2002 Wilcox Etal Coral Reefs
Content
Coral Reefs (2002) 21: 198–204 DOI 10.1007/s00338-002-0221-1
NOTES
T.P. Wilcox Æ M. Hill Æ K. DeMeo
Observations on a new two-sponge symbiosis from the Florida Keys
Received: 5 April 2001 / Accepted: 13 March 2002 / Published online: 24 May 2002 Ó Springer-Verlag 2002
Keywords Symbiosis Æ Sponges Æ Florida Keys Æ Geodia Æ Haliclona
Introduction
The vast majority of sponges are filter-feeding organisms that rely on inward-directed currents for nutrition. As water moves through choanocyte chambers, sponges remove bacterioplankton and dissolved organic material (Reiswig 1971, 1983). Because of the reliance on filtration for feeding, even partial fouling of the sponge surface can exact a physiological cost. For example, several Caribbean sponge species are successfully colonized by the parasitic zoanthid Parazoanthus parasiticus (West 1979; Hill 1998a), and this colonization can reduce pumping rates of the infested sponge (Lewis 1982). Therefore, it is not surprising that sponges have evolved a number of mechanisms to prevent fouling organisms from colonizing their surfaces. One strategy involves the production of chemical anti-fouling compounds. For example, Aplysina fistularis releases the bioactive compounds aeorthionin and homoaerothionin into surrounding water, and these negatively affect spongefouling organisms (Thompson 1985; Walker et al. 1985). Given their reliance on filtration for feeding and their defensive capabilities, symbiotic interactions (sensu de Bary; see Boucher 1985) in which a sponge’s filtering surface is occluded are expected to be rare and/or
T.P. Wilcox (&) Section of Integrative Biology, School of Biological Sciences, University of Texas, Austin, Texas, 78712, USA E-mail: [email protected] Tel.: +1-512-2326283 Fax: +1-512-4713878 M. Hill Æ K. DeMeo Biology Department, Fairfield University, Fairfield, Connecticut 06430, USA
transient. Nonetheless, several sponge–sponge symbioses have been documented where an external sponge largely covers the ostia-bearing surface of an internal sponge. Sponge–sponge symbioses differ from overgrowth and other epizoic associations by the interacting sponges being found nearly exclusively in association with one another. These types of associations have been documented in the Adriatic (Ru ¨ tzler 1970), western Pacific (Kelly-Borges and Bergquist 1988; CerdaGarcia-Rojas et al. 1994; Jung et al. 1995; Sim 1996), and Caribbean (Wulff 1997a). Thus, these interactions, while not common, have been found in several highly dissimilar marine communities. Strong external factors, such as space competition or predation, are likely to play a prominent role in driving the evolution of these symbioses. However, the evolution and ecological stability of these sponge–sponge symbioses are largely unexplored. We recently encountered a two-sponge symbiosis in the Florida Keys that shares many morphological similarities with sponge symbioses of previous studies (Ru ¨ tzler 1970; Sim 1996). This association consists of an inner, spherical, white astropherid sponge covered almost completely by a thin (5-cm thick), green, outer haplosclerid sponge. Generally, only the oscula of the internal sponge are exposed (Fig. 1). This sponge association has only been found in shallow, nearshore, grassbed communities (over the 10 years we have worked on Florida Keys reefs we have never seen this sponge association on the reef proper). Additionally, during extensive surveys of sponge beds, we have never found either sponge growing in isolation. Thus, this association clearly fits the criteria for a symbiosis, and is most likely mutually beneficial. In this note, we present a preliminary taxonomic analysis of the interacting sponges, a description of the habitat where the association is found, and basic demographic data. We also compare this twosponge symbiosis with other sponge epizoic symbioses, and discuss hypotheses concerning the evolutionary and ecological significance of these associations.
199 Survey data In 1997 a single population of the interacting sponges was found and surveyed on the ocean side of the north end of Plantation Key (Fig. 2). In 1999 we conducted a broader search for populations of the interacting species. During 10 years of working on reefs and grass beds around the Florida Keys, neither senior author (T.W. and M.H.) has ever found this association on the reef proper or in deeper grass beds. Therefore, only nearshore grass-bed communities were searched. Eight additional sites, with seemingly identical habitat characteristics to the north Plantation Key site, were searched. Of these areas, only two contained populations of the interacting sponges, and these were surveyed (Fig. 2). Four to six 25-m transects, each separated by approximately 10 m, were run perpendicular to shore at a depth of approximately 1 m. We counted all two-sponge symbioses within 50 cm of either side of the transect. When we encountered an association we estimated the percentage of surface area of the internal sponge that was covered by the external sponge. Additional effort was made to locate any symbiotic sponges living solitarily. We also measured the circumference, at the widest point, of 39–50 randomly chosen symbiont pairs at the three sites with sponge–sponge symbioses. We attempted to not bias our measurements towards larger individuals by randomly choosing points in the grass bed and then measuring the nearest association to that point. The general habitat characteristics for each site were also recorded.
Results and discussion
The color of the external sponge in life is a light green. Ethanol-preserved specimens turned a dull brownishgray, and air-dried specimens were beige and brittle. In life, the external sponge is highly compressible, and often presents a growing edge arched towards the internal sponge’s osculum (Fig. 1). The external sponge is tightly attached to the internal sponge and can only be torn away with some effort. However, when thin sections were made, the external sponge often became detached. External sponge spicular material consisted entirely of hastate oxeas that were slightly bent at the middle (Fig. 3A, B). Oxeas averaged 126 lm long (SD=7.5; n=70) and were 5 lm wide at the center of the spicule (mean=4.9 lm; SD=0.46; n=70). Microscleres were completely absent. In mounted tissue sections, we found ascending spicular tracts separated by approximately one oxea length (i.e., 120–130 lm apart; Fig. 3A). Tangential ectosomal spiculation was absent, and the ends of diactinal spicules appeared to be embedded in spongin, but this was difficult to visualize in ethanolpreserved specimens. Based on these observations, we suggest that the external sponge belongs to the genus Haliclona (Haplosclerida: Chalinidae; van Soest 1980). However, we could not find any reference to a described species that matched the external sponge, indicating that this may be a previously undescribed species. The internal sponge is clearly an astrophorid in the genus Geodia (Astrophorida:Geodiidae). The oscula were organized into a relatively tight cluster located at the top/center of the sponge (Fig. 1). Megascleres and microscleres characteristic of Geodia were abundant and included orthotriaenes (Fig. 3C), sterrasters, strongylasters/strongylospherasters, and oxyasters (Fig. 4).
Fig. 1. Two-sponge association. Above An intact sponge, with only the oscula of the internal sponge visible (arrows); below same sponge in cross section. The external sponge was often found to have an arched growing edge. Diameter of sponge is 15 cm
Methods and materials
General morphological and skeletal characteristics In order to establish that the two sponges are indeed separate species, we collected seven two-sponge symbioses from 1 m depth at Plantation Key (Florida, USA) for morphological examination of the spicules. Specimens were preserved in 90% ethanol within 4 h of collection. Following the techniques described in KellyBorges and Pomponi (1992), nitric acid was used to liberate spicules from approximately 0.5 cm2 of tissue from both the internal (supporting) and external (epizoic) sponges. For our detailed spicule examination, we used a single sponge–sponge pair and sampled three areas from both the external and internal sponges. Spicules were observed and measured using a compound light microscope. To examine spicule organization at the interface of the two sponges, we embedded ethanol-preserved tissues in Paraplast, and sectioned them with a microtome at a thickness of 80–120 lm. Great care was taken to ensure that the external sponge did not float away when the embedding material was removed and sections were transferred to glass slides.
200 Fig. 2. Map of the Florida Keys showing the habitats surveyed (blue-stippled areas). Areas marked with stars represent populations surveyed. Other blue-stippled areas show similar grass-bed habitats that were searched but where no twosponge associations were found
Fig. 3. A Cross section through external sponge illustrating spicule tracts (arrows). Tracts were rarely greater than a single oxea length apart. Scale bar 100 lm. B Hastate oxeas from external sponge. Scale bar 40 lm. C Assorted spicules from internal sponge. Note cladome of orthotriaene in center of figure. Scale bar 50 lm
201
Fig. 4. A Cross section of intact association taken from ethanolpreserved material. A boundary between the two species can be seen at the arrow. B Cross section through interface showing skeletal architecture of two-sponge symbiosis along a zone of contact. a External sponge; b intersponge space; c sterraster-rich cortex of internal sponge. C Section through zone of contact showing area where megascleres from Geodia sp. protrude into body of external sponge. Scale bar in B and C 100 lm
Many sterrasters in early stages of development (exhibiting an oxyspheraster-like morphology) were also found. Spaces ( i.e., subcortical crypts) and clusters of orthotriaenes (i.e., cladomes) were found in the peripheral choanosome. Orthotriaenes and oxeas ranged from 850–1100 lm in length. There were smaller megascleres (primarily oxeas of variable size) and assorted microscleres (primarily oxyspherasters and oxyasters) between the long orthotriaenes. A few raphide and tylostyle spicules were also found. A relatively well-developed hyaline layer was present below the sterrasters. Spiral stretches of oxyasters lined canals in the endosome. Sterrasters averaged 44 lm in diameter (SD=3.5; n=50), oxyspheraster-like sterrasters averaged 27 lm (SD=7; n=20), and oxyaster diameter ranged from 17– 25 lm (n=30). Strongylasters were typically less than 10 lm. The sterrasters were organized in a dense layer (1–2 mm thick) near the surface of the sponge (see Fig. 4B,C).
Based on current descriptions, making a specific assignment for this Geodia is very difficult. Depending upon the particular character examined (such as sterraster size), the internal sponge could be assigned to Geodia gibberosa or G. neptuni. However, when the entire character suite is considered, the internal sponge cannot be unambiguously assigned to either (Wiedenmayer 1977; van Soest and Stentoff 1988; Hadju et al. 1992). This ambiguity may result from a developmental response of the internal sponge to a symbiotic relationship, obscuring its ties to a known Geodia species. Less probably, the internal sponge is a closely related, undescribed congener of G. gibberosa and G. neptuni. Clearly a more thorough taxonomic evaluation is required for both the external and internal sponges. There were also several intriguing aspects of the interface between the two sponges. Sections through ethanol-preserved sponge fragments revealed a distinct border between the interacting species (Fig. 4A). Spaces were observed between the interacting species in cross sections (Fig. 4B), but because the area of interaction may have been modified when sections were made, the actual dimensions of this space are unknown. When the external sponge was pulled away from the internal sponge, we found relatively long megascleres from the internal sponge inserted into the external sponge. This was also observed in several of the sections (Fig. 4C). Three separate populations of the two-sponge symbiosis were surveyed, one in 1997 and two more in 1999 (Fig. 2). All populations were found in shallow (<1.5 m in depth) sea-grass beds dominated by algae (Thalassia, Penicillus, and Halimeda). Anthosigmella varians, Spheciospongia vesparium, Chondrilla nucula, and Scopalina ruetzleri were common sponges. Siderastrea radians was the most common scleractinian, and various zoanthids and gorgonians were also frequently observed. The twosponge association was always found in sand patches free of dense Thalassia growth, though many of the associations were situated near or within clumps of red algae (e.g., Laurencia obtusa), with the algae sometimes obscuring the sponge association. In a few cases, algae were embedded in the oscula of the internal sponge. As can be seen in Fig. 1, the typical shape of the interacting sponges was roughly spherical, but several exhibited irregular shapes. Neither sponge was ever found living without its respective symbiont, though variation did exist in the extent of the internal sponge that was covered by the external sponge. At all three sites we estimated that approximately 45–90% of the internal sponge was covered by the external sponge, with most internal sponges only exposed around the oscula. The average circumference of associations ranged from 32– 45 cm (Table 1). The average density of the symbiosis ranged from 0.91 sponges m–2 at Plantation Key South to 0.08 sponges m–2 at Russel Key (Table 1). No representatives of the sponge–sponge symbiosis were found at six other shallow grass beds surveyed, although these beds were seemingly identical to the beds where we did
202 Table 1. Data from surveys of sponge associations at three sites in the Florida Keys (see Fig. 2). Density data are from replicate transects. Circumference was measured on randomly selected sponges not included in the transects Site Plantation Key North Plantation Key South Russel Key Density Circumference [m–2 ±1 SD (n)] [cm ±1 SD (n)] 0.65±0.24 (5) 0.91±0.52 (6) 0.075±0.03 (4) 32±9 (45) 40±10 (50) 46±16 (39) Damageda N/A 13 (26%) 3 (7.6%)
a Damaged refers to the number of sponges measured for size that exhibited signs of predator damage. We did not assess levels of damage at the Plantation Key North site
find the two-sponge association (Fig. 2). It is interesting to note that we did not find many very small individuals (the smallest association measured was 16 cm in circumference). This could be the result of infrequent recruitment events, or because we missed the recent recruits during our surveys. It is unlikely that we simply missed the smaller sponges, as we made a determined effort to find all individuals near our randomly chosen points for the size surveys. It is unclear, however, as to why recruitment would be so sporadic. Clearly, this is a question that requires further inquiry. Symbiotic sponge–sponge associations have been reported from a variety of habitats. Ru ¨ tzler (1970) documented a large number of seemingly symbiotic associations (which he termed epizoic associations) in a survey of sponge interactions in the Adriatic Sea. The degree of intimacy varied greatly among interacting species, including one association in which the host (internal) sponge (Penares helleri) was 100% covered by a symbiont (external) sponge (Antho involvens). Sim (1996) recently reported four two-sponge symbiotic associations from Komun Island, Korea, that are very similar to those reported by Ru ¨ tzler (1970) and the symbiosis presented here. A common trend in each of these reports is that most of the internal sponges belong to the Order Astrophorida. Some of these sponges have a well-defined cortex that is rich in siliceous spicules, and Ru ¨ tzler (1970) proposed that this trait may be important for the successful establishment of symbiosis. However, a spicule-rich cortex must not be a requirement for establishment of sponge–sponge symbioses, since several sponges lacking spicules were also involved in apparently symbiotic sponge–sponge interactions (Ru ¨ tzler 1970). One of the most striking features of this and other reported sponge–sponge symbioses is the large proportion of the internal sponge that is covered by the external sponge (Ru ¨ tzler 1970; Kelly-Borges and Pomponi 1992; Sim 1996). For most sponges, feeding is completely dependent upon the ability to pump large volumes of water through the choanoderm. Therefore, any obstruction of water flow should negatively affect sponge health. Yet, for the association described here, well over 40% of a typical internal sponge is covered by its external sponge.
Because of both the abundance of this sponge association and the absence of any internal sponges without their concomitant external sponges, it appears that the internal sponge can maintain adequate rates of energy input. If the internal sponge is pumping water, maintenance of flow could be accounted for via several mechanisms. First, it is possible that blockage of 45–100% of the internal sponges’ ostial-bearing surface is not high enough to reduce feeding rates below that required for maintenance, growth and reproduction. Second, it is possible that small interstitial spaces found at the interface of the two sponges might permit the internal sponge to maintain a high enough rate of water filtration. Ru ¨ tzler (1970) proposed a similar mechanism to explain the persistence of symbiotic associations in the Adriatic, particularly the association between Penares helleri and Antho involvens where 100% of the internal sponge was covered by the external sponge. Microscopic examination of the interface between the Florida Keys sponges indicates that a clear border exists between the two species. However, there is evidence of tight integration along much of the interface between the two species, and the space available for filtration must be extremely small. Finally, it is also possible that some of the material captured by the external sponge is translocated to the internal sponge, reducing the need for a large exposed filtering surface in the internal sponge. Resolving the question of host/symbiont feeding, however, must await a detailed examination of the water-flow dynamics between and within these sponge associates. Given the distributions and abundance of these associated sponges, it is clear that the interaction is stable and most likely positive in nature. We found no evidence for either sponge living in isolation, though as associates they can be one of the dominant members of the grassbed community (Table 1). Additionally, the association appeared very specific, with tight tissue integration of the interacting sponges and no evidence for any ‘rejection’ response indicative of parasite infection. Positive interactions have been largely neglected by most community ecologists (reviewed in Bertness and Callaway 1994). However, both theoretical and empirical studies have demonstrated a significant role for positive interactions under stressful or marginal ecological conditions. Under these conditions, ecological success for species that would normally compete may depend upon reciprocal interactions that ameliorate the effects of abiotic and biotic stresses (Bertness and Callaway 1994; Bertness and Hacker 1994). We propose that these stable sponge–sponge symbioses are maintained by ecological factors unique to their habitat. Specifically, extremely limited space available for colonization and an abundance of sponge predators may conspire to promote positive (mutualistic) over negative (competitive) interactions between these two sponge species. Space may be a limiting resource for the external sponge in shallow, sediment-laden grass beds,
203
to these sponge–sponge symbioses. A number of the associations in all three surveyed populations were apparently damaged by spongivores. This damage was always much more severe for internal sponge than the external sponge. In fact, it was not uncommon to find associations that had been effectively ‘cored’, with the internal sponge almost completely consumed while the external sponge was largely intact (Fig. 5). The source of this damage was clearly predation, as most of the recently damaged sponges had clear ‘bite’ marks along the remaining internal sponge. The discovery of this sponge association opens a number of avenues for future research, including examination of nutrient acquisition and studies of both the ecological and evolutionary forces shaping this intriguing symbiosis.
Acknowledgements We would like to thank the Key Largo Marine Research Laboratory for providing outstanding logistical support. C. Bucili, A. Sloan, and M. Reeve provided assistance with collections, and C. Stabile assisted with spicule examinations. The manuscript benefited from comments by A. Hill, C. Stabile, L.A. Dries, D.M. Hillis, and three anonymous reviewers. This is contribution no. 022 from the Key Largo Marine Research Laboratory.
Fig. 5. A two-sponge association damaged by predators. Above Sponge as found in situ, viewed from above; below same sponge in cross section. Approximately 10% of sponges found at the South Plantation Key site and 3% of sponges at Russel Key were clearly missing tissue that was removed by a predator (Table 1). The majority of damage was usually sustained by the internal sponge
References
Atsatt PR, O’Dowd DJ (1976) Plant defense guilds. Science 193:24–29 Bertness MD, Callaway R (1994) Positive interactions in communities. Trends Ecol Evol 9:191–193 Bertness MD, Hacker SD (1994) Physical stress and positive associations among marsh plants. Am Nat 144:363–372 Boucher DH (1985) The idea of mutualism, past and present. In: Boucher DH (ed) The biology of mutualism. Croom Helm, London, pp 1–28 Cerda-Garcia-Rojas CM, Harper MK, Faulkner DJ (1994) Unstable enol sulfates from a two-sponge association. J Nat Prod 57:1758–1761 Chanas B, Pawlik JR (1995) Defenses of Caribbean sponges against predatory reef fish. II. Spicules, tissue toughness, and nutritional quality. Mar Ecol Prog Ser 127:195–211 Green G (1977) Ecology of toxicity in marine sponges. Mar Biol 40:207–215 Hadju E, Muricy G, Custodio M, Russo C, Peixinho S (1992) Geodia corticostylifera (Demospongiae: Porifera), new astrophorid from the Brazilian coast (southwestern Atlantic). Bull Mar Sci 51:204–217 Hay ME (1986) Associational plant defenses and the maintenance of species diversity: turning competitors into accomplices. Am Nat 128(5):617–641 Hill MS (1998a) Sponges harbor genetically identical populations of the zoanthid Parazoanthus parasiticus. Bull Mar Sci 63:513– 521 Hill MS (1998b) Spongivory on Caribbean reefs releases corals from competition with sponges. Oecologia 117(1–2):143–150 Jung JH, Sim CJ, Lee C-O (1995) Cytotoxic compounds from a two-sponge association. J Nat Prod 58:1722–1726 Kelly-Borges M, Bergquist PR (1988) Sponges from Montupore Island, Papua New Guinea. Indo-Malayan Zool 5:121–159 Kelly-Borges M, Pomponi S (1992) The simple fool’s guide to sponge taxonomy. Harbor Branch Oceanographic Institution, Fort Pierce, Florida Lewis SM (1982) Sponge–zoanthid associations: functional interactions. Smithson Contrib Mar Sci 12:465–474
where potential stable attachment sites for larval sponges may be few. Therefore, the external sponge may gain considerable benefit by colonizing an internal sponge that is a better competitor for space. Likewise, sponge predators may prefer the internal sponges, and these sponges would benefit from colonization by a more toxic symbiont. Associational defenses are positive interactions that arise when consumers drive palatable organisms into interactions with less-palatable organisms (Atsatt and O’Dowd 1976; Hay 1986). In the absence of consumers, the interaction between organisms becomes competitive, but in the presence of consumers the interacting species accrue benefits. In the Florida Keys, there are a large number of predators that consume sponges, including seastars (W. Fitt, personal communication) and a variety of fish (Chanas and Pawlik 1995; Pawlik et al. 1995; Pawlik 1998; Hill 1998b). Furthermore, a large body of evidence indicates that vertebrate predators prefer Geodia spp. (Pawlik et al. 1995; Chanas and Pawlik 1995; Pawlik 1998; Wulff 1997b). Thus, the internal sponge may receive considerable benefits from colonization by a chemically defended, encrusting epibiont. The external sponge appears to be a member of the genus Haliclona and several sponges in this genus (as is true of many sponge genera) appear to be chemically defended (Green 1977; Pawlik 1998), thus the external sponge may be chemically defended (work is in progress testing this hypothesis). Further evidence of a protective role for the external sponge comes from observations of spongivore damage
204 Pawlik JR (1998) Coral reef sponges: do predatory reef fishes affect their distributions? Limnol Ocean 43:1396–1399 Pawlik JR, Chanas B, Toonen RJ, Fenical W (1995) Defenses of Caribbean sponges against predatory reef fish. I. Chemical deterrency. Mar Ecol Prog Ser 127:183–194 Reiswig HM (1971) Particle feeding in natural populations of three marine demosponges. Biol Bull 141:568–591 Reiswig HM (1983) Partial carbon and energy budgets of the bacteriosponge Verongia fistularis (Porifera: Demospongiae) in Barbados. Pubbl St Zool Napoli Mar Ecol 2:273–293 Ru ¨ tzler K (1970) Spatial competition among Porifera: solution by epizoism. Oecologia 5:85–95 Sim C (1996) A two-sponge associations from Komun Island, Korea. In: Watanabe YAF (ed) Sponge sciences: multidisciplinary perspectives. Springer, Berlin Heidelberg New York, pp 109–117 Thompson JE (1985) Exudation of biologically active metabolites in the sponge Aplysina fistularis. I. Biological evidence. Mar Biol 88:23–26 van Soest RWM (1980) Marine sponges from Curacao and other Caribbean localities. Part 2. Haplosclerida. Stud Fauna Curacao Other Caribb Isl 191:1–173 van Soest RWM, Stentoff N (1988) Barbados deep-water sponges. Stud Fauna Curacao Other Caribb Isl 70(215):1–175 Walker RP, Thompson JE, Faulkner DJ (1985) Exudation of biologically active metabolites in the sponge Aplysina fistularis. II. Chemical evidence. Mar Biol 88:27–32 West DA (1979) Symbiotic zoanthids (Anthozoa: Cnidaria) of Puerto Rico. Bull Mar Sci 29:253–271 Wiedenmayer F (1977) Shallow-water sponges of the western Bahamas. Experientia Suppl 28. Birkhauser, Basel, 287 pp Wulff JL (1997a) Mutualisms among species of coral reef sponges. Ecology 78:146–159 Wulff JL (1997b) Parrotfish predation on cryptic sponges of Caribbean coral reefs. Mar Biol 129:41–52
NOTES
T.P. Wilcox Æ M. Hill Æ K. DeMeo
Observations on a new two-sponge symbiosis from the Florida Keys
Received: 5 April 2001 / Accepted: 13 March 2002 / Published online: 24 May 2002 Ó Springer-Verlag 2002
Keywords Symbiosis Æ Sponges Æ Florida Keys Æ Geodia Æ Haliclona
Introduction
The vast majority of sponges are filter-feeding organisms that rely on inward-directed currents for nutrition. As water moves through choanocyte chambers, sponges remove bacterioplankton and dissolved organic material (Reiswig 1971, 1983). Because of the reliance on filtration for feeding, even partial fouling of the sponge surface can exact a physiological cost. For example, several Caribbean sponge species are successfully colonized by the parasitic zoanthid Parazoanthus parasiticus (West 1979; Hill 1998a), and this colonization can reduce pumping rates of the infested sponge (Lewis 1982). Therefore, it is not surprising that sponges have evolved a number of mechanisms to prevent fouling organisms from colonizing their surfaces. One strategy involves the production of chemical anti-fouling compounds. For example, Aplysina fistularis releases the bioactive compounds aeorthionin and homoaerothionin into surrounding water, and these negatively affect spongefouling organisms (Thompson 1985; Walker et al. 1985). Given their reliance on filtration for feeding and their defensive capabilities, symbiotic interactions (sensu de Bary; see Boucher 1985) in which a sponge’s filtering surface is occluded are expected to be rare and/or
T.P. Wilcox (&) Section of Integrative Biology, School of Biological Sciences, University of Texas, Austin, Texas, 78712, USA E-mail: [email protected] Tel.: +1-512-2326283 Fax: +1-512-4713878 M. Hill Æ K. DeMeo Biology Department, Fairfield University, Fairfield, Connecticut 06430, USA
transient. Nonetheless, several sponge–sponge symbioses have been documented where an external sponge largely covers the ostia-bearing surface of an internal sponge. Sponge–sponge symbioses differ from overgrowth and other epizoic associations by the interacting sponges being found nearly exclusively in association with one another. These types of associations have been documented in the Adriatic (Ru ¨ tzler 1970), western Pacific (Kelly-Borges and Bergquist 1988; CerdaGarcia-Rojas et al. 1994; Jung et al. 1995; Sim 1996), and Caribbean (Wulff 1997a). Thus, these interactions, while not common, have been found in several highly dissimilar marine communities. Strong external factors, such as space competition or predation, are likely to play a prominent role in driving the evolution of these symbioses. However, the evolution and ecological stability of these sponge–sponge symbioses are largely unexplored. We recently encountered a two-sponge symbiosis in the Florida Keys that shares many morphological similarities with sponge symbioses of previous studies (Ru ¨ tzler 1970; Sim 1996). This association consists of an inner, spherical, white astropherid sponge covered almost completely by a thin (5-cm thick), green, outer haplosclerid sponge. Generally, only the oscula of the internal sponge are exposed (Fig. 1). This sponge association has only been found in shallow, nearshore, grassbed communities (over the 10 years we have worked on Florida Keys reefs we have never seen this sponge association on the reef proper). Additionally, during extensive surveys of sponge beds, we have never found either sponge growing in isolation. Thus, this association clearly fits the criteria for a symbiosis, and is most likely mutually beneficial. In this note, we present a preliminary taxonomic analysis of the interacting sponges, a description of the habitat where the association is found, and basic demographic data. We also compare this twosponge symbiosis with other sponge epizoic symbioses, and discuss hypotheses concerning the evolutionary and ecological significance of these associations.
199 Survey data In 1997 a single population of the interacting sponges was found and surveyed on the ocean side of the north end of Plantation Key (Fig. 2). In 1999 we conducted a broader search for populations of the interacting species. During 10 years of working on reefs and grass beds around the Florida Keys, neither senior author (T.W. and M.H.) has ever found this association on the reef proper or in deeper grass beds. Therefore, only nearshore grass-bed communities were searched. Eight additional sites, with seemingly identical habitat characteristics to the north Plantation Key site, were searched. Of these areas, only two contained populations of the interacting sponges, and these were surveyed (Fig. 2). Four to six 25-m transects, each separated by approximately 10 m, were run perpendicular to shore at a depth of approximately 1 m. We counted all two-sponge symbioses within 50 cm of either side of the transect. When we encountered an association we estimated the percentage of surface area of the internal sponge that was covered by the external sponge. Additional effort was made to locate any symbiotic sponges living solitarily. We also measured the circumference, at the widest point, of 39–50 randomly chosen symbiont pairs at the three sites with sponge–sponge symbioses. We attempted to not bias our measurements towards larger individuals by randomly choosing points in the grass bed and then measuring the nearest association to that point. The general habitat characteristics for each site were also recorded.
Results and discussion
The color of the external sponge in life is a light green. Ethanol-preserved specimens turned a dull brownishgray, and air-dried specimens were beige and brittle. In life, the external sponge is highly compressible, and often presents a growing edge arched towards the internal sponge’s osculum (Fig. 1). The external sponge is tightly attached to the internal sponge and can only be torn away with some effort. However, when thin sections were made, the external sponge often became detached. External sponge spicular material consisted entirely of hastate oxeas that were slightly bent at the middle (Fig. 3A, B). Oxeas averaged 126 lm long (SD=7.5; n=70) and were 5 lm wide at the center of the spicule (mean=4.9 lm; SD=0.46; n=70). Microscleres were completely absent. In mounted tissue sections, we found ascending spicular tracts separated by approximately one oxea length (i.e., 120–130 lm apart; Fig. 3A). Tangential ectosomal spiculation was absent, and the ends of diactinal spicules appeared to be embedded in spongin, but this was difficult to visualize in ethanolpreserved specimens. Based on these observations, we suggest that the external sponge belongs to the genus Haliclona (Haplosclerida: Chalinidae; van Soest 1980). However, we could not find any reference to a described species that matched the external sponge, indicating that this may be a previously undescribed species. The internal sponge is clearly an astrophorid in the genus Geodia (Astrophorida:Geodiidae). The oscula were organized into a relatively tight cluster located at the top/center of the sponge (Fig. 1). Megascleres and microscleres characteristic of Geodia were abundant and included orthotriaenes (Fig. 3C), sterrasters, strongylasters/strongylospherasters, and oxyasters (Fig. 4).
Fig. 1. Two-sponge association. Above An intact sponge, with only the oscula of the internal sponge visible (arrows); below same sponge in cross section. The external sponge was often found to have an arched growing edge. Diameter of sponge is 15 cm
Methods and materials
General morphological and skeletal characteristics In order to establish that the two sponges are indeed separate species, we collected seven two-sponge symbioses from 1 m depth at Plantation Key (Florida, USA) for morphological examination of the spicules. Specimens were preserved in 90% ethanol within 4 h of collection. Following the techniques described in KellyBorges and Pomponi (1992), nitric acid was used to liberate spicules from approximately 0.5 cm2 of tissue from both the internal (supporting) and external (epizoic) sponges. For our detailed spicule examination, we used a single sponge–sponge pair and sampled three areas from both the external and internal sponges. Spicules were observed and measured using a compound light microscope. To examine spicule organization at the interface of the two sponges, we embedded ethanol-preserved tissues in Paraplast, and sectioned them with a microtome at a thickness of 80–120 lm. Great care was taken to ensure that the external sponge did not float away when the embedding material was removed and sections were transferred to glass slides.
200 Fig. 2. Map of the Florida Keys showing the habitats surveyed (blue-stippled areas). Areas marked with stars represent populations surveyed. Other blue-stippled areas show similar grass-bed habitats that were searched but where no twosponge associations were found
Fig. 3. A Cross section through external sponge illustrating spicule tracts (arrows). Tracts were rarely greater than a single oxea length apart. Scale bar 100 lm. B Hastate oxeas from external sponge. Scale bar 40 lm. C Assorted spicules from internal sponge. Note cladome of orthotriaene in center of figure. Scale bar 50 lm
201
Fig. 4. A Cross section of intact association taken from ethanolpreserved material. A boundary between the two species can be seen at the arrow. B Cross section through interface showing skeletal architecture of two-sponge symbiosis along a zone of contact. a External sponge; b intersponge space; c sterraster-rich cortex of internal sponge. C Section through zone of contact showing area where megascleres from Geodia sp. protrude into body of external sponge. Scale bar in B and C 100 lm
Many sterrasters in early stages of development (exhibiting an oxyspheraster-like morphology) were also found. Spaces ( i.e., subcortical crypts) and clusters of orthotriaenes (i.e., cladomes) were found in the peripheral choanosome. Orthotriaenes and oxeas ranged from 850–1100 lm in length. There were smaller megascleres (primarily oxeas of variable size) and assorted microscleres (primarily oxyspherasters and oxyasters) between the long orthotriaenes. A few raphide and tylostyle spicules were also found. A relatively well-developed hyaline layer was present below the sterrasters. Spiral stretches of oxyasters lined canals in the endosome. Sterrasters averaged 44 lm in diameter (SD=3.5; n=50), oxyspheraster-like sterrasters averaged 27 lm (SD=7; n=20), and oxyaster diameter ranged from 17– 25 lm (n=30). Strongylasters were typically less than 10 lm. The sterrasters were organized in a dense layer (1–2 mm thick) near the surface of the sponge (see Fig. 4B,C).
Based on current descriptions, making a specific assignment for this Geodia is very difficult. Depending upon the particular character examined (such as sterraster size), the internal sponge could be assigned to Geodia gibberosa or G. neptuni. However, when the entire character suite is considered, the internal sponge cannot be unambiguously assigned to either (Wiedenmayer 1977; van Soest and Stentoff 1988; Hadju et al. 1992). This ambiguity may result from a developmental response of the internal sponge to a symbiotic relationship, obscuring its ties to a known Geodia species. Less probably, the internal sponge is a closely related, undescribed congener of G. gibberosa and G. neptuni. Clearly a more thorough taxonomic evaluation is required for both the external and internal sponges. There were also several intriguing aspects of the interface between the two sponges. Sections through ethanol-preserved sponge fragments revealed a distinct border between the interacting species (Fig. 4A). Spaces were observed between the interacting species in cross sections (Fig. 4B), but because the area of interaction may have been modified when sections were made, the actual dimensions of this space are unknown. When the external sponge was pulled away from the internal sponge, we found relatively long megascleres from the internal sponge inserted into the external sponge. This was also observed in several of the sections (Fig. 4C). Three separate populations of the two-sponge symbiosis were surveyed, one in 1997 and two more in 1999 (Fig. 2). All populations were found in shallow (<1.5 m in depth) sea-grass beds dominated by algae (Thalassia, Penicillus, and Halimeda). Anthosigmella varians, Spheciospongia vesparium, Chondrilla nucula, and Scopalina ruetzleri were common sponges. Siderastrea radians was the most common scleractinian, and various zoanthids and gorgonians were also frequently observed. The twosponge association was always found in sand patches free of dense Thalassia growth, though many of the associations were situated near or within clumps of red algae (e.g., Laurencia obtusa), with the algae sometimes obscuring the sponge association. In a few cases, algae were embedded in the oscula of the internal sponge. As can be seen in Fig. 1, the typical shape of the interacting sponges was roughly spherical, but several exhibited irregular shapes. Neither sponge was ever found living without its respective symbiont, though variation did exist in the extent of the internal sponge that was covered by the external sponge. At all three sites we estimated that approximately 45–90% of the internal sponge was covered by the external sponge, with most internal sponges only exposed around the oscula. The average circumference of associations ranged from 32– 45 cm (Table 1). The average density of the symbiosis ranged from 0.91 sponges m–2 at Plantation Key South to 0.08 sponges m–2 at Russel Key (Table 1). No representatives of the sponge–sponge symbiosis were found at six other shallow grass beds surveyed, although these beds were seemingly identical to the beds where we did
202 Table 1. Data from surveys of sponge associations at three sites in the Florida Keys (see Fig. 2). Density data are from replicate transects. Circumference was measured on randomly selected sponges not included in the transects Site Plantation Key North Plantation Key South Russel Key Density Circumference [m–2 ±1 SD (n)] [cm ±1 SD (n)] 0.65±0.24 (5) 0.91±0.52 (6) 0.075±0.03 (4) 32±9 (45) 40±10 (50) 46±16 (39) Damageda N/A 13 (26%) 3 (7.6%)
a Damaged refers to the number of sponges measured for size that exhibited signs of predator damage. We did not assess levels of damage at the Plantation Key North site
find the two-sponge association (Fig. 2). It is interesting to note that we did not find many very small individuals (the smallest association measured was 16 cm in circumference). This could be the result of infrequent recruitment events, or because we missed the recent recruits during our surveys. It is unlikely that we simply missed the smaller sponges, as we made a determined effort to find all individuals near our randomly chosen points for the size surveys. It is unclear, however, as to why recruitment would be so sporadic. Clearly, this is a question that requires further inquiry. Symbiotic sponge–sponge associations have been reported from a variety of habitats. Ru ¨ tzler (1970) documented a large number of seemingly symbiotic associations (which he termed epizoic associations) in a survey of sponge interactions in the Adriatic Sea. The degree of intimacy varied greatly among interacting species, including one association in which the host (internal) sponge (Penares helleri) was 100% covered by a symbiont (external) sponge (Antho involvens). Sim (1996) recently reported four two-sponge symbiotic associations from Komun Island, Korea, that are very similar to those reported by Ru ¨ tzler (1970) and the symbiosis presented here. A common trend in each of these reports is that most of the internal sponges belong to the Order Astrophorida. Some of these sponges have a well-defined cortex that is rich in siliceous spicules, and Ru ¨ tzler (1970) proposed that this trait may be important for the successful establishment of symbiosis. However, a spicule-rich cortex must not be a requirement for establishment of sponge–sponge symbioses, since several sponges lacking spicules were also involved in apparently symbiotic sponge–sponge interactions (Ru ¨ tzler 1970). One of the most striking features of this and other reported sponge–sponge symbioses is the large proportion of the internal sponge that is covered by the external sponge (Ru ¨ tzler 1970; Kelly-Borges and Pomponi 1992; Sim 1996). For most sponges, feeding is completely dependent upon the ability to pump large volumes of water through the choanoderm. Therefore, any obstruction of water flow should negatively affect sponge health. Yet, for the association described here, well over 40% of a typical internal sponge is covered by its external sponge.
Because of both the abundance of this sponge association and the absence of any internal sponges without their concomitant external sponges, it appears that the internal sponge can maintain adequate rates of energy input. If the internal sponge is pumping water, maintenance of flow could be accounted for via several mechanisms. First, it is possible that blockage of 45–100% of the internal sponges’ ostial-bearing surface is not high enough to reduce feeding rates below that required for maintenance, growth and reproduction. Second, it is possible that small interstitial spaces found at the interface of the two sponges might permit the internal sponge to maintain a high enough rate of water filtration. Ru ¨ tzler (1970) proposed a similar mechanism to explain the persistence of symbiotic associations in the Adriatic, particularly the association between Penares helleri and Antho involvens where 100% of the internal sponge was covered by the external sponge. Microscopic examination of the interface between the Florida Keys sponges indicates that a clear border exists between the two species. However, there is evidence of tight integration along much of the interface between the two species, and the space available for filtration must be extremely small. Finally, it is also possible that some of the material captured by the external sponge is translocated to the internal sponge, reducing the need for a large exposed filtering surface in the internal sponge. Resolving the question of host/symbiont feeding, however, must await a detailed examination of the water-flow dynamics between and within these sponge associates. Given the distributions and abundance of these associated sponges, it is clear that the interaction is stable and most likely positive in nature. We found no evidence for either sponge living in isolation, though as associates they can be one of the dominant members of the grassbed community (Table 1). Additionally, the association appeared very specific, with tight tissue integration of the interacting sponges and no evidence for any ‘rejection’ response indicative of parasite infection. Positive interactions have been largely neglected by most community ecologists (reviewed in Bertness and Callaway 1994). However, both theoretical and empirical studies have demonstrated a significant role for positive interactions under stressful or marginal ecological conditions. Under these conditions, ecological success for species that would normally compete may depend upon reciprocal interactions that ameliorate the effects of abiotic and biotic stresses (Bertness and Callaway 1994; Bertness and Hacker 1994). We propose that these stable sponge–sponge symbioses are maintained by ecological factors unique to their habitat. Specifically, extremely limited space available for colonization and an abundance of sponge predators may conspire to promote positive (mutualistic) over negative (competitive) interactions between these two sponge species. Space may be a limiting resource for the external sponge in shallow, sediment-laden grass beds,
203
to these sponge–sponge symbioses. A number of the associations in all three surveyed populations were apparently damaged by spongivores. This damage was always much more severe for internal sponge than the external sponge. In fact, it was not uncommon to find associations that had been effectively ‘cored’, with the internal sponge almost completely consumed while the external sponge was largely intact (Fig. 5). The source of this damage was clearly predation, as most of the recently damaged sponges had clear ‘bite’ marks along the remaining internal sponge. The discovery of this sponge association opens a number of avenues for future research, including examination of nutrient acquisition and studies of both the ecological and evolutionary forces shaping this intriguing symbiosis.
Acknowledgements We would like to thank the Key Largo Marine Research Laboratory for providing outstanding logistical support. C. Bucili, A. Sloan, and M. Reeve provided assistance with collections, and C. Stabile assisted with spicule examinations. The manuscript benefited from comments by A. Hill, C. Stabile, L.A. Dries, D.M. Hillis, and three anonymous reviewers. This is contribution no. 022 from the Key Largo Marine Research Laboratory.
Fig. 5. A two-sponge association damaged by predators. Above Sponge as found in situ, viewed from above; below same sponge in cross section. Approximately 10% of sponges found at the South Plantation Key site and 3% of sponges at Russel Key were clearly missing tissue that was removed by a predator (Table 1). The majority of damage was usually sustained by the internal sponge
References
Atsatt PR, O’Dowd DJ (1976) Plant defense guilds. Science 193:24–29 Bertness MD, Callaway R (1994) Positive interactions in communities. Trends Ecol Evol 9:191–193 Bertness MD, Hacker SD (1994) Physical stress and positive associations among marsh plants. Am Nat 144:363–372 Boucher DH (1985) The idea of mutualism, past and present. In: Boucher DH (ed) The biology of mutualism. Croom Helm, London, pp 1–28 Cerda-Garcia-Rojas CM, Harper MK, Faulkner DJ (1994) Unstable enol sulfates from a two-sponge association. J Nat Prod 57:1758–1761 Chanas B, Pawlik JR (1995) Defenses of Caribbean sponges against predatory reef fish. II. Spicules, tissue toughness, and nutritional quality. Mar Ecol Prog Ser 127:195–211 Green G (1977) Ecology of toxicity in marine sponges. Mar Biol 40:207–215 Hadju E, Muricy G, Custodio M, Russo C, Peixinho S (1992) Geodia corticostylifera (Demospongiae: Porifera), new astrophorid from the Brazilian coast (southwestern Atlantic). Bull Mar Sci 51:204–217 Hay ME (1986) Associational plant defenses and the maintenance of species diversity: turning competitors into accomplices. Am Nat 128(5):617–641 Hill MS (1998a) Sponges harbor genetically identical populations of the zoanthid Parazoanthus parasiticus. Bull Mar Sci 63:513– 521 Hill MS (1998b) Spongivory on Caribbean reefs releases corals from competition with sponges. Oecologia 117(1–2):143–150 Jung JH, Sim CJ, Lee C-O (1995) Cytotoxic compounds from a two-sponge association. J Nat Prod 58:1722–1726 Kelly-Borges M, Bergquist PR (1988) Sponges from Montupore Island, Papua New Guinea. Indo-Malayan Zool 5:121–159 Kelly-Borges M, Pomponi S (1992) The simple fool’s guide to sponge taxonomy. Harbor Branch Oceanographic Institution, Fort Pierce, Florida Lewis SM (1982) Sponge–zoanthid associations: functional interactions. Smithson Contrib Mar Sci 12:465–474
where potential stable attachment sites for larval sponges may be few. Therefore, the external sponge may gain considerable benefit by colonizing an internal sponge that is a better competitor for space. Likewise, sponge predators may prefer the internal sponges, and these sponges would benefit from colonization by a more toxic symbiont. Associational defenses are positive interactions that arise when consumers drive palatable organisms into interactions with less-palatable organisms (Atsatt and O’Dowd 1976; Hay 1986). In the absence of consumers, the interaction between organisms becomes competitive, but in the presence of consumers the interacting species accrue benefits. In the Florida Keys, there are a large number of predators that consume sponges, including seastars (W. Fitt, personal communication) and a variety of fish (Chanas and Pawlik 1995; Pawlik et al. 1995; Pawlik 1998; Hill 1998b). Furthermore, a large body of evidence indicates that vertebrate predators prefer Geodia spp. (Pawlik et al. 1995; Chanas and Pawlik 1995; Pawlik 1998; Wulff 1997b). Thus, the internal sponge may receive considerable benefits from colonization by a chemically defended, encrusting epibiont. The external sponge appears to be a member of the genus Haliclona and several sponges in this genus (as is true of many sponge genera) appear to be chemically defended (Green 1977; Pawlik 1998), thus the external sponge may be chemically defended (work is in progress testing this hypothesis). Further evidence of a protective role for the external sponge comes from observations of spongivore damage
204 Pawlik JR (1998) Coral reef sponges: do predatory reef fishes affect their distributions? Limnol Ocean 43:1396–1399 Pawlik JR, Chanas B, Toonen RJ, Fenical W (1995) Defenses of Caribbean sponges against predatory reef fish. I. Chemical deterrency. Mar Ecol Prog Ser 127:183–194 Reiswig HM (1971) Particle feeding in natural populations of three marine demosponges. Biol Bull 141:568–591 Reiswig HM (1983) Partial carbon and energy budgets of the bacteriosponge Verongia fistularis (Porifera: Demospongiae) in Barbados. Pubbl St Zool Napoli Mar Ecol 2:273–293 Ru ¨ tzler K (1970) Spatial competition among Porifera: solution by epizoism. Oecologia 5:85–95 Sim C (1996) A two-sponge associations from Komun Island, Korea. In: Watanabe YAF (ed) Sponge sciences: multidisciplinary perspectives. Springer, Berlin Heidelberg New York, pp 109–117 Thompson JE (1985) Exudation of biologically active metabolites in the sponge Aplysina fistularis. I. Biological evidence. Mar Biol 88:23–26 van Soest RWM (1980) Marine sponges from Curacao and other Caribbean localities. Part 2. Haplosclerida. Stud Fauna Curacao Other Caribb Isl 191:1–173 van Soest RWM, Stentoff N (1988) Barbados deep-water sponges. Stud Fauna Curacao Other Caribb Isl 70(215):1–175 Walker RP, Thompson JE, Faulkner DJ (1985) Exudation of biologically active metabolites in the sponge Aplysina fistularis. II. Chemical evidence. Mar Biol 88:27–32 West DA (1979) Symbiotic zoanthids (Anthozoa: Cnidaria) of Puerto Rico. Bull Mar Sci 29:253–271 Wiedenmayer F (1977) Shallow-water sponges of the western Bahamas. Experientia Suppl 28. Birkhauser, Basel, 287 pp Wulff JL (1997a) Mutualisms among species of coral reef sponges. Ecology 78:146–159 Wulff JL (1997b) Parrotfish predation on cryptic sponges of Caribbean coral reefs. Mar Biol 129:41–52
