Chemical defenses of the tropical green seaweed Neomeris annulata Dickie: effects of multiple compounds on feeding by herbivores

JOURNAL OF ExPERtMENTAL MARINE BIOLOGY AND ECOLOGY

Journal of Experimental Marine Biology and Ecology, 201 (1996) 185-195

ELSEVIER

Chemical defenses of the tropical green seaweed Neomeris anndata Dickie: effects of multiple compounds on feeding by herbivores Wilfred University

Received

A. Lumbang,

Valerie

of Guam Marine Laboratory, Man&o,

6 July 1995; revised

1 November

J. Paul* Guam 96923, USA

1995; accepted

29 November

1995

Abstract The calcareous green alga Neomeris annulata Dickie is widely distributed in shallow inshore waters of many reef habitats in the Caribbean and Pacific. We examined the effects of three brominated sesquiterpenes found in Neomeris annulata on Guam against three reef herbivores that are not strongly deterred by CaCO, in their diets: the parrotfishes Scarus sordidus Forsskal, S. schlegeli (Bleeker) and the sea urchin Diadema savignyi Michelin. The effects of each metabolite were examined separately to compare their effectiveness toward the different herbivores. The major brominated sesquiterpene (compound 1: ca. 1% dry mass in thallus tips) deterred feeding by all three herbivores at or below natural concentrations, as did the other two minor compounds. The terpene most effective at deterring feeding by all three herbivores was a closely related minor compound (compound 2: ca. 0.5% dry mass in thallus tips). We then tested for possible interactions among the three metabolites and synergistic effects by seeing whether natural combinations of the three metabolites enhanced their effectiveness as chemical defenses. The natural mixture of metabolites was never more effective than either compound 1 or compound 2 alone, suggesting no enhanced or synergistic effect of multiple metabolites. In fact, at higher concentrations approaching natural concentrations found in the thallus tips, compound 2 was more effective at deterring feeding by S. schlegeli by itself than in the naturally occurring mixture of terpenes. Keywords:

Neomeris

annulatu;

Brominated

sesquiterpene;

Chemical

defense;

Coral-reef

herbi-

vores; Synergism

1. Introduction Secondary *Corresponding

metabolites

of many tropical seaweeds are effective defenses against.coral

author.

0022.0981/96/$15.00

0

1996 Elsevier Science B.V. All rights reserved

S.SD! 0022-0981(95)00181-6

186

W.A. Lumbang, V.J. Puul I .I. Exp. Mar. Bid. Ed.

201 (19%)

18.5%19_5

reef herbivores (reviewed by Hay and Fenical, 1988; Hay, 1991, 1992; Hay and Steinberg, 1992; Paul, 1992). Individual metabolites can differ in their effectiveness against different reef herbivores, and the defensive function of any compound is highly dependent on its structure, which may affect the way different herbivores perceive it (Hay, 1991; Hay and Steinberg, 1992; Paul, 1992). Even closely related compounds can have quite different effects on consumers (Hay, 1991; Hay and Steinberg, 1992; Meyer et al., 1994). Tropical seaweeds often produce a variety of secondary metabolites and combine chemical and structural defenses to effectively defend against herbivory. These combinations of defenses may be necessary for macroalgae to survive on coral reefs, because shallow coral reefs are among the world’s most heavily grazed habitats (Hay and Steinberg, 1992). Multiple defenses could act additively or synergistically to reduce the ability of herbivores to adapt to seaweed defenses. Combinations of CaCO, and secondary metabolites have been tested as feeding deterrents and both additive (Pennings and Paul, 1992; Schupp and Paul, 1994; Meyer and Paul, 1995) and synergistic (Hay et al., 1994) effects of multiple defenses have been observed. Most studies of seaweed chemical defenses have concentrated on testing the effects of individual metabolites on herbivores; however, many seaweeds produce several structurally related compounds (as do terrestrial plants). Hay (1992) suggests that complex interactions among multiple metabolites and synergistic effects are likely to occur in seaweed-herbivore interactions. To the best of our knowledge, while the combined effects of CaCO, and secondary metabolites have been examined, the combined effects and possible interactions of multiple secondary metabolites have not been experimentally addressed in marine plant-herbivore interactions. Berenbaum (1985) reviews complex interactions known to occur among secondary metabolites in terrestrial plants. Of particular relevance to this study is the potential for analog synergism to occur with mixtures of related metabolites (McKey, 1979; Berenbaum, 1985) whereby slight changes in the structures of related chemicals may produce compounds with different physiological properties (McKey, 1979). For instance, slight structural variation might change a compound to an inhibitor of an enzyme that would otherwise metabolize it. However, only a few examples of synergistic interactions of structurally related plant compounds exist to support the hypothesis of analog synergism (Berenbaum, 1985; Berenbaum et al., 1991; Berenbaum and Zangerl, 1993). The calcareous green alga Neomeris annulata (Dickie) is widely distributed in shallow inshore waters of many reef habitats in the Caribbean and Pacific. Previous studies of the natural products chemistry of N. annulata collected in Bermuda (Barnekow et al., 1989) and Guam and Kwajalein (Paul et al., 1993) show that this species at both locations produces a variety of structurally related brominated sesquiterpenes. The compounds exhibit a variety of biological activities including toxicity to brine shrimp (Barnekow et al., 1989) and the ability to deter feeding by natural assemblages of reef fishes (Paul et al., 1993). Paul et al. (1993) suggested that N. anndata may be particularly well adapted to herbivory because it produces a mixture of several different secondary metabolites, the types of compounds vary across its geographic range, and the alga combines chemical and structural defenses to protect against the variety of herbivores found on coral reefs.

W.A. Lumbang,

VI. Paul I J. Exp. Mar. Bid.

Ed.

201 (1996) 185-195

187

In this study, we examine the effects of three brominated sesquiterpenes found in Neomeris annulatu on Guam against three reef herbivores that are not strongly deterred by CaCO, in their diets: the parrotfishes Scarus sordidus Forsskdl and S. schlegeli (Bleeker) (Schupp and Paul, 1994) and the sea urchin Diudemu suvignyi Michelin (Pennings and Svedberg, 1993). Thus, these herbivores would be more likely to consume calcified seaweeds than some surgeonfishes and rabbitfishes that avoid CaCO, in their diets (Schupp and Paul, 1994). We examine the effects of each metabolite separately to compare their effectiveness toward the different herbivores. Then, as a test for analog synergism, we determine whether natural combinations of the three terpenes enhance their effectiveness as chemical defenses.

2. Methods 2.1. Isolation

of Neomeris

terpenes

Collections of Neomeris unnulutu were made in Agat Bay, Guam (13”25’ N, 144”55’ E) in shallow reef flat habitats (l-2 m depth) as previously described (Paul et al., 1993). The alga was extracted immediately after collection by grinding in a blender with 1: 1 dichloromethane:methanol. After three extractions of the ground algae, the combined solvents were removed by rotary evaporator. Crude extracts were fractionated by vacuum flash silica gel column chromatography with hexanes: ethyl acetate mixtures. Final purification of the Neomeris compounds was accomplished by normal-phase (silica gel) high-performance liquid chromatography (HPLC) of the nonpolar flash column fractions with hexanes:ethyl acetate (9: 1). 2.2. Study organisms Juvenile parrotfishes Scurus schlegeli and S. sordidus (total length 7-10 cm) were captured by barrier net in a shallow bay (l-3 m) adjacent to the seaplane ramp in Apra Harbor, Guam. They were placed in a seawater-filled, aerated cooler for transport to the University of Guam Marine Laboratory where they were transferred into 20 1 flowthrough seawater tanks. Each tank contained two fish of a single species. The two species were distinguished by coloration and shape of the head and held separately. Parrotfishes were caught and fed artificial food for l-3 days prior to the feeding experiments; this allowed enough time for them to begin feeding on artificial food strips. Diudemu suvignyi was collected on the reef flat of Pago Bay (O-l m depth) directly in front of the Marine Laboratory, and held individually in 46 1 plastic tanks. They were used in experiments within 1 wk of collection, after they began feeding on artificial foods. 2.3. Food preparation Parrotfishes and urchins were offered artificial foods similar to those used by Hay et al. ( 1994) to test secondary metabolites from other calcified green algae. We used finely powdered, freeze-dried Enteromorphu cluthrutu (Roth), a green alga that is highly

188

W.A. Lumbang,

VJ. Paul / J. Exp. Mur. Bid

Ed.

201 (1996) 185-19.7

palatable to most herbivores, in an agar base, and this food was molded onto fiberglass window screen that provided a grid for quantifying the amount of food eaten (Hay et al., 1994). Neomeris terpenes were added to the Enteromorpha by placing the freeze-dried Enteromorpha in a small round-bottom flask, adding the appropriate amount of Neomeris metabolites dissolved in approximately 10 ml of diethyl ether (enough to cover all of the Enteromorpha in the flask), and then removing the ether by rotary evaporator to yield dry, powdered Enteromorpha that was evenly coated with compounds. Diethyl ether alone (without metabolites) was added to the control Enteromorpha and evaporated in the same way. The food mixture consisted of 0.72 g agar and 36 ml water stirred together in a beaker and heated in a microwave until boiling (30-45 s). After the agar-water mixture began to cool, 4 g of the rotary-evaporated Enteromorpha was thoroughly stirred into the agar mixture, and the food mixture was poured into a mold on top of the window screen. The mold was based on the design of Hay et al. (1994), and was made from a rectangular piece of 2 mm thick formica with two 2.63 X 25 cm openings cut in it. For our assays, a control food without secondary metabolites was usually poured into one opening of the mold and a treatment food with pure Neomeris terpenes or mixtures of terpenes was poured into the other opening. Neomeris metabolites were tested at a range of natural concentrations occurring throughout the algal thalli (0.2-1.0% dry mass, Fig. 1; see also Meyer and Paul, 1995). In a few experiments we compared two different mixtures of Neomeris terpenes with each other, so both sides of the molds contained food with added compounds. The agar quickly cooled, and the food assumed the shape of the mold and became attached to the window screen. The molds were removed and the screens

mean

Compound 2

+-

SD (% dry mass)

tips: middles: bases:

.95 -+ .16 .36 + .07 .ll * .Ol

tips: middles: bases:

.56 * .37 .22 + 46 .06 k .02

not quantitied but approximately 113 of compound 2 concentration

Fig. 1. Structures and natural concentrations in different parts of the thalli for the three Neomeris sesquiterpenes. Data on natural concentrations are from Meyer and Paul (1995).

annulara

W.A. Lumbang,

Ki. Paul I J. Exp. Mar. Bid.

Ed.

201 (1996) 18%19-T

189

were cut so that a length of screen contained both treatment and control foods (both sides of the mold). Food strips used for the assays measured 17 X 15 square openings in the screen for both treatment and control foods. 2.4. Feeding assays The parrotfish or urchins in each tank were offered a food strip containing treatment and control foods and always had a choice between the two food types. For each species of herbivore, there were 12-19 tanks holding two Scarus sordidus, two S. schlegeli, or individual Diadema savignyi. Food strips were monitored periodically during the assays, and the strips were removed when one-third or more of the total food was consumed or when the experiment was terminated (usually after 24 h). The duration of experiments was l-24 h. If the fish or urchins in a replicate did not feed, or consumed all of both foods, the assay was repeated for those replicates or the replicate was excluded from the data analysis. The same fish and urchins were used for different assays; however, only one assay was conducted per day and fish and urchins were fed plain artificial food in between assays. It is possible that the herbivores became sensitized to the Neomeris compounds after repeated exposures, but the time involved in catching and acclimatizing new animals made it impractical to catch parrotfish and urchins for each assay. Assays with parrotfish were conducted during the day between 1000 and 1800; urchin assays were conducted in the evenings because Diadema did not feed during the day. The amounts of control and treated food eaten at the end of the assay were determined by counting the number of squares of the screen that did not have artificial food covering them (artificial food had been completely removed by the parrotfish or urchins). No-herbivore controls (controls for changes unrelated to consumption, Peterson and Renaud, 1989) were unnecessary because in the absence of herbivores there was no change in the number of squares covered by the artificial food. Based on our observations, the food did not come off of the screens in seawater unless herbivores consumed it. At least once for each compound tested, treated artificial food that was not consumed during the assays was extracted with dichloromethane at the end of the assays and analyzed by thin layer chromatography (TLC) to ensure that the test compound was still present and had not decomposed by the end of the assays. No decomposition of the metabolites was ever observed by TLC. Differences between number of squares consumed for control and treated foods were analyzed with a paired f-test after testing to assure that assumptions of normality were met. Wilk-Shapiro/Rankit Plots (Statistix 4.0) were used to test the differences for normality. All data met the assumption of normality.

3. Results Compounds 1 and 2, the two most abundant secondary metabolites in Neomeris annulata on Guam, deterred feeding by all three species of herbivores at concentrations below those found in the tips of the thalli (Figs. 1,2). The major metabolite, compound

W.A. Lumbang,

190 A.

VJ.

Paul

I J. hp.

Mar.

Bid.

Ed.

XII

(1996)

C.DIADEM4 SAVIGNYI

B. SCARUS SORDIDUS

SCARUS SCIiLEGELI

18.5%19-5

P
E.SCIRUS SORDIDUS

D.SCARUS SCIILECELI

F. 200

DIADBMA SAVICNYI

p-0.023 t N-14 T

P-O.035 p-0.374 N-9 N-13

160

1W

120 60

60

40 40

20 600 I-lI._L

G.

SCARUS SCHLIWELI

0.5x

0

0.2%

KSCARUS SORDIDUS

140

60

120

70

100

60

P-O.011 N-17

0.5x

I.

0.3%

0.2%

DUDEMA SAVICNYI

= 'N-Y" 1

50

66

40

60

3a

40

20

20

10

0 1

0.3x

0

CONTROL

0 i

0.2x

q

CMPD 1

0.45n

0.2%

n

Fig. 2. Results of laboratory feeding trials testing the effects of individually at different concentrations on feeding by the parrotfishes urchin Diadema savignyi. Histogram bars represent means with error replicate tanks with parrotfishes tested in pairs or sea urchins tested paired t-test (two-tailed).

1, significantly

deterred

feeding

by Scat-us schlegeli

CMPD2

CMPD 3

three Neomeris unnukura sesquiterpenes Scarus x/&&i, S. sordidus and the sea bars representing + I SE. N = number of individually. Data were analyzed with a

at 0.5% dry mass (Fig. 2A), by

Scarus sordidus at 1% dry mass (Fig. 2B), and by the urchin Diadema savignyi at 0.5%

dry mass (Fig. 2C). Compound 2 was the most effective of the three metabolites and significantly deterred feeding by the two parrotfishes at 0.2% dry mass (Fig. 2D,E), a relatively low concentration found in the middle of Neomeris thalli (Fig. 1, Meyer and Paul, 1995). Sea urchins were not deterred from feeding by compound 2 at 0.2%, but were strongly deterred at 0.3% dry mass (Fig. 2F). Compound 3, a minor metabolite present at only one-sixth the concentration of compound 1 and one-third the concentration of compound 2 (or < 0.2% dry mass) did not deter feeding by any of the herbivores at natural concentrations (Fig. 2G-I). However, it was as effective at

W.A. Lumbang, KJ. Paul I J. Exp. Mar. Biol. Ecol. 201 (1996)

185-195

191

deterring feeding by the parrotfishes as compound 1, and it significantly deterred them at concentrations of 0.3-0.5% dry mass (Fig. 2G,H). The natural mixture of compounds l-3 (ratio 6:3:1) was also effective at deterring feeding by the parrotfishes at 0.3% dry mass (Fig. 3A,B), which is below natural concentrations of 0.6-1.5% for the middles and tips of thalli (Fig. 1; Meyer and Paul, 1995). The urchins did not avoid the natural mixture at concentrations as high as 0.5% (Fig. 3C), a concentration at which, individually, both compounds 1 and 2 significantly deterred feeding. When pure compounds were compared directly with the mixture at equal con-

A. SCAR.27

SCHLEGELI

B. SCARLJS SORDIDUS

C. DIADEM-A SAVIGNYI

140 120 100 60 60 40 20 w z b 4

O

w

1

D. SCARUS

SCHLEGELI

E. SCARUS 60 50

4 3 Q v1 L 0

100

C

SORDZDUS

P=O.339

40

60

30

60 40

20

20

10

0

0

0.5%

a

0.3%

CMPD

1

CYPD

2

0.3%

ml

F. SCARUS

m 3 3 z

SCHLEGELI

G. SCARUS

160

60

140

70

120

60

100

50

60

40

60

30

40

20

20 0 I----l

SORDIDUS

r

P=O.lll N=ll

T

10 r3.5%

0.2%

0

0.2%

Fig. 3. Results of laboratory feeding trials testing the effects of three Neomrris annulara sesquiterpenes as a natural mixture (6:3: I ratio of compound I :compound 2:compound 3) at different concentrations on feeding by the parrotfishes Scarus schlegefi, S. sordidus and the sea urchin Diadema swignyi. Results are presented for the mixture versus control food (A-C) and versus equal concentrations of compound 1 (D,E) and compound 2 (F,G). Histogram bars represent means with error bars representing + I SE. N = number of replicate tanks with parrotfishes tested in pairs or sea urchins tested individually. Data were analyzed with a paired r-test (two-tailed).

192

W.A.Lumbang,

V.J. Paul

/ J. Exp. Mar.

Bid.

Ed.

201 (1996)

185-19.5

centrations, the mixture was never more effective at deterring feeding by parrotfishes than the individual metabolites. Scarus schlegeli showed a nonsignificant trend toward preferring the mixture over the pure compound 1 at 0.5% dry mass (Fig. 3D). S. schlegeli and S. sordidus did not show any preference between the mixture or pure compound 1 at lower concentrations of 0.3% dry mass (Fig. 3D,E). Scarus schlegeli strongly preferred the mixture over pure compound 2 at a concentration of 0.5% (Fig. 3F), but neither parrotfish species showed a preference between compound 2 and the mixture at a concentration of 0.2% dry mass. We were unable to test the mixtures versus pure metabolites for the urchins because of lack of material; however, comparisons between Fig. 3C and Fig. 2C and Fig. 2F show that the urchins were significantly deterred by feeding on compound 1 at 0.5% dry mass and compound 2 at 0.3% dry mass, but not by feeding on the mixture at either concentration.

4. Discussion The three metabolites from Neomeris annulata were highly effective feeding deterrents against herbivores at concentrations equal to or below natural concentrations. We selected three species of herbivores that do not avoid CaCO, in their diets and will readily consume calcified foods (Pennings and Svedberg, 1993; Schupp and Paul, 1994); therefore, these Neomeris metabolites deter herbivores that are not affected by the mineral defenses of calcified algae. The common co-occurrence of CaCO, and secondary metabolites in tropical seaweeds has been previously reported (Paul and Hay, 1986) and suggested to be adaptive, because the high diversity of tropical herbivores limits the effectiveness of any single defense (Lubchenco and Gaines, 1981; Paul and Hay, 1986; Pennings and Paul, 1992; Hay et al., 1994; Schupp and Paul, 1994; Meyer and Paul, 1995). Halimeda diterpenes can limit feeding by parrotfishes that are not affected by the levels of CaCO, found in Halimeda spp. In contrast, the rabbitfish Siganus spinus and the surgeonfish Naso lituratus were deterred by CaCO, in their diets but were unaffected by Halimeda diterpenes (Schupp and Paul, 1994). Combined defenses (CaCO, and terpenes) increased the number of fish species that were deterred relative to either single defense. A synergistic effect between CaCO, and chemical defenses was demonstrated by Hay et al. (1994) for assays examining feeding behavior of different herbivores which were fed artificial diets containing metabolites of tropical green algae and natural concentrations of CaCO,. The synergistic effects were apparent for some combinations of algal metabolites and herbivores but not others, and the mechanisms producing these effects were not clear. The different species of herbivores varied in their sensitivities to Neomeris chemical defenses. Among the three species we tested, the parrotfishes were deterred by lower concentrations than the sea urchins for compounds 2 and 3 and the natural mixture of metabolites. For compound 1 and the natural mixture, Scarus schlegeli was more strongly affected than S. sordidus. In general, however, all three species we tested

WA. Lumbang, VJ. Paul / J. Exp. Mar. Bid Ed.

201 (1996) 185-19.5

193

responded to the compounds similarly. and were most strongly affected by the same metabolite, compound 2. While all three metabolites deterred feeding at natural concentrations, compound 2 was the most effective feeding deterrent for all three species of herbivores. Our results are consistent with results of field assays testing compounds 1 and 2 individually and in combination against natural assemblages of reef fishes (Meyer and Paul, 1995). In those field assays, compound 2 was a significant feeding deterrent while compound 1 was not. Combinations of the two metabolites were less effective than compound 2 alone. In laboratory assays with the surgeonfish Nuso lituratus, extracts and compound 1 alone significantly deterred feeding, whereas compound 2 and combinations of the two metabolites did not (Meyer et al., 1994). The effectiveness of compound 1 against N. liturutus and compound 2 against two species of parrotfishes suggests that the alga may produce multiple metabolites because these compounds are effective against different herbivores. Other algal metabolites have been shown to effectively deter feeding by some herbivores but not others (reviewed by Hay, 1992; Paul, 1992). Additionally, algal metabolites may serve multiple functions and also deter fouling organisms (Schmitt et al., 1995) or pathogens (reviewed for terrestrial plants by Langenheim, 1994). Natural combinations of the three metabolites did not show enhanced effectiveness as feeding deterrents compared to compound 1 or 2 alone (Fig. 3). In fact, at concentrations of 0.5% compound 2 was significantly more deterrent than the mixture and compound 1 showed a nonsignificant trend of being more deterrent than the mixture. Previous studies with Naso lituratus suggested that combinations of compounds 2 and 3 were more deterrent than compound 2 alone. However, the natural mixture of all three metabolites was not more effective than the individual compounds in deterring feeding by N. liturutus (Meyer et al., 1994). Mixtures have rarely been compared to pure metabolites in other studies of marine plant-herbivore interactions. Enhanced toxicity to structurally related chemical mixtures from terrestrial plants is known to occur (Berenbaum, 1985; Berenbaum and Neal, 1985; Berenbaum et al., 1991; Berenbaum and Zangerl, 1993). Analog synergism may account for the presence of related furanocoumarins in apiaceous plants (Berenbaum and Zangerl, 1993). Growth was significantly slower for larvae of the black swallowtail, Papilio polyxenes, a specialist on furanocoumarin-containing plants, when fed combinations of two widely co-occurring furanocoumarins, angelicin and xanthotoxin, than when fed either individual compound. In vitro studies of cytochrome ~450 metabolism showed that angelicin (an angular furanocoumarin) was metabolized more slowly than xanthotoxin and bergapten (linear furanocoumarins), and when all three compounds were assayed together, overall rates of metabolism were reduced (Berenbaum and Zangerl, 1993). Similarly, a natural mix of furanocoumarins present in parsnip seeds was more toxic to the generalist corn earworm, Heliothus zeu, than an equimolar amount of xanthotoxin (Berenbaum et al., 1991). One difference between the studies with furanocoumarins and our work with the Neomeris metabolites is that toxic effects of the furanocoumarins were observed after the compounds were ingested. It is possible that one reason we did not observe synergistic effects was that we studied feeding deterrence (a behavioral assay) rather than toxicity or other physiological effects. It may be inherently difficult to see synergistic effects when using consumer behavior alone to look for it. Also, if analog

194

W.A. Lumhang,

VJ. Paul

I J. Exp, Mar.

Hiol. Ecol.

201 (1996)

18%195

synergism relies on specific interactions between metabolites and certain enzymes or receptors, then we may not expect to observe it for metabolites that have general physiological effects, such as disruption of cell membranes or interaction with other major cellular components. Although we know little about the mode of action of the Neomeris terpenes as well as most other secondary metabolites, it is possible that analog synergism may only be relevant for certain groups of related metabolites. The actual mode of synergistic interaction has been elucidated in very few studies of analog synergism (Berenbaum et al., 1991). We chose herbivores that were not deterred by CaCO, and therefore would eat the highly calcified Neomeris and found that the alga was chemically defended against herbivores as diverse as fish and sea urchins. However, our study did not suggest that mixtures interacted synergistically to enhance feeding deterrence. Further studies of how concentrations of the three metabolites vary among populations of Neomeris exposed to different levels of grazing or other environmental factors should shed light on other possible roles for these terpenes.

Acknowledgments Financial support was provided by NSF (GER-90233 11) and NIH/MBRS (GM41796). We are grateful to K. Sakamoto, T. Pitlik and G. Paulay who caught fish and urchins. Discussions about synergism with M. Hay and A. Zangerl, discussion about algal biosynthesis with D. Nagle, and critical reading by M. Slattery and two anonymous reviewers greatly improved this manuscript. This is contribution No. 369 from the University of Guam Marine Laboratory.

References Bamekow, D.E., J.H. Cardellina II, AS. Zektzer and G.E. Martin, 1989. Novel cytotoxic and phytotoxic halogenated sesquiterpenes from the green alga Neomeris annulara. J. Am. Chem. Sot., Vol. 11 I, pp. 3511-3517. Berenbaum,

M., 1985. Brementown revisited: interactions among allelochemicals in plants. Rec. Adv. Vol. 19, pp. 139-169. Berenbaum, M. and J.J. Neal, 1985. Synergism between myristicin and xanthotoxin, a naturally cooccurring plant toxicant. J. Chem. Ecol., Vol. 11, pp. 1349-1358. Berenbaum, M.R., J.K. Nitao and A.R. Zangerl, 1991. Adaptive significance of furanocoumarin diversity in Pastinaca sativa (Apiaceae). J. Chem. Ecol., Vol. 17, pp. 207-215. Berenbaum, M.R. and A.R. Zangerl, 1993. Furanocoumarin metabolism in fapilio polyxenes: biochemistry, genetic variability and ecological significance. Oecologia, Vol. 95, pp. 370-375. Hay, M.E., 1991. Fish-seaweed interactions on coral reefs: effects of herbivorous fishes and adaptations of their prey. In, The ecology of$shes on coral reefs, edited by P.F. Sale. Academic Press, San Diego, CA, pp. 96-119. Hay, M.E., 1992. Seaweed chemical defenses: their role in the evolution of feeding specialization and in mediating complex interactions. In, Ecological roles of marine nafural products, edited by V.J. Paul. Comstock Publishing Associates, Ithaca, NY, pp. 93-l 18. Hay, M.E. and W. Fenical, 1988. Marine plant-herbivore interactions: the ecology of chemical defense. Annu. Rev. Ecol. Syst., Vol. 19, pp. I I I-145. Phytochem,

W.A. Lumbang,

VJ. Paul

I J. Exp. Mar.

Biol.

Ecol. 201 (1996)

18%195

195

Hay, M.E., Q.E. Kappel and W. Fenical, 1994. Synergisms in plant defenses against herbivores: interactions of chemical defenses, calcification and plant quality. Ecology, Vol. 75, pp. 1714-1726. Hay, M.E. and P.D. Steinberg, 1992. The chemical ecology of plant-herbivore interactions in marine versus terrestrial communities. In, Herbivores: their interactions with secondary plant metabolites, Volume II: Ecological and evolutionary processes, edited by GA. Rosenthal and M.R. Berenbaum. Academic Press, San Diego, CA, pp. 371-413. Langenheim, J.H., 1994. Higher plant terpenoids: a phytocentric overview of their ecological roles. J. Chm. Ecol., Vol. 20, pp. 1223- 1279. Lubchenco, J. and S.D. Gaines, 1981. A unified approach to marine plant-herbivore interactions. I. Population\ and communities. Annu. Rev. Ecol. Syst., Vol. 12, pp. 405-437. McKey, D., 1979. The distribution of secondary compounds within plants. In, Herbivores: their intrroction.c with secondary plant metabolites, edited by G.A. Rosenthal and J.H. Janzen. Academic Press, New York. NY, pp. 5.5-133. Meyer, K.D. and V.J. Paul, 1995. Variation in secondary metabolite and aragonite concentrations in the tropical green seaweed Neomeris annulata: effects on herbivory by fishes. Mar. Biol., Vol. 122, pp. 5~37~.545. Meyer, K.D., V.J. Paul, H.R. Sanger and S.G. Nelson, 1994. Effects of seaweed extracts and secondary metabolites on feeding by the herbivorous surgeonfish Naso lituratus, Coral Reefs, Vol. 13, pp. 105-I 12. Paul, V.J., 1992. Seaweed chemical defenses on coral reefs. In, Ecological roles of marine natural products, edited by V.J. Paul. Comstock Publishing Associates, Ithaca, NY, pp. 244.50. Paul, V.J., J.M. Cronan and J.H. Cardellina II, 1993. Isolation of new brominated sesquiterpene feeding deterrents from the tropical green alga Neomeris annulata (Dasycladaceae: Chlorophyta). J. Chem. Ecol.. Vol. 19, pp. 1847-1860. Paul, V.J. and M.E. Hay, 1986. Seaweed susceptibility to herbivory: chemical and morphological correlates. Mar. Ec.01. Prog. Ser., Vol. 33, pp. 255-264. Pennings, S.C. and V.J. Paul, 1992. Plant defenses against Dolabella herbivory: the role of chemistry. calcification and toughness. Ecology, Vol. 73, 160661619. Pennings, S.C. and J.M. Svedberg, 1993. Does CaCO, in food deter feeding by sea urchins? Mar. Ecol. Prog. Ser., Vol. 101, pp. 163-167. Peterson, C.H. and P.E. Renaud, 1989. Analysis of feeding preference experiments. Oecologia, Vol. 80, pp. 82-86. Schmitt, T.M., ME. Hay and N. Lindquist, 1995. Constraints on chemically mediated coevolution: multiple functions for seaweed secondary metabolites. Ecology, Vol. 76, pp. 107-123. Schupp, P.J. and V.J. Paul, 1994. Calcium carbonate and secondary metabolites in tropical seaweeds: variable effects on herbivorous fishes. Ecology, Vol. 75, pp. 1172-I 185.