Coral Reefs )2001) 20: 301±308 DOI 10.1007/s003380100179

REPORT

Virginia M. Weis á Wendy S. Reynolds Melissa D. deBoer á Dave A. Krupp Host-symbiont speci®city during onset of symbiosis between the dino¯agellates Symbiodinium spp. and planula larvae of the scleractinian coral Fungia scutaria

Received: 19 September 2000 / Accepted: 1 June 2001 / Published online: 3 October 2001 Ó Springer-Verlag 2001

Abstract Many corals which engage in symbioses with of infection between the partners, and which results in dino¯agellates from the genus Symbiodinium )zooxan- the establishment of a speci®c symbiosis. thellae) produce o€springwhich initially lack zooxan- thellae. These species must choose their symbionts from Keywords Cnidarian á Coral á Planula larva á numerous genetically distinct strains of zooxanthellae Symbiodinium á Symbiosis á Zooxanthellae co-occurringin the environment. In most cases, symbi- osis onset results in an association between a speci®c host coral and a speci®c strain of algal symbiont. This is Introduction the ®rst study to examine host-symbiont speci®city duringsymbiosis onset in a larval cnidarian, and the ®rst Mutualistic endosymbiosis between two unrelated to examine such events in a scleractinian of any life organisms includes a stage where the larger partner or stage. We infected planula larvae of the solitary host ®rst acquires its smaller partner or symbiont Hawaiian scleractinian Fungia scutaria with both )Trench 1993; Douglas 1994, 1998). Symbionts may be homologous zooxanthellae, freshly isolated from transmitted vertically where the symbiont is passed di- F. scutaria adults, and heterologous zooxanthellae, iso- rectly from host parent to o€spring, or horizontally lated from Montipora verrucosa, Porites compressa, and where host sexual progeny must acquire symbionts from Pocillopora damicornis, three species of scleractinians the environment. Horizontal transmission o€ers the po- which co-occur with F. scutaria. We found that homol- tential ¯exibility of choosinga partner, from a variety of ogous zooxanthellae were better able to establish sym- possible candidates, which is best adapted to the condi- bioses with larval hosts than were heterologous isolates, tions at hand. There is a risk, however, that establish- by two separate measures: percent of a larval population ment of a symbiosis could fail, leavingthe non-symbiotic infected, and densities of zooxanthellae per larva. We host with reduced growth and ®tness. In contrast, verti- also measured algal densities in larvae over a 4-day cal transmission ensures that host o€springis provided period until the onset of settlement and metamorphosis. with a complement of symbionts, although these symbi- We found no changes in zooxanthella population den- onts are of limited genetic diversity, a possible dis- sities, regardless of zooxanthella type or the light envi- advantage should environmental conditions change. ronment in which they were incubated. Stronginfection Many members of the phylum Cnidaria engage in an of host larvae with homologous algae compared to endosymbiosis with photosynthetic dino¯agellates from heterologous algae suggests that there is a speci®city the genus Symbiodinium. In most cases, the dino¯agel- process which occurs sometime duringthe early stages lates, also known as zooxanthellae, reside within vacu- oles in the gastrodermal cells of the host cnidarian. Zooxanthellae contribute to host nutrition by providing V.M. Weis )&) á W.S. Reynolds á M.D. deBoer photosynthetically ®xed carbon, whereas the host pro- Department of Zoology, Oregon State University, vides )to the zooxanthellae) inorganic nutrients, a high- Corvallis, OR 97331, USA light environment, and refuge from herbivory )reviewed E-mail: [email protected] in Muscatine 1990; Muscatine and Weis 1992; Muller- Tel.: +1-541-7374359 Fax: +1-541-7370501 Parker and D'Elia 1997). Most species of cnidarians which engage in symbioses with zooxanthellae are obli- D.A. Krupp Department of Natural Sciences, gately symbiotic and have severely reduced growth, Windward Community College, Kaneohe, survivorship, and ®tness in the absence of the symbiosis HI 96744, USA )reviewed in Brown 1997a, 1997b). Despite the obligate 302 nature of these associations, the majority of cnidarian/ gastrodermal cells )Colley and Trench 1983; Fitt and algal symbioses )85%) rely on horizontal transmission Trench 1983a, 1983b). In these studies of infection, )Fadlallah 1983; Babcock and Heyward 1986; Harrison azooxanthellate hosts were challenged with both and Wallace 1990; Richmond and Hunter 1990; Rich- homologous zooxanthellae, i.e., algae obtained from the mond 1997). Many cnidarian hosts spawn azooxan- same host species, and heterologous zooxanthellae, i.e., thellate )i.e., lack zooxanthellae) gametes which are algae isolated from a di€erent species of host. In most fertilized in the water column and in turn develop into cases, some strains of heterologous zooxanthellae were azooxanthellate planula larvae. This new generation of capable of infectinghosts but, by various measures, were hosts must, at some stage, initiate a symbiosis with a new less e€ective than homologous zooxanthellae in estab- complement of zooxanthellae from the environment. lishinga stable association with the host. Other heter- Despite their relatively uniform morphology, zoo- ologous algal strains completely failed to successfully xanthellae are taxonomically highly diverse, as evidenced infect hosts. by varying physiological )e.g., Iglesias-Prieto et al. 1992; Less attention has been devoted to studyingsymbiosis Iglesias-Prieto and Trench 1997), biochemical )e.g., onset in the planula larva stage of cnidarians, despite the Schoenbergand Trench 1980a; Blank and Trench 1985; likelihood that in nature it is this life-history stage where Iglesias-Prieto et al. 1991), and molecular phylogenetic symbiosis is most often initiated. Montgomery and Kre- characteristics )e.g., Rowan and Powers 1991a, 1991b; mer )1995) found that planulae of the scyphozoan Linu- McNally et al. 1994; Baker and Rowan 1997; Wilcox che unguiculata became infected by experimentally added 1998). Recently, zooxanthellae have been divided into homologous zooxanthellae. Schwarz and coworkers four distinct taxa or clades, namely A, B, C, and D, )Schwarz 1996; Schwarz et al. 2001) found that planulae based on ribosomal RNA gene sequences and RFLP of the temperate sea anemone Anthopleura elegantissima patterns )Rowan and Powers 1991a, 1991b; Wilcox 1998; acquired zooxanthellae after feedingon host tissue which Baker 1999; LaJeunesse and Trench 2000). It is dicult contained zooxanthellae recently isolated from an adult. to generalize about the speci®city of algal clades and host We have described larval development and symbiosis taxa, in part because the study of zooxanthellae diversity onset with homologous zooxanthellae in the solitary is still in its infancy. There are numerous examples of scleractinian Fungia scutaria )Krupp 1983; Schwarz et al. host species containing a single algal clade )e.g., Rowan 1999). We determined that larval F. scutaria were com- and Powers 1991a, 1991b; Rowan 1998; Wilcox 1998), as petent to acquire zooxanthellae after development of a well as those harboringmore than one clade )Rowan and mouth, that the zooxanthellae were ingested during the Knowlton 1995; Darius et al. 1998; LaJeunesse and feedingprocess and rapidly phagocytosedby gastroder- Trench 2000). The side-by-side occurrence on reefs of mal cells, and that the zooxanthellae persisted in larvae numerous corals and other cnidarian species harboring through metamorphosis to juvenile polyps. Infection with di€erent algal clades suggests that there is a speci®city zooxanthellae was not required for metamorphosis and process which occurs at some point duringthe initiation further, azooxanthellate juvenile polyps could success- of the symbiosis, where host and symbiont select some fully acquire zooxanthellae. partners and exclude others )Douglas 1994). Indeed, this In this report we extend our studies of symbiosis process of speci®city likely extends beyond the very onset in larval F. scutaria to examine host-symbiont broad level of the clade to genetically distinct popula- speci®city and the kinetics of zooxanthellae infection. tions within single clades. Ongoing work by several We introduced homologous and heterologous zooxan- investigators is beginning to describe within-clade thellae, all members of zooxanthella clade C, to larval genetic diversity )Goulet and Co€roth 1997; Hunter et al. F. scutaria and used the proportion of larvae infected in 1997; Snell and Co€roth 1999; Wilcox et al. 1999), a population and the density of zooxanthellae within the similar to that detailed in earlier classical investigations host larvae as measures of infection success. We exam- )e.g., Schoenberg and Trench 1980a, 1980b; Trench and ined algal density in larvae through time, up to meta- Blank 1987; Banaszak et al. 1993). This diversity pro- morphosis, to determine if algal population size changed vides the opportunity for recognition and speci®city over the duration of the larval phase of the host lifecycle. duringthe initiation of the partnership. We also studied changes in algal density between pop- The role of host-zooxanthella speci®city duringthe ulations of larvae incubated in high-light and low-light onset and establishment of the symbioses has been environments to determine if algal population changes examined in the laboratory in a variety of adult )Kinzie were a€ected by irradiance level. and Chee 1979; Schoenbergand Trench 1980c; Fitt 1984; Davy et al. 1997) and juvenile )Kinzie 1974; Colley and Trench 1983; Fitt and Trench 1983a, 1983b; Colley Materials and methods and Trench 1985; Co€roth et al. 2001) cnidarian hosts. In addition, one study has examined speci®city during Gamete collection and larval cultures symbiosis onset in juveniles in the ®eld )Co€roth et al. 2001). Zooxanthellae establish residence in )infect) these All experiments were performed at the Hawaii Institute of Marine Biology on Coconut Island, Kaneohe Bay, Oahu, Hawaii. Ap- polyp stages of the host by being taken into the animal proximately 50 previously tagged specimens of F. scutaria, around duringthe feedingprocess and beingphagocytosedby 25 cm in diameter, were collected from the reef and placed in 303 seawater tables. Spawningand gametecollection has been described begin ingestion of any material they encounter )Schwarz et al. in detail previously )Krupp 1983; Schwarz et al. 1999). Brie¯y, F. 1999). scutaria generally spawns between 1,700 and 1,900 h, 2±4 days after After 4 h, zooxanthellae were still present in large quantities on the full moon from June through September. We performed our the bottom of the dish. The larvae were removed from the zoox- experiments on larvae resultingfrom spawns in Augustand Sep- anthella isolates, by concentration onto a 60-lm mesh ®lter and tember of 1998. Approximately 1 h before a predicted spawning subsequently the larvae were placed into a clean bowl of FSW. event, the corals were placed in standing, 0.45-lm ®ltered seawater Non-quantitative inspections of larvae under the compound mi- )FSW) in individual glass or plastic bowls. During spawning, ga- croscope immediately after the 4-h incubation revealed that virtu- metes were released directly into the isolated bowls containingsingle ally all larvae, regardless of which algal strain they were challenged adults. After spawning, females were removed from their bowls, with, had gastric cavities full of algae as described previously )see leaving a layer of eggs on the bottom of the dish. Within 30 min Schwarz et al. 1999). The fate of ingested zooxanthellae was not after spawning, water collected from the bowls of all spawning followed at the microscopic level in the study. In previous studies males was combined and a small volume was added to each bowl of on infection with homologous zooxanthellae, however, we found eggs. The bowls were left in seawater tables overnight for fertiliza- that as soon as 1 h post-feeding, algae were present in vacuoles in tion and early larval development. Larvae from all parental crosses the endoderm with very few algae remaining in the gastric cavity were combined and maintained in bowls in FSW which was changed )see Schwarz et al. 1999). every other day. After the summer spawning events, the tagged The followingday, three separate counts of approximately adults were returned to the reef for use in future spawns. 100 larvae each were made from each of the 12 bowls. Approxi- mately 1 ml of larvae in solution was dispensed into a 1.5-ml mi- crofuge tube and spun at 500 g for 1±2 min to gently pellet the Preparation of zooxanthella isolates larvae. The larval pellet was immediately removed with a ®ne-bore pipette and placed on a glass microscope slide. A cover slip was Zooxanthellae were isolated from freshly collected ®st-sized chunks placed on the drop and pressed onto the slide with a pencil eraser. of adult specimens of F. scutaria )de®ned as homologous zooxan- This pressure was sucient to squash the larvae, without completely thellae), Montipora verrucosa, Pocillopora damicornis, and Porites dispersingtheir contents. Dependingon the density of the larval compressa )termed heterologous zooxanthellae) by using the spray culture, each glass slide contained from 50 to 200 larvae. Larvae were from a Water Pik ®lled with FSW to remove and homogenize coral counted and scored as either symbiotic )regardless of the number of tissue. The resulting homogenate was spun in a tabletop centrifuge algae/larva) or aposymbiotic, using a compound microscope. at 2,000 g for 2 min to separate the algae from the animal tissue. The zooxanthella pellet was twice resuspended in FSW and re- pelleted by centrifugation to partially clean the algae of contami- Density of zooxanthellae in larvae over time natinganimal tissue. Followingthis, zooxanthellae were resus- and in di€erent light regimes pended into a very dense suspension in FSW before beingadded to the larval cultures. Zooxanthella isolates were used within 1 h of In a separate experiment, to compare the density of homologous vs preparation. Zooxanthellae from all four coral species, from heterologous zooxanthellae within larvae and to determine changes specimens collected in Kaneohe Bay, have been described as in numbers of zooxanthellae per host larva through time, we members of clade C )Rowan and Powers 1991a, 1991b; Baker and quanti®ed algae in larvae which had been infected with homolo- Rowan 1997). We did not independently con®rm the algal clades gous and heterologous zooxanthellae. We were also interested in for our corals for this study. determining what e€ect light level had on algal density within In preliminary experiments, we attempted to quantify the larvae, hypothesizingthat zooxanthellae in larvae kept in a high- number of algae which we added to a population of larvae. After light environment would increase in density at a faster rate than the two rinses in FSW )described above) the algae from all coral those in larvae incubated in low light. Approximately 3,000 larvae species were highly clumped and contaminated with animal debris, from a single mixed population were divided into nine separate makingquanti®cation impossible. In an attempt to obtain more ®nger bowls. Zooxanthella isolates were prepared from three dif- uniform suspensions of zooxanthella isolates and to remove animal ferent individuals of each of three species of corals, homologous debris, we performed additional and more vigorous washes. In F. scutaria, and heterologous P. compressa and P. damicornis. Each doingso, we considerably reduced the ability of the algaeto infect of these isolates was added to one of the nine bowls of larvae as the host, which resulted in inconsistent and greatly reduced per- described above. centages of larvae infected )data not shown). We therefore decided Followingrinsingand placement into fresh FSW, each of the to provide non-quanti®ed but saturatingconcentrations of algaeto nine bowls of infected larvae was further subdivided into two bowls, the larvae )see below), which resulted in consistent and repeatable one destined for a high-light treatment and the other for a low-light infection results. treatment. The ®nal experimental design therefore involved 18 bowls: zooxanthella isolates from three coral species ´ algae from three individuals per coral species ´ two light treatments. Percent of F. scutaria larvae infected with homologous High-light treatment bowls were placed in a seawater table and heterologous zooxanthellae which was exposed in the mornings to full ambient light, but which was in the shade of an overhanging eave in the afternoons. Low- In order to compare the abilities of di€erent zooxanthella types to light treatment bowls were in the same seawater table but were take up residence in F. scutaria larvae, we measured the proportion covered by dark plastic tubs which allowed in virtually no light. of larvae infected with di€erent algal types one day after an ex- The number of zooxanthellae per larva was quanti®ed for 25 perimental infection event. Approximately 10,000 larvae from a symbiotic larvae per bowl on each of 4 days followinginfection, and single mixed population were divided into 12 separate ®nger bowls. an average algal density per larva was calculated per bowl per day. Zooxanthella isolates were prepared from three di€erent individ- uals of each of four species of corals, homologous F. scutaria, and heterologous P. compressa, P. damicornis, and M. verrucosa. Statistical analyses Each of the zooxanthella isolates was added to one of the 12 bowls of larvae )i.e., three replicates of each species). An even layer Analysis of variance )ANOVA), nested ANOVA, and two-tailed of freshly isolated algae was added to the bottom of the bowls unpaired student t-tests were performed usingSAS 6.12 software. containing4-day-old larvae, which are competent to feed and ac- Data expressed as percentages were arcsin transformed before t- quire symbionts )Schwarz et al. 1999). Several drops of homoge- tests were performed. The kinetics data were logtransformed to nized Artemia were added to the bowls to stimulate a feeding meet the assumptions of homogeneity of variance, and a nested response as described in Schwarz et al. )1999). The Artemia slurry ANOVA was performed with the followinge€ects: zooxanthella prompts the swimminglarvae to drop to the bottom of the dish and isolate type, bowls of larvae nested within isolate type, light 304 treatment, and time. Student t-tests were corrected for multiple within larvae infected with homologous vs heterologous comparisons usingthe Bonferroni procedure. zooxanthella isolates. F. scutaria larvae containingho- mologous F. scutaria algae had much higher algal den- Results sities than those larvae with heterologous P. compressa or P. damicornis zooxanthellae )Fig. 2), despite very high variability in the homologous algal densities on all days Percent of Fungia scutaria larvae infected with and in both high-light and low-light treatments. The av- homologous and heterologous zooxanthellae erage number of F. scutaria zooxanthellae resident within larvae )n=24 total bowls of larvae counted, across all To compare infection success in F. scutaria larvae three replicates, four times and both light treatments) was infected with homologous vs heterologous algae, we 22.2‹5.5 )mean ‹ SD), almost 3 times higher than the sampled subsets of larval populations 1 day after expo- 8.5‹1.7 for P. compressa algae, and 5.5 times higher than sure to di€erent zooxanthella isolates, and determined the 3.9‹1.2 for P. damicornis algae. A nested ANOVA of the percentage of larvae which contained symbionts. algal density data comparing zooxanthella isolate type vs These treatments were highly signi®cantly di€erent from time vs light treatment showed a highly signi®cant one another )ANOVA, p<0.001). Nearly all larvae in- zooxanthella isolate type e€ect )p<0.001). Further, t- cubated with F. scutaria algae contained zooxanthellae, tests showed a signi®cant di€erence between the density with an average of 96‹1.5% )mean‹SD; n=three in larvae of F. scutaria algae vs P. compressa algae and vs bowls of larvae) of larvae being symbiotic )Fig. 1). High P. damicornis algae )both p<0.001). percentages of larvae incubated with P. compressa and P. damicornis algae were also symbiotic )84‹5.2 and 83‹4.4%, respectively). However, these values were Changes in algal density with time signi®cantly lower )t-test, p<0.05) than those from lar- and between light treatments vae with F. scutaria algae. In contrast, larvae incubated with M. verrucosa zooxanthellae were nearly all apo- We were interested in determiningwhether algalpopu- symbiotic, with only 3‹3.7% of larvae beinginfected lation densities within larvae changed over the duration )di€erent than F. scutaria by p<0.001). of the larval life-history stage, and whether altering light regime would a€ect algal densities. As shown in Fig. 2, Density of homologous and heterologous surprisingly, algal densities remained very stable over zooxanthellae in larvae the 4-day samplingperiod, regardless of zooxanthella isolate or treatment type )nested ANOVA time e€ect, To compare the density of di€erent algal types within F. scutaria larval hosts, we quanti®ed the number of algae

Fig. 1 Percentage of Fungia scutaria larvae symbiotic with homol- ogous F. scutaria algae and heterologous Porites compressa, Pocillopora damicornis, and Montipora verrucosa zooxanthellae Fig. 2 Densities of homologous and heterologous zooxanthellae in 1 day after experimental infection. Bars represent means ‹ SD, Fungia scutaria larvae over time and in di€erent light regimes. Data n=three bowls of larvae with 100 larvae counted per bowl. * and points represent means ‹ SD, n=three bowls of larvae with 25 *** indicate values di€erent from homologous F. scutaria percent- larvae counted per bowl. Values for the Fungia zooxanthella type age, p<0.05 and p<0.001, respectively dark treatment are slightly o€set for clarity 305 p>0.1). Similarly, there were no di€erences in zooxan- same clade, there were clear di€erences in their abilities thellae densities between larvae incubated in high-light to establish a symbiosis in F. scutaria larvae. These vs low-light treatments within any species or on any day within-clade di€erences in speci®city contribute to )nested ANOVA light treatment e€ect, p>0.1). overwhelmingbiochemical, physiological,and molecu- lar evidence )reviewed in Trench 1993, 1997; Rowan 1998) of the taxonomic diversity of zooxanthellae resi- Discussion dent within cnidarians. Further, this speci®city helps to explain the observed pattern of speci®c hosts harboring This study is the ®rst to examine speci®city events during speci®c symbionts in the ®eld )Rowan and Powers the early stages of symbiosis between scleractinian larvae 1991a, 1991b) despite the likely availability of multiple and symbiotic dino¯agellates. By two separate measures, symbionts in the surroundingenvironment. percent of larvae infected and density of zooxanthellae Our ®ndings of host-symbiont speci®city in F. scu- per larva, we found that homologous algae were better taria larvae agree with numerous previous reports of able to colonize larvae than were heterologous algae. In speci®city with algal infection in the polyp stage of the addition, we found no changes in algal densities per host lifecycle. These studies examined both repopulation larva over 4 days, regardless of the type of algae added of adult polyps which had been rendered aposymbiotic, or light regime. speci®cally several species of the anemone Aiptasia )Kinzie and Chee 1979; Schoenbergand Trench 1980c; Fitt 1984) and the temperate anemone Cereus pedun- Recognition and speci®city culatus )Davy et al. 1997), and symbiosis onset in juve- nile polyps, speci®cally scyphistomae of C. xamachana This investigation of symbiosis onset in F. scutaria lar- )Colley and Trench 1983; Fitt and Trench 1983b; Fitt vae provides evidence of recognition and speci®city be- 1984; Colley and Trench 1985) and newly settled pri- tween host and symbiont, two processes which are mary polyps of the gorgonians Pseudopterogorgia central to the biology of cooperative associations bipinnata )Kinzie 1974), P. porosa, and Plexaura kuna )Trench 1993; Douglas 1994). Under saturating condi- )Co€roth et al. 2001). It is dicult to generalize about tions, that is, with abundant zooxanthella isolates these studies, as widely varyingmeasures of infection available for infection, F. scutaria larvae became in- success were used, ranging from densities of algae per fected in higher proportions )Fig. 1) as well as with unit animal )similar to this study) to growth rates of, or greater numbers of zooxanthellae )Fig. 2) when chal- developmental changes in the host. In addition, infec- lenged with homologous compared to heterologous tions were monitored over periods ranging from a few zooxanthellae. Whereas homologous F. scutaria algae days )similar to this study) to 36 weeks. In all cases, infected nearly 100% of larvae, heterologous zooxan- however, homologous zooxanthellae appeared to es- thellae from M. verrucosa established populations in tablish stronger symbioses with hosts than did heterol- only 3% of larvae. In addition, algae from P. compressa ogous ones. For example, in repopulation studies of and P. damicornis were, respectively, 2.7 and 5.5 times Aiptasia tagetes, algae from 14 species of hosts were less numerous within larvae than were homologous al- introduced )Schoenbergand Trench 1980c). Algae from gae from F. scutaria. This di€erential ability to establish three species failed to infect anemones, results similar to a partnership suggests that there is a recognition process our infections with M. verrucosa algae in F. scutaria. The between the partners which plays a role in the ultimate remainingheterologousstrains infected, but at lower establishment of a speci®c combination of host and rates and lower densities than the homologous strain, symbiont )Trench 1987, 1993; Douglas 1994). which is again similar to our infections with P. com- This is the ®rst study which documents in detail the pressa and P. damicornis zooxanthellae. In studies of speci®city of symbiosis onset in the planula larva stage symbiosis onset in scyphistomae of C. xamachana, four of cnidarians. We reported previously on the ability of heterologous strains of algae were initially able to infect F. scutaria larvae to initiate a symbiosis with zooxan- hosts, but two of these subsequently failed to establish thellae from Aiptasia pallida )Symbiodinium pulchrorum, stable symbioses, and the other two infected but at a clade B) and Cassiopeia xamachana )S. microadriaticum, slower rate than did homologous S. microadriaticum clade A), zooxanthellae which are distantly related to )Colley and Trench 1983). In studies of symbiosis onset those from F. scutaria )clade C, Rowan and Powers in Plexaura kuna and Pseudoptergorgia porosa, Co€roth 1991a, 1991b; Baker and Rowan 1997). Based on these and coworkers )2001) found that juveniles of both gor- ®ndings, we suggested that F. scutaria exhibited a low gonian species could be infected with Symbiodinium spp. degree of speci®city )Schwarz et al. 1999). We did not, from clades A, B, and C. After beingplaced for however, quantify the infections or compare them with 3 months in the ®eld, however, the vast majority of those achieved with homologous zooxanthella isolates. juvenile polyps contained only clade B zooxanthellae, In contrast, this study quanti®es symbiosis onset when the clade which is found in ®eld-collected juveniles and usingfour zooxanthella isolates grouped in clade C and adults. The similarity in speci®city phenomena between obtained from four species of scleractinians which co- larval and adult stages of cnidarians suggests that these occur on reefs in Hawaii. Despite their occurrence in the events are not particular to a host developmental stage, 306 rather that the mechanisms drivinginitial symbiosis fornium; Banaszak et al. 1993; LaJeunesse and Trench onset in planula larvae are the same as those which oc- 2000) and rejected all other algal species as well as cul- cur in juvenile and adult polyps. tured S. californium. The very early cellular events in host-symbiont rec- Our insight into recognition and speci®city events in ognition and speci®city were not examined in this study. algal±cnidarian symbioses is further limited by our lack Future work on these initial cellular events between the of information on the nature of infective zooxanthellae partners will be required before we can gain a complete in the ®eld. For example, we do not know if source understandingof these early events in this symbiosis. It zooxanthellae are naked motile swarmers or, alterna- is worthwhile here to brie¯y discuss some of the exper- tively, freshly expelled coccoid symbionts surrounded by imental constraints which were encountered in the exe- host tissue. Despite these limitations in our under- cution of this study, how they limit our interpretation of standingof the mechanisms of recognition and speci- the data, and how they relate to other studies on the ®city, our data suggest these events occur in F. scutaria cellular nature of the zooxanthella±cnidarian recogni- and result in discrimination between algal types. tion process. In this study, all zooxanthella isolates which were used to infect larvae were highly contaminated with Kinetics of algal infection animal debris. Indeed, when we attempted to produce cleaner, less contaminated isolates, as described above in Densities of zooxanthellae in larval F. scutaria did not the Materials and methods, we found that the infection change over the 4-day sampling period, regardless of the success decreased dramatically, even in homologous al- algal type infecting the larval populations, or of the light gal infections. There are three possible explanations for level at which the larvae were incubated )Fig. 2). This these observations of decreased infection with cleaner was surprising, as we hypothesized that algae resident in zooxanthella isolates. The ®rst possibility is that the al- larvae incubated in high light would be more productive, gae are somehow compromised by vigorous cleaning, and therefore grow faster than those in low-light con- and therefore exhibit reduced infection success because ditions. Other studies measuringalgalpopulation fewer algae are viable. The second possibility is that changes in hosts over time have reported the occurrence recognition events occur not between host and symbiont of initial lagperiods, lastingseveral days after inocula- but between host and contaminatinghost source debris, tion, before algal populations started to increase )Colley for example, an animal vacuole membrane surrounding and Trench 1983; Berner et al. 1993; Davy et al. 1997). It the algae. The third possibility is that algae are ingested is possible that our short samplingperiod of 4 days, the non-speci®cally duringfeeding,and that host debris acts duration of the larval stage before the initiation of set- as a non-speci®c feedingcue which increases infection tlement, falls within such a lagphase of population success of highly contaminated algal isolates by in- growth in the algae. The duration of the planula larva creasingthe number of algaewhich enter the host. stage of F. scutaria in the ®eld is unknown. If it is similar Evidence from other studies on host feedingduring in length to that of larvae in the laboratory, our data symbiosis onset suggests that our data are explained, at indicate that signi®cant changes in symbiont±host bio- least in part, by the third possibility, i.e., zooxanthellae mass ratios do not occur in F. scutaria planulae. Static are taken up non-speci®cally duringfeedingand host zooxanthellae densities in larvae could also be explained debris acts as a feedingcue. In studies of early events in by the existence of a maximum symbiont±host biomass the symbiosis between zooxanthellae and C. xamachana, ratio, a carryingcapacity for algaewithin the host. If algal infections of host scyphistomae were less success- this were the case, zooxanthellae densities could only ful, regardless of host source, when using naked, motile increase with the growth of the juvenile polyp after cultured algae, or when using freshly isolated algae metamorphosis. Our data do not necessarily support this which had been treated with a detergent to remove host hypothesis, as variability in homologous algal densities membranes, compared to untreated, freshly isolated al- remained high throughout the 4-day sampling period. If gae )Colley and Trench 1983). These results led the au- densities reached a carryingcapacity, one would predict thors to contend )1) that the contaminatinganimal that the variability would decrease over the 4-day peri- debris was important in a non-speci®c feedingresponse od. Further investigations which monitor mitotic index which resulted in algal phagocytosis, and )2) that spec- of algae in larvae would give more information on algal i®city events between the partners are likely to happen growth and host regulation. after non-speci®c phagocytosis of potential algal sym- This study enhances our understandingof the speci- bionts. Similarly, Schwarz and coworkers )Schwarz ®city events occurringduringthe initial establishment of 1996; Schwarz et al. 2001) describe non-speci®c feeding the symbiosis between a larval scleractinian and Sym- by planula larvae of the anemone Anthopleura elegan- biodinium spp. Complementary ®eld studies, which tissima on zooxanthellae and non-symbiotic species of monitor long-term dynamics between developing corals algae. Larvae were shown to ingest all algal types in the and zooxanthella strains, such as those recently under- presence of Artemia slurry used as a feedingcue. They taken on gorgonians )Co€roth et al. 2001), would pro- ultimately retained only freshly isolated zooxanthellae vide additional information to help explain the complex from A. elegantissima adult hosts )Symbiodinium cali- patterns of speci®city observed in nature. 307

Acknowledgements This work was supported by a grant from the Ecosystems of the world: coral reefs. Elsevier, Amsterdam, pp National Science Foundation )IBN:9728405) to V.M.W. We thank 133±207 W. Mizuno for help with larval husbandry, C. Mitchelmore, Hunter CL, Morden CW, Smith CM )1997) The utility of ITS M. Bracken, and K. Nielsen for statistical support, and members sequences in assessingrelationships amongzooxanthellae and of the Weis laboratory group, three anonymous reviewers and the corals. Proc 8th Int Coral Reef Symp 2:1599±1602 editor for comments on the manuscript. Iglesias-Prieto R, Trench RK )1997) Photoadaptation, photoacc- limation and niche diversi®cation in invertebrate-dino¯agellate symbioses. Proc 8th Int Coral Reef Symp 2:1319±1324 Iglesias-Prieto R, Govind NS, Trench RK )1991) Apoprotein References composition and spectroscopic characterization of the water- soluble peridinin-chlorophyll a-proteins from three symbiotic Babcock RC, Heyward AJ )1986) Larval development of certain dino¯agellates. Proc R Soc Lond 246:275±283 gamete-spawning scleractinian corals. Coral Reefs 5:111±116 Iglesias-Prieto R, Matta JL, Robbins WA, Trench RK )1992) Baker AC )1999) The symbiosis ecology of scleractinian corals. Photosynthetic response to elevated-temperature in the symbi- University of Miami, Miami otic dino¯agellate Symbiodinium microadriaticum. Proc Natl Baker AC, Rowan R )1997) Diversity of symbiotic dino¯agellates Acad Sci USA 89:302±305 )zooxanthellae) in scleractinian coral of the Caribbean and Kinzie RA )1974) Experimental infection of aposymbiotic gorgo- eastern Paci®c. Proc 8th Int Coral Reef Symp 2:1301±1306 nian polyps with zooxanthellae. J Exp Mar Biol Ecol 15:335± Banaszak AT, Iglesias-Prieto R, Trench RK )1993) Scripsiella 345 velella sp. nov. )Peridiniales) and Gloeodinium viscum sp. nov. Kinzie RA, Chee GS )1979) The e€ect of di€erent zooxanthellae on dino¯agellate symbionts of two hybrids )Cnidaria). J Phycol the growth of experimentally reinfected hosts. Biol Bull 27:428±436 156:315±327 Berner T, Baghdasarian G, Muscatine L )1993) Repopulation of a Krupp DA )1983) Sexual reproduction and early development of sea anemone with symbiotic dino¯agellates: Analysis by in vivo the solitary coral Fungia scutaria )Anthozoa: Scleractinia). ¯uorescence. J Exp Mar Biol Ecol 170:145±158 Coral Reefs 2:159±164 Blank RJ, Trench RK )1985) Speciation and symbiotic dino¯a- LaJeunesse TC, Trench RK )2000) Biogeography of two species of gellates. Science 229:656±658 Symbiodinium )Freudenthal) inhabitingthe intertidal sea Brown BE )1997a) Coral bleaching: causes and consequences. Proc anemone Anthopleura elegantissima )Brandt). Biol Bull 8th Int Coral Reef Symp 1:65±74 199:126±130 Brown BE )1997b) Disturbance to reefs in recent times. In: McNally KL, Govind NS, Thome PE, Trench RK )1994) Small- Birkeland C )ed) Life and death of coral reefs. Chapman and subunit ribosomal DNA sequence analyses and a reconstruc- Hall, New York, pp 354±379 tion of the inferred phylogeny among symbiotic dino¯agellates Co€roth MA, Santos SR, Goulet TR )2001) Early ontogenetic )Pyrrophyta). J Phycol 30:316±329 expression of speci®city in a cnidarian-algal symbiosis. Mar Montgomery MK, Kremer PM )1995) Transmission of symbiotic Ecol ProgSer )in press) dino¯agellates through the sexual cycle of the host scyphozoan Colley NJ, Trench RK )1983) Selectivity in phagocytsosis and Linuche ungiculata. Mar Biol 124:147±155 persistence of symbiotic algae by the scyphistoma stage of the Muller-Parker G, D'Elia CF )1997) Interaction between corals and jelly®sh Cassiopeia xamachana. Proc R Soc Lond 219:61±82 their symbiotic algae. In: Birkeland C )ed) Life and death of Colley NJ, Trench RK )1985) Cellular events in the reestablishment coral reefs. Chapman and Hall, New York, pp 96±113 of a symbiosis between a marine dino¯agellate and a coelen- Muscatine L )1990) The role of symbiotic algae in carbon and terate. Cell Tissue Res 239:93±103 energy ¯ux in reef corals. In: Dubinsky Z )ed) Ecosystems of the Darius HT, Dauga CD, Grimont PAD, Chungue E, Martin PMV world: coral reefs. Elsevier, Amsterdam, pp 75±87 )1998) Diversity in symbiotic dino¯agellates )Pyrrhophyta) Muscatine L, Weis VM )1992) Productivity of zooxanthellae and from seven scleractinian coral species: restriction enzyme anal- biogeochemical cycles. In: Falkowski PG, Woodhead A )eds) ysis of small subunit ribosomal RNA genes. J Euk Microbiol Primary productivity in the sea. Plenum Press, New York, 45:619±627 pp 257±272 Davy SK, Lucas IAN, Turner JR )1997) Uptake and persistence of Richmond RH )1997) Reproduction and recruitment in corals: homologous and heterologous zooxanthellae in the temperate Critical links in the persistence of reefs. In: Birkeland C )ed) sea anemone Cereus pedunculatus )Pennant). Biol Bull 192:208± Life and death of coral reefs. Chapman and Hall, New York, 216 pp 175±197 Douglas AE )1994) Symbiotic interactions. Oxford Science Publi- Richmond RH, Hunter CL )1990) Reproduction and recruitment cations, Oxford of corals: comparisons amongthe Carribbean, the tropical Douglas AE )1998) Host bene®t and the evolution of specialization Paci®c and the Red Sea. Mar Ecol ProgSer 60:185±203 in symbiosis. Heredity 81:599±603 Rowan R )1998) Diversity and ecology of zooxanthellae on coral Fadlallah TH )1983) Sexual reproduction, development and larval reefs. J Phycol 34:407±417 biology in scleractinian corals. A review. Coral Reefs 2:129±150 Rowan R, Knowlton N )1995) Intraspeci®c diversity and ecological Fitt WK )1984) The role of chemosensory behavior of Symbiodi- zonation in coral algal symbiosis. Proc Natl Acad Sci 92:2850± nium microadriaticum, intermediate hosts, and host behavior in 2853 the infection of coelenterates and molluscs with zooxanthellae. Rowan R, Powers D )1991a) A molecular genetic classi®cation of Mar Biol 81:9±17 zooxanthellae and the evolution of animal-algal symbioses. Fitt WK, Trench RK )1983a) Endocytosis of the symbiotic dino- Science 251:1348±1351 ¯agellate Symbiodinium microadriaticum Freudenthal by endo- Rowan R, Powers D )1991b) Molecular genetic identi®cation of dermal cells of the scyphistomae of Cassiopeia xamachana and symbiotic dino¯agellates )zooxanthellae). Mar Ecol Prog Ser resistance of the algae to host digestion. J Cell Sci 64:195±212 71:65±73 Fitt WK, Trench RK )1983b) Infection of invertebrate hosts with SchoenbergDA, Trench RK )1980a) Genetic variation in Symbi- the symbiotic dino¯agellate Symbiodinium microadriaticum. odinium )=Gymnodinium) microadriaticum Freudenthal, and Endocyto Cell Res 2:675±681 speci®city in its symbiosis with marine invertebrates. I. Isoen- Goulet TL, Co€roth MA )1997) A within colony comparison of zyme and soluble protein patterns of axenic cultures of Sym- zooxanthella genotypes in the Caribbean gorgonian Plexaura biodinium microadriaticum. Proc R Soc Lond 207:405±427 kuna. Proc 8th Int Coral Reef Symp 2:1331±1334 SchoenbergDA, Trench RK )1980b) Genetic variation in Symbi- Harrison PL, Wallace CC )1990) Reproduction, dispersal and odinium )=Gymnodinium) microadriaticum Freudenthal, and recruitment of scleractinian corals. In: Dubinsky Z )ed) speci®city in its symbiosis with marine invertebrates. II. Mor- 308

phological variation in Symbiodinium microadriaticum. Proc R Trench RK )1987) Dino¯agellates in non-parasitic symbioses. In: Soc Lond 207:429±444 Taylor F )ed) The biology of dino¯agellates. Blackwell Scien- SchoenbergDA, Trench RK )1980c) Genetic variation in Symbi- ti®c, Oxford, pp 530±570 odinium )=Gymnodinium) microadriaticum Freudenthal, and Trench RK )1993) Macroalgal-invertebrate symbioses: A review. speci®city in its symbiosis with marine invertebrates. III. Endocyto Cell Res 9:135±175 Speci®city and infectivity of Symbiodinium microadriaticum. Trench RK )1997) Diversity of symbiotic dino¯agellates and the Proc R Soc Lond 207:445±460 evolution of microalgal-invertebrate symbioses. Proc 8th Int Schwarz JA )1996) Feedingbehavior and acquisition of zooxan- Coral Reef Symp 2:1275±1286 thellae by the planula larvae of the temperate sea anemone Trench RK, Blank RJ )1987) Symbiodinium microadriaticum Anthopleura elegantissima. Santa Cruz, CA. MSc Thesis, Uni- Freudenthal, S. goreauii sp. nov., S. kawagutii sp. nov. and S. versity of California pilosum sp. nov.: gymnodinioid dino¯agellate symbionts of Schwarz JA, Krupp DA, Weis VM )1999) Late larval development marine invertebrates. J Phycol 23:469±481 and onset of symbiosis in the scleractinian coral Fungia scutaria. Wilcox TP )1998) Large-subunit ribosomal RNA systematics of Biol Bull 196:70±79 symbiotic dino¯agellates: Morphology does not recapitulate Schwarz JA, Weis VM, Potts DC )2001) Feedingbehavior and phylogeny. Mol Phylogen Evol 10:436±448 aquisition of zooxanthellae by planula larvae of the sea anem- Wilcox TP, Hickok J, Sloan A )1999) Genotypic diversity among one Anthopleura elegantissima. Mar Biol )in press) algal symbionts isolated from Cassiopeia xamachana. Am Zool Snell TL, Co€roth MA )1999) Use of intersimple sequence repeats 39:121A in analyzingintra- and interspeci®c variability in scleractinian corals. Am Zool 39:122A