|
|
||||||||
1 Trondhjem Biological Station, Department of Biology, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway
2 Department of Earth Sciences, University of North Carolina-Wilmington, Wilmington, North Carolina 28403
3 Department of Environmental Sciences, University of Virginia, Charlottesville, Virginia 22903
4 Oregon Institute of Marine Biology, University of Oregon, PO Box 5389, Oregon 97420
* To whom correspondence should be addressed. E-mail: Johanna.Jarnegren{at}bio.ntnu.no
The large bivalve Acesta bullisi lives permanently attached around the anterior tube opening of the cold-seep tubeworm Lamellibrachia luymesi, where it uses byssal threads and an unusually formed shell to hold the aperture and plume of the tubeworm inside its inhalant mantle cavity. The nature of this association has provoked much speculation, yet it has never been fully explained. Experiments and stable isotope data strongly suggest that Acesta bullisi is oophagous, preying upon the lipid-rich eggs (zygotes) released by the host tubeworms. Moreover, virtually all mature individuals are found on female rather than male worms.
At hydrothermal vents and hydrocarbon seeps, a variety of organisms that lack microbial symbionts feed opportunistically on other organisms that possess symbionts (1). Here we report on a non-chemoautotrophic seep bivalve, Acesta bullisi (Bivalvia; Limidae), that has evolved specialized mechanisms to exploit a unique source of nutrition derived ultimately from chemosynthesis: the energy-rich eggs of giant tubeworms. Opportunistic egg or embryo predators are common among terrestrial vertebrates (2,3) and marine invertebrates (4). Egg specialists are found in some groups of insects, including ichneumonid wasps (5) and phorid flies (6) and at least one genus of nemertean (7), but we know of no marine animals other than A. bullisi that possess extreme morphological modifications to permit egg predation as the major mode of nutrition.
Acesta bullisi Vokes, 1963, is a large (up to 11 cm in length) bivalve that surrounds the distal tube aperture of Lamellibrachia luymesi van der Land and Norrevang, 1975, a vestimentiferan polychaete (family Siboglinidae) that inhabits methane seeps between 400 and 700 m depth on the Louisiana slope, northern Gulf of Mexico (8,9). Like other siboglinids, L. luymesi has no gut, mouth, or anus; it obtains its nutrition from internal sulfur-oxidizing chemosynthetic bacteria (1013). L. luymesi is a dominant species in the seep community, forming dense bush-like aggregations, living for more than 200 years, and growing to over 2 m in length (8,14,15). Oocytes produced in the ovaries are inseminated internally and stored, pending release, in paired ovisacs at the distal ends of the oviducts (16). However, germinal vesicle breakdown, nuclear fusion, and embryonic development all occur after release from the mother (16), so the released propagules may be referred to as either "oocytes" (the nucleus is arrested at the primary oocyte stage), "zygotes" (the cytoplasm contains both egg and sperm nuclei), or "eggs" (the haploid pronuclei have not yet united to form a diploid nucleus). For the sake of clarity, we call the newly released propagules "eggs" throughout the present paper.
On dives of the submersibles Johnson Sea-Link I and II during October 2002, February and November 2003, and July 2004, we sampled the bivalves and recorded several hours of video showing the natural relationship of live A. bullisi and L. luymesi (Fig. 1A). After gentle collection with the submersible, specimens were carried to the surface in closed plastic containers, then transferred to aquaria in a 7 °C shipboard cold room, where the worms were used within 2 days and the bivalves were held for experimentation no longer than 14 days.
|
|
Instantaneous fecundities of female tubeworms were determined by dissecting the terminal portion of the paired gonoducts, removing eggs by pipette, and counting quantitative subsamples resuspended in known volumes of seawater. In a collection of 22 females of L. luymesi taken from sites where A. bullisi occurs, we found as many as 59,583 eggs (mean = 18,455, s.d. = 19,484, n = 22) stored in the ovisacs and available for release.
We determined the sexes of host worms by examining the vestimentum region for a well-known sexual dimorphism (16) and confirmed genders by microscopic examination of gametes. A. bullisi was more frequently associated with female tubeworms than with males. Although the sex ratio of the worms is unbiased (22), 86% (n = 37) of A. bullisi individuals collected more-or-less at random were found on female worms. Of the five bivalves found on male worms, all were rather small (5890 mm height), and two of them had not yet started to mold their shells around the apertures of the worms. We have noted in the laboratory that juveniles of A. bullisi are apt climbers and move about often, using their well-developed foot and byssus, and can probably change host tubeworms quite easily. Swimming has not been observed.
The eggs of L. luymesi average 105 µm in diameter (23), a full order of magnitude larger than the phytoplankton cells and seston that dominate the diets of typical filter-feeding bivalves. To determine whether A. bullisi is able to capture, manipulate, select, and ingest tubeworm eggs, we measured clearance rates of eggs by individual bivalves. Five bivalves were placed in individual aquaria, each containing 6.6 l of seawater at 7 °C and L. luymesi eggs at concentrations ranging from 8 to 18 eggs/ml. A control with no bivalve present was run simultaneously. The eggs were kept suspended by magnetic stir bars, with the speed adjusted as high as possible without disturbing the bivalves. Two 20-ml samples of seawater were taken from each tank every 30 min, and eggs were counted with a binocular dissecting microscope. After 2 to 3 h of feeding, the stomach contents and pseudofeces of the bivalves were removed and checked for the presence of eggs. Individual bivalves grazed, on average, 22% of the suspended eggs per hour (s.d. = 5.8, n = 5). Very few eggs (<2% of consumption) were found in the pseudofeces, but large numbers were found in the stomach contents of all experimental individuals, indicating that eggs were retained and ingested.
Evidence that the bivalves obtain nutrition from the tubeworms in nature comes from the stable isotopic composition of A. bullisi relative to host L. luymesi and to another locally abundant Acesta species (A. cf. excavata) not associated with the tubeworms. Stable isotope analysis was performed on several tissues collected from A. bullisi, A. cf. excavata, and adult L. luymesi, and also on L. luymesi eggs. Muscle, digestive gland, mantle, and gonad were analyzed from the two species of Acesta, and both gonads and eggs were analyzed from L. luymesi. All A. bullisi individuals analyzed for isotopes were attached to female tubeworms. The L. luymesi eggs analyzed for
34S analysis were taken from several individuals and composited to produce enough sample for analysis. The eggs were rinsed with deionized water, but the introduction of some seawater sulfate into the composite egg
34S sample proved unavoidable. All isotopes (
34S,
15N, and
13C) were determined using continuous flow isotope ratio mass spectrometry (CF-IRMS; GV Instruments) on dried and homogenized tissue samples (1). The stable isotope abundances are reported as
![]() |
iE is expressed in
(per mil). Carbon, nitrogen, and sulfur standards were calibrated relative to Peedee Belemnite (PDB), atmospheric N2, and Cañon Diablo Troilite (CDT), respectively. Analytical precision was typically better than ±0.2
.
Dietary contributions to A. bullisi from L. luymesi and from planktonic seston (a common source of nutrition to filter-feeding bivalves) were calculated from S and N isotope compositions simultaneously, according to the following equations:
|
| (1) |
![]() | (2) |
![]() | (3) |
fractionation in
15N typically observed between trophic levels (1). The "seston" isotopic endmembers for
34S and
15N were derived from tissues collected from the filter-feeding A. cf. excavata. This endmember assignment was justified because of previous determination that A. cf. excavata filter feeds on seston (24). The
34Sseston was set equal to the mean
34S found in A. cf. excavata. The
15Nseston was set equal to the mean
15N found in A. cf. excavata minus the 3
fractionation noted above. These
34S and
15N seston values were very close (0.8
and 3
, respectively) to values reported for the Gulf of Mexico (1,25) and other open-ocean sites (24). The
34S and
15N values of L. luymesi gonadal tissue were used as the "worm" isotope endmembers. Worm gonad
15N was indistinguishable from that of egg
15N (Fig. 3B). Egg
34S deviated from gonad
34S by about 12
owing to the presence of seawater sulfate introduced during the sample composite process, but were still clearly depleted in
34S and distinct from the seston and the A. cf. excavata
34S (Fig. 3A). The worm gonad isotopes were used in the calculations in lieu of direct egg values because the
34S egg values were skewed by the addition of isotopically enriched seawater sulfate that would not be readily assimilated as a sulfur nutrition source. The "worm" and "seston" endmember isotopes, and the mean isotope values of A. bullisi gonad, digestive gland, muscle, and mantle and of L. luymesi gonads were used to simultaneously solve equations 13 for fseston and fworm, using an optimization routine that minimized the difference between the observed mean isotope values for A. bullisi and those generated by solving the mixing equation (Microsoft Solver, Excel 2002). Residuals were weighted evenly between N and S isotopes. Standard deviations of the resulting fractional dietary contributions were estimated through Monte Carlo simulations (n = 100) of the solution, using one standard deviation in the observed seston, worm, and A. bullisi isotopes as the boundaries for the simulations.
|
34S,
15N, and
13C values of A. bullisi were similar to those of host tubeworms, but they were isotopically depleted relative to the typically observed range of marine seston (1,24) and the "free-living" A. cf. excavata (Fig. 3A). Oxidation of sulfide by symbiotic chemoautotrophic bacteria (13) yielded the expected, lower
34S values in L. luymesi (Fig. 3A). Similarly depleted
34S (depleted 15
to 40
relative to seawater-sulfate, seston, or A. cf. excavata) characterized the L. luymesi eggs and A. bullisi tissues. For
13C, the worms, eggs, and A. bullisi were isotopically similar to each other and distinct from A. cf. excavata (Fig. 3C). For
15N, there was good isotopic fidelity between worms and eggs, and the A. bullisi values were more enriched than the worm and egg values but isotopically depleted relative to A. cf. excavata (Fig. 3B).
The sulfur and nitrogen isotopes provided the most distinct measures of specific source contributions from L. luymesi and planktonic seston to the overall nutrition of A. bullisi. The mean and standard deviations of the measured 34S and 15N isotope compositions of A. bullisi suggest that eggs were the likely conduits transferring worm-derived biomass to this bivalve. Because of the considerable variation in egg and A. bullisi
13C, we were unable to use carbon values to calculate the dietary contribution from eggs more exactly. Nevertheless, the carbon isotopes provided some further evidence of linkage between A. bullisi and L. luymesi eggs. Worm, egg, and A. bullisi tissues had
13C isotope values that were all similar, but depleted relative to those for both common planktonic seston (1,24) and A. cf. excavata (Fig. 3B). The higher variability in egg and A. bullisi
13C is likely due to variation in content of isotopically depleted lipids in the eggs. Egg samples were harvested from L. luymesi gonads in various stages of development and fatty acid composition. Lipid extraction from a lipid-rich egg subsample enriched the overall
13C of the lipid-extracted eggs by about 7
, suggesting that lipid content is significant in determining the value and variability of
13C in the eggs. Although we have not been able to observe oophagy in situ, the stable isotopic evidence suggests that egg releases are sufficiently large or frequent to supply 72% (± 15%) of the metabolic requirement of A. bullisi individuals associated with L. luymesi.
Variability in all A. bullisi isotopes (
13C,
15N, and
34S) is influenced by the overall number of eggs consumed and the timing of egg release, neither of which is likely to be constant. We do not know the range of egg concentrations that might be present inside the mantle cavity during an egg release, what proportion of eggs escape from the mantle cavity, or how often eggs are available to be eaten. The large numbers of eggs (tens of thousands) stored in the ovisacs of most females of L. luymesi suggest that eggs might be released at high concentrations. We used fairly high concentrations of eggs in our clearance experiments, but we had no way of knowing if these concentrations were realistic. Thus, the main value of our clearance experiments was to establish that A. bullisi is able to ingest the eggs of its hosts.
The overwhelming majority of A. bullisi individuals are attached to female tubeworms, but a few are found on males. These latter individuals are generally smaller and often are not firmly attached, suggesting that they might still be capable of moving to another worm. Indeed, these bivalves are very mobile, especially as juveniles. We do not know whether individuals attached to males could obtain nutrition from sperm bundles released into the water column.
Multiple lines of evidence suggest that the Lamellibrachia/Acesta association is fueled by oophagy: (1) L. luymesi produces large numbers of lipid-rich eggs that are available throughout the year, (2) A. bullisi appears to prefer female worms over males, (3) tubeworm eggs are readily captured and ingested by the bivalves, and (4) stable isotope data indicate that A. bullisi obtains a large part of its nutrition from tubeworms. The oophagy hypothesis is also supported by recent work showing that stable isotope composition in the shell of A. bullisi shifts between the juvenile and adult stages (López Correa et al., University of Erlangen, Germany; unpubl. data); the
13C value is significantly more depleted in adults than in juveniles, whose
13C is similar to that of normal filter-feeding bivalves. A feeding shift is also apparent between the juvenile and adult portions of shells from single individuals.
Because most marine animals breed seasonally, eggs represent only a small fraction of the annual diet for most planktivorous animals (4,26). The intimate association between Acesta bullisi and Lamellibrachia luymesi was probably able to evolve because L. luymesi breeds throughout the year (16) and lives for a very long time (8,14,15). To our knowledge, the only other marine animals known to specialize in egg predation are nemertean ribbon worms of the genus Carcinonemertes that consume the brooded eggs of decapod crustaceans (7). We note that at least some of the crab species that host these nemertean egg predators are also year-round breeders (27).
| Acknowledgments |
|---|
| Footnotes |
|---|
| Literature Cited |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |