG, triacylglyceride; TCA, tricarboxylic acid. Different MEs possess different localizations (plastidial, mitochondrial). For simplicity, ME is placed in the cytosol. Modified after [22,65,99].

Figure 5

Overview of FA and LC-PUFA synthesis in P. tricornutum. Shown are the hypothetical de novo fatty acid synthesis (FAS) and the HTA (16:3n-4) synthesis plastidial (green) and the EPA (20:5n-3) synthesis at the ER membrane (blue) with further incorporation at the sn-1 and sn-2 position of glycosylglycerides (in plastid or Kennedy pathway in ER). Purple: Long chain acyl-coenzyme A (CoA) synthetases (LACS), lysophospholipid acyltransferases (LPLAT), acyl-CoA:glycerol-3-phosphate acyltransferase (GPAT) and acyl-CoA:lysophosphatidic acyltransferase (LPAAT), phosphatidic acid phosphatase (PAP), elongases (Elo) and desaturases (ΔD or ωD); Red: putative genes, enzymes encoded by genes that have been identified are marked in bold. The other genes are predicted for EPA synthesis from transcriptional data. The co-factors for the desaturases are not indicated. At the ER, the FA are available as acyl-CoA or linked to a glycerol-backbone such as PC indicated by . MGDG is indicated by a glycerol-backbone with a framed G. Before and after elongation, the FA has to be de-linking and re-linking from the glycerol-backbone indicated with two consecutive arrows.Question marks indicate that the reaction and the involved enzymes are not predicted. Modified after [28,113].