A Three-Stage Symbiosis Forms the Foundation of Seagrass Ecosystems

See allHide authors and affiliations

Science  15 Jun 2012:
Vol. 336, Issue 6087, pp. 1432-1434
DOI: 10.1126/science.1219973


Seagrasses evolved from terrestrial plants into marine foundation species around 100 million years ago. Their ecological success, however, remains a mystery because natural organic matter accumulation within the beds should result in toxic sediment sulfide levels. Using a meta-analysis, a field study, and a laboratory experiment, we reveal how an ancient three-stage symbiosis between seagrass, lucinid bivalves, and their sulfide-oxidizing gill bacteria reduces sulfide stress for seagrasses. We found that the bivalve–sulfide-oxidizer symbiosis reduced sulfide levels and enhanced seagrass production as measured in biomass. In turn, the bivalves and their endosymbionts profit from organic matter accumulation and radial oxygen release from the seagrass roots. These findings elucidate the long-term success of seagrasses in warm waters and offer new prospects for seagrass ecosystem conservation.

Seagrass meadows are important ecological and thus economic components of coastal zones worldwide (1, 2). In many areas, coral reefs and seagrass meadows are tightly linked habitats that form the basis for marine biodiversity (3). Seagrasses serve as a keystone habitat for migrating coral reef species as well as thousands of other animals, including waterbirds, fish, dugongs, manatees, and turtles; are important carbon and nutrient sinks; and are important to fisheries and coastline protection (13). Dense seagrass meadows attenuate currents and waves and trap pelagic and benthic organic matter in the sediment (2, 4, 5). Owing to a lack of oxygen in many coastal marine sediments, an important fraction of organic matter is decomposed by bacteria that use the abundant sulfate in seawater as an electron acceptor instead of oxygen and produce toxic sulfide as a metabolic end product (6). Although seagrasses transport oxygen into their roots and the surrounding rhizosphere (radial oxygen release) (2, 7), sulfide production outpaces oxygen release under warmer conditions, resulting in sulfide accumulation and seagrass mortality (2, 7, 8). Seagrass beds tend to accumulate organic matter, and so it is expected that seagrass beds would build up toxic sulfides and hence have a limited productivity and diversity (2). But this is not the observed case, and the underlying reason for the long-term persistence of seagrass ecosystems is an enigma (fig. S1A).

We tested the hypothesis that a three-stage symbiosis between seagrasses, associated burrowing lucinid bivalves, and their symbiotic gill bacteria contribute to reducing the cyclic build-up of sulfide (fig. S1, B to D). Paleo records suggest that the Lucinidae and their endosymbiotic relation date back to the Silurian (911), but that they increasingly diversified since the evolutionary emergence of seagrasses in the late Cretaceous (2, 12, 13). Seagrass communities later became widespread in the Eocene, and lucinid remains frequently occur in association with their deposits since (13, 14). Lucinids and their gill-inhabiting bacteria have a symbiosis in which the bivalves transport sulfide and oxygen to their gills (fig. S1D), where the bacteria oxidize sulfide for synthesizing sugars that fuel growth of both organisms (1519). We hypothesized that seagrass meadows may provide an optimal habitat for these bivalves and their symbionts by indirectly stimulating sulfide production through high organic matter input and by providing oxygen through radial oxygen release from the roots. In turn, lucinids remove sulfide, which could relieve any stress caused to seagrass growth by sulfide accumulation as organic matter is degraded (fig. S1, A and B).

Indirect support for our hypothesis was provided by a worldwide meta-analysis of 84 studies describing the fauna of seagrass beds in 83 sites covering the entire climatic distribution of seagrasses, combined with a 110-point field survey that we conducted at Banc d’Arguin, Mauritania (20). The meta-analysis reveals a relationship that covers 11 out of 12 seagrass genera (and Ruppia spp.) and at least 18 genera of Lucinidae (Fig. 1 and table S1). Only meadows of Phyllospadix spp., a seagrass genus that grows on bare rock, do not associate with Lucinidae. The association spans six out of seven continents, with bivalve densities ranging from 10 to over 1000 individuals per square meter. The bivalves were present in 97% of the tropical seagrass sites, 90% of the subtropical meadows, and 56% of the temperate seagrass beds surveyed, indicating that the association may be dependent on temperature-related sulfide production (8). Furthermore, results from our field study showed a positive correlation between seagrasses and lucinids that explained 42% of their respective variation [Pearson’s correlation coefficient (r) = 0.65] (fig. S2).

Fig. 1

Presence (green; dark points are quantitative, light points are qualitative) and absence (red) of lucinids in seagrass ecosystems based on our meta-analysis. The bivalves were present in 97% (93% of the quantitative sites) of all tropical seagrass beds, 90% (83% of the quantitative sites) of the subtropical beds, and 56% (50% of the quantitative sites) of the temperate seagrass meadows. The seagrass-lucinid association spans six out of seven continents, at least 18 genera of lucinids, and 11 out of 12 seagrass genera (and Ruppia spp.). Only meadows of Phyllospadix spp., a seagrass genus that grows on bare rock, did not contain Lucinidae. The analyzed ecosystems generally contained high (~100 individuals per square meter) to extremely high densities (>1000 individuals per square meter) of lucinids (table S1).

To experimentally test our hypothesis (fig. S1B), we investigated the effects of sulfide oxidation by the lucinid bivalve Loripes lacteus on the production of the seagrass species Zostera noltii and the potential reciprocal benefits for Loripes in a full factorial experiment under controlled conditions (20). We set up Zostera, Loripes, Zostera-Loripes, and bare sediment treatments in the top sections of 40 two-compartment columns (fig. S3), which were placed in a large seawater basin. The lower compartment of each column contained anaerobic seawater and an injection tube through which sulfide was added twice a week in half of the columns. The injected sulfide was allowed to diffuse into the top section through a porous membrane.

The presence of Loripes, and to a lesser extent of Zostera, decreased sediment sulfide levels. After 5 weeks, pore water sulfide concentrations in the top sections of the sediment controls reached about 400 μM, whereas the semiweekly addition of sulfide caused levels to increase to nearly 2700 μM (Fig. 2A). The presence of Zostera decreased sulfide levels to ~200 μM in the controls and 2200 μM in the sulfide addition treatments. In contrast, sulfide levels remained low when Loripes was present (~15 μM), even in the sulfide addition treatments. As expected, the oxygen detection depth was reduced when sulfide was added but increased when only Loripes, but not Zostera, was present because of sulfide-oxidation and intake of surface water (Fig. 2B). Zostera alone did not significantly affect sediment oxygen conditions. The joint presence of Zostera and Loripes enhanced oxygen detection depth beyond that of their separate effects.

Fig. 2

(A) Pore water sulfide concentrations and (B) oxygen detection depth after 5 weeks; error bars represent SEM (n = 5 replicates). Oxygen detection depth decreased as sulfide was added [analysis of variance (ANOVA) F1,32 = 8.9, P < 0.006]. The presence of Loripes reduced sulfide levels (repeated measures ANOVA: F1,32 = 268.8, P < 0.001) and increased oxygen detection depth (F1,32 = 125.0, P < 0.001). Reduction of the sulfide concentration by Zostera alone was less, but still significant (F1,32 = 6.8, P = 0.014). That interactions occurred between Zostera and Loripes was apparent in the oxygen measurements (F1,32 = 48.3, P < 0.001) but was also significant in the sulfide data (F1,32 = 7.8, P = 0.009). The interaction between Loripes and sulfide was significant for the sulfide measurements (F1,32 = 102.7, P < 0.001) but not for the oxygen data (F1,32 = 0.3, P = 0.578).

Our experiment showed that Zostera production is facilitated by Loripes, both in the control and in the sulfide-addition treatments. In the treatments without Loripes, sulfide addition reduced Zostera shoot biomass to 50% of the controls (Fig. 3A). Reduced shoot biomass was accompanied by decreased root biomass (Fig. 3B) and impaired phosphate uptake (20). In contrast, the addition of Loripes increased Zostera shoot biomass 1.9-fold and root weight 1.5-fold, as seen in the sulfide-addition treatments. In the treatments without additional sulfide, the presence of Loripes increased both shoot and root weight by 1.4-fold and 1.3-fold, respectively.

Fig. 3

(A) Zostera shoot and (B) root dry weight biomass per column and (C) Loripes condition expressed as the dry weight flesh/shell ratio after 5 weeks; error bars represent SEM (n = 5 replicates). Zostera biomass was reduced by means of sulfide addition (ANOVA: shoots F1,16 = 72.6, P < 0.001; roots F1,16 = 12.0, P = 0.003), whereas the presence of Loripes had a positive effect on both shoot (F1,16 = 61.3, P < 0.001) and root biomass (F1,16 = 50.2, P < 0.001). We found no significant effects on rhizome biomass. Loripes condition was positively affected by both sulfide addition (ANOVA: F1,16 = 37.3, P < 0.001) and Zostera presence (F1,16 = 9.0, P = 0.008). We also found a significant positive combined effect of the presence of Zostera and sulfide on Loripes condition (F1,16 = 5.4, P = 0.034).

Loripes condition, expressed as the flesh/shell dry weight ratio, was positively affected by sulfide addition (Fig. 3C). Furthermore, the addition of Zostera did not affect Loripes in the units to which no sulfide was added but improved the bivalve’s condition in the sulfide treatments. As hypothesized, the positive effect of Zostera on Loripes seems to result from radial oxygen release from the seagrass roots (fig. S1B). Although sulfide was almost completely removed in all Loripes treatments (Fig. 2A), the bivalve was less able to profit from the addition of sulfide in the absence of Zostera (Fig. 3C). This indicates that at least in the Loripes units without seagrass, sulfide was not completely oxidized by the symbiotic bacteria because of oxygen limitation.

Overall, our results confirm our hypothesis that a three-stage symbiosis between seagrass, lucinids, and sulfide-oxidizing bacteria reduces sulfide stress in seagrass meadows. Even though radial oxygen release by Zostera noltii and of seagrasses in general is limited (21, 22), Loripes in our experiment clearly benefitted from the increased oxygen input in the sediment. In the field, the positive effects of seagrasses on lucinids are not confined to sediment oxygenation alone but also by indirectly stimulating sulfide production and releasing dissolved organic molecules (2, 18). The positive effects of Loripes on Zostera in our experiment could not be explained by differences in nutrient availability (20). Plants were not nutrient-limited, but both Zostera and Loripes significantly lowered dissolved ammonium and phosphorus in the sediment pore water, whereas sulfide addition increased nutrient availability (fig. S4). We found that in our experiment, the negative effects of sulfide addition on Zostera biomass could not fully be prevented by Loripes addition (Fig. 3A), despite the removal of almost all sulfide by Loripes after 3 days. As the observed experimental effects could not be attributed to differences in nutrient availability, this is most likely caused by the pulsed nature of our sulfide supply. This may have led to short periods of exposure of Zostera to toxic sulfide levels.

Coastal ecosystems, and seagrass meadows in particular, are currently declining at an alarming and increasing rate worldwide, leading to loss of biodiversity (1). Extensive restoration efforts have had little success so far (<30%), despite their extremely high costs (±$100,000 per hectare) (23). Similar to the function of mycorrhizae, pollinators, or seed dispersers in terrestrial systems (2426), our findings indicate that restoration efforts should not only focus on environmental stressors such as eutrophication, sediment run-off, or high salinity as a cause of decline but should also consider internal ecological interactions, such as the presence and vigor of symbiotic or mutualistic relations. Breakdown of symbiotic interactions can affect ecosystem functioning, with bleaching events in coral reefs as a clear example (27). Similar to the well-known symbiosis between corals and their unicellular algal endosymbionts (28), we conclude that symbioses, rather than one defining species, forms the foundation of seagrass ecosystems.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S4

Tables S1 and S2

References (29119)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank G. Quaintenne and H. Blanchet for their help with the collection of Loripes; J. Eygensteyn and E. Pierson for technical assistance; and G. J. Vermeij, H. de Kroon, T. J. Bouma, E. J. Weerman, and C. Smit for their comments on the manuscript. T.v.d.H. was financially supported by the “Waddenfonds” program; M.v.d.G. and T.P. by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO)–WOTRO Integrated Programme grant W. awarded to T.P.; and J.d.F. and J.v.G. by the NWO–VIDI grant 864.09.002 awarded to J.v.G. B.S. was supported by an NSF CAREER award, the Andrew Mellon Foundation, and the Royal Netherlands Academy Visiting Professorship. The authors declare no conflicts of interest. A detailed description of all materials and methods, sources, as well as supplementary information are available as supplementary materials. The data are deposited in DRYAD at
View Abstract

Navigate This Article