Hepoxilins and trioxilins in barnacles: an analysis of their potential roles in egg hatching and larval settlement

Barnacles form a major constituent of the hard fouling organisms that encase man-made structures submerged in the marine environment(Christie and Dalley, 1987). Understanding the factors that regulate their life cycle is crucial towards the development of non-toxic or environmentally benign antifoulants to replace toxic tributyl tin-containing compounds. Following fertilization, barnacle eggs (embryos) are brooded in the mantle cavity of the sessile adults. Upon hatching, the larvae are liberated into the open water and proceed through six planktonic instars and one cyprid instar. During the cyprid or settlement stage, the larvae are believed to use their antennules to detect a suitable substratum upon which to attach and metamorphose into the sessile juvenile barnacle (Crisp, 1976).

The polyunsaturated fatty acid arachidonic acid [AA; 20:4(n-6)] is the precursor of two major families of bioactive lipids, the prostaglandins (PGs),arising through the action of cyclooxygenase(s) (COX), and a series of lipoxygenase (LOX) products that include the leukotrienes and(poly)hydroxylated eicosanoids (Stanley,2000). Whereas AA is found mainly in mammalian phospholipids(Kühn and Borngraber,1998), the equivalent predominant C20 polyunsaturated fatty acid present in aquatic invertebrates is usually eicosapentaenoic acid [EPA;20:5(n-3); Ackman, 1980]. The elongated polyunsaturated fatty acid docosahexaenoic acid [DHA; 22:6(n-3)] is also present in many aquatic animals and acts as a substrate for LOX(Holland et al., 1999) and aspirin-acetylated COX in terrestrial mammals(Serhan et al., 2002).

Within the barnacle life cycle there are two key stages where C20 lipids(eicosanoids) and C22 fatty acid derivatives are thought to be involved in signalling pathways that may influence barnacle development (see review byClare, 1995). Firstly, the LOX products 8-hydroxy-5,9,11,14,17-eicosapentaenoic acid(Hill and Holland, 1992;Hill et al., 1988) and a range of tri-hydroxy compounds (Hill et al.,1993; Song et al.,1990a) have been implicated as barnacle egg-hatching factors in both Elminius modestus and Semibalanus balanoides. In vivostudies, using extracted seawater in which barnacles (S. balanoidesand Chirona hameri) had previously liberated their larvae, found trace levels of tri-hydroxy fatty acids, suggesting that these, rather than the mono-hydroxy fatty acid compounds, were the biologically active compounds(Song et al., 1990b). Secondly, Knight et al. (2000)demonstrated that the COX products, prostaglandins E₂ and E₃, caused a dose-dependent inhibition of larval settlement in Balanus amphitrite, while indomethacin, a selective COX inhibitor,stimulated the settlement process.

Hepoxilins are epoxide metabolites of AA, EPA and DHA derived through the LOX pathway (Pace-Asciak,1986; Pace-Asciak et al.,1983; Reynaud and Pace-Asciak,1997; Fig. 1). In mammals, hepoxilins act primarily on calcium and potassium channels in membranes, which ultimately results in a wide range of functions such as insulin secretion, vascular permeability and vasoconstriction within the aorta and the activation of neutrophils and platelets (for reviews, see Pace-Asciak,1993,1994;Pace-Asciak et al., 1999). In invertebrates, their only reported function is in membrane depolarisation, via a potassium channel-mediated effect in the neurons of the marine mollusc Aplysia californica(Piomelli et al., 1989). Trioxilins are the biologically inert hydrolysis products of the hepoxilins(Fig. 1;Pace-Asciak et al., 1987;Pace-Asciak and Lee, 1989). Hill et al. (1993) reported a number of trioxilins with an 8,11,12-trihydroxy configuration (A-series)produced by S. balanoides that commonly appeared in fractions with egg-hatching activity.

The current study investigates the generation of trioxilins by E. modestus and B. amphitrite using a combination of high-performance liquid chromatography (HPLC) and gas chromatography (GC),both linked to mass spectrometry (MS). The potential functions of their precursors, hepoxilins, as signalling molecules within the egg-hatching and settlement processes of barnacles are also described.

Mussels covered by the test barnacle Elminius modestus Darwin were removed from the shores around Mumbles Pier, Swansea, UK on spring low tides. Upon return to the laboratory, the animals were maintained at approximately 15°C and 35‰ salinity in a seawater-circulating aquarium for up to 2 weeks prior to use in barnacle egg-hatching activity assays.

Adult Balanus amphitrite Darwin (from Beaufort, NC, USA) were maintained in aerated, 10 μm-filtered seawater at 22°C, which was changed every other day, and were fed daily on newly hatched Artemia(Sanders, Great Salt Lake, UT, USA). Following larval release, nauplii were cultured in batches (∼5000; at 1 larva ml^–1) in aerated,∼0.7 μm-filtered seawater at 28°C on a 16 h:8 h L:D cycle and were fed daily on Skeletonema costatum (Seasalter Shellfish Ltd,Whitstable, UK) in 1 litre at∼2×10⁵ cells ml^–1. Antibiotics were administered to the cultures as described by Rittschof et al.(1992). Under these conditions, development to the settlement stage cyprid took 4–5 days. Cyprids were obtained by filtration (retained on a 250 μm mesh) and stored in 0.45 μm-filtered seawater at 6°C until use in larval settlement assays.

Calcium ionophore A23187, arachidonic acid (AA), eicosapentaenoic acid(EPA), methoxylamine hydrochloride, bis (trimethylsilyl) trifluoroaceamide and dodecane were purchased from Sigma-Aldrich (Poole, UK). Hepoxilin A₃ was supplied by Biomol Research Laboratories (Plymouth Meeting,PN, USA). The stable hepoxilin B₃ analogue, PBT-3, was prepared as described in Demin and Pace-Asciak(1993). All solvents and acids used in the preparation of material for HPLC-MS and GC-MS were either AnalaR(BDH, Lutterworth, UK) or HPLC grade (Fisher Scientific UK, Loughborough). Ethereal diazomethane was prepared from N-methylnitrosotoluene sulphonamide and alcoholic KOH at 65°C using an Aldrich kit(Sigma-Aldrich).

Trioxilin A₃ used in hatching assays and as HPLC-MS and GC-MS standards was prepared by acidification of hepoxilin A₃. Briefly,100 μl portions of hepoxilin standard (5 μg) were dried under nitrogen,resuspended in 20 μl acetonitrile (20%) and acidified using 50 μl of 0.1% formic acid. Samples were incubated for 20 min, freeze-dried and stored at –80°C in 100 μl methanol until use.

The barnacle extracts used in larval settlement assays and those that were routinely run on HPLC-MS and GC-MS were prepared from known masses(0.7–1.6 g) of soft tissue of either adult barnacles (E. modestus or B. amphitrite) or known numbers (1000–10 000)of larvae (B. amphitrite). Individual species were disrupted on ice in 5 ml of filtered (0.22 μm) seawater. The crude cell preparations were stimulated, for 20 min at 16°C, with 5 μmol l^–1calcium ionophore A23187 and either 10 μmol l^–1 AA, 10μmol l^–1 EPA or an equal volume of the fatty acid carrier solvent, ethanol. Debris was removed by centrifugation (2 min, 10 000 g, 4°C) before acidifying the supernatants to pH 3.5 with 10% acetic acid. Samples (4 ml) were loaded onto pre-washed C₁₈Sep-Pak solid phase extraction cartridges (Waters Chromatography, Watford,UK), washed with 5 ml ultra-pure water followed by 1 ml hexane and eluted with methanol (5 ml). Eluants were dried under nitrogen, resuspended in 200 μl methanol and stored at –80°C until preparation for separation by HPLC-MS or GC-MS or use in settlement assays.

Crude extracts for use as positive controls in hatching assays were obtained by macerating each adult barnacle species (E. modestus or B. amphitrite;∼50 g) in 25 ml 0.22 μm-filtered seawater. After 1 h, the liquid fraction was decanted off and the solid fraction homogenised twice (1 h each wash) with 25 ml seawater. Seawater fractions were combined and centrifuged to remove barnacle debris (15 min, 500 g,4°C). An equal volume of acetone was added to the supernatant, and the precipitated protein was pelleted by centrifugation (10 min, 500 g, 4°C). The supernatant was centrifuged at 14 000 g for 30 min at 4°C before being filtered (Whatman #1) to remove any remaining particulates. The supernatant was then rotary evaporated(40°C) to remove all acetone, and the aqueous fraction adjusted to contain a final volume of 15% ethanol. The sample was acidified to pH 3.9 with glacial acetic acid, loaded onto a pre-washed (5 ml methanol followed by 5 ml ultra-pure water) C₁₈ Sep-Pak, washed with 20 ml ethanol (15%)followed by 20 ml petroleum ether and finally eluted with ethyl acetate. The ethyl acetate eluants were stored at –20°C until use in egg-hatching assays and product identification studies by electron impact gas chromatography–mass spectrometry (GC-EIMS).

Samples were separated using a Jasco HPLC system on a Spherisorb ODS2 (5μm, 25 cm×0.46 cm) column (Hichrom) at 1 ml min^–1with a 45 min linear gradient of acetonitrile:0.01% formic acid (20:80 to 80:20 v/v). Prior to injection onto the column, samples were dried under N₂ and resuspended in 100 μl of acetonitrile:0.01% formic acid(20:80 v/v) and 5 μl 10% formic acid. The solvent flow passed through a Jasco diode-array detector and was then split; 1/13th of the flow was directed into the electrospray source of a Quattro II triple quadrupole mass spectrometer (Waters Ltd, Elstree, UK) operated in the negative ion mode, and the remainder went to a fraction collector. Fractions (1 min) were collected for 40 min of the HPLC run and, where necessary, were derivatized for GC-EIMS.

Samples were converted to their methoxine O-trimethylsilyl ether methyl esters for GC-EIMS as described by Knight et al.(1999). Briefly,N₂-dried samples were methoximated with 50 μl of 2%methoxylamine hydrochloride for 30 min at room temperature (RT). After two steps of ethyl acetate extraction, esterification was carried out using ethereal diazomethane (1 h at RT) and the esters treated with 50 μl bis(trimethylsilyl) trifluoroaceamide overnight at RT in an N₂atmosphere. Samples were resuspended in 20 μl dodecane for GC-EIMS on a Trio 100 instrument (Thermoquest, Hemel Hempstead, UK). Chromatography was carried out on a 30 m DB5 capillary column (Jones Chromatography, Hengoed, UK)using helium as the carrier gas and a temperature gradient of 10°C min^–1 over 150–320°C. The GC column was routed into the mass spectrometer for analysis by EIMS.

The promotion or inhibition of B. amphitrite larval settlement by ionophore-challenged B. amphitrite tissue, hepoxilin A₃and PBT-3 was investigated using freshly moulted cyprids (day 0) and 3-day-old cyprids, respectively. All larval settlement assays were performed using flat-bottomed 24-well culture plates. Each well contained 2 ml of 0.45μm-filtered seawater, 10 cyprids (day 0 or day 3), and 10 μl of hepoxilin or PBT-3 dissolved in 20% methanol so as to give a final concentration of 0.25, 0.5, 5 or 30 μmol l^–1. Control plates were run in parallel, each containing either 10 μl methanol (20%) or 10 μl seawater without hepoxilin A₃ or PBT-3. All plates were incubated at 28°C in the dark for up to 48 h and were scored for the percentage attachment/metamorphosis.

Immediately prior to their use, E. modestus were carefully removed from the host mussel shells using a sharp seeker. For each assay, fully developed paired egg masses were removed intact from a single adult barnacle and washed for 1 h inrunning seawater. One of the egg masses was transferred into 250 μl of filtered (0.45 μm) seawater containing a known concentration of each test compound in an ethanol vehicle. The other egg mass was placed in 250 μl of a control solution containing an equal amount of ethanol (<5%) in seawater. All egg masses were incubated for 15 min at 12°C in separate wells of 48-well culture plates and then fixed by the addition of∼100 μl of 10% seawater formalin. Numbers of hatched (stage 1 nauplii) and unhatched larvae were counted using a Sedgewick–Rafter counting chamber. The number of hatched larvae was expressed as a percentage of the total number of larvae. Any undeveloped larvae found in the centre of the egg mass were excluded from the count. Any unhatched egg masses were tested for hatching viability by addition of the crude hatching factor.

Ionophore challenge of both adult B. amphitrite and E. modestus soft tissues resulted in the formation of a number of A₂₀₀ UV-absorbing species(Fig. 2A). Online analysis by negative ion electrospray MS revealed the presence of EPA, AA and DHA eluting at 34.08 min (M-H^– at m/z 301), 37.66 min(M-H^– at m/z 303) and 36.15 min (M-H^– at m/z 327), respectively, together with some LOX products:mono-hydroxyeicosatetraenoic acids (28.5 min; λ_max 234 nm;M-H^– at m/z 319), mono-hydroxyeicosapentaenoic acids (26.2 min; λ_max 234 nm; M-H^– at m/z 317) and monohydroxydocosahexaenoic acids (28.0 min; λ_max 234 nm;M-H^– at m/z 343). Another component present in the mixture that eluted from HPLC at 16 min generated an intense M-H^– ion at m/z 351, corresponding to a molecular mass of 352. The methoxime O-trimethylsilylether methyl ester derivative of this compound eluted from the GC at 15.05 min and gave an EI mass spectrum consistent with the EPA-derived trihydroxy fatty acid trioxilin A₄(8,11,12-trihydroxy-5,9,14,17-eicosatetraenoic acid), with ions at m/z 567(M-CH₃)⁺, 477 (567-TMSOH)⁺, 444(characteristic rearrangement and loss of ω chain), 441[M-CH₂.CH=CH.(CH₂)₃.COOCH₃]⁺,383(M-CH₃CH₂CH=CH.CH₂.CH=CH.CH₂-TMSOH)⁺,371, 282, 281, 243[TMSO=CH.CH₂.CH=CH.(CH₂)₃.COOCH₃]⁺and 211[TMSO=CH.CH=CH.(CH₂)_.CH=CH.CH₂.CH₃]⁺(Fig. 2B). The AA-derived product trioxilin A₃ (8,11,12-trihydroxy-5,9,14-eicosatrienoic acid) was also present in the preparation of E. modestus hatching factor. It was identified by co-elution with authentic trioxilin A₃on HPLC-MS, eluting at 18.0 min with an M-H^– ion at m/z 353,and by GC-EIMS, eluting at 14.4 min with ions at m/z 569(M-CH₃)⁺, 479 (569-TMSOH)⁺, 444, 441[M-CH₂.CH=CH.(CH₂)₃.COOCH₃]⁺,383(M-CH₃CH₂CH=CH.CH₂.CH₂CH.CH₂-TMSOH)⁺,371, 282, 281, 243[TMSO=CH.CH₂.CH=CH.(CH₂)₃.COOCH₃]⁺and 213 [TMSO=CH.CH=CH.(CH₂)_.CH₂CH₂ CH₂.CH₃]⁺. Incubation of E. modestus crude cell extract with either exogenous EPA or AA prior to ionophore challenge caused a marked increase in the production of trioxilin A₄ and trioxilin A₃, respectively. There was no evidence of trioxilin B₃/B₄ production in either barnacle species(based on the reported mass spectra); however, small amounts of an isomericω-3 trioxilin A₃, 8,11,12-trihydroxy-9,14,17-eicosatrienoic acid, were observed in the GC-MS profile (14.6 min) and identified as arising from ω-3 AA by characteristic ions at m/z 569, 479, 463, 446, 422, 385,371, 351, 282, 281, 243, 211, 147, 121 and 73. The DHA-derived product 10,13,14-trihydroxy-4,7,11,16,19-docosapentaenoic acid was also present,eluting from GC at 17.3 min and generating a spectrum with ions at m/z 608(M⁺), 593, 470, 439, 409, 351, 230, 211, 121 and 73.

Immediately prior to settlement, the free-swimming cypris larvae of B. amphitrite were observed searching the plate surface with their antennules (Fig. 3A). During the settlement process, larvae initially attached to the substratum by their antennules, then head-planted down onto the surface and eventually metamorphosed (moulted) into the new spat(Fig. 3B).

For all experiments, the use of methanol as the solvent carrier (at 0.1%final incubation concentration) had no significant effect on larval settlement of B. amphitrite when compared to seawater controls (unpaired t-test, P>0.05; data not shown). Crude larval (equivalent to products from 1000 cyprids well^–1) and adult (equivalent to products from 0.5 adults well^–1) extracts from B. amphitrite both had a stimulatory effect on larval settlement (unpaired Student's t-tests, P=0.004 and P=0.01 for larvae and adults, respectively; Fig. 3C). Extracts equivalent to the products from ≤200 larvae well^–1 and ≤0.1 adults well^–1 had no effect on the settlement of B. amphitrite larvae (unpaired Student's t-tests, P>0.05). Neither hepoxilin A₃ nor a stable analogue of hepoxilin B₃, PBT-3, had a stimulatory(Fig. 3D and 3E, respectively)or inhibitory (data not shown) effect on the settlement of B. amphitrite larvae.

Post-removal from adult E. modestus, the pair of egg masses each appeared as a tight aggregation of over 1000 torpedo-shaped embryos embedded in a membrane-bound structure (Fig. 4A,B). Upon addition of the crude egg-hatching factor, the nauplii larvae rapidly emerged (<5 min) from the egg mass with limbs outstretched and actively swimming (Fig. 4C,D). Crude hatching extract derived from E. modestuscaused 98.17±0.44% (mean ± 1 s.e.m., N=3) of larvae in the egg mass to hatch. Similar levels of hatching (>95%) were observed when E. modestus egg masses were exposed to B. amphitrite crude hatching extract (data not shown). Authentic trioxilin A₃ had no effect on hatching at any of the concentrations tested(Fig. 5A), while hepoxilin A₃ caused significant numbers of larvae to hatch at the 10^–6 mol l^–1 concentration only (paired t-test, P=0.0037; Fig. 5B). The stable hepoxilin B₃ analogue PBT-3 at a concentration of 10^–7 mol l^–1 also had a stimulatory effect on hatching (paired t-test, P=0.0121;Fig. 5C). A concentration of 10^–6 mol l^–1 PBT-3 had no significant effect on larval hatching, although the larvae released were observed to exhibit an inhibition in their swimming behaviour that was not seen with the other concentrations of PBT-3.

EPA is the major C20 polyunsaturated fatty acid in the barnacles B. amphitrite and E. modestus and is metabolised by 8- and 12-LOX activities to form 12-hydroxyeicosapentaenoic and 8-hydroxyeicosapentaenoic acids (Knight et al., 2000). As was seen in S. balanoides by Hill et al.(1993), the current study found that, following acidification, B. amphitrite tissue extracts contained trioxilin A₃, trioxilin A₄ and 10,13,14-trihydroxy-4,7,11,16,19-docosapentaenoic acid. However, unlike Hill et al. (1993), no evidence was found to support the generation of trioxilin B₄(10,11,12-trihydroxy-5,8,14,17-eicosatetraenoic acid). The production in B. amphitrite of trioxilins A₃ and A₄, together with minor components of ω-3 trioxilin A₃ and 10,13,14-trihydroxy-4,7,11,16,19-docosapentaenoic acid, demonstrates that the 12-LOX is active in barnacles across a range of polyunsaturated fatty acids including AA, EPA, DHA and ω3 AA. The absence of the epoxide precursors(hepoxilins) is not unexpected as hydrolysis (both enzymatic and non-enzymatic), particularly under the acidic conditions of HPLC, occurs rapidly (Demin et al.,1995).

It has previously been suggested that the trioxilins exhibit biological activity as egg-hatching factors (Hill et al., 1993; Song et al.,1990a). The commercial availability of hepoxilin A₃allowed us to investigate the effects of hepoxilin A₃ and trioxilin A₃ on egg hatching. In contrast with earlier reports, trioxilin A₃ exhibited no effects in either assay at concentrations up to 10^–6 mol l^–1. The epoxide precursor,hepoxilin A₃, was active at 10^–6 mol l^–1. Hepoxilin A₃ and the stable analogue for hepoxilin B₃, PBT-3, both stimulated the hatching of nauplius larvae at 10^–6 mol l^–1 and 10^–7 mol l^–1, respectively. When standardised for the levels of hatching in the ethanol controls, the amount of hatching in the 10^–7 mol l^–1 PBT-3 treatment(34.67±3.84%; mean ± 1 s.e.m., N=3) was approximately four times greater than that observed for the 10^–6 mol l^–1 hepoxilin A₃treatment (8.83±1.72%, N=6). This difference was presumably a reflection of the rapid enzymatic breakdown of hepoxilin A₃ to the inert trioxilin A₃. Despite the evidence for the involvement of hepoxilin in the egg-hatching process, the levels of hatching for the 10^–7 mol l^–1 PBT-3 treatment(62.33±10.11%, N=3) were much lower than those observed when the egg mass was exposed to the crude hatching factor (98.17±0.44%, N=6). The reason for this discrepancy may be due to a number of possibilities. Firstly, it may be that the process of barnacle egg hatching is triggered by a number of factors, such as 8-hydroxyeicosapentaenoic acid(Hill and Holland, 1992;Hill et al., 1988), that may act synergistically to cause larval release. Secondly, given that the major precursor fatty acid in barnacles is EPA, it is possible that the EPA product,hepoxilin A₄, or indeed the DHA-derived 10-hydroxy-13,14-epoxydocosapentaenoic acid, might result in higher levels of larval hatching activity than those observed with the equivalent AA-derived product, hepoxilin A₃. Unfortunately, hepoxilin A₄ (or a stable analogue) is not commercially available, thus the relative effects of the EPA-derived product versus the equivalent AA-derived product remain to be determined. Finally, PBT-3 used in the current study is an analogue for hepoxilin B₃, a product that was not observed in E. modestus. The use of a stable analogue for hepoxilin A₃in larval hatching assays might show greater activity than observed in the present study.

Mention should also be made of the concentration of hepoxilins required in the current study to potentiate larval hatching. It is questionable whether a concentration as high as 10^–6–10^–7 mol l^–1 hepoxilin would exist naturally in the mantle cavity of barnacles where these unhatched larvae normally reside. However, if synergism exists, where a range of bioactive eicosanoids contribute to the natural barnacle hatching-factor activity, then a cumulative effect could lead to concentrations as high as those used in the current study.

No effects on B. amphitrite settlement were seen with either hepoxilin A₃ or PBT-3, suggesting that these compounds are not involved in the communication pathways for settlement. However, the involvement of hepoxilins and trioxilins within the barnacle settlement process cannot be completely ruled out. If these compounds function as internal signallers, the concentrations required to permeate the cyprid exoskeleton (cuticle) may exceed those used in the current study (maximal concentration used was 30 μmol l^–1). Indeed, Knight et al.(2000) required 50 μmol l^–1 prostaglandin E₂ or prostaglandin E₃ before significant effects on B. amphitrite settlement were observed.

The mechanism of action of hepoxilins and other inducers of larval egg hatching is likely to differ greatly from their potential involvement in the settlement process. Unlike cyprids, which have a thick impervious cuticle,nauplii have only a thin cuticle that has been shown to be highly porous to compounds with molecular masses of <800 Da(Stuart et al., 2002). Thus,in addition to the possibility of interaction with external receptors,compounds such as eicosanoids (∼350 Da) may also naturally diffuse through the outer epidermis and make contact with internal receptors. By contrast, the external cues for settlement are generally thought to be peptidic or proteinaceous in character and potentially detected by specific receptors on the cyprid antennules (Clare,1995). Rather than acting as exogenous signalling compounds, the involvement of eicosanoids in larval settlement is more likely to be within internal signal transduction pathways. An internal role for endogenous eicosanoids is supported by the finding that indomethacin, a selective COX inhibitor, stimulates larval settlement(Knight et al., 2000).

In summary, trioxilins derived from AA, EPA, ω3 AA and DHA were produced following ionophore challenge in both E. modestus and B. amphitrite. Although trioxilin A₃ did not affect egg hatching,the parent epoxide, hepoxilin A₃, was active at 10^–6 mol l^–1. The stable analogue PBT-3 showed activity at 10-fold lower concentrations than hepoxilin A₃,presumably due to its stability to hydrolysis. The egg-hatching activity of barnacle extracts can be ascribed, in part at least, to the hepoxilins. As the extraction protocol involved acidification, which would hydrolyse hepoxilins,it is also likely that some bioactivity may reside with some of the numerous other eicosanoid-like components produced by E. modestus and B. amphitrite.

We should like to thank Dr Dan Rittschof, Duke University Marine Laboratory, for supplying adult Balanus amphitrite. This work was supported by NERC grant GR3/12765 and NERC studentship NER/S/A/2000/03648.