Tasieae, Vol. 32, No . 6, pp . 639-655, 1994 Elsevia Scirnex Ltd ~~-~1~1~94~E001~ Printed in Geest B~tain 0041-0101J94 526.00+ 0.00

REVIEW ARTICLE

DEFENSIVE ROLES FOR SECONDARY METABOLITES FROM MARINE AND -FEEDING NUDIBRANCHS

PeTEtt PRORSCH Julius-von-Sechs-Institut f'ttr Biowissenschaften, Universität Wiuzburg, Mittlerer Dallcnbelweg 64, D-97082 Wtitzburg, Germany

(Received 19 November 1993 ; accepted 3 February 1994)

P. PROKSCH. Defensive roles for secondary metabolites from marine sponges and sponge-feeding nudibranchs. Toxicon 32, 63955, 1994.-In the marine environment sponges (Porifera) constitute one of the most interesting sources of bioactive natural products. The high frequency of bioactive components in these primitive filter-feeders is interpreted as chemical defence of sponges against environmental stress factors such as predation, overgrowth by fouling organisms or competition for space. The highest incidence of toxic or deterrent sponge metabolites is consequently found in habitats such as coral reefs that are characterized by intense competition and feeding pressure due, for example, to carnivorous fish. Further support for the adaptive significance of sponge constituents is derived from the observation that sponges which are growing exposed are usually more toxic than those growing unexposed . Whereas the chemical defence of sponges seems to be highly effective against most species of fish, a group of shell-less gastropods, the nudibranchs, has specialized on sponges. While feeding on sponges the nudibranchs sequester the effective chemical armoury of their prey, which is subsequently employed for their own protection. Some nudibranchs, however, have become independent of this interspecific flow of natural products and are able to accumulate defensive compounds through de nova synthesis.

INTRODUCTION Evarr though the search for bioactive natural products from marine organisms is still a young field when compared to the long tradition of terrestrial pharmacognosy, the last two or three decades of marine pharmacognosy have yielded a fascinating array of natural products with pronounced pharmacological activities, which in turn have prompted major screening efforts by academic institutions, pharmaceutical companies, and public health organizations such as the NCI, in the search for new drugs from the sea (e.g. voN BERr ePSCx, 1980; MUNRO et al., 1987; 1ZINEHART er al., 1993). Promising compounds that originated from these surveys include the powerful antitumour compounds didemnin B (1, Fig. 1) from of the genus Trididemntvn and bryostatin 1 (2, Fig. 1) from the bryozooan Bugula neritina (ScFnen-rz et al., 1993). Bryostatin 1 is expected to enter clinical

639

640 P. PROKSCH

trials shortly (IREi,nxn et al., 1993). Other pharmacologically active marine natural products, such as tetrodotoxin (3, Fig. 1) or okadaic acid (4, Fig. 1), have been established as indispensable tools for physiological studies of Na+ channels (C.v and Joxnnrr, 1990) or of cellular processes involving phosphorylation and dephosphorylation (I'IAYSTEAD et al., 1989).

Cli=OH

Fya. 1 . C~c~u. srxucrux~s (Cwetinued ). 1, Dimemnin B; 2, bryostatin 1; 3, tetrodotoxin; 4, okadaic acid.

Defensive Roles for Secondary Metabolites 641

CHO

e

H~C CH~ FhC CHs

Z fl

OH

~a. 1. S, Latrunculin-A; 6, latrunculin-B; 7, hetetonemin; 8, acalnr+cü~l: g, manoalide ; 10, seco-manoalide; 11, siphonodictidine .

Marine invertebrates have yielded a larger number of bioactive natural products than algae haveIRELAND et al., 1993; FAULICIVER, 1993), which is in sharp contrast to the terrestrial envirotunent where plants are by far the richest source of secondary metabolites

642 P. PROKSCH

NH OH

Fya. 1. 12, Siphonodictyal-A; 13, aiphonodictyal-H ; 14, 1-methyladenine; I5, serothionin; _16, homoaerothionin ; 17, isofistularin-3 ; 18, acrophobin-2 .

Defensive Roles for Secondary Metabolites 643

CN

CONFIA

CFI,OH

F~c. 1 . l9, Fistularin-l ; 20, aeroplysinin-1 ; 21, dienone; 22, halichondrin-H ; 23, ciguatoxin .

(Lucxrmt, 1984). Among the numerous groups of marine invertebrates the sponges (Porifera) and coelenterates (Coelenterata) are by far the leading sources of natural products in terms of the sheer number of compounds isolated (F~ul.xxeR, 1993). With regard to the biogenetic diversity of secondary constituents, sponges rank first. At least

644 P. PROKSCH

10(H1-1500 different natural products representing terpenoids, polyketides, and amino acid derivatives have been reported from sponges ~FALJLKNER, 1993 ; IRELAND et al., 1993 ; SARMA Bt al., 1993).

H~C

24 2~

u

H~

ÇHa

H~C CH~

F~a. 1. 24, A diketopiperatine ; 25, 9-isocyanopupukeanane ; 26-_28, seaquiterpene and savterterpcne furane; 29-31, eeequiterpenéi~onitrilee, -isothiocyanates, and -formamide; 32, olepupuane .

Defensive Roles for Secondary Metabolites 645

CHO

H,C CFh FIaC CH~

Fta. 1 . 33, 7-Deaoetoxyolepupuane; 34, polygodial; 35, limatulone .

The high incidence of bioactive natural products in sessile marine invertebrates, such as sponges that are frequently exposed to numerous stress factors such as predation, overgrowth by fouling organisms or competition for space without effective morphological defence mechanisms, is certainly not incidental. It is rather considered to have evolved as chemically mediated defence mechanisms protecting these soft-bodied invertebrates from environmental dangers. Research on the ecological importance of marine natural products is still in its infancy when compared to the long tradition of terrestrial chemical ecology which can be traced back at least 100 years (e.g. STAHL, 1904; ZuxnL, 1895). Ample experimental evidence in favour of the defensive role of secondary metabolites from sponges, however, already exists, although it is scattered in the literature. It is the aim of this report to compile this information and to give a comprehensive view of our current knowledge on the adaptive significance of secondary metabolites from sponges with other marine organisms.

BIOLOGY OF SPONGES Sponges (Porifera) are primitive multicellular organisms that originated some 1 .6 billion years ago in the Paleozoic and are distinguished from the true Eumetazoa by the lack of organized tissues (Mïn.LER, 1982; $ARMA et al., 1993). All 5000 or so currently known sponge taxa are aquatic, with the majority of them restricted to the marine environment. Sponges are filter-feeders which inhale water through numerous pores (ostia) piercing the external body wall. These pores lead into a system of channels which, after permeating almost the whole body, open to the exterior by a few larger exhalant openings called oscula. A continual flow of water through this channel system is maintained by ciliary

646 P. PROKSCH

action. From this stream of water microscopic food particles such as , microalgae or detritus are phagocytozed . Sponges show a circumpolar distribution and have adapted to a diverse array of marine habitats (MESS, 1989; SAxMA et al., 1993). They may be found beneath the antarctic ice-cover as well as in temperate or tropical oceans . The size, shape and colour of sponges may also vary extensively. Some species may reach a height of more than 2 m, whereas others can be found as tiny encrusting sponges on shells with diameters seldom exceeding more than a few centimeters. Some sponges are cryptic, living secluded in caves or niches, whereas others are found exposed, often signalling their presence by a bright and conspicuous coloration . Sponges contribute significantly to the biomass of most marine habitats, sometimes as dominant elements, for example on coral reefs (SARMA et al., 1993). In spite of their sessile existence, their longevity and their potential nutritional importance, sponges are only rarely attacked by predators such as carnivorous fish (BRAEKMAN and DALOZE, 1986; MESS, 1989). In a survey of the dietary habits of Caribbean reef fish only 11 out of a total of 200 different species analysed were found to feed on sponges, suggesting that sponges are avoided by the majority of the usually polyphagous reef fish (RANDALL and HARTMANN, 1968). Sponges in the classes Calcarea or Sclerospongiae may derive morphological protection from predators by their sharp silica or calcium carbonate spicules which are imbedded in the soft tissue. However, many sponges from the largest class Demospongiae, such as the bath sponge Spongia o~cinalis, have no morphological defence mechanisms since protective spicules are replaced by proteinaceous fibres made from spongin, which are unlikely to deter predators. Why are these seemingly defenceless sponges avoided by most marine predators?

Chemical defence of marine sponges against fish Chemical defence through accumulation of sometimes copious quantities of toxic or deterrent natural products is commonly found within the Porifera (BRAEKMAN and DALOZE, 1986; MESS, 1989; PAwc.nc, 1993). A striking example of the significance of secondary products from sponges for the chemical defence against fish is provided by the vividly red-coloured sponge Latrunculia magn~ca that is found in the Gulf of Aquaba (Red Sea). Latrunculia magnfica is one of a few sponge species from this habitat that grow exposed, whereas the majority of sponges in the GulfofAquaba are cryptic and are usually found below coral growth or rocks (NEEMAN et al., 1975). Field observations indicated that specimens of L. magnifica were avoided by fish that readily accepted some of the cryptic sponges when exposed. When L. magnifica was squeezed, the sponge was found to exude a reddish fluid that caused fish to flee immediately from the vicinity. When the juice was squeezed into aquaria, fish (Gambusia a~nis) died within a few minutes. Symptoms developing following contact with the sponge toxins included excitation of the fish after X10 min, jumping, partial paralysis, turning over the back, haemorrhage of the gills and finally death (NEEMAN et al., 1975). Chemical analysis of the sponge afforded the macrocyclic latrunculin-A (5, Fig. 1) and -B (6, Fig. 1) that were responsible for the strong ichthyotoxicity of the sponge (KASHMAN et al., 1980; GROWEiSS et al., 1983). The toxicity of latrunculin-A or latrunculin- B was shown to be at least partly due to inhibition of acetylcholinesterase . In addition to an immediate inhibitory effect on the esterase an almost complete inhibition of the enzyme

Defensive Roles for Secondary Metabolites 64 7

was observed upon incubation of the toxin for 1 hr with cholinesterase followed by addition of the substrate (i~IEEMAN et al., 1975). In addition to latrunculin-A or latrunculin-B numerous other natural products from marine sponges have been shown to be effective deterrents to marine fish occurring with the respective sponges in the same habitat. Examples include heteronemin (7, Fig. 1), scalardial (8, Fig. 1), manoalide (9, Fig. 1), and seco-manoalide (10, Fig. 1). In field studies conducted at Guam all of the latter sponge constituents were found to deter feeding of generalist fish at or below their natural concentrations as present in the respective sponges, thereby indicating their protective nature under ecologically relevant conditions (PnU~, 1992).

Geographical correlation of sponge toxicity Most reports on toxic or deterrent sponge constituents are from tropical sponges, whereas accumulation of bioactive natural products seems to be less pronounced in sponges from temperate or cold waters. Screening experiments that were conducted along a north-south geographic gradient from Washington (U.S.A.) to Vera Cruz (Mexico) indicated geographical patterns of sponge toxicity (GR1~N, 1977). In the respective study, goldfish (Carassius auratus) that were kept in aquaria were exposed to solvent extracts from sponges, whereas field studies employed wrasses (Halichoeres bivittatus) that were force-fed with sponge tissue. The total number of sponge species tested for ichthyotozicity varied from 34 at San Juan Island off the coast ofWashington, to 36 at Vera Cruz, Mexico. The percentage of toxic sponges vs. total number of sponges tested showed a clear increase from north to south. Whereas only 9% of all sponges from San Juan Island were toxic to fish, the number of toxic species increased to 75% at Vera Cruz (GR~rr, 1977). The observed increase of toxic sponges from north to south is paralleled by an increase of species diversity of marine fish. In a survey conducted at the Pacific coast of Canada, for example, a total of 227 species of fish was reported from nearshore and deep waters (CLEMENS and WILBY, 1946), whereas in the Gulf of Mexico the number of fish species in nearshore and deep waters amounted to c. 440, with more than 300 species being found on coral reefs (STARCK, 1968). Surveys conducted in Queensland, Australia, reported more than 1000 species of near-shore and deep-water fish (Scoz-r, 1962). It is assumed that the increase of species diversity of marine fish in tropical waters will result in higher competition for food resources which may lead to more specialized feeding habits. As a consequence the grazing pressure of fish on invertebrates such as sponges is likely to increase in tropical habitats, which in turn will result in optimized defence strategies of prey organisms, including the accumulation of highly effective toxins (GR~rr, 1977). The proposed adaptive significance of natural products for the defence of sponges is corroborated by comparative studies on the toxicity of exposed and unexposed sponges conducted in several tropical habitats. At Vera Cruz, Mexico, approximately 80% of all toxic sponges were found living exposed to potential fish predators, whereas of the nontoxic sponges c. 70% were cryptic, for example under corals (Bnxus and GREEN, 1974). Similar trends were observed in a field study on Lizard Island at the Great Barrier Reef (Australia), which focused on sessile or slow-moving invertebrates from several taxonomic groups including sponges, soft corals, gorgonians, holothurians and ascidians. In this study 73% of all exposed coral reef invertebrates were found to be toxic to fish, whereas 75% of all unexposed species were nontoxic (Batas, 1981).

648 P. PROKSCH

Chemical defence of marine sponges against spatial competition andfouling Natural products from sponges are certainly also involved in spatial competition as well as in the inhibition of fouling by epibionts, even though comparatively few studies have been conducted on these latter aspects. Spatial competition is pronounced on coral reefs that are characterized by an exceedingly high species diversity and a density of population that is unmatched in any other ecosystem, with the possible exception of tropical rain forests (MESS, 1989; SALE, 1991). Homogenates of sponges have been shown to inhibit the growth of other sessile invertebrates including bryozoans, bivalves and brachiopods, thereby providing preliminary evidence that natural products from sponges may play a role in competition with other invertebrates (GREEN et al., 1990; JACxsox and Buss, 1975; THOMPSON et al., 1985). This suggestion, which was based on laboratory studies, is corroborated by the occurrence of distinct bare zones around sponges that are frequently observed in the natural habitat (JACxsoN and Buss, 1975; THOMPSON et al., 1985). The actual compounds that may be responsible for the observed allelopathic effects of sponges, however, have only been characterized in a few cases. A vivid example of the allelopathic importance of natural products from sponges is provided by burrowing sponges of the genus Siphonodictyon. Sponges of this genus, including S. coralliphagum from the Caribbean, burrow into the heads of living corals. In order to survive, these sponges must be able to prevent overgrowth by the coral polyps. Consequently, 1-2 cm zones of dead coral polyps are observed around the base of each oscular chimney (SULLIVAN et al., 1981). Sponges maintain the dead zones around the chimneys by the production of mucus which apparently acts as a carrier for the toxic compounds siphonodictidine (11, Fig. 1) as well as siphonodictyal-A (12, Fig. 1) and -B (13, Fig. 1) (SuLLivnN et al., 1981). In laboratory experiments siphonodictidine was shown to inhibit the rate of photosynthesis of endosymbiotic algae in the hard coral Acropora Formosa at concentrations > 10 ppm, whereas at a concentration of 100 ppm lysis of the coral cells was observed (SuLLtvAN et al., 1983). The sponge S. mucosa grows partially buried in coral debris, thereby differing from S. coralliphagum which burrows into live corals. Interestingly, S. mucosa was also shown to produce copious quantities of mucus. The toxic metabolites siphonodictidine and siphonodictyal, however, were not detected in the mucus of S. mucosa, thereby strengthening the supposed allelopathic role of the respective natural products from S. coralliphagum (SuLLtvAN et al., 1983). Similar allelochemical interactions have been reported between the liver sponge Plakortis halichondroides and the stony coral Agaricia lamarcki in the Caribbean . The sponge was shown to kill the coral on direct contact and also indirectly through waterborne metabolites by reducing the photosynthetic rates and increasing the nocturnal respiration rates of the symbiotic zooxanthellae of the corals (PORTER and TARGETT, 1988). Most marine surfaces are covered with bacteria, diatoms and protozoa that form the `primary film' which has been shown to be a prerequisite for the colonization by higher epibionts such as macrophytes or invertebrates (DAVrs et al., 1989). The process of`fouling' may be dangerous, especially for filter-feeders such as sponges, since pores are easily obstructed thereby reducing the feeding efficiency ofthe sponge; alternatively, filter-feeding epibiontc invertebrates such as may even compete directly with the sponge for food. Even though epibionts are virtually omnipresent in the marine environment, the surfaces of most sponges are usually remarkably free from fouling macrophytes or invertebrates . Given the great number of antibiotically active natural products that have been isolated from sponges (BURKHOLDER, 1973; BRAEKMAN and DALOZE, 1986; McCAr~REY et al., 1985;

Defensive Roles for Secondary Metabolites 649

Mss, 1989), an inhibition of primary film formation by sponge metabolites seems likely. However, most of the experiments on the antibacterial activity of sponge chemical constituents have been conducted with terrestrial rather than with marine bacteria, which must bring into question the ecological relevance of data obtained in the respective experiments (B$RCQutsr and BEDFORD, 1978; Bnxus et al., 1990). Several recent studies, however, demonstrated the antibacterial activity of sponge metabolites against ecologically relevant marine bacteria. The purine derivative 1-methyladenine ~, Fig. 1) isolated from the Caribbean sponge Aplysilla glacialis, for example, was shown to inhibit the growth of several marine bacteria from the same habitat including spp. in the genera Vibrio, Acinetobacter and Flavobacteritun (Bosznv and F~uL.Ktv~t, 1992). A further obstacle to the ecological evaluation of activities of sponge constituents is the lack of knowledge with regard to the localization of the respective compounds within the sponges. For an optimized defence against epibionts an exudation of sponge constituents capable of interfering with settlement and growth of epibionts in seawater surrounding the sponge should be expected. Only a few studies, however, have so far addressed the localization of natural products within sponges. The sponge Aplysina fistularis has been shown to release the biologically active metabolites serothionin (15, Fig. 1) and homoaerothionin (16, Fig. 1) into seawater at a rate sufficient to affect the behaviour and survival of several invertebrates capable of colonizing the sponge, such as hydroids, bryozoans and limpets (1~oamsox et al., 1985; Wai.~cER et al., 1985). Addition- ally it was found that mechanical injury ofthe sponge stimulated the release of aerothionin and homoaerothionin at rates 10-100 higher than controls (Wwt.xER et al., 1985). However, clearly more studies on the localization and possible exudation of sponge constituents have to be conducted before more generalized conclusions on the significance of natural products from sponges as a means to prevent fouling can be drawn.

Evidence for an induced chemical defence of the marine sponge Verongia aerophoba The sulphurous yellow-coloured sponge Verongia aerophoba, which is found frequently in the Mediterranean as well as around the Canary Islands is exceptionally rich in brominated compounds presumably derived from dibromotyrosine (T~Exnp~v~r et al., 1993a ; and references cited therein). Strikingly different patterns of brominated constitu- ents were obtained upon addition of McOH to fresh sponge (resulting in approximately 50% aqueous McOH due to the high water content of fresh sponge) or to freeze-dried sponge as revealed by HPLC analysis ('I~YnrANT and PRORSCH, 1993). Isofistularin-3 (17, Fig. 1) and aerophobin-2 (18, Fig. 1) were the dominating brominated constituents in extracts prepared from freeze-dried V. aerophoba, whereas extracts obtained following immersion of fresh sponge into solvent were characterized by increasing amounts of fistularin-1 (19, Fig. 1), aeroplysinin-1 (20, Fig. 1) and dienone (21, Fig. 1) (the latter compounds were almost completely missing in the extract of freezedried sponge) and declining amounts of aerophobin-2 and isofistularin-3. The pronounced differences between both extracts are the result of enzymatically catalysed biotransformations involving degradation of serophobin-2 (18) and isofistularin-3 (17) (via fistularin-1) to seroplysinin-1 (~, which in turn give rise to the dienone (21~ as shown in in vitro studies (T~t~wx~r et al., in preparation). The enzymatic conversion of precursors such as aerophobin-2 or isfistularin-3 to aeroplysinin-1 or dienone is of special interest with regard to the biological activities of the biotransformation products. Aeroplysinin-1 (20), as well as the dienone (21), show

650 P. PROKSCH

antibiotic activity against a plethora of Gram-positive and Gram-negative bacteria, whereas the substrates aerophobn-2 (18) and isofistularin-3 (17) are inactive (TEEYAPANT et al., 1993b). Furthermore, aeroplysinin-1 (20) and the dienone (21) showed pronounced cytotoxicity in vitro towards several human carcinoma cell lines with Ic y (inhibitory concentrations which reduce cellular division by 50% compared to controls) comparable to those of clinically used anticancer drugs such as cisplatin (TEEYAPANT et al., 1993b). Even though no bioassays have yet been performed with ecologically relevant epibionts, it is possible that the potent cytotoxicity of aeroplysinin-1 and of the dienone, as well as their antibacterial activity, form an obstacle to the colonization of V. aerophoba by fouling organisms. Chemical defence through an enzymatic conversion of performed biologically inactive storage compounds into active defence metabolites following disruption of cells (wounding) has been demonstrated, for example, for cyanogenic glycosides or glucosino- lates, both from higher terrestrial plants (LUCKNER, 1984). It is possible that V. aerophoba represents the first example of a similar defence mechanism in marine sponges.

Origin of sponge metabolites Even though mârine sponges accumulate a plethora of bioactive natural products, it is by no means clear whether all of these compounds are also synthesized by sponges, or whether they originate from other sources such as endosymbionts or filtered food particles. Detailed studies on the origin of sponge metabolites are difficult, since axenic cultivation of marine sponges has not yet been successfully established. Evidence for a sequestration of natural products by sponges either through the food chain or through endosymboontc microorganisms is therefore indirect in most cases and relies on chemical comparison of sponge constituents to those from bacteria or microalgae (KosnYnsxl and IsHIBasI-n, 1993). An example is provided by the polyether macrolides from the sponge Halichondria okadai that include halichondrin-B (22, Fig. 1) which shows powerful cytotoxic properties against human cells. The macrolides have been suggested to originate from bacteria or microalgae (possibly endosymbionts) due to their striking structural similarities to other macrolides such as okadaic acid (4) or ciguatoxin (23, Fig. 1) that are synthesized by dinoflagellates (SHIMIZU, 1993). More direct evidence for an endosymbiotic origin of sponge metabolites has been obtained for the sponge Tedania ignis, which is the source of several diketopiperazines (e.g. _24, Fig. 1) . From specimens of T. ignis from Bermuda several undescribed species of bacteria identified as members of the genus Micrococcus have been isolated. When grown in batch cultures in the laboratory the bacteria were found to produce the same diketopiperazines as present in the sponge, proving that the respective compounds are indeed bacterial metabolites ($TIERLE et al., 1988). Isolation and in vitro cultivation of endosymbionts seem to be promising for closer insights into the origin of sponge constituents in the future (KOBAYASHI and Isl~nst-u, 1993).

BIOLOGY OF SPONGE-FEEDING NUDIBRANCHS Nudibranchs are exclusively marine, slug-like invertebrates in the molluscan subclass Opisthobranchia, which have undergone complete detorsion of the visceral mass and lost the protective shell characteristic of other gastropods (KnRUSO, 1987). Nudibranchs are among the most vividly coloured organisms found in the sea. Even though they possess no protective shell and seem to be easy prey for fish or other predators such as crabs, reports of predatoon on nudibranchs are virtually nonexistent (except by other nudi-

Defensive Roles for Secondary Metabolites 651

branchs) (KARUSO, 1987; FAULKNER, 1992). Rejection ofnudibranchs by predatory fish (or other predators) is due to various elaborate defence mechanisms employed by these molluscs. Several nudibranchs including members of the genera Pleurobranchus and Philine produce sulfuric acid which is released from acid glands when the molluscs are molested (TxoMrsorr, 1983). Other nudibranchs are specialized to feed on coelenterates and incorporate intact nematocysts (stinging cells) from their prey (THOMP50N and BROWN, 1984). These stinging cells are transported through the gut of the nudibranchs to cnidosacs at the cerata (wing-like appendices of the mantle) where they are stored intact and released upon touch (the mechanisms by which nematocysts are transported through the body of the nudibranchs without firing are unknown). Other nudibranchs feed selectively on invertebrates, thereby sequestering the el%ctive arsenal of chemical weapons from their prey.

Sequestration of bioactive sponge metabolites by nudibranchs Of the few sponge-feeding organisms known from the marine environment nudibranchs are among the most important. In contrast to other potential sponge feeders, such as specialized fish, the sponge-feeding nudibranchs are apparently not only immune to hazardous sponge constituents but also able to sequester natural products from their sponge diet and to employ them for their own chemical defence (KARUSO, 1987; ScxEUat, 1990; FAULKNER, 1992). A vivid example was provided for the Phyllidia varicosa in one of the pioneering studies on sponge-feeding nudibranchs. An early report from 1963 (JoxnxNFS) indicated that crustaceans kept in an aquarium with specimens of P. varicosa died within 30 min, due to the poisonous mucus exuded by the nudibranch which proved also toxic to fish. Attempts to isolate and identify the toxin from specimens of P. varicosa from Hawaii, however, failed due to the small number of nudibranchs available. By chance, a specimen of P. varicosa was observed feeding on a sponge of the genus Ciocalypta from which the toxin 9-isocyan- opupukeanane (25, Fig. 1) and an isomer were isolated and identified (BURRESON et al., 1975). The nudibranch apparently had sequestered the toxin from its food source. Numerous studies on sponge-feeding nudibranchs that succeeded the pioneering inves- tigation on P. varicosa clearly established a dietary origin for most of the natural products found in the molluscs (KARUSO, 1987; FAULKNER, 1992). Some ofthese comparative studies indicated a selective uptake of sponge metabolites by the molluscs. The nudibranch Cadlina luteomarginata from California yielded a complex series of sesquiterpene and sesterterpene furans (_20_28, Fig. 1) as well as sesquiterpene isonitriles, -isothiocyanates, and -formamide (e.g. 29-_31, Fig. 1) (WALKER, 1981, Ph.D. dissertation, University of California, San Diego; THOIKPSON et al., 1982). All of the latter compounds that were stored exclusively in the dorsal mantle of the molluscs could be traced to sponges (such as Axinella spp.) found in the vicinity of the molluscs. However, it was apparent that sponge constituents were selectively sequestered by the nudibranch since sometimes only one or two com- pounds from an array of similar metabolites produced by the sponge were found in the nudibranch. In laboratory bioassays all of the sequestered sponge constituents inhibited feeding of fish (Carassius auratus). Of the various compounds present in the nudibranch the `natural mixture' of isothiocyanates (as occumng in the nudibranchs) was most active, inhibiting feeding of goldfish in an aquarium bioassay at a concentration of l0 kg/mg.

652 P. PROKSCH

De novo synthesis of natural products by nudibranchs Even though a dietary source could be traced to most of the secondary compounds isolated from nudibranchs, there are exceptions to the rule. The nudibranch Dendrodoris limbata from the Mediterranean is able to synthesize various drimane sesquiterpenoids including olepupuane (32, Fig. 1), as shown by feeding experiments with '4 C mevalonic acid (CIMINO et al., 1983). Within the molluscs, olepupuane (32) is confined exclusively to external tissues (border of the mantle), whereas other drimane esters such as 7-deacetoxy- olepupuane (33, Fig. 1) are present in the gills or in the egg masses (AVILA et al., 1991). The dialdehyde polygodial (34, Fig. 1) previously isolated from specimens of D. limbata is suspected to arise from olepupuane (32) (by enzymatic transformation?) upon disrupture of mantle cells. Polygodial (34) acts as a strong feeding deterrent to fish, presumably due to the interaction of the aldehydic groups of primary amines present in target molecules (AVILA et al., 1991). Due to its reactivity, however, polygodial (34) is also toxic to the nudibranch. The suspected wounding-related transformation of olepupuane (32) into polygodial (~ (which resembles the induced bioactivation of brominated metabolites in the sponge Yerongia aerophoba) would not only result in a highly effective chemical defence precisely at the site of attack, but would furthermore limit the nudibranch's risk of intoxication by its own metabolites. It is interesting to note that polygodial (34) is not restricted in its natural distribution to molluscs such as D. limbata, but is also present in trees of the East African genus Warburgia (Canellaceae), where it serves as chemical defence against herbivorous insects such as Spodoptera exempta (Noctuidae) (NAKANISHI and Kueo, 1977).

Chemical defence and evolution of nudibranchs It has been proposed that the shell-less nudibranchs have evolved from shelled ancestral molluscs by a process that involved gradual loss of the shell and detorsion of the visceral mass (FAULKNER and GHISSLnv, 1983). Where does the chemical defence fit into this evolutionary scheme? Is the chemical defence a preadaptation that preceded the gradual loss of the protective shell as a necessary requirement for the origin of morphologically defenceless nudibranchs, or did chemical defence evolve after the shells had been reduced? A final answer to this intriguing evolutionary problem will probably never be found. However, the occurrence of bioactive secondary metabolites, such as limatulone (35, Fig. 1), in shelled molluscs (believed to be closer to the ancestral forms) such as Collisella limatula (PAWLIK et al., 1986), indicates that secondary defence metabolites may be of broad occurrence within the molluscs (including shelled species) and hence may represent an ancestral character that is likely to have originated prior to the more recent reduction of the shell. In evolutionary terms nudibranch species such as D. limbata, which employ de novo synthesized defence metabolites, may have advantages over those that rely on chemicals derived from prey (e.g. sponges), since the former will be able to utilize food resources irrespective of a sequestration of defence metabolites, which is a limiting factor for the latter.

CONCLUSION The long evolutionary history of sponges, as well as their circumpolar distribution, demonstrates the successful adaptation of this ancient group of organisms. Even though our knowledge on the ecological roles ofsponge metabolites is still limited, the present data

Defensive Roles for Secondary Metabolites 653

indicate that the remarkable success of sponges is at least partially due to their impressive chemical armoury which apparently protects these invertebrates from numerous environ- mental dangers. The efficiency of the chemical defence of sponges is highlighted by the sequestration of these secondary chemicals by specialized sponge-feeding nudibranchs which utilize sponge-derived compounds for their own defence against predators. It is, however, still unclear whether secondary compounds isolated from sponges are also synthesized by these invertebrates or whether they are derived through filter-feeding or from endosymbionts or originate through `mixed biosynthesis' involving, for example, biotransformation of extraneous precursors within the sponge. These fundamental, as well as applied, aspects (for example, the possible production of secondary compounds of biomedical importance through endosymbionts grown in vitro) will offer new and exciting research avenues for bioactive compounds from marine sponges.

Acknowledgements-Financial support of our study on the brominated constituents of Verongia oerophoba through a grant of the DFG (Schwerpunkt `Chemische Ökologie') is gratefully acknowledged. Further thanks are due to R. EHSt, (Wiuzburg) for the drawing of the figure.

REFERENCES Avae, C., Ct~to, G., CAtsraro, A. and Srtrm r e, A. (1991) Drimane sesquiterpenoida in Mediterranean Dendrodoris nudibranchs: anatomical distribution and biological role. Experientia 47, 307-310. Betcus, G. J. (1981) Chemical defense mechanisms on the Great Barrier Reef, Australia. Science 211, 49799. Betzus, G. J. and GxEa;tv, G. (1974) Toxicity in sponges and holothurians: a geographic pattern. Science 1~, 951-953. Hexus, G. J., Scxut.rt:, B., Wxtoxz, M., Ga~x, G. and Goa~z, P. (1990) Antibiosis and antifouling in marine spon~s: laboratoryversus field studip. In: New Perspective In Sponge Biology, pp. 102-108 (Rürzr ne, K., Ed.). Washington: Smithsonian Institution Press. Ht~rscx, K. voN (1980) Drugs from marine organisms. Naturwissensehajten 67, 338-342. Boaznv, S. C. and Feut..t riEtt, D. J. (1992) Chemistry and chemical ecology of the Bahamian sponge Aplysilla glacialis. J. them. Ecol. 18, 309-332. BAe~xex, J. C. and Det.oza, D. (1986) Chemical defence in sponges. Pure Appl. Chcm. S8, 357-364. Buxn;xota~t, P. R. (1973) Ecology of marine antibiotics and coral reefs. In: Biology andGeology of Coral Reefs, II, pp. 117-I82 (JOxFS, O. A. and ErroFwtv, R., Eds). New York: Academic Press. Buatteox, B. J., Cl,eanY, J., Fn~a, J. and Sc~, P. J. (1975) 9-Isocyanopupukeanane, a marine invertebrate allomone with a new seaquiterpene skeleton. J. Am. them. Soc. 97, 4763764. Cet, M. and Joauex, P. C. (1990) How does vestibule surface charge affect ion conduction and toxin binding in a sodium channel? Blophys . J. S7, 88391. Ci~to, G., De Rose, S., DE S~exo, 3., SoneNO, G. and Vn.uxt, G. (1983) Dorid nudibranch elaborates its own chemical defense. Science 219, 1237-1238. C~vs, W. A. and Wu av, G. V. (1946) The fishes of the Pacific coast of Canada. Bull. Fuh . Res. Bd Can. 68, 1-368. Devts,A. R., TexaErr, N. M., McCoxxEr-r.,O. J. and Yourro, C. M. (1989) Epibiosis of marine algae and benthic invertebrates: natural products chemistry and other mechanisms inhibiting settlement and overgrowth . In: Bioorganic Marine Chemistry, pp. 8 114 (Scrums, P. J., Ed.). Berlin: Springer. Feu~xraae, D. J. and Gtr, M. T. (1983) Chemical defense and evolutionary ecology of dorid nudibranchs and some other opiathobranch gastropods. Mar. Ecol. Progr. Ser. 13, 29301. Feut.w~t, D. J. (1992) Chemical defenses of marine molluscs . In: Ecological Roles ofMarine Natural Products, pp. 119-163 (PAUL, V. J., Ed.). Ithaca: Comstock. Feutxt~rate, D. J. (1993) Marine natural products. Nat. Prod. Rep. 9, 323-539. (Preceding reviews by same author.) Gtv?Etv, G. (1977) Ewlogy of toxicity in marine sponges. Mar. Biol. 40, 207-215. G~rr, G., Go~z, P. and Beaus, G. J. (1990) Antimiaobial and ichthyotoxic properties ofmarine sponges from Mexican waters . In: Ntw Perspectives in Sponge Biology, pp. 109-114 (Rt7xzt~tt, K., Ed.). Washington: Smithsonian Institution Press. GAOWHSa, A., Sts U. and KesEn~tet~, Y. (1983) Marine toxins of l atrunculia magnlfica. J. org. Chcm. 48, 3512-3516. HAYSrEAD, T. A. J., Sna, A. T. R., Ceat.ttva, D., Hotvivox, R. C., Tsuxtrext, Y., Coax, P. and Beam, D. (1989) Effects of the tumor promoter okadaic acid on intracellular phosphorylation and metabolism . Nature 337, 78~1 .

654 P. PROKSCH

IREtaNO, C. M., COPP, B. R., FosrER, M. P., McDoxAt,n, L. A., RADISKY, D. C. and SWEasEV, J. C. (1993) Biomedical potential of marine natural products. In : Marine Biotechnology, pp. 1-43 (ATTAWAY, D. H. and ZAeoRStcv, O. R., Eds). New York: Plentun Press. JACtcsox, J. B. C. and Buss, L. (1975) Allelopathy and spatial competition among coral reef invertebrates . Proc. natn . Acad. Sci. U.S.A. ?2, 51605163. JOHANNES, R. E. (1963) A poison-secreting nudibrattch (Molhtsca :Opisthobranchia) . Veliger S, 104-105 . KARUSO, P. (1987) Chemical ecology ofthe nudibranchs. In : Bioorganic Marine Chemistry, I, pp. 320 (Sct~uER, P. J., Ed.). Berlin : Springer . KwsttMwx, Y., GROwetss, A. and Srtt~tueu, U. (1980) Latrunculin, a new 2-thiazolidinone macrolide from the marine sponge Latrunculia magn~ca. Tetrahedron Lett . 21, 3629-3632 . Kotü+vASttt, J. and Ist-ne~stu, M . (1993) Bioactive metabolites of symbiotic marine microorganisms . Chem . Rev. 93, 1753-1769. LuctctvER, M. (1984) Secondary Metabolism in Microorganisms, Plants and Animals. Berlin : Springer. McCAtFttEV, E. J. and Ext)EAN, R. (1985) Antimicrobial activity of tropical and subtropical sponges. Mar. Biol. 89, 1~. MEas, D. (1989) Gijte im Riff Stuttgart : WissenschaRliche Verlagsgesellschaft . MULLER, W. E. G. (1982) Cell membranes in sponges. Ins. Reu. Cytol. 77, 129-181 . MUNRO, M. H. G., LUIBRAND, R. T. and Bt.ux'r, J. W. (1987) The search for antiviral and anticancer compounds from marine organisms. In: Bioorganic Marine Chemistry, pp. 93-185 (Sct~ttER, P. J., Ed.). Berlin: Springer. NAtcANtsxt, K. and Kuso, I. (1977) Studies on warburganal, muzigadial and related compounds . Israel J. Chem . 16, 28-31 . N~~tAtv, L, Ftst~tsox, L. and KASxntAN, Y. (1975) Isolation ofa new toxin from the sponge l atrunculia magnifica in the gulf of Aquaba (Red Sea). Mar. Biol. 30, 293-296. PAUL, V. J. (1992) Chemical defenses of benthic marine invertebrates. In: Ecological Roles of Marine Natural Products, pp. 164-188 (PAUL, V. J., Ed.). Ithaca: Comstock . P~wtatc, J. R. (1993) Marine invertebrate chemical defenses . Chem. Reu. 93, 19l 1-1922. PAWLIK, J. R., ALetzArt, K. F. and FAULKNER, D. J. (1986) Evidence of a defensive role for limatulone, a novel triterpene from the intertidal limpet Collisella limatula. Mar. Ecol . Progr. Ser . 30, 251-260 . PORTER, J. W. and TARasrr, N. M. (1988) Allelochemical interactions between sponges and corals . Biol . Bul/. 175, 230-239 . RANDALL, J. E. and HARrMAxN, W. D. (1968) Sponge-feeding fishes of the West Indies . Marine Biol . 1, 216-225. RtxEttARr, K. R., $WELD, L. S. and COI~1-PARSONS, M. (1993) Antiviral substances. In : Marine Biotechnology I, pp. 309-342 (ATTAWAY, D. H. and ZAnORSKY, O. R., Eds) . New York : Plenum Press. SALE, P. F., Ed. (1991) The Ecology of Fishes on Coral Reejs. San Diego: Academic Press. $ARMA, A. S., DAUA1, T. and MüLt .ER, W. E. G. (1993) Secondary Metabolites from Marine Sponges . Berlin : Ullstein Mosby. Sct~uER, P. J. (1990) Some marine ecological phenomena: chemical basis and biomedical phenomena. Science 248, 173-177 . ScFtMtrz, F. J., BOWDEN, B. F. and TorH, S. I. (1993) Antitumor and cytotoxic compounds from marine organisms . In : Marine Biotechnology, I, pp. 197-308 (ATTAWAY, D. H. and Z.AeoRStcv, O. R., Eds) . New York: Plenum Press. ScoTT, T. D. (1962) The Marine and Fresh Water Fishes of South Australia . Adelaide : W. L. Hames. StnMtzu, Y. (1993) Dinoflagellates as sources of bioactive molecules . In: Marine Biotechnology, I, pp . 39I~10 (ATTAWAY, D. H. and ZAHORSKY, O. R., Eds) . New York : Plenum Press. STAHL, E. (1904) Die Schutzmittel der Flechten gegen Tierfraß . In : Festschrift zum 70. Geburtstag von Ernst Haeckel, pp. 357-375 . Jena: G. Fischer. STARCR, W. A. (1968) A list offishes ofAlligator Reef, Florida with comments on the nature of the Florida reef fish fauna. Undersea Biol. 1, 1~0. Srmet.E, A. C., CARDELLINA, J. H. II and Stxot.erox, F. L. (1988) A marine Micrococcus produces metabolites ascribed to the sponge Tedania ignis. Experientia 44, 1021 . SULLIVAN, B., DtuRA, P., McIxrvRE, D. E. and FAULKNER, D. J. (1981) Antimiaobial constituents of the sponge Siphonodictyon coralliphagum . Tetrahedron 37, 979-982 . SULLIVAN, B., FAULKNER, D. J. and WERe, L. (1983) Siphonodictidine, a metabolite of the burrowing sponge Siphonodityon sp. that inhibits coral growth. Science 221, 1175-1176 . TEEYAPANT, R. and PROICSCH, P. (1993) Hiotransformation of brominated compounds in the marine sponge Verongia aerophoba : evidence for an induced chemical defense? Natwwissenschaften 80, 369-370 . TEEYAPANT, R., KRns, P., WRAV, V., WITTE, L. and PROICSCH, P. (1993a) Brominated secondary compounds from the marine sponge Verongia aerophoba and the sponge feeding gastropod Tylodina peruersa . Z. Naturforsch. 48c, 640-644 . TPEYAPANi, R., WoERDExeeo, H. J., KRI73, P., HACKER, J., WRAV, V., WrrrE, L. and PROKSCtI, P. (19936) Antibiotic and cytotoxic activity of brominated compounds from the marine sponge Verongia aerophoba . Z. Natwforsch. 48c, 939J945. THOMPSON, T. E. (1983) Detection of epithelial acid secretions in marine molluscs : review of techniques, and new analytical methods. Comp . Biochem. Physiol. A 74, 61521 .

Defensive Roles for Secondary Metabolites 655

THOMPSON, J. E. (1985) Exudation ofbiologically-active metabolites in the sponge Aplysinafrstularis. I. Biological evidence . Mar. Biol. 88, 23-26. Txo~r'sox, T. E. and BROWN, G. H. (1984) Biology of Opisthobranch Molluscs, II, pp. 1-224. London: The Ray Society. THOMPSON, J. E., WALKER, R. P., Wiurt~v, S. J. and FAULKNER, D. J. (1982) A chemical defense mechanism for the nudibranch Cadlirra luteomarginata . Tetrahedron 38, 186.1873. Txoe~soN, J. E., WALKER, R. P. and FAULKNER, D. J. (1985) Screening and bioassays for biologically active substances from forty marine sponge species from San Diego, California, USA. Mar. Biol. 88, 11-21. WALKER, R. P., TxoMesox, J. E. and FAULKNER, D. J. (1985) Exudation of biologically-active metabolites in the sponge Aplysina frstularis. II. Chemical evidence. Mar. Biol. 88, 27-32. ZIJKAL, H. (1895) Morphologische and biologische Untersuchungen ûber die Flechten. Sber. K Böhm. Wiss. Math.-Nat. Kl. 104, 1303-1395.