Brown algae (Phaeophyta) make up most of the benthic biomass on temperate and polar rocky shores. Kelp forests are key ecosystems of such coastal areas and they constitute habitat and breeding grounds for a very large diversity of marine life forms - besides offering a physical protection by providing a wave energy-absorbing buffer for coasts. Brown algae belong to a lineage that has been evolving independently of other major photosynthetic lineages, in particular the green plants (Chlorophyta) and red algae (Rhodophyta). Instead, they are classified within the Stramenopiles and Chromalveolates together with diatoms, golden-brown algae and oomycetes (Baldauf 2003). As a consequence of this singular evolutionary history, brown algae exhibit many unusual, and often unique, features. These features are adaptations to the marine coastal environments in which brown algae are usually the dominant organisms in terms of biomass, in particular by forming extensive kelp forests structuring these ecosystems. The key role of kelp forests, effectively constituting an interface between the ocean, the atmosphere and land masses, in the biogeochemical cycle of halogens is well established (Carpenter et al. 2000, Kupper et al. 2008).

Ectocarpus siliculosus is a filamentous brown alga with a worldwide distribution along temperate coastlines, and is a nuisance as a “fouling” organism on many man-made surfaces in the sea. It has some significant advantages as an experimental model and constitutes one of the best-studied seaweeds (Peters et al. 2004, Charrier et al. 2008); it can easily be cultivated in small volumes of seawater media both axenically and with associated bacteria (Muller et al. 2008); it belongs to a sister group of the ecologically and economically very important kelps; its entire, well-known life cycle can be completed within a few months in culture; many molecular tools are available, including mutant collections (Peters et al. 2004), microarrays (Dittami et al. 2009) and proteomics (Ritter et al. 2010). It has also recently become the first seaweed of which the entire genome has been sequenced and thus offers unprecedented opportunities for study (Cock et al. 2010).

Like all other marine organisms, marine algae are subject to a virtually permanent onslaught of a broad spectrum of pathogens - including viruses (Muller and Knippers 2001), chytrid fungi (Gleason et al. 2011), oomycetes (Strittmatter et al. 2009) and endophytic / parasitic algae (Ellertsdottir and Peters 1997). While some algal pathogens have been known since the 19th century (e.g., Wright 1877), mechanistic aspects have been studied only for little more than a decade (e.g., Potin et al. 2002). A significant finding of recent years was that brown and red algae share key defense mechanisms with animals and higher plants (Kupper et al. 2001, Potin et al. 2002).

The mammalian response to microbial attack involves the activation of immune response genes, including those involved in the arachidonic acid cascades, leading to the synthesis of leukotrienes and prostaglandins (Funk 2001, Gachon et al. 2010). In comparison, plants synthesize compounds of direct antimicrobial toxicity such as reactive oxygen species (ROS), phytoalexins and pathogenesis-related proteins. Higher plants are also known to activate genes involved in the oxylipin cascade such as lipoxygenases, in a process that appears similar to the arachidonic acid cascade in mammals (Andreou et al. 2009). Many inducible defense genes are regulated by signalling pathways involving one or more low molecular weight signal molecules, such as jasmonic acid, salicylic acid, ethylene, and possibly H₂O₂ (Reymond and Farmer 1998, Orozco-Cardenas et al. 2001).

The oxidative burst was initially discovered in human macrophages (Baldridge and Gerard 1933). It is now recognized also as a key element of plant (Wojtaszek 1997, Mittler et al. 2004) and algal (Potin et al. 2002, Dring 2005, Cosse et al. 2007) defense, where it is considered both as a rapid defense response and an internal emergency signal. It consists of the rapid and massive production of ROS such as superoxide (O₂^-), hydrogen peroxide (H₂O₂), and hydroxyl radicals (？OH)which are likely produced via the interaction of plasma membrane-associated NAD(P)H oxidases (Wojtaszek 1997). Hence, pathogen attack triggers both the production of ROS and the oxidation cascade of fatty acids leading to the production of bioactive oxylipins in plants. In algal systems, the oxidative burst was initially discovered in response to injury in a red alga (Collen et al. 1994). Over the last decade, oxidative bursts in response to oligosaccharide elicitors were observed in a number of red and brown algal models, in most cases to control the growth of bacterial biofilms (Weinberger et al. 1999, Weinberger and Friedlander 2000a, 2000b, Kupper et al. 2001, 2002, 2006, 2009) or the attack of eukaryotic endophytes (Bouarab et al. 1999, Kupper et al. 2002). More recently, the involvement of an oxidative burst in the wound healing response of a siphonalean green alga was highlighted (Ross et al. 2005), including its crosscommunication with nitric oxide signalling (Ross et al. 2006).

Most polyunsaturated fatty acids (PUFAs) in higher plants belong to the C₁₈ family, whilst polyunsaturated C₂₀ fatty acids seem to occur only in very few species. In contrast, algal lineages such as the Phaeophyceae (Hofmann and Eichenberger 1997) produce both C₁₈ and C₂₀ PUFAs in significant amounts. A number of algae have become industrial sources of biotechnologically important PUFAs such as arachidonic acid (C20:4) or eicosapentaenoic acid (C20:5) (Colombo et al. 2006, Spolaore et al. 2006).

Oxylipins and prostaglandins in particular are important intra- and intercellular messengers in all living organisms. Among the cellular signalling responses which ROS generation can trigger is the activation of fatty acids by the production of oxygenated PUFAs called oxylipins which play a key role against abiotic and biotic stress in plants and metazoans. In mammals, these compounds (which include leukotrienes and prostaglandins) are produced from the oxidation of C₂₀ PUFAs. Prostaglandins mediate pivotal functions in immunity processes and cell development, allergic reactions and, in a broader sense, defensive stress responses (Funk 2001). Furthermore, in plants, their recruitment during wounding stress, pathogen and herbivore invasions has been fully confirmed in recent years (Farmer et al. 2003, Farmer and Schulze-Lefert 2005). In our previous work, we had shown that bacterial surface macromolecules (especially lipopolysaccharides) can trigger both an oxidative burst and induce the release of free fatty acids and synthesis of oxylipins in Laminaria (Kupper et al. 2006). More recently, we observed similar effects following treatment of Laminaria with linolenic and arachidonic acid as well as methyl jasmonate (Kupper et al. 2009): besides triggering an oxidative burst, these compounds led to a release of free fatty acids and oxylipin synthesis and, significantly, to an induced resistance against the kelp endophyte Laminariocolax tomentosoides. However, in neither of these studies the formation of cyclopentenones such as prostaglandins or jasmonate was observed. However, cyclopentenones have recently been found to be synthesized in Laminaria under copper stress (Ritter et al. 2008) constituting the only report of cyclopentenones in brown algae to date. Other than that, all reports of cyclopentenones reported from seaweeds originate from red, not brown algae - e.g., methyl jasmonate in Chondrus crispus (Bouarab et al. 2004), PGA₂, PGE₂, PGF₂ and 15-keto-prostaglandins in Gracilaria verrucosa (Sajiki 1997, Imbs et al. 2001, Dang et al. 2010) and PGE₂ in Gracilaria tenuistipitata (Hsu et al. 2008) and G. rhodocaudata (Sajiki 1997). In none of these cases, a function could be associated with these compounds yet, even though in several cases prostaglandin production was found to be upregulated by some form of stress (e.g., Sajiki 1997, Imbs et al. 2001, Bouarab et al. 2004, Gaquerel et al. 2007, Hsu et al. 2008). Interestingly, prostaglandin content (or production, respectively) has been associated with occasional cases of toxicity of seaweeds to humans (Higa and Kuniyoshi 2000). Contrasting with these, the observation of phlorotannin induction by airborne methyl jasmonate in Fucus by Arnold et al. (2001) and our finding of the activity of exogenous methyl jasmonate in Laminaria (Kupper et al. 2009) so far remain the only cases of a cyclopentenone showing a biological effect in a brown alga. Contrasting with these studies, Wiesemeier et al. (2008) did not observe defense-related effects of jasmonate in other species of brown algae.

Considering the paramount importance of prostaglandins in the defense of mammalian cells, the scope of this study was to screen 5 prostaglandins as potential defense elicitors in brown algae. In analogy to our previous studies, we monitored Laminaria digitata for an oxidative burst response and changes in its fatty acid and oxylipin profile. Then, following the identification of PGA₂ as a potent trigger of an oxidative burst in Laminaria and taking advantage from the availability of the Ectocarpus genome (Cock et al. 2010), we investigated prostaglandin effects on the transcriptional induction of 5 candidate genes potentially related with oxidative stress, defense and cell rescue.

Prostaglandins (PGE₁, PGE₂, PF₂α, PGB₁, PGA₂) were obtained from FLUKA and Cayman Europe (Tallinn, Estonia), dissolved in dimethyl sulfoxide and stored at -20℃ until further use. These prostaglandins had been preselected among a multitude of compounds due to their established, stress-related roles in other organisms.

For experiments involving Laminaria, essentially the same procedures as described previously for studying the effects of lipopolysaccharides (Kupper et al. 2006) and free fatty acids and methyl jasmonate (Kupper et al. 2009) were used. Briefly, young L. digitata sporophytes [produced from gametophyte strains CCAP 1321/1 (♀) and CCAP 1321/2 (♂) from Helgoland] were grown in half-strength Provasoli ES (PES) medium (Starr and Zeikus 1987) prepared from natural Atlantic open-ocean water. They were illuminated with daylight-type fluorescent lamps at an irradiance of 9 μE m^-2 s^-1 for 10 h per day and kept at 10 ± 1℃. Sporophyte thalli were transferred to fresh medium at one-or two-week intervals. They were used for experiments when they had reached a length of 5 mm to 3 cm.

For assessing the effects of prostaglandins on oxidative metabolism in Laminaria, thalli were incubated in 10 mL filtered seawater in petri dishes at 12℃, to which dilutions of PGE₁, PGE₂, PF₂α, PGB₁ or PGA₂ were added at final concentrations of 100 μM. The concentration of hydrogen peroxide in the medium around algal thalli was determined using the luminol chemiluminescence method (Glazener et al. 1991) with a LUMAT LB 9507 luminometer (EG&G Berthold, Bad Wildbach, Germany), as described before (Kupper et al. 2001).

For all further studies following the finding of PGA₂ as a potent inducer of the oxidative burst, a dose-response curve for the oxidative burst was measured using the aforementioned chemiluminescence assay, with final PGA₂ concentrations of 0.1, 1, 3, 6 and 10 μM, respectively. For analyses of free fatty acids and oxylipins, Laminaria thalli were incubated with PGA₂ at a final concentration of 10 μM. Tissue samples were subsequently frozen and ground in liquid nitrogen. The frozen powder was transferred into borosilicate tubes and homogenized with 2 mL of ethyl acetate. Lipids were extracted by mixing on a rotary shaker for 1 h at 4℃ and then 1 mL of ice cold water was added. The mixture was kept 5 min on ice and then centrifuged at 4℃ and 3,500 g. After recovery of the organic phase, ethyl acetate was evaporated under a stream of argon. Lipid analyses were performed as described previously (Kupper et al. 2006). Oxylipins were identified by reverse phase-high performance liquid chromatography coupled to a mass spectrometer (HPLC-MS) (Finnigan, Manchester, UK) in atmospheric pressure ionisation (APCI) and negative ion mode. Structural assignments were confirmed by gas chromatography (GC) (HP 5890

Series II; Hewlett Packard, Palo Alto, CA, USA) coupled to mass spectrometry (HP 5971A; Agilent Technologies, Santa Clara, CA, USA) after appropriate derivatizations. Fatty acids were quantified by GC-MS as methyl esters from standard curves obtained by measuring the peak surfaces of authentic standards. Statistic significance for increases of PUFAs and oxylipins amounts after treatments compared to controls was tested by a Student t-test.

All experiments with Ectocarpus used the genomesequenced strain E. siliculosus CCAP 1310/4 (Cock et al. 2010). Monoeukaryotic, clonal cultures of this strain were grown in 650 mL filter cap suspension flasks (175 cm²; Greiner Bio-One, Courtaboeuf, France) at 15℃ using half-strength modified Provasoli-enriched sea water (West and McBride 1999), at 10 μE m^-2 s^-1 (daylight-type fluorescent lamps, 12 h photoperiod).

Treatments with PGA₂ (obtained from FLUKA and Cayman Europe, as above) were performed at a final concentration of 10 μM in an experimental procedure developed by Zambounis, Strittmatter and Gachon (unpublished / personal communication). Briefly, algal cell cultures were treated as follows: PGA₂ or ethyl ester (controls; 10 μL per 15 mL of culture) was added for 2 h each day during seven consecutive days. Following incubation the material was rinsed and placed in fresh half-strength modified Provasoli-enriched seawater in 650 mL filter cap suspension flasks (175 cm²; Greiner Bio-One). The experimental procedure included two biological and three technical replicates each, respectively. Cultures were harvested for RNA extraction by concentrating the culture on 70 μm nylon mesh filters (Cell Strainers™; BD Biosciences, Oxford, UK), quickly dried with an autoclaved paper towel, and immediately immersed in RNALater™ solution (Ambion, Austin, TX, USA), left for perfusion for 1 day at 4℃, followed by storage at -80℃.

Total RNA was extracted from approximately 70 mg (fresh weight) of tissue using the Qiagen RNeasy^® Plant Mini-kit (Qiagen, Crawley, UK) and DNase-treated (RNAse-free DNAse I; Invitrogen, Paisley, UK), according to the manufacturer’s instructions with the following modifications: Two volumes (900 μL) of buffer RLC was added to the disrupted tissue (Step 2; see RNeasy^® Plant Mini-kit protocol; Qiagen) which resulted in cleaner, higher yield RNA extraction. Quality and integrity assessments as well as quantification of the RNA were performed using a Bioanalyzer 2100™ (Agilent Technologies) with the corresponding Agilent RNA 6000 Nano Kit. From each RNA sample, 0.5 μg was reverse transcribed to cDNA using SuperScript^® III First-Strand Synthesis System (Invitrogen) according to the manufacturer’s instructions. Subsequently, first strand cDNAs were diluted with nuclease-free water to a final concentration of 1 ng μL^-1.

Potential molecular responses of E. silicilosus to PGA₂ treatment were monitored with five stress marker genes (Table 1), following the protocol of Zambounis, Strittmatter and Gachon (unpublished/personal communication). The specificity, efficiency and dynamic range of the primers were tested on 1/5 serial dilutions of cDNA (range, 0.24-150 ng for elongation factor 1α [EF1a], vanadium-dependent bromoperoxidase [vBPO], manganese superoxide dismutase [MnSOD] and heat shock protein 70 [HSP70]; 0.12-75 ng for glutaredoxin [GRX]; 0.048-30 ng for glutathione S-transferase [GST]). The absence of genomic DNA contamination was examined in each sample with the intron-specific primer mN (Table 1) (Le Bail et al. 2008).

Quantitative expression analysis was performed using a QUANTICA™ real-time PCR system (Techne-Barloworld, Stone, Staffordshire, UK). Quantitative polymerase chain reactions (qPCRs) were run in a final volume of 20 μL with 2 ng cDNA and 300 μM of each primer with MesaGreen™ reaction mix (Eurogentec, Southampton, UK). Reactions were run in triplicates. The cycle parameters were 10 min at 95℃, followed by 45 runs of 15 s at 95℃ and 1 min at 60℃. Dissociation curves were obtained by melting curve analysis from 65 to 95℃ with readings every 0.5℃ and 10 s hold between them.

Relative variation in gene expression ratios was calculated as fold changes relative to the control experiments using REST 2009 software (Pfaffl et al. 2002). EF1a was used as internal reference gene. Expression ratios of tested genes were determined between the control samples and the corresponding, treated samples using the geometric means of each group. Statistical significance of the results was tested [p(H1)] with pairwise fixed reallocation randomization tests using the REST 2009 software (M. Pfaffl, Munich; Qiagen, Hilden, Germany).

Among the five prostaglandins (PGE₁, PGE₂, PF₂α, PGB₁, PGA₂) tested in Laminaria, only PGA₂ was found to be active (Fig. 1). The threshold concentration was observed to be between 0.1 and 1 μM, and the response occurred within seconds after adding the compound (Fig. 1). The strongest response was in the range of 5 μmol hydrogen peroxide per gram fresh weight. Also, diphenylene iodonium (DPI) at 10 μM did not inhibit the response of L. digitata to PGA₂ (data not shown).

Fig. 2 provides an overview of free saturated and unsaturated fatty acid contents in Laminaria upon incubation with PGA₂. No significant differences in the amounts of individual and total free fatty acids, compared to controls, were observed after this treatment.

Three oxylipins, previously identified as relevant markers of fatty acid oxidation in Laminaria (Kupper et al. 2006), were quantified (Fig. 3). When compared to control conditions, PGA₂ treatment resulted, after 10 min, in significantly (p < 0.001) increased levels of free 15-hydroxyeicosatetraenoic acid (15-HETE). In contrast, no significant differences were observed for 13-hydroxyoctadecatreienoic (13-HOTE) and 15-hydroxyeicosapentaenoic acids (15-HEPE).

In order to monitor the gene expression profiles obtained after PGA₂ treatment in Ectocarpus, five candidate genes (vBPO, MnSOD, HSP70, GRX, GST) were selected

based on our results in a parallel proteomics study of Ectocarpus subject to biotic stress (infection by Eurychasma dicksonii; Strittmatter, Gachon, Grenville-Briggs, van West, Potin and Kupper unpublished) in order to represent different functional categories related to oxidative stress, defense and cell rescue. Their respective transcript accumulations were quantified by a qPCR approach; in both biological repetitions after PGA₂ treatment the relative expression ratios of all target genes did not exhibit any significant differences compared to control samples [p(H1) > 0.05] (Table 2).

Significantly, this study reports the first case of an oxidative burst triggered by a prostaglandin in an algal or plant-like organism. The finding is very remarkable even with regard to animal models, where prostaglandins are known as modulators rather than actual inducers of oxidative bursts (Vaidya et al. 1999). To our knowledge, only one case of an oxidative burst triggered by a prostaglandin has been reported in eukaryotes, that of 15-deoxydelta(12,14)-prostaglandin J₂, an agonist of peroxisome proliferator activated receptor gamma, which induces superoxide anion production in human cells, with an efficacy similar to its action on ERK phosphorylation (Huang et al. 2002). In contrast, some prostaglandins have been reported to inhibit the oxidative burst (e.g., Ottonello et al. 1995). Also, the oxidative burst observed here is among the strongest in amplitude ever observed in a brown algal system: The PGA₂-triggered oxidative burst is considerably stronger with up to 5 μmol H₂O₂ g^-1 FW observed, in contrast to around 1 mol H₂O₂ g^-1 FW for oligoguluronates (Kupper et al. 2001, 2002), lipopolysaccharides (Kupper et al. 2006), polyunsaturated free fatty acids and methyl jasmonate (Kupper et al. 2009). In contrast to arachidonic acid and methyljasmonate (MeJA) (Kupper et al. 2009), the oxidative burst response to PGA₂ in Laminaria increased in a dose-response-like manner towards higher concentrations, with the strongest response observed at 100 μM. Also, contrary to the response triggered by lipopolysaccharides and methyl jasmonate, the oxidative burst in response to PGA₂ occurs within seconds after treatment.

Whilst the oxidative burst triggered by arachidonic acid, MeJA (Kupper et al. 2009), oligoguluronates (Kupper et al. 2001, 2002) and lipopolysaccharides (Kupper et al. 2006) can be inhibited by DPI, this does not apply to the response to PGA₂. This finding indicates that at least two different sources of ROS co-operate in oxidative burst responses in brown algae. This is consistent with certain results in higher plants (Bolwell et al. 1998, Bestwick et al. 1999) which suggest that both an NAD(P)H oxidase, a neutrophil-like superoxide-generating enzyme, susceptible to inhibition by DPI, a suicide substrate of flavin-containing oxidases (O’Donnell et al. 1993), and peroxidases should be involved in the oxidative burst.

Moreover, it is noteworthy that besides activating an oxidative burst, lipopolysaccharides (Kupper et al. 2006), polyunsaturated free fatty acids, MeJA (Kupper et al. 2009) and PGA₂ (this study) also induce fatty acid oxidation in Laminaria leading to the synthesis of hydroxyl-PUFAs - even though in the case of PGA₂ only one product (15-HETE) was present in significantly increased amounts. The major compounds observed in these studies, 13-HOTE, 15-HETE and 15-HEPE, have been described in Laminaria (Gerwick 1994) as putative ω-6 lipoxygenase metabolites. These three compounds (derivatives of C18:3, C20:4 and C20:5, respectively) have been reported to be anti-inflammatory mediators in mammals (Miller et al. 1990) but their physiological relevance in brown algae remains enigmatic. In our recent studies, we demonstrated that levels of these fatty acid hydroxides were highly increased after treatment of Laminaria plantlets with either bacterial lipopolysaccharides (Kupper et al. 2006), PUFAs or arachidonic acid (Kupper et al. 2009) and excess of copper (Ritter et al. 2008). The activation of such an intensity of PUFA oxidation was not reproduced here after treatment with PGA₂ - in fact, the low levels of free fatty acid release and fatty acid oxidation after PGA₂ treatment compared to the effects of lipopolysaccharide, MeJA and PUFAs are quite remarkable. This would support the notion that fatty acid oxidation is either not essential for the establishment of an oxidative burst when thalli are treated with PGA₂, i.e., that this prostaglandin activates signalling events considerably down-stream of PUFA oxidation - or that PGA₂ itself is a major agonist in defense reactions not recognized in previous studies. Further studies will be needed to address this question.

Furthermore and like for the peroxisome-proliferator activated receptors α and γ (PPAR α and γ) in mammals (Kliewer et al. 1995), arachidonic acid could be recognized by the same target receptors in L. digitata as its prostaglandin derivative: The absence of effects of inhibi

tors of cyclooxygenase and prostaglandin biosynthesis, indomethacin and aspirin (Jaworek et al. 2001), allene oxide synthase, aspirin (Pan et al. 1998) and of lipoxygenase (salicylhydroxamic acid) on the oxidative burst triggered by arachidonic acid in Laminaria which we have recently reported (Kupper et al. 2009) may be considered as first indications in this aspect. But neither in terrestrial plants nor in algae, any evidence exists so far for PPAR-like receptors. On the other hand, the possibility that arachidonate and PGA₂ selectively activate the production of two independent sources of ROS could explain the lack of inhibitor effects.

Evidence is growing that if terrestrial plants recognize their own lipid-based signals, they may also be able to recognize those of herbivores. For instance, caterpillar regurgitant contains aminated eicosanoids (peptidolipids) that are perceived by higher plants and induce elevated defenses (Arimura et al. 2000, Turlings et al. 2000). In this respect, it is noteworthy that prostaglandins have been described as elicitors from the oral secretions of herbivores (Schultz and Appel 2004) recognized by plants. A similar interpretation might be considered for the role of prostaglandins in the defense of Laminaria - even though PGA₂ has been shown to be synthesized in brown algae (Ritter et al. 2008), it cannot be ruled out to be produced by grazers and pathogens attacking brown algae as well.

The substantial knowledge about the important model Ectocarpus (Peters et al. 2004, Charrier et al. 2008), including the recently-accomplished genome sequencing project (Cock et al. 2010), prompted us to investigate this alga for potential molecular responses to PGA₂ treatment. However, our results with Ectocarpus showed no expression response to PGA₂ treatment for all five genes tested, highlighting that at most levels including gene expression, the overall effects of PGA₂ remain unclear at the current stage. Nevertheless, the unprecedented finding of a strong oxidative burst triggered by PGA₂ in Laminaria, suggesting that this compound is a potent defense inducer, warrants more in-depth mechanistic investigations going beyond the scope of this manuscript.

In conclusion, the discovery of an oxidative burst in an alga triggered by a prostaglandin is intriguing. Prostaglandins - in particular, PGA₂ and its derivatives - are well established as feeding deterrents in corals (Gerhart 1984, Whalen et al. 2010): the findings presented in this study highlight new perspectives of PGA₂ being a more widespread defense molecule and inducer in the marine environment, in very distantly related lineages.