sign in sign up
<<
Home , Larva

Hypotheses

Evolution of direct-developing larvae: selection vs loss Margaret Snoke Smith,1 Kirk S. Zigler,2 and Rudolf A. Raff1,3*

Summary , urchins, reveal that a majority of sea urchin Observations of a sea urchin larvae show that most species exhibit one of two life history strategies (Fig. 1).(10) species adopt one of two life history strategies. One strategy is to make numerous small , which develop One strategy is to make many small eggs that develop into a into a with a required feeding period in the water larva that must feed and grow for weeks or months in the water column before . In contrast, the second column before metamorphosis. In the water column, these strategy is to make fewer large eggs with a larva that does larvae are dependent on abundance for food and are not feed, which reduces the time to metamorphosis and subject to high levels of predation.(11) The other strategy thus the time spent in the water column. The larvae associated with each strategy have distinct morpholo- involves increasing size and maternal provisioning, which gies and developmental processes that reflect their decreases the reliance on planktonic food sources and the feeding requirements, so that those that feed exhibit time to metamorphosis, thus minimizing larval mortality. indirect development with a complex larva, and those that However, assuming finite resources for egg production, do not feed form a morphologically simplified larva and increasing egg size also reduces the total number of eggs exhibit direct development. Phylogenetic studies show that, in sea urchins, a feeding larva, the pluteus, is the produced. ancestral form and the morphologically simplified direct- Both of these life history strategies are associated with developing larva is derived. The current hypothesis for developmental differences resulting in distinct larval forms. evolution of the direct-developing larval form in sea Generally the strategy of producing many small eggs and a urchins suggests that major developmental changes feeding larva is characteristic of indirect development. In sea occur by neutral loss of larval features after the crucial transition to a nonfeeding life history strategy. We urchins, indirect developers produce a complex pluteus larva, present evidence from Clypeaster rosaceus, a sea urchin which has arms bearing a ciliary band, in which nest a mouth with a life history intermediate to the two strategies, and a gut—all structures associated with feeding and survival which indicates that major developmental changes for in the water column. In contrast, the strategy of producing accelerated development have been selected for in a larva fewer larger eggs and nonfeeding larva is characteristic of that can still feed and maintains an outward, pluteus morphology. We suggest that transformation of larval direct developers. These organisms produce a barrel-shaped form has resulted from strong selection on early initiation larva that has reduced or no arms, an incomplete or no ciliary and acceleration of adult development. BioEssays band, no mouth and no functional gut. One of the striking 29:566–571, 2007. ß 2007 Wiley Periodicals, Inc. differences between direct- and indirect-developing larvae is the timing of the formation of the large left —the tissue Evolution of marine larvae that gives rise to much of the adult rudiment. Indirect Marine larval forms can evolve rapidly and dramatically, but the developers form a large left coelom and initiate adult selection pressures involved in the evolution of development development in advanced larvae, weeks after fertilization, are poorly understood. There is much debate among theories whereas direct developers form a large left coelom before the about what the important selective pressures are and how end of gastrulation and initiate adult development hours after these forces shape the evolution of life histories of marine fertilization. Phylogenetic studies show that indirect develop- (1–9) organisms. Observations in one prominent group of ment is the ancestral form of sea urchin development, but direct development has evolved independently multiple times.(2,12,13) 1Department of Biology, Indiana University, Bloomington, IN. 2Department of Biology, The University of the South, Sewanee, TN. Although much is known about the differences between 3School of Biological Sciences, University of Sydney, Sydney, Australia. direct and indirect developers, the evolutionary transition Funding agency: NSF. between the two modes is poorly understood because there is *Correspondence to: Rudolf A. Raff, Department of Biology, Indiana no fossil record of echinoid larvae. The current hypothesis for University, Myers Hall 150, 915 East Third Street, Bloomington, IN the evolution of direct-developing larvae is based on living, 47405. E-mail: [email protected] (13) DOI 10.1002/bies.20579 apparently intermediate larval forms and suggests that Published online in Wiley InterScience (www.interscience.wiley.com). major developmental changes happen relatively late in the process. Wray(13) has proposed, albeit recognizing that there

566 BioEssays 29.6 BioEssays 29:566–571, ß 2007 Wiley Periodicals, Inc. Hypotheses

Figure 1. Larval development of a large number of sea-urchin species have been examined and can be sorted into one of three categories: small eggs with obligately feeding larvae, larger eggs with facultative feeding larvae, and still larger eggs with nonfeeding larvae. The distrubutions of sea urchin species into these catagories are listed here. This figure is based on Emlet 1990.(10)

is selection for new features such as rapid metamorphosis, ment of these species allows us to test basic premises of the that the barrel-shaped morphology of many direct developers current model. The developmental shift to early formation of a is largely a result of relaxed selection on structures associated large left coelom, as seen in nonfeeding direct-developing with feeding.(14) According to this scenario, there is first an larvae, should occur relatively late in the evolution of direct- increase in egg size and provisioning that allows for the loss of development if features that speed development evolve after feeding ability (Fig. 2, steps 1 and 2). After this trophic change, feeding is lost and larval features are reduced. the selective pressures for maintenance of feeding structures We examined a potential intermediate, the larvae of the are relaxed and there is selection to reduce the time to facultative feeder, Clypeaster rosaceus, a sea biscuit urchin metamorphosis. It is not until the third step, after the ability to found in the Caribbean and southern US. Its larvae have an feed has been lost, that Wray suggests that major develop- exterior pluteus morphology and may feed, but they are not mental changes occur, indicating that the trophic switch is required to do so in order to metamorphose. The egg diameter crucial for the transition to direct development and that the of C. rosaceus, 275 microns(15,16) and time to metamorpho-  developmental changes are secondary and potentially by- sis, 5–7 days(15) are also intermediate between direct and  products of relaxed selection. indirect developers. C. subdepressus, a relative of C. rosaceus, is a typical obligate feeding, indirect developer. Intermediates and ancestors Internally, the pluteus of C. rosaceus presents a surprise. To evaluate Wray’s model and the idea that major develop- Although C. rosaceus maintains an outward pluteus morpho- mental changes occurred neutrally due to the change in logy there has been a heterochronic acceleration in the feeding ability near the end of the transition to direct formation of a large left coelom reminiscent in form, size and development, we can examine, against the predictions of the timing of direct-developing larvae (Fig. 3). Early onset and model, one of the few species that represents a transitional acceleration of the formation of a large left coelom in C. state, facultative feeding. Although such intermediate forms rosaceus suggests two major changes in the current thinking are only proxies for the fossil record, examining the develop- on how the direct development evolved (Fig. 2). First, major

BioEssays 29.6 567 Hypotheses

C. subdepressus eggs and plutei are clear as usual with indirect developers. Second, the transition to a barrel-shaped larva may be driven by selection for streamlined development of the adult and rapid metamorphosis. Subsequent loss of some feeding structures may be due to selection for allocation of the resources to rapid development of adult structures, which suggests a different course of evolution than expected, and that adaptive selection may play a much stronger role than previously thought in shaping the evolution of the direct-developing larval form.

Is Clypeaster rosaceus anomalous? As direct development has arisen independently in several echinoid lineages, it is possible that each species takes a different developmental path. However, early formation of a large left coelom is common in nonfeeding echinoids. For example, the taxonomically distant direct developers, Per- onella japonica,(17) H. erythrogramma,(18) Asthenosoma ijimai(19) and Phyllacanthus parvispinus(20) all form large left just following gastrulation. Without fossil larval intermediates, we may never know exactly how direct development evolved in each lineage of echinoids, but C. rosaceus and other species suggest that the heterochrony in large left coelom formation could be a general and early step of the transition of developmental mode.

Selection on rapid development of adult features We suggest a new hypothesis of how direct development evolved in sea urchins (Figs. 3,4). First, following the transition to larger eggs, the observation from C. rosaceus suggests that the next step is to accelerate the timing of left coelom formation, not loss of the ability to feed. C. rosaceus can still feed, yet it vastly accelerates the development of a large left coelom before the end or just after gastrulation is complete. Secondly, we suggest that reducing the time to metamorphosis is not a passive step happening only after the ability to feed has been lost. Rather, the speed of adult development prior to metamorphosis is arguably the crucial target of selection in the transition to direct development. C. rosaceus already shows a Figure 2. Wray’s theory of the transitional steps during the dramatic reduction in time to metamorphosis (5–7 days) evolution of direct-developing larvae. This model suggests that relative to C. subdepressus (16–28 days).(15) We suggest that developmental changes occur neutrally after feeding is lost and that the reduction in time to metamorphosis is a relatively late the last modification during the transition to direct development step in the process. (Copyright 1996 from Parallel Evolution of may be re-allocation of cells and other resources that would Nonfeeding Larvae in Echinoids by Gregory A. Wray. Repro- otherwise go to arms and other larval structures, likely driven duced by permission of Taylor & Francis Group, LLC, http:// by selection rather than drift. This hypothesis is supported by www.taylorandfrancis.com) the highly modified embryonic cell lineages of the direct developer H. erythrogramma in which a large proportion of blastomeres are allocated to adult fates.(21) Analogous modi- developmental changes occur before the trophic transition to fications of blastomere allocation occurs in the fresh water nonfeeding is complete. Likely other changes such as egg clam, Unio, where increased provisioning to the 2d blastomere content have occurred given that C. rosaceus eggs and plutei contributes to the formation of a relatively large larval shell are opaque, reminiscent of the direct developers, whereas needed for the parasitic lifestyle of the Unio larva.(22,23)

568 BioEssays 29.6 Hypotheses

Figure 3. Clypeaster rosaceus maintains the same elaborate larval outward morphology and ability to feed of the typical indirect developer C. subdepressus (D,E) Surprisingly though C. rosaceus forms a large left coelom (H) before the end of gastrulation, which is remarkably similar to the development of the direct developer H. erythrogramma (I). Regardless of the larval development, all three species form relatively similar adults (A,B,C). First row: adults; second row: larvae; third row: larvae showing the left coelom. G is an image from the whole embryo, and H and I are stained sections of the larvae. Plc, presumptive left coelom; lc, left coelom. Pictures D, E were taken by Richard Emlet and reprinted from the Journal of Experimental Marine Biology and Ecology, vol. 95, titled Facultative Planktotrophy in the Tropical Echinoid Clypeaster rosaceus (Linnaeus) and a Comparison with Obligate Planktotrophy in Clypeaster subderpressus (Gray) (Clypeaste roida:Echinoidea) pp188–189, copyright 1986, with permission of Elsevier.

The evolution of direct-developing larvae reflects a similar abundance or both, this would provide a fitness advantage scenario to blind cave . Yamamoto et al.(24) have shown over indirect developers.(14) that the loss of functional in these fish is not simply a result of relaxed selection for the maintenance of eyes due to Does selection on provisioning matter? the fish living in the dark, but rather more of a result of selection Besides changes in larval morphology, other modifications are for a modified jaw morphology. The loss of eyes was due to the associated with the evolution of direct development in sea negative correlation with the formation of a larger jaw. urchins. The eggs of the direct developer H. erythrogramma Similarly, in echinoid larvae, the evolution of direct-developing are 100 times the volume of the eggs of its indirect-developing larva is driven by selection for rapid development of adult sister species, H. tuberculata. Similarly, C. rosaceus eggs features and metamorphosis once there has been an have also increased in volume and have 30 times the volume of increased investment in egg provisioning, which would H. tuberculata eggs (6 times the volume of C. subdepressus increase fitness by reducing the time to metamorphosis. In eggs). In addition, direct developers contain qualitatively an environment with high levels of predation or low food different stores of lipids and .(25,26) Concurrent with the

BioEssays 29.6 569 Hypotheses

of about 3 myr or less. Such rapid larval evolution is known for larvae and gastropod shells.(29,30) Thus, it may be possible to decrease the egg size again after the transition to larger eggs and move back from facultative feeing to obligate feeding. At the extreme end, individual features such as arm growth could be reinitiated given re-activation of critical genes, but the entire reversal likely involves a sufficient number of changes to be improbable.

Conclusion Often the evolution of simplified morphologies is attributed to the loss of complexity due to a lack of selection to maintain the complex structures. Such thinking is part of the current theory of how barrel- shaped, direct-developing sea urchin larvae evolved from ancestors with elaborate, indirect-developing larvae. We suggest a different hypothesis in which following an increase in egg provisioning, there are major developmental changes, such as acceleration of the timing of coelom formation and initiation of adult development driven by selection for rapid metamorphosis. Our hypothesis is supported by recent evidence from C. rosaceus, which still has the ability to feed, but it exhibits a heterochronic shift in large left coelom formation. As shown here with C. rosaceus as well as results reported in re-investigation of the classic case of loss in blind cave fish, loss of features is not always associated with relaxed selection, but may be driven by selection on other traits.(24,31) The acceleration of large left coelom formation in C. rosaceus and other nonfeeding sea urchins is a key innovation that makes more rapid development to metamorphosis possible, and thus opens other downstream changes that can also be selected for. Detailed understanding of such organisms will shed light on the developmental pathways underlying the Figure 4. Based on the observation that C. rosaceus, a origin and evolution of novel features. facultative feeder, forms a large left coelom relatively early in C. rosaceus offers a compelling case for selection for rapid development, this shows a new theory for the evolution of direct metamorphosis driving the formation of barrel-shaped larvae. development. We suggest that developmental changes happen However, to establish the generality of the accelerated before the ability to feed is lost and these developmental development of adult features, study of other facultatively changes may be driven by selection for rapid metamorphosis. feeding larvae will need to be done. Only one other such other faculatatively feeding sea urchin is currently known. There are also examples of pluteus-like nonfeeding larvae. Additionally, increase in egg size and content, sperm size has also examination of the cell lineage in C. rosaceus will show increased in direct developers.(27) Egg provisioning is crucial whether modification of blastomere allocation have occurred for rapid development of the adult without the need for a larval in C. rosaceus. A change in blastomere allocation, as seen in feeding machine. the highly modified H. erythrogramma embryo cell lineage, Revisiting Dollo’s Law would be consistent with our hypothesis that loss of larval As apparent intermediates exist in the sea urchin fauna, it may features is actively driven by selection and seems unlikely to be suggested that larval evolution is not necessarily a one-way occur due to relaxed selection. Lastly, work is needed to street. Evolution in reverse to restore pluteus features might understand the molecular mechanisms underlying left also be occurring. Our earlier studies of the implication of coelom formation, at the level of gene pathways as well as in Dollo’s Law(28) indicate that some features might be reversible, cellular level processes. Comparison of such work to that of if evolutionary steps have taken place in a relatively short time H. erythrogramma will show whether gene expression and

570 BioEssays 29.6 Hypotheses

cell-level mechanisms are more characteristic of direct 14. Wray G. 1995. Evolution of Larvae and Developmental Modes. In: developers or maintain similarity to indirect developers. Such McEdward L, editor. Ecology of Marine Invertebrate Larvae. Boca Raton: CRC Press. comparisons will also allow us to evaluate whether early, large 15. Emlet RB. 1986. Facultative planktotrophy in the tropical echinoid left coelom formation generally involves the same genes or Clypeaster rosaceus (Linnaeus) and a comparison with obligate whether there are different molecular ways to arrive at this planktotrophy in Clypeaster subdepressus (Gray). J Exp Mar Biol Ecol 95:183–202. phenotypic outcome. 16. Miner BG, Cowart JD, McEdward LR. 2002. Egg energetics for the facultative planktotroph Clypeaster rosaceus (Echinodermata:Echinoi- dea), revisted. Biol Bull 202:97–99. Acknowledgments 17. Okazaki K. 1975. Normal development to metamorphosis. In: Czihak G The authors would like to thank Harilaos Lessios, Smithsonian (with help from R. Peter), editors. The Sea Urchin Embryo. Berlin, Tropical Research Institute, Panama, for vital help with this project. Heidelberg, New York: Springer-Verlag. Chapter 7, pp 177–232. 18. Ferkowicz MJ, Raff RA. 2001. Wnt gene expression in sea urchin development: Heterochronies associated with the evolution of develop- References mental mode. Evo Dev 3:24–33. 1. Vance RR. 1973. On reproductive strategies in marine benthic 19. Amemiya S, Emlet RB. 1992. The development and larval form of an invertebrates. Am Nat 107:339–352. Echinothurioid Echinoid Asthenosoma ijimai, revisited. Biol Bull 182: 2. Emlet RB, McEdward LR, Strathmann R. 1987. larval 15–30. ecology viewed from the egg. Echinoderm Studies 2:55–136. 20. Parks AL, Bisgrove BW, Wray GA, Raff RA. 1989. Direct development in 3. Emlet RB, Hoegh-Guldberg O. 1997. Egg size and post-larval the sea urchin Phyllacanthus parvispinus (Cidaroidea): Phylogenetic performance: experimental evidence from a sea urchin. Evolution 51: history and functional modification. Biol Bull 177:96–109. 141–152. 21. Wray GA, Raff RA. 1990. Novel origins of lineage founder cells in the 4. McEdward LR. 1997. Reproductive strategies of marine benthic direct developing sea urchin Heliocidaris erythrogramma. Dev Biol invertebrates revisited: facultative feeding by planktotrophic larvae. Am 141:41–54. Nat 150:48–72. 22. Lillie FR. 1895. The of Unionidae. J Morphol 10:1–100. 5. Hart MW. 2000. Evolutionary loss of larval feeding: Development, form 23. Lillie FR. 1898. in cleavage. In: Biological Lectures of Marine and function in a facultative feeding larva, Brisaster latifrons. Evolution Biological Laboratory of Woods Hole, Mass. Boston: Ginn and Company. 50:174–187. pp 43–67. 6. Levitan DR. 2000. Optimal egg size in : Theory 24. Yamamoto Y, Stock DW, Jeffery WR. 2004. Hedgehog signaling controls and phylogenetic analysis of the critical relationship between egg size eye degeneration in blind cavefish. Nature 432:844–847. and developmental time in echinoids. American Naturalist 256:175– 25. Byrne M, Villinski JT, Cisternas P, Popodi E, Raff RA. 1999. Maternal 192. factors and the evolution of developmental mode: evolution of oogenesis 7. McEdward LR. 2000. Adaptive evolution of larvae and life cycles. in Heliocidaris erythrogramma. Devel Genes Evol 209:275–283. Seminars in Cell and Dev Bio 11:403–409. 26. Villinski JT, Villinski JC, Byrne M, Raff RA. 2002. Convergent maternal 8. Hart MW. 2002 Life history evolution and the comparative developmental provisioning and life-history evolution in Echinoderms. Evolution biology of echinoderms. Evol Development 4:62–71. 56:1764–1775. 9. Pechenik JA. 2006. Larval experience and latent effects-metamorphosis 27. Raff RA, Herlands L, Morris VB, Healy J. 1990. Evolutionary modification is not a new beginning. Integrative Comp Biol 46:323–333. of echinoid sperm correlates with developmental mode. Dev Gro Diff 10. Emlet RB. 1990. World patterns of developmental mode in echinoid 32:283–291. echinoderms. In: Hoshi M, Yamashita O, editors. Advances in inverte- 28. Marshall CR, Raff EC, Raff RA. 1994. Dollo’s law and the death and brate reproduction 5. Amsterdam: Elsevier. resurrection of genes. Proc Nat Acad Sci 91:12283–12287. 11. Lamare MD, Barker MF, 1999. In situ estimates of larval development 29. Raff RA, Byrne M. 2006 The active evolutionary lives of echinoderm and mortality in the New Zealand sea urchin Evechinus chloroticus larvae. Heredity 97:244–252. Echinodermata:Echinoidea). Marine Ecology Progress Series 180:192– 30. Collins R, Cipriani R. 2003. Dollo’s law ans the re-evolution of shell 211. coiling. Proc Roy Soc Lond B 270:2551–2555. 12. Strathman R. 1978. The evolution and loss of feeding larval strategies of 31. Yamamoto Y, Espinasa L, Stock DW, Jeffery WR. 2003. Development marine invertebrates. Evolution 32:894–906. and evolution of craniofacial patterning is mediated by eye-dependent 13. Wray GA. 1996. Parallel evolution of nonfeeding larvae in Echinoids Syst and -independent processes in the cavefish Astyanax. Evo Devo 5:435– Biol 45:308–322. 446.

BioEssays 29.6 571