REpROdUCTIvE ISOLATION BETwEEN SpECIES OF SEA URChINS

Existing knowledge on reproductive isolating barriers between sea urchin species is reviewed. Experiments involving artificial production of hybrids between congeneric echinoid species have shown that in most cases hybrids are viable and capable of backcrossing. Only species separated for > 5 million yrs show complete post-zygotic isolation. Each potential prezygotic isolating barrier appears to be incapable of completely preventing gene flow between sympatric species. different habitat preferences exist in many, but not all, sympatric species. Annual reproductive cycles are too environmentally labile to isolate entire species. Lunar reproductive rhythms may be a form of temporal isolation in some diadematid species, but they are generally lacking in other echinoids. Gametic isolation is bidirectional and complete in a few pairs of congeneric species, but as a rule it allows one-way gene flow between congeneric species. There is no correlation between pre-zygotic isolation and the time since separation of the species. Bindin, a reproductive molecule involved in gamete incompatibility, shows evidence of strong selection in genera that contain sympatric species, but appears to be evolving neutrally in genera that do not. however, the cause of selection, where it exists, is more likely to be some form of intraspecific process, such as sexual selection, rather than reinforcement to avoid hybridization. Even though no single barrier seems to be either absolute or universal, the combination of several barriers is potentially capable of reducing the probability of hybrid production in nature, which may explain why there is little credible evidence of natural hybridization or introgression between sea urchin species. John pearse has made major contributions to the study of reproductive ecology of sea urchins, particularly contributions that have focused on factors that govern the timing of reproduction of these animals. The reproductive ecology of sea urchins has also received a fair amount of attention recently from the point of view of reproductive isolation, i.e., the intrinsic biological barriers that maintain species as independent gene pools. Emergence of reproductive isolation converts geographic isolates into separate species, which will then evolve independently even if they were to come into contact with each other. Before we attempt to characterize the evolution of reproductive barriers we must first know what they are. In this article, I review existing knowledge of reproductive ecology as it pertains to the emergence of reproductive isolation, and thus to speciation. For more than a century sea urchins have been a model organism in embryology and developmental biology; the same cannot be said for their role in speciation research. As it will become obvious in this paper, we do not have the data to draw firm conclusions about the full trajectory of the evolution of reproductive isolation in any pair of sea urchin sister species. what we do know pertains to specific aspects of the biology of certain species that are relevant to speciation. with few exceptions, there is no comprehensive body of knowledge that would allow one to determine which reproductive barrier accounts for complete genetic isolation between two sympatric sea urchin species, let alone permit firm conclusions of how reproductive isolation arises in this class of echinoderms. Yet, as Mayr (1954) perceived early on, and as BULLETIN OF MARINE SCIENCE, vOL. 81, NO. 2, 2007 192 more recent treatments of speciation—such as that of Coyne and Orr (2004)—have also made clear, sea urchins do play a role in speciation theory (see also palumbi and Lessios, 2005): They provide a test of whether ideas developed from the study of arthropods and vertebrates (for which information regarding speciation is more extensive, but far from complete) also apply to organisms with starkly different fertilization systems and reproductive ecologies. From this point of view, that fertilization in sea urchins is external and that the echinoid behavioral repertoire is limited actually represent advantages for the study of the emergence of reproductive isolation, in that they greatly simplify the possibilities of potential reproductive barriers that can isolate species. Reproductive barriers between species are usually classified into two major categories, prezygotic and postzygotic (Mayr, 1963; dobzhansky, 1970; Coyne and Orr, 2004). prezygotic isolation barriers drastically reduce the ability or the opportunity of two species to mate with each other, even when their geographic ranges overlap. postzygotic barriers consist of some aspect of the biology of hybrids that reduces their fitness, either because they survive less well than non-hybrids, or because they have a lower probability of mating successfully. Reproductive ecology of each species is obviously more closely related to prezygotic barriers, but post-zygotic barriers are not irrelevant. Low fitness of hybrids can lead to selection that will shift the reproductive ecology of the hybridizing species away from each other, the phenomenon of reproductive character displacement (Brown and wilson, 1956), through the process of reinforcement (dobzhansky, 1940; Butlin, 1989; Servedio and Noor, 2003). Indeed, in the view of one of the most prominent figures in speciation research, the emergence of postzygotic isolation is a stage of speciation that precedes the emergence of prezygotic isolation (dobzhansky, 1940). I will, therefore, begin by examining what is known regarding postzygotic reproductive barriers in sea urchins. postzygotic Barriers in Sea Urchins Fitness of F1 hybrids in the Laboratory.—Because generation time in echinoids is usually on the order of a year or more, and because sea urchin larvae and adults are not always easy to keep alive in captivity, assessments of the fitness of hybrids and their ability to back-cross to their parents are few. To my knowledge, fitness of hybrids with congeneric parental species has been determined in only five genera, Echinometra, Pseudechinus, Strongylocentrotus, Heliocidaris, and Diadema (Table 1). The most complete data, with quantified survival rates, come from the newly discovered and as yet unnamed Indo-west pacific species of Echinometra. These are recently diverged species, thus it may not be surprising that their hybrids generally develop normally. The survivorship of larvae to metamorphosis in the cross between eggs of Echinometra mathaei (Blainville) and sperm of Echinometra sp. C is significantly lower than those of conspecific parents, but it is still > 60% (Rahman and Uehara, 2004), which would indicate that, by itself, this difference could not account for reproductive isolation between these two species. A slight depression in the settlement rates of hybrids between E. mathaei and Echinometra sp. A is compensated by higher growth rates of the hybrids, relative to non-hybrid offspring (Rahman et al., 2005). That the cross between Heliocidaris tuberculata Lamarck eggs and Heliocidaris erythrogramma (valenciennes) sperm is lethal, even though these species have been isolated for only 4–5 million yrs (my), may arise from the shift LESSIOS: REpROdUCTIvE ISOLATION IN EChINOIdS 193 Ta bl e 1. S ur vi vo rs hi p of h yb rid s i n ex pe rim en ta l c ro ss es b et w ee n co ng en er ic sp ec ie s o f s ea u rc hi ns , r el at iv e to ti m e si nc e di ve rg en ce . Pa re nt al sp ec ie s c on tri bu tin g: Ti m e* (m y) eg gs sp er m D ev el op m en t R ef er en ce 1. 1– 1. 5 Ec hi no m et ra m at ha ei × sp . C su rv iv or sh ip sl ig ht ly d ep re ss ed R ah m an a nd U eh ar a, 2 00 4 sp . C × m at ha ei no rm al 1. 1– 1. 3 Ec hi no m et ra m at ha ei × sp . A su rv iv or sh ip sl ig ht ly d ep re ss ed R ah m an e t a l., 2 00 5 sp . A × m at ha ei su rv iv or sh ip sl ig ht ly d ep re ss ed 1. 1– 1. 5 Ec hi no m et ra sp . A × sp . C su rv iv or sh ip sl ig ht ly d ep re ss ed R ah m an e t a l., 2 00 1 sp . C × sp . A no rm al 1. 1– 1. 3 Ec hi no m et ra sp . A × E. o bl on ga no rm al A sl an a nd U eh ar a, 1 99 7 E. o bl on ga × sp . A no rm al 7 Ps eu de ch in us al bo ci nc tu s × no va ez ea la nd ia e no rm al M cC la ry a nd S ew el l, 20 03 no va ez ea la nd ia e × al bo ci nc tu s lo w se ttl em en t s uc ce ss 1. 6 Ps eu de ch in us al bo ci nc tu s × hu tto ni no rm al M cC la ry a nd S ew el l, 20 03 hu tto ni × al bo ci nc tu s no rm al 7 Ps eu de ch in us hu tto ni × no va ez ea la nd ia e di es a t t he 8 a rm p lu te us st ag e M cC la ry a nd S ew el l, 20 03 no va ez ea la nd ia e × hu tto ni lo w se ttl em en t s uc ce ss 1. 2– 1. 4 St ro ng yl oc en tro tu s dr oe ba ch ie ns is × pa lid us no rm al St ra th m an n, 1 98 1 pa lid us × dr oe ba ch ie ns is no rm al 4. 4 S tro ng yl oc en tro tu s fr an ci sc an us × p ur pu ra tu s di es a t g as tru la tio n N ew m an , 1 92 3 p ur pu ra tu s × fr an ci sc an us no rm al 4– 5 H el io ci da ri s er yt hr og ra m m a × tu be rc ul at a vi ab le R af f e t a l., 1 99 9 tu be rc ul at a × er yt hr og ra m m a di es a t g as tru la tio n 7– 14 D ia de m a sa vi gn yi × se to su m vi ab le U eh ar a et a l., 1 99 0 * A pp ro xi m at e tim es a re b as ed o n m ito ch nd ria l D N A d at a, ta ke n fr om L au nd ry e t a l. (2 00 3) fo r E ch in om et ra , f ro m L ee (2 00 3) a nd B ie rm an n et a l. (2 00 3) fo r S tro ng yl oc en tro tu s, fr om Je ffe ry e t a l. (2 00 3) fo r P se ud ec hi nu s, fr om Z ig le r e t a l. (2 00 3) fo r H el io ci da ri s, an d fr om L es si os e t a l. (2 00 1) fo r D ia de m a BULLETIN OF MARINE SCIENCE, vOL. 81, NO. 2, 2007 194 of H. erythrogramma to direct development, a highly atypical evolutionary change among echinoids. however, the cross between eggs of Strongylocentrotus franciscanus (A. Agassiz, 1863) and sperm of Strongylocentrotus purpuratus (Stimpson, 1857) (both of which have planktonic larvae and split from each other at roughly the same time as the two species of Heliocidaris) is also lethal, which may indicate that by 5 my dobzhansky-Muller incompatibilities (i.e., developmental difficulties arising from incompatible alleles in at least two loci of the F1 hybrids) begin to render sea urchin hybrids inviable, even when no developmental shifts are involved. The low survivorship (in one direction) of offspring from the cross between Pseudechinus novaezealandiae Mortensen and Pseudechinus albocinctus hutton indicates that such unidirectional developmental difficulties extend to species that have been separated for 7 my. however, the cross of Pseudechinus huttoni Benham and P. novaezealandiae, which is lethal in both directions, suggests that this is the period of time in which complete post-zygotic isolation becomes evident. One would not expect such an estimate to hold universally true in all echinoids, and indeed, data from Diadema appear to suggest that hybrids between Diadema savignyi (Michelin) and Diadema setosum (Leske), which diverged approximately 7–14 my ago, are completely viable in at least one direction. Uehara et al. (1990) reported producing adult hybrids with eggs of D. savignyi and sperm of D. setosum, but did not mention whether the offspring from the reciprocal cross survived. To the extent that such limited data can be generalized, it would appear that post-zygotic isolation in echinoids is absent in species that have diverged < 5 my ago, may be present in species that diverged between 5 and 10 my, but not in all cases. Ability of F1 hybrids to Backcross.—Survivorship of F1 hybrids would mean little if they were not able to backcross to the parental species. For obvious reasons, the data that could address this question are even scantier than those on survivorship (Table 2). Rahman and Uehara (2004) found that backcross fertilizations at limiting sperm concentrations of 10–5 are successful in all possible directions between hybrids and parentals and between all types of F1 hybrids of E. mathaei and Echinometra sp. C. Survival rates of larvae and juveniles arising from backcrosses were as high as those of non-hybrid ones. The same result was obtained by Rahman et al. (2001) about hybrids of the cross between Echinometra sp. C and Echinometra sp. A. Aslan and Uehara (1997) found fertilization rates to be low in the backcross of sperm from one type of hybrid and eggs of Echinometra oblonga (Blainville), but backcrosses involving other types of hybrids proceeded at high rates. In Strathmann’s (1981) study with Strongylocentrotus, all three hybrid individuals that survived to sexual maturity were female. These hybrids could be fertilized by sperm of Strongylocentrotus palidus (Sars), but not of Strongylocentrotus droebachiensis (O. F. Müller). Thus, these limited data would suggest that problems with backcrossing may start to arise when echinoid species have remained separated for more than a million yrs, but, even at 2–3 my, they are not sufficient by themselves to genetically isolate sympatric species. prezygotic Barriers to hybridization habitat Separation.—It could be argued that habitat separation is a geographic, rather than reproductive, isolating barrier, but a case can also be made that habitat preferences are a barrier intrinsic to the organisms themselves, because of genetic prefLESSIOS: REpROdUCTIvE ISOLATION IN EChINOIdS 195 erences for settling in a particular habitat. Either way, the evolution of such preferences will maintain isolation between sympatric species if it renders them incapable of exchanging genes because of their different ecological requirements. This is particularly true in the marine realm in which depth zonation between broadly sympatric congeneric species is frequent. how effective is this kind of barrier in echinoids? Because there are rarely more than 10 extant species in each echinoid genus, it is not particularly common for two closely related species to inhabit the same geographic region. when such range overlap exists, congeneric species often show separation in depth, or some other aspect of preferred habitat. This is the case for Echinometra both in the Indo-pacific (Tsuchiya and Nishihira, 1984; Nishihira et al., 1991; Rahman and Uehara, 2004) and in the Caribbean (Lessios et al., 1984; hendler, et al., 1995); it is also the case for most of the sympatric species of Strongylocentrotus on the west coast of North America (Lillie, 1921; Newman, 1923; pearse, 1981, 2006; Strathmann, 1981; Rogers-Bennett, 2007), for the Indo-pacific species of Diadema (McClanahan, 1988; pearse, 1998; Muthiga and McClanahan, 2007), and for the Caribbean species of Lytechinus (Lessios et al., 1984; hendler et al., 1995). however, occurrence of individuals of one species in the habitat preferred by the other is not uncommon (Lessios and Cunningham, 1990; Rogers-Bennett et al., 1995; Levitan, 2002; McCartney and Lessios, 2004; Rahman and Uehara, 2004). what is more, different habitat preferences of sympatric congeneric species are not always the rule. For example, there appears to be no obvious habitat separation among the two IndoTable 2. Rates of fertilization between hybrids, and between hybrids and pure individuals. In the designarion of the hybrids, the maternal species is listed first. Approximate times are based on mitochndrial DNA data, taken from Laundry et al. (2003). Time (my) Sperm from Eggs from Reference 1.1–1.5 Echinometra mathaei (Em) vs Echinometra sp. C (Ec) Rahman and Uehara, 2004 Em × Em Em × Ec Ec × Em Ec × Ec Em × Em high normal normal high Em × Ec normal normal normal normal Ec × Em normal normal normal normal Ec × Ec low normal normal high 1.1–1.3 Echinometra oblonga (Ed) vs Echinometra sp. A (Ea) Ed × Ed Ed × Ea Ea × Ed Ea × Ea Aslan and Uehara, 1997 Ed × Ed high high high low Ed × Ea low high high high Ea × Ed very low low very low low Ea × Ea low low low high 1.1–1.5 Echinometra sp. C (Ec) vs Echinometra sp. A (Ea) Rahman et al., 2001 Ea × Ea Ea × Ec Ec × Ea Ec × Ec Ea × Ea high normal normal high Ea × Ec normal normal normal normal Ec × Ea normal normal normal normal Ec × Ec low normal normal high 1.2–1.4 Strongylocentrotus droebachiensis (Sd) vs Strongylocentrotus palidus (Sp) Strathmann, 1981