Accumulation of Dimethylsulfoniopropionate in Geukensia demissa Depends on Trophic Interactions.

Abstract

Dimethyl sulfide (DMS) constitutes up to half of the atmospheric sulfur produced biogenically (1,2) and may affect global climate (3). A major source of atmospheric DMS is the enzymatic cleavage of dimethylsulfoniopropionate (DMSP), which is synthesized by many phytoplankters (4, 5) and a few vascular plants, including Spurtinu a/tern$oru (5). Most DMSP is released following rupture of cell walls (6, 7) and is then subject to microbial degradation to DMS (8, 9). Little attention has been given to salt marsh DMSP fluxes outside the autotrophic and microbial components of the food web. The purpose of this study was to explore DMSP pathways within the Great Sippewissett Marsh (Falmouth, Massachusetts), determining in particular whether tissue concentrations of DMSP in Geuketxiu demissu vary with food resources. G. demissa, the ribbed mussel, is the dominant animal in salt marshes in the eastern United States (10). Ribbed mussels in the Great Sippewissett Marsh filter most of the marsh water during each tidal cycle (1 1) and thus are likely to play a key role in marsh DMSP fluxes. G. demissu can directly consume S. alternijloru detritus in addition to plankton and bacteria ( 12). Peterson et al. showed that the isotopic composition of G. demissu in the Great Sippewissett Marsh follows a horizontal gradient reflecting a shift in available food resources, from a diet high in phytoplankton (up to 70%) near the bay, to mostly Spartina detritus (80%) in the marsh interior (I 3). Delta values for 34S isotope indicate that much of this shift occurs in a relatively short region along the tidal channel between sites 3 and 4 (Fig. 1). We hypothesized that mussels near the bay would have relatively high levels of DMSP due to a diet rich in live phytoplankton. Conversely, mussels in the interior of the marsh were expected to have lower levels of DMSP. Their diet is dominated by Spartina detritus, which, if directly consumed, would probably be depleted of DMSP by leaching, and if indirectly consumed via bacteria and nanozooplankton filtration, would likely be depleted of DMSP by microbial decomposition. We expected this shift in DMSP concentration to be most dramatic between site 4 and all others, following the shift in 34S delta values (Fig. 1). The mussel collection sites in Figure 1 were chosen to correspond to isotope study sites (13). Five mussels (5-7.5 cm long) were collected from each site in August 1994, and the digestive glands were analyzed separately from the rest of the body. The samples were incubated in 2N KOH in sealed vials at 25°C for 24 h, allowing DMS from hydroxide decomposition of DMSP to partition to equilibrium. Mussel DMSP content was calculated from determinations of DMS in headspace samples by gas chromatography (Chromosil 330 column, Sievers 350B Sulfur Chemiluminescence Detector). To test for differences among collection sites, the KruskalWallis nonparametric procedure was used (SPSS/PC+). Variances for DMSP/g in mussels are often high and nonnormal (14). In our study, they were also nonhomogeneous. The Kruskal-Wallis test eliminates dependency on normal distributions, homogeneity of variances, and other parametric test assumptions. In spite of small sample sizes and high variation within sites-both decreasing the likelihood of detecting differencesthe DMSP concentration in digestive glands (P = 0.01 I) and