SEARCH

SEARCH BY CITATION

Keywords:

  • phenolics;
  • dinoflagellate toxin;
  • diatom;
  • cyanobacteria;
  • ecology;
  • genome mining

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Bacteria, fungi, algae and higher plants are the most prolific producers of natural products (secondary metabolites). Compared to macroalgae, considerably fewer natural products have been isolated from microalgae, which offer the possibility of obtaining sufficient and well-defined biological material from laboratory cultures. Interest in microalgae is reinforced by large-scale data sets from genome sequencing projects and the development of genetic tools such as transformation protocols. This review highlights what is currently known about the biosynthesis and biological role of natural products in microalgae, with examples from isoprenoids, complex polyketides, nonribosomal peptides, polyunsaturated fatty acids and oxylipins, alkaloids, and aromatic secondary metabolites. In addition, we introduce a bioinformatic analysis of available genome sequences from totally 16 microalgae, belonging to the green and red algae, heterokonts and haptophytes. The results suggest that the biosynthetic potential of microalgae is underestimated and many microalgal natural products remain to be discovered.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Natural products and their biosynthesis

Natural products (or secondary metabolites) are compounds with an immense diversity of biological activities and have therefore been fundamental in drug discovery (Li & Vederas, 2009). Natural products may already be predisposed with desirable pharmacological properties such as cellular uptake or binding to biological targets. In contrast to primary metabolism, secondary metabolism is not vital to basic cellular functions (growth, development). At the same time, primary metabolism tends to be conserved among all organisms, or at least throughout a group of organisms, whereas products of secondary metabolism are exceedingly diverse and often specific for certain species or even isolates. These natural products often mediate interactions between an organism and its (biotic or abiotic) environment, for example as signal molecules, and they are of potential pharmaceutical value (Davies, 2007; Maschek & Baker, 2007).

Bacteria, fungi and plants are the most predominant producers of natural products. While actinomycetes such as Streptomyces species have been the richest source of new antibiotics (Dairi, 2005), higher plants are, for example, well known for their vast phenylpropanoid metabolism that leads to flavonoids/flavanoids, coumarins, some alkaloids, lignin and a myriad of other compounds (Hahlbrock & Scheel, 1989; Vogt, 2010). In the classical approach for drug discovery, natural products are isolated and purified from their natural source, often by an activity-guided purification strategy. Many natural products are only formed in trace amounts by their producers; therefore, large amounts of biomass are sometimes required to obtain sufficient material for structure determination and thorough activity studies. An important aspect in the development of novel bioactive compounds is the construction of derivatives of the original natural product. Derivatives can be either produced by choosing a suitable synthetic strategy that can easily be altered, by chemical derivatization of the isolated natural product (semi-synthesis), or by modification of the biosynthetic pathway (von Nussbaum et al., 2006; Kirschning et al., 2007). The last method strongly benefits from the identification of genes that encode the biosynthetic enzymes because it allows for pathway engineering. For example, natural product derivatives can be obtained by combinatorial biosynthesis or mutasynthesis (Kirschning et al., 2007). In addition, knowledge of biosynthetic genes and their regulation can be exploited to increase production. These approaches are supported by systems biology, which can combine large-scale data sets to functionally manipulate an organism.

Interestingly, analyses of complete genome sequences during the last decade have shown that many genomes contain more gene clusters coding for the biosynthesis of natural products than the natural products that have actually been isolated from the same species (Corre & Challis, 2009; Winter et al., 2011). Even in well-investigated genera such as Streptomyces and Aspergillus, genome information suggests that many natural products remain to be discovered (Sanchez et al., 2008). One reason for this observation is the fact that often natural products are only produced under very specific conditions, and the responsible genes may be silent under standard laboratory conditions. One of the big challenges for natural product discovery is to activate such silent genes. Different approaches have been successfully applied, such as varying the growth conditions in a systematic way, overexpressing activator genes or removing epigenetic silencing (Scherlach & Hertweck, 2009; Cichewicz, 2010), but it remains to be seen what the most generally applicable strategies are.

Algae: evolution, significance and available genome sequences

This review focuses on natural products from microalgae. Algae can be defined as photosynthetic eukaryotes that lack roots, leaves and other organs characteristic of higher plants (Parker et al., 2008). Marine photoautotrophs, which are mostly algae and cyanobacteria, account for roughly half of global carbon fixation (Field et al., 1998). Algae can be divided into microalgae (microscopic algae) and macroalgae. However, it should be kept in mind that algae are evolutionarily quite heterogeneous. Like higher plants, some algae are derived from a eukaryotic ancestor that acquired a photosynthetic cyanobacterium in a single endosymbiotic event, resulting in green algae (chlorophyte and streptophyte algae, the latter being most closely related to higher plants), red algae (rhodophytes) and glaucophytes (Fig. 1). In other algae, a secondary endosymbiotic event has occurred whereby a red or green alga was taken up by a eukaryotic ancestor (Keeling et al., 2005; Bowman et al., 2007). In this way, the chromalveolates are derived from the uptake of a red alga, and chlorarachniophytes as well as euglenophytes are derived from the uptake of a green alga by another eukaryotic ancestor. The plastids in secondary endosymbiotic algae are enclosed by three or four membranes. Another peculiarity is the presence of a remnant nucleus, called nucleomorph, in the plastids of some of these algae such as the cryptophytes (Moreira & Philippe, 2001; Armbrust et al., 2004). The very diverse chromalveolates include, for example cryptophytes, haptophytes, dinoflagellates as well as diatoms and the multicellular brown algae. The latter two are also members of the heterokont algae (or stramenopiles).

image

Figure 1. Evolutionary origins of algae. Microalgae can be found in all groups listed apart from brown algae. Algae are phylogenetically highly heterogeneous, and many lineages also contain nonphotosynthetic species. (*) It is believed that additional events of plastid replacement occurred in dinoflagellates. The scheme was adopted from Keeling (2010) and simplified. Micrographs courtesy of Jens Bösger (Chlamydomonas reinhardtii), Richard Pipe and Glen L. Wheeler (remaining three algae).

Download figure to PowerPoint

Microalgae have traditionally been used to address questions in ecology, biogeochemistry, systematics and evolution, cell biology, and as model systems for higher plants. Flagellated algae have even been used to model ciliated mammalian cells along with their impact on human diseases. Because of broad interest, microalgal genomes are being sequenced, and at present, 16 genome sequences from ongoing or completed genome projects are available (Table 1), and information from another dozen projects is not available yet (Lü et al., 2011). The first microalgal genome sequence in 2004 belongs to the rhodophyte Cyanidioschyzon merolae, an inhabitant of sulphur-rich, acidic hot springs (Matsuzaki et al., 2004). Some 3 years later, the genome sequence of a main model green-algal species was published, Chlamydomonas reinhardtii (Merchant et al., 2007; Fig. 1). For decades, this freshwater and soil-dwelling chlorophyte has been very important, for example for research on photosynthesis and eukaryotic flagella (Mitchell, 2000; Rochaix, 2002). The related species Volvox carteri is a simple multicellular microalga that consists of c. 2000 biflagellate Chlamydomonas-like somatic cells and about 16 gonidia, and its genome sequence gives important hints on the evolution of multicellularity (Prochnik et al., 2010). Genomes of the chlorophytes Ostreococcus (Derelle et al., 2006; Palenik et al., 2007) and Micromonas (Worden et al., 2009) that are ubiquitous members of marine phytoplankton followed. With a cell diameter of 1–2 μm, they belong to the smallest eukaryotes.

Table 1. Available microalgal genome sequences
SpeciesDescriptionGenome size (Mb)No. of chromosomesGene number% Genes with intronsGC content (%)Referencesa
  1. JGI, Joint Genome Institute.

  2. a

    A list of websites is provided in the Supporting Information.

Chlorophyta
Chlamydomonas reinhardtiiModel species, freshwater1211715 1439264Merchant et al. (2007)
Chlorella variabilisParamecium symbiont, model for viral-algal interactions46.2129791 67Blanc et al. (2010)
Coccomyxa sp. C-169Chlorella relative4916962796 JGI
Micromonas sp. CCMP1545Marine picoeukaryote21.91910 5755065Worden et al. (2009)
Micromonas sp. RCC299Marine picoeukaryote20.91710 0563764Worden et al. (2009)
Ostreococcus lucimarinusMarine picoeukaryote13.221765120 Palenik et al. (2007)
Ostreococcus tauriMarine picoeukaryote12.62078922558Derelle et al. (2006)
Ostreococcus sp. RCC809Marine picoeukaryote13.3207492  JGI
Volvox carteriSimple multicellular relative of C. reinhardtii1381414 5209256Prochnik et al. (2010)
Rhodophyta
Cyanidioschyzon merolaeThermo-acidophile16.52053310.555Matsuzaki et al. (2004)
Galdieria sulphurariaThermo-acidophilec. 11.42   Muravenko et al. (2001), Barbier et al. (2005)
Heterokontophyta
Aureococcus anophagefferensToxic bloom-forming marine pelagophyte57 11 501  Gobler et al. (2011)
Fragilariopsis cylindrusMarine psychrophilic diatom     JGI
Phaeodactylum tricornutumMarine pennate diatom27.4 10 402  Bowler et al. (2008)
Thalassiosira pseudonanaMarine centric diatom32.42411 776 47Armbrust et al. (2004)
Haptophyta
Emiliania huxleyiBloom-forming marine coccolithophore     JGI

Diatoms are typically unicellular and represent the most common type of marine phytoplankton. As a major bloom-forming marine diatom, Thalassiosira pseudonana is also of high ecological relevance (Armbrust et al., 2004), while the sequence of the pennate diatom Phaeodactylum tricornutum revealed additional information about diatom evolution and function (Bowler et al., 2008). Together, these two diatoms serve as models for the cell biology of complex plastids surrounded by four membranes. While diatoms possess a complex silica cell wall, some haptophytes surround themselves by calcium carbonate discs, so-called coccoliths. The ubiquitous marine coccolithophore Emiliania huxleyi (Fig. 1), whose genome is currently being sequenced, also causes frequent algal blooms. On the other hand, the pelagophyte Aureococcus anophagefferens is responsible for toxic blooms in estuaries (Gobler et al., 2011). Recently, the first genome of a multicellular seaweed was reported, giving novel insight into the biology of the brown alga Ectocarpus siliculosus (Cock et al., 2010).

Natural products from algae

The majority of algal natural products has been isolated from marine macroalgae, with red algae representing the most generative source (Maschek & Baker, 2007). However, relatively few studies have investigated the biosynthetic pathways of macroalgal secondary metabolites.

Regarding microalgae, certain species of dinoflagellates such as Karenia mikimotoi (Fig. 1) are notorious producers of toxins that can have detrimental effects during blooms. Toxic algal blooms are generally called red tides because some of them lead to a red colour of the water (Anderson, 1994). Algal toxins can accumulate through the food chain. This for example can lead to high toxin concentrations in clams, which are filter feeders that live on plankton. In this way, blooms of toxin-producing dinoflagellates can lead to seafood poisoning in humans and massive mortalities in fish and other animals, which also has significant impact on fisheries, aquaculture and tourism. On the other hand, while dinoflagellates are responsible for the majority of harmful marine blooms (Kellmann et al., 2010), dinoflagellates such as Symbiodinium species also comprise many symbionts of marine invertebrates (Wakefield et al., 2000).

Knowledge on toxin biosynthesis on the genetic and biochemical levels in dinoflagellates is scarce because laboratory work with these algae encompasses a series of major challenges. Firstly, it is an almost insurmountable task to sequence the huge dinoflagellate genomes. Karenia brevis, for example, has an estimated genome size of around 1011 bp, 30 times the size of the human genome (Van Dolah et al., 2009). Moreover, dinoflagellate genes can be present in hundreds of copies, first shown with the gene encoding a luciferin-binding protein in Gonyaulax polyedra (also called Lingulodinium polyedrum; Lee et al., 1993). Secondly, dinoflagellate transformation is to date not routinely possible (Walker et al., 2005). Thirdly, many dinoflagellate species are difficult to culture, and it is rarely possible to maintain them axenically in the absence of (extracellular or endosymbiotic) bacteria, which can be tightly associated (Piel, 2009; Kellmann et al., 2010). Karenia brevis, for example, is believed to depend on unidentified factors or capabilities provided by bacteria (Monroe & Van Dolah, 2008).

The presence of bacteria in a dinoflagellate culture makes it difficult to identify the true source of a toxin. For this reason and other problems mentioned above, the assumption that a compound originates from dinoflagellates still needs to be proven in many cases. But also for many other natural products described, the identity of the producer organism may not be entirely clear, and its unambiguous identification is a persistent problem. For example, it is suspected that some natural products isolated from marine invertebrates (e.g. sponges, snails, coral polyps, worms) actually stem from associated microorganisms (Hildebrand et al., 2004). A microbial natural product can accumulate in a higher organism through the food chain or a symbiotic relationship. In principle, identification of the natural product-forming organism can be accomplished either by axenic cultivation or by cloning and heterologous expression of the biosynthetic genes (cf. Hildebrand et al., 2004). In the course of this review, the producer issue will be addressed in several examples.

In contrast to macroalgae and some dinoflagellates, many other microalgae can be grown in clonal populations under well-defined, axenic conditions in the laboratory. In addition, stable transformation has been reported for several species (Walker et al., 2005; Parker et al., 2008). Particularly, microalgae with sequenced genomes may open up new avenues in natural product research. This review will focus on microalgae, illustrated by examples from the literature, with additional data from bioinformatic analyses of available genome sequences. Pathways and metabolites from microalgae will be compared to those from higher plants, cyanobacteria and macroalgae. In the following, this review will be organized into sections on isoprenoids, complex polyketides, nonribosomal peptides, polyunsaturated fatty acids (PUFAs) and oxylipins, alkaloids and aromatic secondary metabolites.

Isoprenoids

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Isoprenoids (terpenoids) represent one of the largest groups of natural products. They are found in all living organisms and comprise both primary and secondary metabolites (Rodríguez-Concepción & Boronat, 2002; Kirby & Keasling, 2009). Sterols, phytohormones as gibberellins or abscisic acid, phytol (e.g. as a side chain in chlorophyll a), prenylated quinones and many carotenoids are examples of primary metabolites that are isoprenoids or contain isoprenoid moieties. In higher plants, the majority of isoprenoids have secondary functions in the protection against pathogens or herbivores, or in the competition with other plants (Rodríguez-Concepción & Boronat, 2002). Isoprenoid biosynthesis occurs in three stages: (1) the biosynthesis of C5 precursors, (2) the formation of polyprenyl pyrophosphates and (3) their further processing.

In macroalgae, isoprenoids constitute the majority of natural products (Maschek & Baker, 2007). Here, the most important subclasses are sesquiterpenes (C15) and diterpenes (C20). In addition, triterpenes (C30) are found in the form of polyethers such as thyrsiferol, enshuol or teurilene (Fernández et al., 2000). In multicellular red algae, such as the prolific genus Laurencia, many isoprenoid natural products are halogenated. Algal polyether toxins can be formed by two different biosynthetic routes, originating either from triterpenes or from polyketides (Vilotijevic & Jamison, 2009; see also section on complex polyketides). Isoprenoid secondary metabolites are less common in microalgae than macroalgae (Maschek & Baker, 2007).

Biosynthesis of activated isoprenoid precursors

In this first stage of isoprenoid biosynthesis, the activated C5 precursor isopentenyl pyrophosphate (IPP) and its isomer dimethylallyl pyrophosphate (DMAPP) can be formed via the classical mevalonate pathway or the alternative methylerythrol phosphate (MEP) pathway (Bouvier et al., 2005). In higher plants, cytosol and plastid play complementary roles in isoprenoid biosynthesis: For example, the mevalonate pathway in the cytosol is involved in the formation of sterols and sesquiterpenes, while the MEP pathway in the plastid leads to phytol and carotenoids (Fig. 2a; Lichtenthaler, 1999). Coordinate regulation of the two pathways and exchange of intermediates between the compartments is poorly understood and seems to depend on the conditions and tissue (Bouvier et al., 2005). Experiments with inhibitors specific for mevalonate or MEP enzymes and the characterization of pathway mutants indicate that limited amounts of isoprenoid precursors can be exchanged between plastid and cytosol (Rodríguez-Concepción, 2010). There is biochemical evidence for transport of IPP, geranyl pyrophosphate and farnesyl pyrophosphate across plastid envelope membranes (Bick & Lange, 2003; Flügge & Gao, 2005), but so far, no prenyl pyrophosphate transporter has been identified in any organism. Both mevalonate and MEP pathways are also present in the brown alga E. siliculosus (Cock et al., 2010). In microalgae, the situation is somewhat different, however, and Table 2 summarizes data on the distribution of the two pathways in different species. To evaluate available genome sequences, we used blastp (Altschul et al., 1997) to search for HMG-CoA reductase and DXP reductoisomerase, enzymes of the mevalonate pathway and the MEP pathway, respectively. For the most part, the data yielded a consistent picture: Algae generally have both pathways as do higher plants, with the exception of chlorophytes, which only have the MEP pathway (Table 2). The proposal that only a single MEP pathway operates in the chloroplast of chlorophytes (Schwender et al., 2001) is supported by the fact that all genome sequences listed in Table 1 only encode a single DXP reductoisomerase. Moreover, for C. reinhardtii, V. carteri, Ostreococcus lucimarinus and Ostreococcus tauri, the other enzymes of the MEP pathway have also been shown to be encoded by unique genes (Frommolt et al., 2008). In one plausible, yet speculative scenario, cytosolic biosynthesis of phytosterols and other isoprenoids in chlorophytes may depend on the export of prenyl pyrophosphates from the plastid. An isotope labelling experiment suggests that the mevalonate pathway is also absent from the chlorophyte Caulerpa taxifolia, a macroalga that consists of a single multinucleate cell: 13CO2 was found to be incorporated into the backbone of the sesquiterpene caulerpenyne, a major secondary metabolite of C. taxifolia, whereas no enrichment was observed when the alga was fed with 1-13C-acetate (Pohnert & Jung, 2003).

image

Figure 2. Biosynthesis and examples of isoprenoids in algae. (a) Simplified scheme of isoprenoid biosynthesis and its compartmentation in higher plants and algae. Note that chlorophytes likely possess only a plastidic MEP pathway for the formation of DMAPP and IPP, while most other algae and higher plants have both mevalonate and MEP pathways (see Table 2 and main text for details). (b) Examples of algal secondary isoprenoids. FPP, farnesyl pyrophosphate; GPP, geranyl pyrophosphate; GGPP, geranylgeranyl pyrophosphate; G3P, glyceraldehyde 3-phosphate. (a) Adopted from Rodríguez-Concepción & Boronat (2002), and modified.

Download figure to PowerPoint

Table 2. Isoprenoid precursor biosynthesis in microalgaea
SpeciesPathwayEvidence
MevalonateMEPGenome sequenceIsotope labellingInhibitor study
  1. +, evidence for the presence of the pathway.

  2. a

    Studies were selected that are able to provide evidence both for a pathway or against a pathway. For example, as sterols are synthesized via the mevalonate pathway in higher plants, isotope labelling experiments that indicate biosynthesis of sterols via the MEP pathway in chlorophytes suggest that the mevalonate pathway is absent from this group of algae.

  3. b

    See Huss et al., 1999 for problems with Chlorella taxonomy.

  4. c

    HMG-CoA reductase is not among the proteins predicted from the A. anophagefferens genome, but HMG-CoA synthase, mevalonate kinase, diphosphomevalonate decarboxylase and IPP isomerase are present.

  5. d

    High concentrations of lovastatin (10–20 μg mL−1), an inhibitor of the mevalonate pathway, decreased cell growth.

Streptophyta
Klebsormidium flaccidum++ Schwender et al. (2001) 
Mesostigma viride++ Schwender et al. (2001) 
Spirogyra sp.++ Schwender et al. (2001) 
Chlorophyta
Botryococcus braunii + Sato et al. (2003) 
Chlamydomonas reinhardtii +Merchant et al. (2007)Lichtenthaler (1999), Schwender et al. (2001) 
Chlorella sp. +Blanc et al. (2010)Lichtenthaler (1999), Schwender et al. (2001)b 
Dunaliella salina +  Capa-Robles et al. (2009)
Gloeotilopsis planctonica + Schwender et al. (2001) 
Micromonas sp. +Worden et al. (2009)  
Ostreococcus sp. +Derelle et al. (2006), Palenik et al. (2007)  
Scenedesmus obliquus + Lichtenthaler (1999), Schwender et al. (2001) 
Tetraselmis striata + Schwender et al. (2001) 
Trebouxia asymmetrica + Schwender et al. (2001) 
Volvox carteri +Prochnik et al. (2010)  
Rhodophyta
Cyanidioschyzon merolae +Matsuzaki et al. (2004)  
Cyanidium caldarium++ Lichtenthaler (1999) 
Galdieria sulphuraria++Barbier et al. (2005)  
Heterokontophyta
Aureococcus anophagefferens(+)c+Gobler et al. (2011)  
Fragilariopsis cylindrus++JGI  
Haslea ostrearia +  Massé et al. (2004)d
Nitzschia ovalis++ Cvejić & Rohmer (2000) 
Ochromonas danica++ Lichtenthaler (1999) 
Phaeodactylum tricornutum++Bowler et al. (2008)Cvejić & Rohmer (2000) 
Rhizosolenia setigera++ Massé et al. (2004) 
Thalassiosira pseudonana++Armbrust et al. (2004)  
Haptophyta
Emiliania huxleyi++JGI  
Euglenophyta
Euglena gracilis++ Kim et al. (2004) 

The mevalonate pathway may also be absent from Haslea ostrearia and C. merolae, which may therefore represent exceptions among the heterokonts and red algae, respectively (Table 2). From the genome sequence, it is predicted that C. merolae only contains HMG-CoA synthase (and IPP isomerase), but none of the other four enzymes of the mevalonate pathway (Matsuzaki et al., 2004). Similarly, it is currently unclear whether the mevalonate pathway is present at all in H. ostrearia. While isotope labelling experiments were unsuccessful, this diatom was also treated with fosmidomycin. This inhibitor of the MEP pathway inhibited the biosynthesis of several isoprenoids including a sterol, which usually emerges from the mevalonate pathway (Massé et al., 2004). Observations from chlorophytes, H. ostrearia and C. merolae thus suggest that loss of the mevalonate pathway is not uncommon in algae. A model for the evolution of isoprenoid precursor biosynthesis in land plants and algae was presented recently (Grauvogel & Petersen, 2007).

Biosynthesis of polyprenyl pyrophosphates and isoprenoid end products

In the second stage of isoprenoid biosynthesis, linear C10–C25 polyprenyl pyrophosphates are synthesized from IPP and DMAPP by a family of closely related short-chain prenyltransferases (Vandermoten et al., 2009). The 1′-4 condensation of DMAPP with a single IPP molecule by geranyl pyrophosphate synthase (GPPS) yields the monoterpene precursor geranyl pyrophosphate, while condensation of two molecules of IPP with DMAPP by farnesyl pyrophosphate synthase (FPPS) results in the sesquiterpene precursor farnesyl pyrophosphate. Geranylgeranyl pyrophosphate synthase (GGPPS) from fungi and animals needs farnesyl pyrophosphate as the acceptor substrate for IPP, whereas the plastid-localized GGPPS in plants catalyses the successive addition of three IPP molecules to DMAPP (Vandermoten et al., 2009). The occurrence and distribution of the different short-chain prenyltransferases in algal genomes has not yet been examined. The fundamental role of farnesyl pyrophosphate and geranylgeranyl pyrophosphate as precursors of primary metabolites (sterols and carotenoids), however, suggests that all algae contain orthologues of FPPS and GGPPS. In agreement with this assumption, we detected candidates for FPPS, GGPPS and also for GPPS in the currently available algal genomes (Table 1) by blastp searches with the corresponding protein sequences from Arabidopsis.

In the third stage, the various polyprenyls can be further metabolized to a vast array of isoprenoids. For example, farnesyl pyrophosphate and geranylgeranyl pyrophosphate can be dimerized to yield squalene (the precursor of sterols and triterpenes) and phytoene (the precursor of carotenoids), respectively (Bouvier et al., 2005). While the genetic basis of the biosynthetic pathways of sterols and carotenoids – the two major classes of primary isoprenoids in plants and algae – has been examined in detail by comparative genomics (Frommolt et al., 2008; Desmond & Gribaldo, 2009), much less is known about the occurrence and formation of secondary isoprenoids in microalgae. Emission of volatile monoterpenes (C5) from higher plants has been known for several decades and some of its ecological consequences are well characterized, whereas emission of monoterpenes from axenic cultures of unicellular marine algae has only recently been demonstrated (Yassaa et al., 2008). Highest monoterpene emissions were found in the chlorophyte Dunaliella tertiolecta and in the diatoms P. tricornutum and Fragilariopsis kerguelensis, whereas emissions from three other diatoms, E. huxleyi and two cyanobacteria were low or insignificant. Shipboard measurements confirmed the marine production of monoterpenes and isoprene, even though emissions were over an order of magnitude lower than in terrestrial forests (Yassaa et al., 2008). These findings and their physiological and ecological significance call for more detailed investigation.

Monoterpenes and sesquiterpenes have also been isolated from strains of the chlorophyte genus Chlorella for the purpose of biochemical taxonomy (Liersch, 1976 and references therein; some of the strains are now classified as Scenedesmus species (Huss et al., 1999)). The diatom Pseudonitzschia multiseries, which is closely related to Nitzschia species (cf. Table 2), is assumed to be the producer of the neurotoxin domoic acid (see Fig. 6), although a bacterial origin can at present not be ruled out (Bates et al., 2004). Labelling studies with 13C-acetate suggest that domoic acid is formed by condensation of geranyl pyrophosphate with an activated glutamate derivative (Douglas et al., 1992). As it is known from higher plants that labelled acetate is readily incorporated into mevalonate-derived cytosolic isoprenoids, but poorly incorporated into MEP-derived plastidic isoprenoids (Lichtenthaler, 1999), the low incorporation of the 13C-label from acetate into the isoprenoid moiety of domoic acid is in agreement with biosynthesis of geranyl pyrophosphate via the MEP pathway in P. multiseries.

A restricted number of diatoms synthesize highly branched isoprenoids (HBIs) that belong to the sesterterpene (C25) and triterpene subclasses (Fig. 2b; Damsté et al., 2004). HBI alkenes are considered specific diatom markers in marine sediments, where they are widely found. Interestingly, the diatom Rhizosolenia setigera makes C25 haslenes and C30 rhizenes via the mevalonate route, whereas H. ostrearia appears to use the MEP route (Massé et al., 2004). Apart from information on precursor biosynthesis, little is known about pathway intermediates or enzymes involved in HBI biosynthesis, and enzymes responsible for the formation of the additional T branch in these molecules have not been identified yet.

The molecular details of the biosynthesis of secondary isoprenoids in algae are essentially unknown. In land plants, their formation is initiated by members of a large family of terpene synthases (Bohlmann et al., 1998; Trapp & Croteau, 2001). Using protein sequences of three different terpene synthases from Arabidopsis for blastp searches, we were not able to find candidates for algal terpene synthases in any of the algal genomes listed in Table 1. As the diatom P. tricornutum was one of the microalgae reported to emit various monoterpenes, albeit at low rates (Yassaa et al., 2008), P. tricornutum and other algae may have evolved terpene synthases unrelated to the enzymes from land plants. Alternatively, the low rates of monoterpene production in P. tricornutum may result from nonenzymatic reactions (Wise et al., 2002). Notably, an aqueous protein extract from the red macroalga Ochtodes secundiramea was shown to catalyse the formation of the monoterpene myrcene from geranyl pyrophosphate, and the enzyme could be partially purified (Wise et al., 2002). However, no algal monoterpene synthase has yet been identified (Wise, 2003).

Carotenoids are tetraterpenes (C40) best known as primary metabolites that are produced by all photosynthetic organisms, and that are involved in light harvesting and photoprotection (Grossman et al., 2004). In addition, carotenoids are abundant within lipid globules in the eyespot apparatus, a primitive visual system that exists in most flagellate chlorophytes (Kreimer, 2009). Secondary carotenoids not required for photosynthesis and localized either in plastoglobules or in cytosolic lipid droplets are produced by many green microalgae under stress conditions and can be accumulated to high levels (Lemoine & Schoefs, 2010). Some species are even used to produce carotenoids industrially. For example, the unicellular chlorophytes Haematococcus pluvialis and Dunaliella salina are used for the production of astaxanthin (Fig. 2b) and β-carotene, respectively (Del Campo et al., 2007). These high-value products are used as additives in food and animal feed. While β-carotene in D. salina accumulates inside the plastid in interthylakoid lipid globules (Del Campo et al., 2007), astaxanthin is usually found in cytosolic lipid droplets and is esterified with fatty acids (Lemoine & Schoefs, 2010). A variety of functions have been suggested for astaxanthin, ranging from storage of carbon and energy to protection against reactive oxygen species or UV irradiation (Lemoine & Schoefs, 2010).

A genome-based examination of carotenoid biosynthetic genes indicated that C. reinhardtii may also be able to synthesize ketocarotenoids (Lohr et al., 2005). Experiments subsequently revealed that the formation of ketocarotenoids in C. reinhardtii is confined to maturing zygospores (Lohr, 2008). Under nutrient deprivation, the zygotes develop into thick-walled resting spores that accumulate ketocarotenoids and massive amounts of neutral lipid. Unlike H. pluvialis, zygospores of C. reinhardtii contain 4-ketolutein (Fig. 2b) as the major ketocarotenoid, but astaxanthin was also detected (Lohr, 2008; S. Werner, M. Bauch, A. Hallmann, V. Schmitt and M. Lohr, unpublished results). The availability of the complete genome sequence and an elaborate molecular toolbox for C. reinhardtii will facilitate proteomic, transcriptomic and reverse genetic approaches to gain a deeper understanding of the molecular mechanisms of the concomitant accumulation of ketocarotenoids and lipids in green algae.

Algae are currently being explored as a renewable source of biofuel (Radakovits et al., 2010; Lü et al., 2011). While triacylglycerides attract considerable interest in this regard, the isoprenoid pathway also delivers compounds that could be used as biodiesel (Fortman et al., 2008). For example, the unicellular chlorophyte Botryococcus braunii accumulates large amounts of fatty acids and isoprenoids, with hydrocarbon contents of up to 60% dry weight (Metzger & Largeau, 2005). B. braunii strains are grouped into three chemical races. In race B, triterpenes such as the botryococcenes (Fig. 2b) are the predominant hydrocarbons, whereas race L produces lycopadiene, a tetraterpene hydrocarbon. In addition to these hydrocarbons, B. braunii makes related ether lipids. For example, twelve lycopanerols derived from lycopadiene have been described in race L (Metzger & Largeau, 2005). In B. braunii, the majority of hydrocarbons is located in the cell wall and extracellular globules derived therefrom (Largeau et al., 1980; Weiss et al., 2010).

Understanding the regulation of isoprenoid metabolism is the key to its potential manipulation

Sufficient knowledge on how the routing of polyprenyl precursors at critical branch points of the biosynthetic network of isoprenoid formation is regulated will be a prerequisite for increasing the yield of desired isoprenoids in biotechnological applications. For cytosolic isoprenoids, the triterpene secondary metabolism competes with sterol metabolism for the common precursor farnesyl pyrophosphate. Similarly, the formation of mono- and diterpenes in the plastid competes with carotenoid biosynthesis. One mechanism of controlling the allocation of precursors to different pathways may be the formation of multi-enzyme complexes that channel the substrate across the branch point towards a specific class of isoprenoids. There is experimental evidence that phytoene synthase forms a complex with GGPPS (Cunningham & Gantt, 1998 and references therein; Welsch et al., 2010), thereby shunting isoprenoid precursors to the carotenoid pathway. Another way to achieve a higher flux through certain pathways could be an increased expression of key pathway genes. In Arabidopsis, five of the 40 putative terpene synthase genes are arranged in tandem with genes encoding bona fide GGPPS (Aubourg et al., 2002). As GGPPS provides the substrate of the terpene synthases, the parallel duplication of genes for terpene synthase and GGPPS in Arabidopsis supports a central role of GGPPS and potentially other polyprenyl synthases in controlling the flux towards different branches of the isoprenoid pathway. Coordinate regulation between mevalonate and MEP pathways in species with both pathways is another aspect that needs to be examined more thoroughly. Details of the regulation of the two pathways in microalgae may differ from that in land plants, because isoprenoids like gibberellins and strigolactones, which serve as phytohormones in land plants, appear to be absent from unicellular algae (Blanc et al., 2010) or are likely to have different functions in these species.

In summary, many interesting questions about isoprenoid biosynthesis in algae and other plants remain unresolved. As described, pathway regulation and exchange of compounds between compartments are incompletely understood. Furthermore, enzymes responsible for the formation of secondary isoprenoids in algae are still largely unidentified. Research on isoprenoid biosynthesis may help to exploit algae as source of both natural products and biofuel.

Complex polyketides

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

The biosynthesis of polyketides involves repeated cycles of condensation, reduction and dehydration steps using carboxylic acids as building blocks (Fischbach & Walsh, 2006; Hertweck, 2009). In contrast to the related biosynthesis of fatty acids, the three-step processing of the β-keto group after each condensation step can be incomplete. As for fatty acid synthase (FAS), there are two main types of polyketide synthase (PKS): in type I PKS, the various enzymatic functions are performed by different domains within large multifunctional proteins. In contrast, monofunctional proteins form noncovalent complexes in type II PKSs. PKSs that resemble chalcone synthase, which catalyses the committed step in flavonoid biosynthesis in higher plants and some bryophytes, are classified as type III PKSs (Hertweck, 2009). In type I FAS, each catalytic domain operates in every round of chain elongation (Smith & Tsai, 2007), whereas both iterative and noniterative versions of type I PKSs exist (Hertweck, 2009). Noniterative type I PKSs contain sets of domains, termed modules, for each elongation cycle. As a consequence, the structure of the polyketide usually corresponds to the number and architecture of modules. This assembly line mechanism, with its principle of colinearity, renders modular PKSs a prime target for protein engineering.

Polyketides from dinoflagellates

As for microalgae, little is known about polyketides. Exceptions are toxin-producing dinoflagellate species. In a few examples, stable isotope labelling experiments have shown that toxins and related compounds from dinoflagellates are of polyketide origin (Rein & Snyder, 2006; Kellmann et al., 2010). A major group of dinoflagellate polyketides are polyether compounds. Macrolides (macrolactones) are another group of well-known dinoflagellate polyketides, exemplified by the amphidinolide toxins from Amphidinium species (Kobayashi & Kubota, 2007). Strikingly, isotope labelling patterns often show that the C1 atom from acetate is lost at some positions of the polyketide product. For example, such deletions are observed in amphidinolides (Rein & Snyder, 2006), and they are a hallmark of dinoflagellate polyketides. Various chemical mechanisms have been put forward to explain these C1 deletions (Rein & Snyder, 2006; Kellmann et al., 2010). One plausible mechanism involves the formation of an α-diketo group by a flavin-dependent monooxygenase, followed by a Favroskii rearrangement, a second oxidation step, and finally loss of a carbon atom by decarboxylation (Wright et al., 1996). While more experimentation is required to support a Favorskii/decarboxylation mechanism in C1 deletions, a Favorskii rearrangment (without subsequent decarboxylation) has been found to be catalysed by a bacterial type II PKS (Xiang et al., 2004). The 1,4-diketo moieties of amphidinoketides (Fig. 3), also isolated from Amphidinium species (Bauer et al., 1995), may also be explained by C1 deletions, although this has not been substantiated by incorporation of stable isotopes.

image

Figure 3. Polyketides from microalgae and their (putative) producer. Apart from hormothamnione, all compounds are believed to originate from dinoflagellate species.

Download figure to PowerPoint

In the biosynthesis of polyethers, the formation of the backbone is believed to be more complex in dinoflagellates than in bacteria or higher plants and may employ unusual mechanisms (Gallimore, 2009). Among polyether natural products from dinoflagellates, the structurally most intriguing group is formed by the polyether ladder toxins. The first example of a polyether ladder toxin was brevetoxin B (Fig. 3), which was isolated from K. brevis (formerly Gymnodinium breve) and reported in 1981 (Lin et al., 1981). Brevetoxins are neurotoxins that act on voltage-gated sodium channels (Kellmann et al., 2010). It is assumed that the polyketide ladder toxins are formed by epoxidation of trans-double bonds, followed by a stunning cascade of epoxide openings that propagates over the whole molecule to yield an extensive system of fused ether rings (Gallimore & Spencer, 2006; Kellmann et al., 2010). According to a current working model, the epoxide opening cascade proceeds in the direction opposite to the direction of polyketide extension (Van Wagoner et al., 2010). Two natural products recently isolated from K. brevis, brevisamide (Satake et al., 2008) and brevisin (Satake et al., 2009), may help to elucidate mechanistic aspects of polyether biosynthesis: brevisamide only contains a single ether ring, whereas brevisin is the first example where two polycyclic polyether moieties are interrupted by a methylene group (Fig. 3). Epoxide opening cascades may also take part in the biosynthesis of the cyclic ether moieties of triterpenoid polyethers mentioned in the section on isoprenoids (Vilotijevic & Jamison, 2009). The biosynthesis of spiroacetal polyketides, which are related to the polyethers, can also proceed via dehydrative cyclization as an alternative to epoxide opening. Biosynthetic steps for this alternative route, including a spiroacetal synthase, have recently been identified in bacteria (Takahashi et al., 2011). Note that brevisamide and other nitrogen-containing polyketides are discussed in more detail in the section on alkaloids.

As dinoflagellates are difficult to obtain in pure cultures and have unusually large genomes, as discussed in the Introduction, virtually no complete dinoflagellate PKS genes have been cloned and sequenced so far, and it has not been possible to unambiguously link a PKS gene to a corresponding polyketide product. As gene knockout methods are not yet established in dinoflagellates and many other algae, one possibility to infer the function of a gene is to combine physiological evidence with sequence information (Monroe & Van Dolah, 2008). For example, the production of a specific compound may correlate with the presence of a gene in a series of dinoflagellate isolates. Once PKS genes or cDNAs have been identified, heterologous expression may be a rewarding option, but could be difficult if the genes are big or if the polyketide products are toxic to the heterologous host.

Recently, the cloning of a 16-kb hybrid nonribosomal peptide synthetase (NRPS)/PKS gene cluster from K. brevis was reported (López-Legentil et al., 2010). For this purpose, almost 4000 fosmid clones were screened by PCR. The gene cluster harbours three NRPS genes and one PKS gene, each gene encoding a single module. PCR with plastidic DNA suggested that the gene cluster is localized in the chloroplast and may have been acquired through endosymbiosis with ancestral cyanobacteria (López-Legentil et al., 2010). However, contamination of the sample with nuclear or cyanobacterial DNA cannot be ruled out completely. It would now be interesting to learn which natural products are encoded by the discovered gene cluster and what their biological significance is.

Another important step towards the understanding of toxin biosynthesis in dinoflagellates is the first sequences of full-length transcripts of putative dinoflagellate PKS genes (Monroe & Van Dolah, 2008). To this end, PKS- and FAS-related sequences were identified in a K. brevis expressed sequence tag (EST) library, and rapid amplification of cDNA ends (RACE) was used to obtain full-length sequences. The presence of spliced leader (SL) sequences (a common feature added to nuclear-encoded dinoflagellate transcripts by trans-splicing), 3′-untranslated regions (UTRs) and polyadenylation confirmed that full-length transcripts were obtained. Surprisingly, each transcript only encoded a single PKS function, with one exception that encoded both an acyl carrier protein (ACP) and a ketosynthase (KS) domain. While this organization resembles type II PKSs, a phylogenetic analysis grouped these KS sequences in the protist type I PKS/FAS clade (Monroe & Van Dolah, 2008). One interesting possibility the authors envisaged is a process where after transcription of type I genes, SL trans-splicing leads to the formation of transcripts that encode single domains (Monroe & Van Dolah, 2008). To support this hypothesis, the genomic regions encoding the identified transcripts will have to be identified. As fatty acid biosynthesis is generally carried out by type II enzymes in plants (including algae) (Leibundgut et al., 2008), the identified transcripts may alternatively be involved in fatty acid biosynthesis in K. brevis.

In summary, research on dinoflagellate polyketides is at an early stage and faces many challenges. There is very little unambiguous information on toxin biosynthesis in these organisms, and in many cases, the suspected dinoflagellate origin of a toxin still needs to be proven. First steps to understand the molecular genetic details behind dinoflagellate blooms have been made in the last few years. High-throughput sequencing of transcriptomic libraries is emerging as an informative approach for research on dinoflagellates, and another example of this strategy will be discussed in the section on alkaloids.

Genome mining for microalgal PKS genes

To explore whether modular PKSs occur in microalgae other than dinoflagellates, we searched protein complements of available algal genome sequences for KS sequences using blastp. Hits were obtained for the majority of species, indicating that many microalgae from diverse lineages indeed possess one or several modular PKSs (Table 3). Interestingly, modular PKSs appear to be present in eight of nine chlorophyte species investigated, whereas the only two rhodophyte species sequenced to date (C. merolae and Galdieria sulphuraria) seem to lack these enzymes. A patchy distribution of modular PKS genes in protists was noted earlier, but the available evidence did not favour one of two evolutionary scenarios that either involve multiple events of gene acquisition or multiple events of gene loss (John et al., 2008). However, KS sequences from protists (chlorophytes, haptophytes, dinoflagellates, apicomplexans) form distinct clades that are separate from bacterial and fungal clades (Snyder et al., 2005; John et al., 2008; Monroe & Van Dolah, 2008). In the four microalgal heterokonts sequenced so far, the presence of modular PKSs correlates with the ability to cause harmful blooms. This is true even if no direct connection can yet be drawn between the production of secondary metabolites and the ability to form harmful blooms. In contrast to A. anophagefferens, which is the causative agent of so-called brown tides, T. pseudonana, P. tricornutum and Fragilariopsis cylindrus do not seem to possess modular PKS genes. The annotated genome of the brown alga E. siliculosus, which also belongs to the heterokonts, does not harbour a type I PKS gene, but has a type I FAS gene (Cock et al., 2010). Whereas most chlorophytes may only possess one or a few modular PKS genes, Coccomyxa sp. C-169, but also the heterokont A. anophagefferens and the haptophyte E. huxleyi, may have ten or possibly more modular PKS genes (Table 3). While we are not aware of any toxins reported from E. huxleyi, another haptophyte known to form blooms, Prymnesium parvum, is known to produce polyketide toxins called prymnesins (Manning & La Claire, 2010). The polycyclic polyethers prymnesin-1 and prymnesin-2 further contain two chlorine substituents, an acyclic amino group and a diyne moiety, and it has been proposed that glycosylation might be used as a localization signal (Manning & La Claire, 2010). It was noticed previously that the toxic bloom former A. anophagefferens has an increased number of genes associated with secondary metabolite biosynthesis, although no specific toxin has been identified in this alga (Gobler et al., 2011).

Table 3. Number of genes that encode modular PKSs and NRPSs estimated from microalgal genome sequencesa
SpeciesType I PKS genesbNRPS genes
  1. a

    blastp was used to search predicted proteins for ketosynthase and condensation domains, and genomic regions of found hits were further inspected. A high uncertainty is associated with this prediction because of possible inaccuracies with genome assembly, gene prediction and domain identification.

  2. b

    Some of the PKSs may contain additional NRPS modules.

Chlorophyta
Chlamydomonas reinhardtii10
Chlorella variabilis11
Coccomyxa sp. C-169101
Micromonas sp. CCMP154530
Micromonas sp. RCC29920
Ostreococcus lucimarinus30
Ostreococcus tauri40
Ostreococcus sp. RCC80900
Volvox carteri10
Rhodophyta
Cyanidioschyzon merolae00
Galdieria sulphuraria00
Heterokontophyta
Aureococcus anophagefferens101
Fragilariopsis cylindrus00
Phaeodactylum tricornutum00
Thalassiosira pseudonana00
Haptophyta
Emiliania huxleyi100

We chose C. reinhardtii, Micromonas sp. RCC299, O. lucimarinus and E. huxleyi and subjected select type I PKSs to domain analysis (Fig. 4 and Supporting Information). These predicted proteins have remarkably large subunit sizes between 14 000 and 22 000 residues (1.5–2.2 MDa) and possess approximately eight to 14 extension modules. Some microalgal PKSs may also contain NRPS domains: While the functionality of the C domain at the N terminus of the PKS from Micromonas sp. RCC299 is unclear, the identity of the subsequent A domain is better supported (see Supporting Information). The functions of predicted AMP-binding domains in the enzymes from C. reinhardtii and O. lucimarinus (Fig. 4) are less certain; in case of the putative initiation module of the C. reinhardtii PKS, this domain might be an acyl-coenzyme A ligase (acyl-CoA synthetase) as was suggested for the initiation module of the PKS from the apicomplexan Cryptosporidium parvum (Zhu et al., 2002).

image

Figure 4. Examples of modular PKSs predicted from available algal genome sequences. Knowledge about conserved residues was used to verify domain predictions from various programs (see Supporting Information for details). A question mark indicates that it is unclear whether a domain is functional. Domain abbreviations: KS, ketosynthase; DH, dehydratase; ER, enoylreductase; KR, ketoreductase; ACP, acyl carrier protein; MT, methyltransferase; C, condensation domain; A, adenylation domain.

Download figure to PowerPoint

The examined proteins from Micromonas sp. RCC299 and O. lucimarinus seem to lack a designated domain for off-loading of the polyketide, and the sequences of the last KS domains show alterations in conserved motifs (Fig. 4 and Supporting Information). Possibly, the terminal KS domain is involved in the interaction with a distinct but yet unidentified subunit that will continue polyketide biosynthesis. Other PKS genes from algal genomes are predicted to encode noncanonical domains. For example, another putative PKS gene from E. huxleyi (protein identifier 632088) encodes HMG-CoA synthase domains, which are known to direct β-alkylations in polyketide biosynthesis (Calderone, 2008).

It should be noted that during genome sequencing, the assembly of eukaryotic genomic regions that encode modular proteins is complicated by the large size of these genes that may also contain repetitive sequences. Gene prediction can also be difficult because splicing specificity may vary from species to species, and some transcripts may be subjected to alternative splicing. Correct annotation of the large genes that encode modular proteins will require careful manual curation, which is facilitated by extensive transcriptomic data such as ESTs or RNA-Seq data. Despite several uncertainties, the domain analysis presented in Fig. 4 highlights some important points: Firstly, modular PKSs that appear to lack acyltransferase (AT) domains seem to be common in microalgae. These so-called trans-AT PKSs, which require a free-standing AT protein for ACP loading, were only discovered a few years ago (Piel, 2010). An analysis with Pfam (Finn et al., 2010) showed that of the ten type I PKSs predicted in each Coccomyxa sp. C-169 and E. huxleyi, three and seven are trans-AT PKSs, respectively. We did not find any hints for genes encoding possible trans-AT proteins when examining the genomic regions around the PKS genes in Coccomyxa sp. C-169 and E. huxleyi, or the genes for the other PKSs described in Fig. 4. Secondly, while known toxin producers such as dinoflagellates or A. anophagefferens are rich sources of natural products, even microalgae traditionally not known to produce polyketides seem to possess corresponding genes. The challenge is now to characterize these genes and link them to their corresponding biosynthetic products. Information obtained on polyketide biosynthesis in more tractable algae may also help to refine hypotheses on toxin production in dinoflagellates and thereby guide experimental strategies.

Nonribosomal peptides

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

NRPSs resemble noniterative type I PKSs in their modular organization, but have different domains and use amino acids as monomers instead of carboxylic acids (Fischbach & Walsh, 2006; Marahiel & Essen, 2009). NRPS core functions are carried out by a condensation (C) domain, an adenylation (A) domain and a peptidyl carrier protein (PCP). Monomers are selected by the A domain, and it is possible to predict the substrate specificity from the sequence of an A domain with some confidence, at least in bacteria (Rausch et al., 2005). In cyanobacteria, nonribosomal peptides represent a major class of natural products, and cyanobacterial NRPSs very often occur in mixed NRPS/PKS systems (Welker & von Döhren, 2006).

To find NRPSs in microalgae, we searched proteins predicted from microalgal genomes for C domains using blastp. While the number of proteins predicted to contain both thiolation (PCP or ACP) and A (or AMP-binding) domains is higher, only three predicted proteins were found with C domains (Table 3). In addition to potential mixed NRPS/PKSs described in the previous section, evidence for NRPSs was found in Chlorella variabilis, Coccomyxa sp. C-169 and A. anophagefferens. The dimodular NRPS from C. variabilis (protein identifier 144896) consists of 2635 amino acid residues and might catalyse the formation of a diketopiperazine. Coccomyxa sp. C-169 is predicted to make a protein that only consists of an A domain followed by a C domain (1370 residues, protein identifier 68188). The predicted NRPS from A. anophagefferens (protein identifier 70689) contains three modules with 4374 amino acid residues in total. Program NRPSpredictor (Rausch et al., 2005) suggests that the product would be a tripeptide consisting of proline and two hydrophobic amino acids (valine, leucine, isoleucine, 2-aminobutyric acid or isovaline). In all these cases, it is difficult to predict the structures of the final NRPS products because the predicted open reading frames might be inaccurate, NRPSs might be used iteratively, in conjunction with an additional subunit, or the nonribosomal peptide might be further tailored after release from the NRPS, similar to the modification of the penicillin precursor l-δ-(α-aminoadipoyl)-l-cysteinyl-d-valine found in bacteria and fungi (Nolan & Walsh, 2009). While only three NRPS genes were detected in the 16 genomes listed in Table 3, a preliminary analysis of the genome of Bigelowiella natans predicts seven NRPS genes in this chlorarachniophyte (E. Sonnenschein, personal communication). Considering the cyanobacterial origin of chloroplasts, it is perhaps surprising that NRPS genes seem so sparse in most microalgae. It can be speculated that cyanobacteria acquired NRPS genes after the first endosymbiotic events that led to formation of algae, or algae may subsequently have lost NRPS genes. Alternatively, it is possible that microalgal NRPSs have diverged too strongly from known NRPSs to be identified reliably by the methods used here.

Polyunsaturated fatty acids and oxylipins

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Polyunsaturated fatty acids

PUFAs possess multiple double bonds that are typically separated by a methylene group (Fig. 5a). PUFAs are found throughout microalgae and accumulate through the marine food web to end up for example as health-promoting constituents of fish oils (Rogalski & Carrer, 2011). PUFA abundance affects the food quality of microalgae for organisms at higher trophic levels (zooplankton, fish) (Brett & Müller-Navarra, 1997), and herbivorous zooplankton may have the ability to selectively feed on algae with a higher PUFA content. PUFAs have a broad physiological activity; for example, they are important for the maintenance of membrane fluidity at low temperatures. In the conventional aerobic biosynthesis pathway, which is well known for animals and plants, standard fatty acids from FAS such as stearic acid are further elongated, and double bonds are introduced by position-specific desaturases between condensation steps catalysed by elongases (Napier, 2002). Genes for oxygen-dependent desaturases are widespread in microalgal genomes (Chi et al., 2008; Cock et al., 2010), suggesting that PUFA biosynthesis by the elongase–desaturase pathway is common.

image

Figure 5. Examples of PUFAs and (putative) PUFA synthases in algae and bacteria. (a) Examples of PUFAs. Linoleic acid is an essential fatty acid that is abundant in plant oils, but is also found in algae. Docosahexaenoic acid is a healthy PUFA found in fish oil, but is absent from crop plants. (b) (Putative) bacterial and algal PUFA synthases. While the PUFA synthase from the marine bacterium Shewanella sp. was characterized experimentally, similarity in sequence and domain organization was used as evidence to infer the same function for the protein from the alga Schizochytrium sp. (Metz et al., 2001). A similar organization of a predicted protein from Emiliania huxleyi now suggests that also this protein is involved in PUFA biosynthesis. To prepare this figure, the architecture of the (putative) PUFA synthases from Shewanella sp. and Schizochytrium sp. was adopted from Metz et al. (2001), and the domain organization of the potential PUFA synthase from E. huxleyi was derived from sequence similarity.

Download figure to PowerPoint

A decade ago, an alternative, anaerobic pathway to PUFAs was discovered (Metz et al., 2001; Kaulmann & Hertweck, 2002). This novel pathway employs an iterative PUFA synthase, which can be considered an intermediate between FAS and PKS. PUFA synthase is believed to insert double bonds by KR and DH domains without further reduction of the enoyl group. Whereas known PUFA synthases and type I FASs function iteratively, both iterative and noniterative versions of type I PKSs exist. This new pathway for PUFA biosynthesis has been described for the marine bacterium Shewanella sp. and Schizochytrium sp., a marine heterokont alga (Metz et al., 2001). A distinctive feature of the characterized PUFA synthase from Shewanella sp. and its homolog from Schizochytrium sp. is the occurrence of a series of ACP domains, with up to nine domains in tandem as found in Schizochytrium sp. (Fig. 5b). Mutants of the bacterial PUFA synthase from Shewanella japonica showed that the different ACP domains are functionally equivalent and that the number of functional ACP domains correlates with the amount of PUFA produced (Jiang et al., 2008b). Interestingly, one of the approximately ten predicted PKSs in the coccolithophore E. huxleyi contains four identical ACP domains in tandem (Fig. 5b). This protein is predicted to have a subunit size of 0.62 MDa (5874 residues), and our analysis indicates that the organization of the 14 domains resembles the organization of PUFA synthases from marine bacteria and Schizochytrium, even though the subunit organization is different (Fig. 5b). This suggests that this protein functions as a PUFA synthase in E. huxleyi, which is a known producer of PUFAs. When grown at 15 °C, total fatty acids of eight different E. huxleyi isolates were shown to contain c. 50–70% PUFAs (Pond & Harris, 1996).

Some mechanistic aspects concerning the arrangement and configuration of double bonds in PUFA synthase-derived PUFAs remain unsolved: Firstly, while PKSs insert double bonds in the α,β-position, the typical homoconjugated (‘skipped’) arrangement of double bonds in PUFAs requires a special mechanism. In bacterial trans-AT PKSs, two different ways have recently been discovered that enable double bond shifts in the course of polyketide elongation, catalysed by either a DH domain or a specialized shift module (Kusebauch et al., 2010; Moldenhauer et al., 2010). It is possible that these mechanisms are also employed by PUFA synthases. Secondly, dehydration most often leads to trans double bonds in polyketides, and in principle, cis double bonds could be formed by various mechanisms (Hertweck, 2009). In most PUFAs, exclusively cis double bonds are found, and it is unclear how they arise. For example, it is conceivable that a double bond shift from the α,β-position to the β,γ-position is accompanied by a trans-to-cis shift.

Oxylipins

PUFAs are the substrates for various oxidative enzymatic transformations. Generally, these products are termed oxylipins. At least in diatoms, oxylipins are widely distributed, and it can be estimated that roughly 30% of marine diatoms are capable of oxylipin production (Wichard et al., 2005). In diatoms, the class of α,β,γ,δ-unsaturated aldehydes has attracted much attention because of their potential ecological role. Even if the ecological impact of these metabolites is still under discussion, data point towards a potential involvement in chemical defence and chemical communication within the plankton (Ianora et al., 2007). Little is known about the enzymes actually involved in these transformations. Stable isotope-labelled precursors revealed that as in higher plants, lipoxygenases initiate fatty acid oxidation (Pohnert, 2005). The resulting intermediary hydroperoxy fatty acids can then be transformed via several different pathways to further downstream products. Cleavage of the hydroperoxides can be mediated by the action of hydroperoxide lyase-type enzymes that presumably work in a hydrolytic mode (Barofsky & Pohnert, 2007) or via unique halolyases, where carbon–carbon bond cleavage is assumed to be assisted by the nucleophilic attack of a halide ion (Wichard & Pohnert, 2006). However, reduction of the intermediate hydroperoxides to alcohols, transformations to ketones or epoxidation to epoxyalcohols has also been observed in marine diatoms (Fontana et al., 2007; d'Ippolito et al., 2009). These reactions lead to the formation of a multitude of oxygenated fatty acids of different reactivity that might also mediate ecological interactions in the plankton. The unsaturated aldehyde 2,4-decadienal has also been observed in the prymnesiophyte Phaeocystis pouchetii where it may fulfil similar ecological functions as discussed for diatoms (Hansen et al., 2004).

Alkaloids

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Whereas approximately 20% of plants produce alkaloids (Ziegler & Facchini, 2008), they are relatively rare in algae, even in marine macroalgae (Güven et al., 2010). Alkaloids are defined as nitrogen-containing compounds usually derived from amino acids, with the nitrogen atoms mostly part of heterocycles, but the distinction between alkaloids and other nitrogenous natural products is not always well defined. Alkaloids do not originate from a single biosynthetic pathway, rather, different alkaloid subclasses are made by many different pathways. For most alkaloids discussed below, the biosynthesis was shown or can be assumed to involve PKSs.

Genomic analysis indicated that the pelagophyte A. anophagefferens possesses five proteins with similarity to berberine bridge enzymes, which are involved in the biosynthesis of isoquinoline alkaloids (Gobler et al., 2011). So far, these enzymes have not been characterized experimentally, and no alkaloids have been isolated from A. anophagefferens. The dinoflagellate Alexandrium ostenfeldii produces macrocyclic imines called spirolides. These fast-acting toxins exert their activity by targeting acetylcholine receptors and calcium channels (Kellmann et al., 2010). Spirolides additionally contain spiroacetal groups. Stable isotope labelling of 13-desmethyl spirolide C (Fig. 6) from A. ostenfeldii showed incorporation of both acetate and a single glycine unit, indicating that spirolides are made by hybrid NRPS/PKS enzymes (MacKinnon et al., 2006). Glycine incorporation has also been demonstrated for dinophysistoxins 5a and 5b (Macpherson et al., 2003), and it has been proposed for the recently isolated brevisamide (Satake et al., 2008; see Fig. 3), which was mentioned in the section on complex polyketides. Whereas dinophysistoxins 5a and 5b and brevisamide contain a nitrogen atom as part of an amide function, azaspiracids contain a piperidine moiety (Fig. 6). No isotope labelling experiments have been performed to shed light on azaspiracid biosynthesis, and the responsible biosynthesis genes also remain to be identified for the spirolides, brevisamide and the dinophysistoxins.

image

Figure 6. Alkaloids from microalgae and their (putative) producer. PKSs were shown or can be supposed to be involved in the biosynthesis of all compounds above except domoic acid, which may arise from the condensation of geranyl pyrophosphate with an activated glutamate (Douglas et al., 1992). For azaspiracid-1, the corrected structure was taken from Nicolaou et al. (2006).

Download figure to PowerPoint

Saxitoxin (Fig. 6) is a highly potent alkaloid neurotoxin that acts on voltage-gated sodium channels. Natural products of the saxitoxin group are produced by both marine dinoflagellates and freshwater cyanobacteria, and precursor labelling suggests that the biosynthetic route is identical in both cases (Kellmann et al., 2010). Until recently, the genes responsible for saxitoxin biosynthesis evaded elucidation despite considerable efforts. Stimulation of saxitoxin biosynthesis in crude enzyme extracts by addition of carbamoylphosphate suggested that O-carbamoyltransferase might be involved (Kellmann & Neilan, 2007), serving as a starting point for targeting the genes by PCR with degenerate primers. This approach ultimately led to the discovery of a 35-kb gene cluster, comprising 31 open reading frames, in the cyanobacterium Cylindrospermopsis raciborskii (Kellmann et al., 2008). At the beginning of the pathway proposed for C. raciborskii, the c. 1200-residue PKS SxtA would catalyse an unusual Claisen condensation between arginine and propionyl-ACP. According to this model, all six guanidinium nitrogen atoms of saxitoxin would be derived from arginine: one of them from the α-amino group and the other five from the guanidinium groups of two different arginine molecules. The seventh nitrogen atom, in the carbamoyl group of saxitoxin, is likely derived from carbamoylphosphate (Kellmann et al., 2008). In subsequent work with two Alexandrium species, 454 sequencing was employed to obtain 1.2 million ESTs, and for many cyanobacterial saxitoxin biosynthesis genes, similar sequences were found in the dinoflagellate transcriptomes (Stüken et al., 2011). In case of the ESTs from sxtA, the presence of SL sequences and polyadenylation indicates that the gene is located in the nucleus. Finally, saxitoxin production strongly correlated with the presence of sxtA (as judged by PCR) in an examination of 28 dinoflagellate species from six different genera (Stüken et al., 2011). Taken together, genes responsible for saxitoxin biosynthesis have been identified in cyanobacteria and dinoflagellates, and it has become clear that at least for sxtA, Alexandrium is independent from any associated bacteria.

Zoanthamine alkaloids are a group of natural products with various structures and activities mostly isolated from soft corals of the order Zoantharia (Behenna et al., 2008). Zoanthamine, the founder of this group, was isolated from Zoanthus sp. (Rao et al., 1984). Later, the related alkaloid zooxanthellamine was isolated from a culture of the dinoflagellate Symbiodinium sp. Y-6 (Nakamura et al., 1998). The striking structural similarity of zooxanthellamine to other zoanthamine alkaloids, including stereochemical properties, suggests that the coral obtains these natural products through the food chain or a symbiotic relationship (Nakamura et al., 1998). Interestingly, Symbiodinium sp. Y-6 used for zooxanthellamine purification (Nakamura et al., 1998) was originally isolated from a marine flatworm and described to be uni-algal, but not axenic (Nakamura et al., 1993). Therefore, the culture arguably contained bacteria that in principle may also be the alkaloid producers. Zoanthamines nicely illustrate the difficulty in tracing the source of the natural product, particularly when dinoflagellates are involved because these algae on one hand may occur as symbionts of marine invertebrates and on the other hand are difficult to culture axenically.

Aromatic secondary metabolites

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

In higher plants, the phenylpropanoid pathway represents the major route to aromatic secondary metabolites. Complex phenylpropanoid metabolism seems to be confined to higher plants, ferns and mosses (Emiliani et al., 2009). In higher plants, structure and function of phenylpropanoids are extremely diverse. For example, flavonoids can serve as UV protectants, anthocyanins as pigments, stilbenoids as antibiotics (so-called phytoalexins) and lignin as a cell wall component (Hahlbrock & Scheel, 1989). Phenylpropanoids are derived from phenylalanine, an end product of the shikimate pathway (Hahlbrock & Scheel, 1989; Vogt, 2010). The committed step that converts phenylalanine to cinnamate is catalysed by phenylalanine ammonia-lyase (PAL; Logemann et al., 1995), and this reaction constitutes the transition from primary to secondary metabolism. While PALs occur ubiquitously in terrestrial plants, this enzymatic function is also utilized by many fungi and bacteria, including some cyanobacteria (Emiliani et al., 2009). In some basidiomycete and ascomycete fungi, PAL is involved in the catabolism of phenylalanine. So far, no evidence for the presence of PAL has been found in algae (Emiliani et al., 2009).

Our own blastp searches in conjunction with knowledge on specific residues that may allow distinguishing between PAL and the homologous histidine ammonia-lyase (HAL; Williams et al., 2005) further support the absence of PAL from algae. While it was reasoned that HAL may accept phenylalanine as a substrate (Seyedsayamdost et al., 2011), genomic data provide evidence for this enzyme in only two species: A. anophagefferens and E. huxleyi. The (multicellular) brown alga E. siliculosus also seems to lack PAL (Cock et al., 2010). However, one cannot fully rule out a functionally analogous PAL variant that cannot be identified by sequence similarity. At the same time, there is evidence for aromatic secondary metabolites in algae, for example in Chlamydomonas. In the snow alga Chlamydomonas nivalis, the Folin-Ciocalteu colorimetric method has been used as a measure of total phenolics. In one experiment, UV irradiation of cells resulted in an up to 18% increase in total phenolics compared to control cells treated with white light (Duval et al., 1999). At the moment, it is unknown which pathway is responsible for the production of phenolic compounds in C. nivalis. In a recent metabolomics approach in C. reinhardtii, compounds were identified by matching mass spectra and retention times with entries in a database of compounds. This GCxGC/MS (tandem gas chromatography coupled with mass spectrometry) experiment supports the formation of the phenylpropanoid caffeate (3,4-dihydroxycinnamate) in C. reinhardtii (May et al., 2008). Similarly, p-coumarate was identified as a major secreted compound in E. huxleyi (Seyedsayamdost et al., 2011). Furthermore, even though there is no evidence for a PAL gene in C. reinhardtii, this alga is predicted to contain at least two proteins (protein identifiers 206056 and 285680 in annotation version 4 by the Joint Genome Institute) with sequence similarity to oxidoreductases involved in phenylpropanoid metabolism (Merchant et al., 2007). However, as it is not possible to predict the substrate specificity of oxidoreductases with high reliability, these genes have to be investigated experimentally. It is also possible that these genes are involved in catabolic processes, such as the detoxification of aromatic aldehydes. Another interesting example is the detection of lignin, generally considered a hallmark of vascular plants, in the seaweed Calliarthron cheilosporioides, a coralline red alga (Martone et al., 2009).

Flavonoid biosynthesis has also been observed in algae that apparently lack PAL. It has therefore been suggested that the flavonoid pathway is more ancient, and PAL was later integrated by land plants into the pre-existing pathway (Cock et al., 2010). In Chlamydomonas eugametos, flavonoid compounds are used as sex pheromones (Birch et al., 1953), while a whole series of flavonoid biosynthesis genes is predicted in E. siliculosus (Cock et al., 2010). Possibly, aromatic secondary metabolism in algae is not initiated by the action of PAL, but is organized in a different fashion. For example, almost 60 years ago Birch et al. (1953) proposed a route from tyrosine via 3,4-dihydroxyphenylalanine and 3,4-dihydroxyphenylpropionate to caffeate in C. eugametos.

Furthermore, polyketide biosynthesis is an alternative route to aromatic compounds. In higher plants and many bryophytes, a chalcone synthase catalyses the first committed step in flavonoid biosynthesis (Jiang et al., 2006). For example, naringenin chalcone synthase is a type III PKS that extends the coenzyme A-activated phenylpropanoid p-coumarate by three C2 units from malonyl-coenzyme A to yield a second aromatic moiety. It is possible that hormothamniones (Fig. 3), styrylchromone toxins isolated from the marine cryptophyte Chrysophaeum taylori (Gerwick et al., 1986; Gerwick, 1989), are formed by a type III PKS.

Many studies have analysed the formation of phlorotannins, which are phenolic compounds unique to brown algae (Amsler & Fairhead, 2006). Phlorotannins comprise both oligomers and polymers that consist of phloroglucinol moieties. They can act as both primary and secondary metabolites. Phlorotannins are cell wall components and play a role in wound healing in brown algae (Amsler & Fairhead, 2006). It has so far not been possible to experimentally identify genes or enzymes responsible for phlorotannin biosynthesis. Indeed, whereas higher-plant tannins are synthesized by the phenylpropanoid pathway, phlorotannins from brown algae are generally believed to be of polyketide origin; however, a phenylpropanoid pathway has been proposed as well (Amsler & Fairhead, 2006). Interestingly, while there is no experimental evidence for enzymes responsible for phlorotannin biosynthesis, the E. siliculosus genome was predicted to encode three type III PKSs, which may direct phloroglucinol biosynthesis (Cock et al., 2010). It should be noted, however, that the prediction of a type III PKS function is not trivial. The sequences of type III PKSs and other enzymes from the thiolase superfamily show high similarity, and enzymatic function may be determined by the identity of a few diagnostic residues (Jiang et al., 2008a). After heterologous expression of a type III PKS gene from the brown alga Sargassum binderi in Escherichia coli, the major product in enzyme assays with cell lysates was found to be a triketide pyrone (Baharum et al., 2011). Identification of the biosynthetic pathway should help to understand the regulation of phlorotannin biosynthesis, together with their biological role in brown algae. In summary, it needs to be examined to what extent the phenylpropanoid and polyketide pathways are responsible for the biosynthesis of aromatic secondary metabolites in algae. A cooperation of both pathways to form phenylpropanoid-derived polyketides is another possibility. While brown algae produce a vast variety of phlorotannins, it seems clear that aromatic secondary metabolism in microalgae does not generate as many compounds as in higher plants, but it is possible that research in this area may lead to the discovery of currently unknown natural products and shed light on ecological aspects.

Perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Many questions on microalgal lifestyle and secondary metabolism remain unanswered. In addition to the dinoflagellates, the cultivation of many other microalgae is still a big challenge to scientists. It has been speculated that cryptophytes and euglenophytes may be as rich sources of natural products as the dinoflagellates if culture conditions in the laboratory can be improved (Shimizu, 1996). Further, it is unclear to what extent secondary metabolism in algae is regulated at the epigenetic level. Evidence for epigenetic regulation of secondary metabolism in fungi has only been obtained in the last few years (Cichewicz, 2010). In one example, addition of epigenetic modifiers to fungal cultures elicited the production of three related oxylipins that were undetectable in the control cultures (Williams et al., 2008). Regulation, time frame and ecological relevance of epigenetic modifications in fungi have not been examined much however, and even less is known in algae.

The availability of entire genome sequences from several microalgae provides important clues on algal metabolism. Strategies for mining genomes for the discovery of new natural products have recently been described in a comprehensive way (Zerikly & Challis, 2009). As in bacteria and fungi, it needs to be kept in mind that natural products can be specific for a certain algal strain (Shimizu, 1996). Another open question concerns clustering of functionally related genes in algal genomes. Clustering is common for metabolic genes in general in bacteria, and for secondary metabolic genes in filamentous fungi (Osbourn, 2010). In contrast, it may be less frequent in algae, which would make it difficult to infer the function of a gene from its genomic context. An example of functional clustering of genes in primary metabolism is nitrate assimilation in C. reinhardtii (Fernández et al., 1989). The genome of the chlorophyte Coccomyxa sp. C-169 possibly contains a rare example of functional clustering of secondary biosynthetic genes in algal genomes: This gene cluster contains four putative PKS genes with sizes between 30 and 40 kb, together with a desaturase gene (Fig. 7). Clustering of secondary metabolic genes is a field of active research in higher plants, where it may be an exception rather than a rule (Osbourn, 2010). Similarly, more research is necessary to understand functional clustering in algal genomes.

image

Figure 7. Putative secondary metabolic gene cluster in Coccomyxa sp. C-169. Numbers in brackets indicate the corresponding protein identifiers in version 2.0 of the genome annotation (http://genome.jgi-psf.org/Coc_C169_1).

Download figure to PowerPoint

The availability of genome sequences has also paved the way for systems biology. For example, large-scale proteome analysis can be conducted with methods such as liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS). Thereby, the experimentally measured MS/MS spectra are compared with MS/MS spectra of candidate peptides generated bioinformatically from the available databases. In C. reinhardtii, several sub-proteomes have been analysed this way, including the flagella, centriole, chloroplast ribosomal proteins, mitochondria and the eyespot (Rolland et al., 2009; Wagner et al., 2009). Such analyses allow conclusions about the subcellular localization of specific metabolic pathways. For example, the eyespot apparatus of C. reinhardtii consists of carotenoid-rich globules and contains the retinal-based photoreceptors attached to its specialized plasma membrane part. The eyespot proteome revealed proteins involved in the general biosynthesis pathway of carotenoids, in lipid metabolism and in retinal biosynthesis (based on homology searches), indicating that at least some of the relevant biosynthesis steps occur in close vicinity of the eyespot (Schmidt et al., 2006).

Recently, a proteomics method was developed to discover natural products and their biosynthetic pathways even without requiring genome sequence information (Bumpus et al., 2009). This approach focuses on high molecular weight bands in SDS gels, bigger than approximately 150 kDa, often formed by PKS or NRPS proteins. Interesting bands are, for example, analysed by in-gel digestion before LC-Fourier transform MSn is conducted on the resulting peptide mixture. Peptides carrying an active site serine from ACP or PCP domains are then identified in a phosphopantetheinyl ejection assay: The phosphodiester linkage of the phosphopantetheinyl cofactor is labile during MS/MS, similar to phosphoproteomics with serine and threonine. Peptide sequences obtained by this method are then used to design degenerate PCR primers to amplify portions of the expressed gene cluster (Bumpus et al., 2009). This genome sequence-independent approach was successfully used to discover novel nonribosomal peptides from Bacillus species (Bumpus et al., 2009; Evans et al., 2011).

The biological function of many algal secondary metabolites is unclear. To properly address functional aspects, reverse genetic methods are indispensable. During nuclear transformation of microalgae, transgenes usually integrate in random fashion, with homologous recombination occurring at very low frequencies (Parker et al., 2008). In some algae such as C. merolae, Chlorella species or V. carteri, homologous recombination may occur more frequently, and integration of a marker gene via homologous recombination has been observed (Walker et al., 2005). However, homologous recombination is regularly used as a tool for gene targeting only in C. merolae (Imamura et al., 2010; Kobayashi et al., 2011). Transformation methods have been developed for several chlorophytes (Walker et al., 2005; Parker et al., 2008). For example, transformation of nuclear and both organellar genomes is possible in C. reinhardtii (Harris, 2001). Efforts are currently being made to establish and improve transformation protocols for diatoms. Here, protocols are most advanced for P. tricornutum, although transformation efficiencies are still low (Parker et al., 2008). Gene silencing strategies have been established in C. reinhardtii, O. tauri and P. tricornutum (Rohr et al., 2004; Corellou et al., 2009; De Riso et al., 2009; Molnar et al., 2009). In contrast to C. reinhardtii, whose vegetative cells are haploid, diatoms grow as diploid vegetative cells (Chepurnov et al., 2008). To recombine useful traits or mutations, it would be crucial to control the sexual cycle of diatoms in the laboratory, but so far, this is generally not possible. A valuable exception is Seminavis robusta, which has been proposed as a model species for this and other reasons (Chepurnov et al., 2008). The next step to a model diatom for S. robusta would be to sequence its genome. While a large part of algal research is currently funded with the goal of exploring new energy sources, this research is at the same time expected to greatly improve the genetic tool set required to address other questions that are of ecological or biotechnological significance.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Available genome sequences are an invaluable resource for algal research and reveal that a lot is yet to be learned about secondary metabolism. Genome sequences aid the identification of individual genes or entire pathways to known compounds, thereby greatly facilitating regulatory and functional investigations. Even more, genome sequences indicate that microalgae may produce compounds that are still undiscovered. The importance of mining the genomes of new groups of organisms is illustrated by the recent discovery of the first natural product from an obligately anaerobic bacterium (Lincke et al., 2010). Together with continuously improving analytical and genetic methods, genome sequences thus help to better understand algal physiology and interactions of algae with their environment, and they may lead to the discovery of new compounds with useful properties. Therefore, we can hope that available genome sequences will help to lift new treasures to the surface of turbid algal reservoirs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

We are grateful to Dr Eva Sonnenschein (University of California, San Diego) and Dr Jaclyn Winter (Leibniz Institute for Natural Product Research and Infection Biology, Jena) for helpful discussions, to Dr Richard Pipe and Dr Glen L. Wheeler (Marine Biological Association, Plymouth UK) and Dr Jens Bösger (Friedrich Schiller University, Jena) for the provision of micrographs, and to Nico Überschaar (Leibniz Institute for Natural Product Research and Infection Biology, Jena) for help with the preparation of Fig. 1. Our primary research has been supported by the Deutsche Forschungsgemeinschaft (DFG), the Bundesministerium für Bildung und Forschung (BMBF), the Pakt für Forschung und Innovation and the Swiss National Science Foundation (SNF).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Isoprenoids
  5. Complex polyketides
  6. Nonribosomal peptides
  7. Polyunsaturated fatty acids and oxylipins
  8. Alkaloids
  9. Aromatic secondary metabolites
  10. Perspectives
  11. Conclusions
  12. Acknowledgements
  13. References
  14. Supporting Information
FilenameFormatSizeDescription
fmr304-sup-0001-SupportingInformation.pdfapplication/PDF20K 

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.