Marine-terrestrial contrasts in the ecology of plant chemical defenses against herbivores

TREE vol. 6, no. 17, November

Marine-Terrestrial Contrasts in the Ecologyof PlantChemical Defenses AgainstHerbivores Mark E. Hay Small marine herbivores tkat live on the plank lheg consume often selectively eat seaweeds fhaf are thernicullg defended from fishes. Their feeding is unaffected or stimulated by the plant meta6olites that deter lishes, and these small herbivores dramatically reduce their susceptibility to predation by associating with hosf plants that are noxious to fishes. Ecological similarities betweerr these small marine herbivores and nulverous terrestrial insects suggest that herbivorous insects also may have evolved a preference for toxic plants Gecause [his dinrinishes fheir losses to predators, parasites and pathogens. Although marine and terrestrial planls arrd her6ivores evolved in strihinglg different environments, the ease of experimentalion in some marine sgslems ma&s them ideal for addressing certain questions of furrdamental importance to 6olh terrestrial and marine workers. Certain fundamental differences between plants and herbivores in marine and terrestrial communities must be discussed before these systems can be productively compared. Terrestrial plants allocate biomass to below-ground structures that are unavailable to many herbivores, and much of the aboveground mass may be woody structural tissue that is of little nutritional value. Because seaweeds live in a buoyant medium and take up nutrients through their photosynthetic thalli, they don’t make roots or allocate as much mass to nonphotosynthetic structural materials. Seaweeds are therefore all foliage, completely above ground and very available to herbivores. On herbivore-rich portions of coral reefs, seaweeds commonly lose f-3.5% of their total mass each day or 60-100% of their total production, with herbivorous fishes alone taking 40 000 to I56 000 bites/ ml/day (Ref. I I. The most intensively grazed terrestrial systems appear to be African grasslands grazed by large herds of ungulates that re-

Mark Hay is at the Institute of Marine Sciences, University of North Carolina at Chapel Hill, Morehead City, NC 28557. USA.

362

move about 66% of above-ground production2. Given the large allocation to below-ground structures, this may be no more than about 33% of total production, about r/3 to ‘12 of the rates in marine communities. In both marine and terrestrial systems, herbivory is generally greater in tropical than in temperate areas, but the highest rates of herbivory on tropical coral reefs can be ten to IO0 times greater than the highest rates in tropical forestst,s. Most herbivory in terrestrial systems is due to insects and vertebrate homeotherms4,5. Insects are largely absent from marine systems, and most herbivory is due to vertebrate ectotherms (fishes1 and a diverse array of invertebrates that are present all year and rarely undergo diapause (sea urchins, gastropods, polychaetes and a variety of marine crustaceansP. Because virtually all marine herbivores are ectotherms, changes in temperature may affect herbivory rates more in marine than in terrestrial systems’. With decreasing temperature, herbivory decreases in marine systems, but may increase in terrestrial communities because of the increased energy demands that low temperatures place on vertebrate homeothemrs. Terrestrial herbivores cope with the large amounts of relatively indigestible cellulose, hemicellulose and lignin that occur in terrestrial plants by f I ) timing their life cycles such that the feeding stages cooccur with the seasonal flush of new, high-quality leaves; (2) sucking plant sap and thus avoiding the structural materials; and (3) harboring microbial gut symbionts that make cellulose energetically available to the herbivore3,4. These strategies appear rare among marine herbivores, although sap sucking occurs among the ascoglossan sea slugs. These slugs feed primarily on chemically-rich green seaweeds from which they sequester both defensive secondary metabolites and functional chloroplasts (these continue to fix carbon for up to three

1997

months and pass much of this energy on to the slug9~rol. The apparent absence of cellulose-digesting microbial symbionts among most marine herbivores may be an artifact of the minimal effort marine investigators have devoted to their study. There can also be dramatic differences between marine and terrestrial herbivores in how food quality and abundance affect basic population biology. As an example, the tremendous physiological and morphological plasticity of sea urchins results in their population density and survivorship being largely unaffected by resource levels’~“. When algal resources decline, individual urchins shrink in size by resorbing skeletal material and differentially allocating resources to body parts used for food gathering. This allows urchins to overexploit local resources, turning kelp beds into urchin barrens, while remaining at high densities for many years thereafter, neither dying nor dispersing. This pattern contrasts with terrestrial systems where changes in plant resources appear to have large effects on the population biology of herbivoresr2. Reviewing this broad, conceptually rich and complex topic in a short article forces crude generalizations and mandates that intellectually intriguing subtleties and important specifics be omitted in favor of stressing selected general points. I hope this comparison stimulates communication among marine and terrestrial workers, but it cannot adequately represent the depth and richness of either field. Interested readers should therefore consult the more lengthy articles listed in the references. Selection for plant defenses In marine communities, most selection for plant defenses is generated by a diverse assemblage of generalistgrazers consisting primarily of fishes, sea urchins and gastropods6,7. Because these grazers are generalist feeders, they can occur at high densities relative to the abundance of their favored food plants. They often drive these to local extinction, thus having a profound effect on marine CommunitiesQ. In contrast, many terrestrial workers explicitly state or assume that most herbivory is due to

TRfE vol. 6, no. ‘II, November

1991

insects3,‘1-‘J, most of which are relatively specialized feeders”,“. If this apparent difference is real (but see Ref. 5 for a persuasive argument that mammalian herbivores are more important), then terrestrial plants would be under greater selection for traits that deter a limited number of specialists, and marine plants’ for traits that deter a broad range of taxonomically diverse generalists. Thus, the potential for closely linked coevolution between pairs of interacting species would be lower in marine than in terrestrial communities9. Although there has been a long history of considering insectplant relationships as coevolved, recent studies suggest that true coevolution is unlikely and apparently uncommon in both mariney,16 and terrestrial systems13,‘4,‘7-‘9 (see also Ref. 201. Coevolution of insects and plants may be less common than was once thought for several reasons, including f I I the limited impact of insect herbivores; (2) the great present-day, and possibly greater historic, effect of mammalian herbivores; and (3) the importance of physical stresses in interrupting coevolutionary processes’.J.5.‘2-1’ Is2’. Interactions that appear to be coevolved often may not be. For example, encrusting coralline algae and their associated gastropod herbivores provide an exceptional fossil record of a plant-herbivore interaction. However, although the characteristics of both the plants and herbivores suggest a strong history of coevolution, the fossil record demonstrates that the major ‘herbivore deterrent’ characteristics of the algae evolved many millions of years before the herbivores’6. The supposedly coevolved characteristics of these plants and herbivores appear to be based primarily on fortuitous preadaptations rather than reciprocal evolution. Terrestrial studies show similar patterns, with some closely associated plants and insects showing little, if any, evidence of coevolution”. It is clear, nevertheless, that the prevalence of specialized herbivore species differs dramatically between marine and terrestrial systems. Insects make up the huge majority of herbivore species in terrestrial communities (herbivorous insects and the plants they eat

may comprise 50% of the world’s speciesi31, and it has been estimated that about 90% of herbivorous insect species feed on three or fewer families of plants”. In contrast, most marine herbivores are very opportunistic, commonly feeding on ten to more than 20 families of plants and often supplementing their diets with animal materiallO. Some of the differences in feeding specificity between terrestrial and marine habitats could result from basic differences in life history and recruitment modes of common herbivore species9. In terrestrial communities, adult insects in some speciose groups spend minimal time feeding and are often short lived, mobile, widely dispersing and equipped primarily for reproduction and careful placement of eggs on appropriate host plants; iuvenile stages are often longer lived, relatively immobile and do most of the feeding. (These generalizations hold for diverse orders, such as the Lepidoptera and Diptera, but not for others, such as the Coleoptera and Hemiptera.1 The inconsistent and sometimes very poor correlation between the rank order of host species selected by ovipositing females and the performance of the larval stages on those hosts suggests that host specificity is often determined more by the behavior of the female than by the physiological needs of the feeding stage’8.11,2i. For example, female Drosophila magnaquinaria normally oviposit only on skunk cabbage (Lysichitom americanum), but their larvae grow significantly better on tomato (Lycopersicum esculentum) plants*‘. Most marine herbivores contrast with terrestrial Lepidoptera and Diptera in that adults do most of the feeding and are usually less mobile than the iuveniles9,25. The relatively short-lived juvenile stages may move over great distances because adults liberate larvae into the plankton rather than placing them carefully on appropriate hosts. Marine herbivores, therefore. have less ability than most insects to choose the host plant onto which they will settle, because larvae are usually returned to coastal habitats by the actions of physical oceanographic processes (currents, internal waves,

etc.) rather than by directional swimming. Additionally, within a habitat, approaching the bottom often entails large losses to predators. This should select against specialists that would need to approach the bottom many times in order to encounter their particular host, and for generalists that would be able to feed on whatever they encountered following settlement9. Specialist herbivores occur in marine communities, but they are uncommon and appear to have relatively minimal effect on their host plants’,‘O. Regardless of the relative importance of different herbivores, it is clear that selection has favored both marine and terrestrial plants having a variety of chemical, structural or life history characteristics that diminish losses to herbivores”,‘. Focusing on the similarities and differences between plants and herbivores in marine versus terrestrial communities may provide clues to factors affecting the evolution of feeding specialization and to the ecology and evolution of plantherbivore interactions in general!,?“.

Effects of plant chemical defenses on herbivore feeding patterns Both marine and terrestrial plants produce a diversity of secondary metabolites that deter feeding by herbivores; differences occur primarily due to the near absence of alkaloid production by seaweeds and the common addition of halogens to seaweed compounds’. In contrast to terrestrial studies, many experiments assessing the effectiveness of seaweed secondary metabolites have been conducted in the field (usually on relatively undisturbed coral reefs), assaying the compounds against the summed feeding of the diverse array of herbivores attacking the plant6. These types of field experiment minimize the need to assume which herbivores are, or are not, important, and allow a rigorous test of specific metabolites and the selective advantage they would confer under natural conditions. Assays can then be run in the lab to assess the effects against particular herbivores. The data from these studies show that a large number of chemically diverse secondary metabolites significantly reduce plant 363

TREE

>

7007

2

560-

z+ 2

420-

0 zi 2

140-

DIPLODZ

fl=0

o-

Ii 5

t?ttOh48~t~S

PC001

3

2

(9)

0

dI

5

p = 0.60

4c 4

I

30 2 ‘O M 0 48

\

1

(IO

cl=017

I

=AChYCkXYCL-A

Fig. I. (al Feeding preferences (mean 2 ISI) of two common omnivorous fishes and two common mesograzers. all of which co-occur in the temperate Atlantic along the southeastern coasts of the United States. The brown seaweed Dictyora menstrualis lformerly Dictyota dichotomaj was preferred by the mesograzers butavoided by thefishes.Ibt Feedingpreferenceson palatableseaweedstreatedtshadedt and untreated (hatched) with two diterpene alcohols that are produced by D. menstrualis. The metabolites deterred feeding by both fishes but either stimulated or did not affect feeding by the mesograzers. Numbers in parentheses at the bases of the hatched histograms on the right give the number of paired samples used in each assay. Data were compiled Irom Refs 33 and 34

losses in both the laboratory and the field; however, there is often tremendous variance in the effects of a given compound on different herbivores and of related compounds on the same herbivore’,9. Such differences may also be seen among geographic regions. For instance, brown algae in the north Pacific produce polyphenolics at concentrations of O-6% of algal dry mass, and the palatability of these algae to invertebrate herbivores is broadly predictable on the basis of polyphenolic content2’. These same types of alga in Australia and New Zealand produce polyphenolics at up to three times the concentration of those in the north Pacific. However, at these sites, polyphenolic content is not a predictor of algal

364

6, no.

11,

November

7997

p = 0.02

.HOLdROOKl

280-

LAGODON

vol.

palatability and polyphenolics that significantly deter herbivores in the north Pacific have little if any effect on Australasian herbivores28,29. There also appear to be dramatic differences in polyphenolic production between seaweeds in temperate and tropical communities. Polyphenolic production is common among temperate brown algae, but on herbivore-rich tropical reefs where virtually all plants are ‘bound to be found’ by generalist grazers, polyphenolics are rarely produced in significant quantities’,29. This is opposite to the pattern that would be predicted by land-based theoriesr.3n.31 and to the pattern seen in comparisons of temperate and tropical forests3. There appears to be no funda-

mental difference in the effectiveness or mode of action of tannins versus other types of compounds2A,29,32, and there is clearly little if any validity in assessing allocation to chemical defenses by clumping compounds into crude categories such as total terpenes or total phenolics. Small alterations of chemical structure such as the location of one hydroxyl group on a diterpene molecule can change a compound from a feeding deterrent to a stimulant’. Although terrestrial ecologists have commonly grouped plant secondary metabolites into broad categories for quantification or for discussing their general mode of action30,31, this now appears no more appropriate for terrestrial than for marine compounds32. What selects for resistance to plant chemical defenses? Terrestrial ecologists debate which factors select for resistance to plant chemical defenses and what role this resistance plays in the evolution of host-plant specialization’R,19.22.23.33. Early assertions that insects have specialized to partition resources or to increase their metabolic efficiency have been generally discounted’ 3 18,22-24. Present efforts are focused more on the role host specialization may play in allowing herbivores to diminish the impact of their natural enemies, to find mates, to identify appropriate physical habitats and to avoid making feeding mistakes that involve consuming plant toxins; however, the relative importance of these factors is debated”3

15.18.21-24.33

Marine studies support the hypothesis that small herbivores such as insects and marine mesograzers (herbivores 0.2 to 20 mm in length, such as amphipods and small crabs) that have limited mobility and are subject to high potential rates of predation may preferentially feed on chemically-rich plants in order to minimize their susceptibility to natural enemies25.26.3rr37. Figure I shows one example of this from the temperate Atlantic. The common omnivorous fishes Diplodus holbrookiand Lagodon rhomboides both avoid feeding on the brown alga Dictyota menstrualis due to its production of diterpene alcohols that significantly deter their feeding34,35. In contrast, the polychaete

TREE vol. 6, no. 17, November

7997

Platynereis dumerilii and a mixed species group of amphipods lprimarily Ampithoe longimanal, both of which live in nonmobile tubes that they build on the seaweeds from which they feed, selectively consume the chemically-rich seaweed Dictyota. Feeding by these mesograzers is either not affected or is stimulated by the same compounds that deter the fishes. By selectively living on a seaweed that is chemically defended from, and thus seldom visited by, omnivorous fishes, these small relatively sedentary herbivores appear to diminish their losses to consumers. During seasons when fishes feed most actively, abundance of Ampithoe longimana declines on species of algae that are palatable to fishes but increases significantly on Dictyota plants in the same habitatis. Experiments on herbivorous crabs, amphipods and ascoglossan gastropods in the Caribbean and tropical Pacific show similar patternslo. These small sedentary herbivores selectively feed on seaweeds that are chemically repellent to fishes; they are unaffected or stimulated by the compounds that deter fishes, and in field and laboratory assays, they experience dramatically reduced fish predation when they associate with these plants25,26,‘“.3’, This host-associated refuge works even when the attacking fishes are carnivores rather than herbivores or omnivores2i~2”,S7. Losses to enemies were diminished by host-specific crypsis2h,i7, host-specific decreases in encounter frequency3’-j6 and host-derived chemical defenses2’,26,37. These same mechanisms operate to diminish the susceptibility of insects to their enemies in present-day terrestrial communities’S.‘A.22.23 , and it has been suggested that toxic plants would have been even more refuges before the important Pleistocene extinctions removed common megaherbivores lmammoths, camels, horses, etc.1 from several continents’. By living on hosts that were unpalatable to megaherbivores, insect larvae may have reduced indirect consumption by these large mammals. Predator avoidance or deterrence may be a major factor selecting for both insects and marine mesograzers that feed from, or become specialists on,

chemically-rich plants. However, in marine communities, this method of minimizing predation is used by generalists as well as specialists and evolving resistance to a seaweed defensive metabolite rarely leads to specialization, even if the metabolite is selectively sequestered as a predator deterrentlO. It appears that several generalist mesograzers are resistant to a broad range of unrelated lipophilic metabolites that deter fish feeding. Although these mesograzers prefer to feed from plants producing these compounds, they also continue to feed from many other unrelated plants’.‘9. Thus, although acquisition of ‘enemy-free space’ may be very important in selecting for host use in both marine mesograzers and terrestrial insects, it does not sufficiently explain the differing degrees of specialization between these groups. Both generalist insects’0 and marine mesograzers’O can selectively feed on toxic plants and reap the rewards of decreased predation without becoming as specialized as is common among insects in general. The unresolved question of why insects are so commonly host-plant specialists and why this occurs so rarely among other herbivores, either terrestrial or marine, is one of the most vexing questions to arise from contrasts of marine and terrestrial systems and deserves increased attention. Acknowledgements This work was supported by NSF grants OCE 89-0013 I and 89-I 1872 Comments from Patricia Hay and four anonymous reviewers improved the manuscript

References I Hay, M.E. and Steinberg, P.D. in Herbivores: Their Interaction with Secondary Plant Metaboliles: Vol. II, Ecological and Evolutionary Processes IRosenthal. G. and Berenbaum, M.R., edsl. Academic Press (in press) 2 McNaughton. S.I. 11985) Ecoi. Monogr. 55, 259-294 3 Coley. P.D. and Aide, T.M. II9901 in P/antAnimal Interactions: Evolutionary Ecology in Tropical and Temperate Regions (Price. P.W.. Lewinsotin, T.M.. Femandes, C.W. and Benson, W.W., edsl, pp. 25-49# john Wiley 4 Howe, H F. and Westley, L.C. ( I9881 Ecological Relationships of Plants and Animals. Oxford University Press 5 Crawley. M.J. (19891 in Insect-P/ant Interactions. Vol. I (Bemays. E.A.. edl, pp. 45-71, CRC Press

6 Lubchenco. I. and Gaines. S D 11981 I Annu. Rev. Ecol. Syst. 12. 405-437 7 Duffy, I.E. and Hay, M.E. II9901 BoScience 40. 368-375 8 Hay, M.E. ( I99 I I in The Ecology of Fishes on Corai Reefs (Sale, P.F., ed.1. pp. 96-l 19. Academic Press 9 Hay, M.E. and Fenical, W I19681 Annu. Rev. Ecol. Syst 19. I I l-145 IO Hay, M.E. in Ecologica/ Roles of Marine Natural Products (Paul. V.1, ed.1. Comstock Publishing Associates lin press) I I Levitan, D.R. I !9891 Ecology 70. 1414-1424 I2 Denno, R.F. and McClure, M.S. II9831 Variable Plants and Herbivores in Natural and Managed Systems, Academic Press I3 Strong. D.R, Lawton. J.H. and Southwood, R. I I9841 Insects on P/ants. Community Patterns on P/an&, Harvard University Press I4 Spencer, K.C. I I9881 Chemica! Mediation of Coevolution, Academic F’ress I5 Bemays. E.A. II9891 Evol. Ecol 3.299-j II 16 Steneck. R.S. in P/an&Animal lnteracfions in the Marine Benthos Ilohn, D.. Hawkins, S.I. and Price. I.. edsl, Oxford University Press (in press1 I7 Futuyma. D.I. and Slatkin. M II9831 Coevolution, Sinauer Associates I8 Futuyma. D.I. and Moreno. C. ! 19881 Annu. Rev. Ecol. Syst. i9. 207-233 19 Futuyma. D.I. and McCafferty. S S. II9901 Evolution 44, I88ii- I9 I 3 20 Mitter. C , Farrell, B. and Futuyma. D I ( I99 I I Trends Eco/. Evoi. 6. 290-293 21 lermy. T. ( 19841 Am Nat I24,609-b30 22 Courtney. S.P. apd Kibota. 1.1’. i 19901 m Insect-P/ant Interacfions, Vol. Ii ;Bcrnays. EA.. ed.1. pp 161-188. CRC P:ess 23 laenike. I. 11990) Annu. Rev. Ecoi. Syst. 2 I, 243-273 24 Kibota. T.T. and Courtney, S P. f I991 I Oecoiogia 86. 25 I-260 25 Hay, M.E.. Duffy. j E. and Fenical. W l I9901 Ecology 7 I, 733-743 26 Hay, M.E., Pawlik. j.R, Duffy. 1.E and Fenical,W.119891 Oecolog~a81.418-427 27 Steinberg. P.D. (19851 Eco;. Monogr. 55, 33’3-349 28 Steinberg. P.D. and Van Aitena. I. Ecol. hlonogr. I in press1 29 Steinberg, P.D. in Ecoiogical Roles ol Marine Natural Producb (Paul. V.I.. ed I, Comstock Publishing Associates fin press1 30 Feeny. P. ( I9761 Recent Adv. Phytochem. IO. I-40 31 Coley. P.D., Bryant. I I? and Chapin. F.S.. Ill 11985) Science 230,895-899 32 Bernays, E A, Driver, GC and Bilgener. M. i 19891 Adv. Eco/. Res. 19. 263-332 33 Bernays. E.A. and Graham, ,M I i9881 Ecology 69,88&892 34 Hay, M.E., Duffy, I.E., Pfister, C.A and Fenical. W. II9871 Ecology 68, 1567.-1580 35 Hay, M E . Renaud. P E. and Fenical, W ! I9881 Oecologia 75. 246-252 36 Hay, M.E.. Duffy. I.E.. Fenical. W. and Gustafson. K. I I9881 Mar. Eco!. Prog. Ser. 48. 185-192 37 Hay, M.E.. Duffy. I.E.. Paul. \/.I.. Renaud. P.E and Fenical. W. (I9901 Llmno/ Oceanogr. 35, 1734-l 743 38 Duffy. I.E. and Hay. M.E. I I 991 I Ecology 72. 12-1298 39 Duffy. I E. and Hay. M.E II991 1 Ecology 72, 354-358 40 Blum. MS. ef a/ Ii99OI 1. Chem Eco/. 16. 223-244

365