INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Traditionally, the only components of the dimethylsulfide (DMS) emission process in the marine environment studied seriously by biological oceanographers were the phytoplankton and macroalgae (1, 3, 6, 14, 18, 19). Over the past few years, the roles of bacteria (11, 12, 20, 28), protozoa (microzooplankton), and metazoans (50, 52, 54) have been viewed as important to this process. More recently it was discovered that a fungus, Fusarium lateritium, also degrades dimethylsulfoniopropionate (DMSP; [CH₃]₂S⁺---CH₂---CH₂---COO) and releases DMS (2, 5). The enzyme responsible for this degradation is DMSP lyase, and its activity in a fungus adds another major phylogenetic group to the participants in the DMS-emitting portion of the global sulfur cycle. This DMSP lyase-producing fungus was isolated from the salt marsh ecosystem, where more DMS is emitted per unit area than from any other ecosystem (44). DMS emissions from the salt marsh occur mainly over the areas populated by cordgrasses, such as Spartina alterniflora (44), which contains high DMSP levels (80 to 280 µmol g¹ [dry weight]) (9, 27, 41). While the details of the DMS emission process in salt marshes remain unclear, it is known that the rate of emissions increases during leaf decay (9). Furthermore, it has been well established that the main mediators of cordgrass shoot decay are fungi (32). These Spartina decomposers contribute to nutrient cycling in the salt marsh via lignocellulolysis and immobilization of plant nitrogen during the decay process (32, 33). Although it has been suggested that lignocellulolytic fungi are capable of using organic material found in living Spartina shoots (33), no studies have been done to investigate the possibilities that these fungi also degrade DMSP and cause the increased DMS emissions during leaf senescence and decay. This paper reports the distribution of DMSP lyase-producing fungi associated with DMSP-producing, halophytic plants in the salt marsh and suggests that fungi play a significant role in DMS emissions in decaying cordgrass.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolates of fungi and oomycotes. Nineteen species of true fungi (eumycotes; kingdom Fungi) and four species of oomycotes (species of the genus Halophytophthora; kingdom Protoctista or Chromista) (Table 1) were obtained from ascospores, conidia, or zoospores isolated from decaying salt marsh grass or mangrove leaves (25). All of the fungi obtained from marsh grasses were tested for DMSP lyase activity immediately following isolation (except Phaeosphaeria spartinicola from Spartina anglica, isolated in 1991, and Buergenerula spartinae I, isolated in 1982). The oomycotes and fungi from mangrove and other tree leaves (i.e., the mangroves Rhizophora mangle, Rhizophora stylosa, and Avicennia germinans and wild cherry, Prunus serotina) were isolated from 1985 to 1996 and maintained on pieces of autoclaved natural substrate atop dilute (2% [vol/vol]) V-8 agar (25), with transferring no more than twice per year.

Screening for fungal DMS emissions. The fungi and oomycotes were inoculated into a basal salt medium (2) containing either 1 mM DMSP or DMSP supplemented with either 2 mg of -sitosterol (Sigma) per liter alone or -sitosterol and 20 mg of yeast extract per liter. The fungi and oomycotes were inoculated on slants of this medium in 15-ml vials that were sealed with Teflon-coated rubber caps and aluminum crimps once the fungi had begun to grow into the agar. Gas samples (50 µl) from these vials were removed 4, 5, 6, and 8 days after inoculation and injected into a gas chromatograph (GC-8A flame ionization detector; Shimadzu) to test for the presence of DMS (11). F. lateritium cultures (2) and uninoculated media served as positive and negative controls, respectively. The initial screening of fungal species for DMSP lyase activity was performed in a blind assay in which the individual performing the analysis was unaware which strains came from DMSP-producing plants.

DMSP lyase activity. To challenge the screening results for those strains testing positive for DMSP lyase activity, cultures were incubated in liquid medium (tryptic soy broth [TSB] diluted 1:1 with seawater and containing 25 µg of chloramphenicol ml¹) for 1 to 2 weeks in order to obtain sufficient biomass for the assays. During the growth period, the cultures were regularly checked for the presence of bacteria by plating out aliquots onto tryptic soy agar. For routine DMSP lyase assays, fungi were washed to remove all traces of medium and resuspended in buffer (50 mM PO₄ [pH 7.0] diluted 1:1 in seawater). To vent any ethanol produced by these strains (ethanol produces a peak on our gas chromatograph close to that of DMS), it was necessary to incubate them for 1 h, and then rinse and resuspend the biomass in buffer. DMSP (1 mM) was added to these fungal suspensions and DMS emissions were monitored. Suspensions with no added DMSP served as controls; this control was particularly important because fungi can produce and release DMS via pathways that do not include cleavage of DMSP (46).

Acrylate mineralization. The fungi were inoculated onto slants of the basal salt medium (2) containing 20 mg of yeast extract per liter plus or minus 5 mM acrylate in 15-ml vials that were sealed with Teflon-coated rubber caps and aluminum crimps once slight growth was apparent. After 1 week of incubation, gas samples were removed and carbon dioxide was analyzed on a Carle Instruments gas chromatograph as described previously (13). Integrator counts of CO₂ from vials containing only yeast extract were subtracted from those supplemented with acrylate before conversions to molar concentrations of CO₂ were calculated. One-milliliter aliquots of analytical grade CO₂ served as standards.

DMSP lyase induction. Either DMSP or glycine betaine (2 mM) was added to buffered cell suspensions. After 1.5 h, the fungal mycelia were filtered, thoroughly washed, and resuspended in fresh buffer and 1 mM DMSP was added as a substrate to measure the rate of lyase activity. DMS emissions were monitored every 15 min for 2 h.

Protein. At the end of each experiment the fungal biomass from each vial was removed by filtration, dried, put into 1 ml of 2 mM NaOH, and boiled for 5 min. Protein was estimated with the Bradford reagent (BioRad, Hercules, Calif.).

DMS uptake. To test all the strains shown in Table 1 for DMS uptake activity, DMS was added to 1 ml of buffered mycelial suspensions in 15-ml vials, which were immediately sealed. DMS was allowed to equilibrate with the liquid phase for 1 h, and then gas samples were removed every 2 h and analyzed. The two fungi, Mycosphaerella sp. II strain FSU and D. salina, which evolved DMS from added DMSP and took up DMS were further analyzed. Because a good kinetic analysis of DMS emission and DMSP uptake was difficult to obtain with these strains, a sulfur balance experiment was performed. DMSP (2 mM) was added to buffered suspensions of the fungi, and DMS emissions were monitored every 15 min until the DMS in the gas phase disappeared. Three hours after DMS emissions ceased, the fungi were separated from the suspension medium by filtration and 4 M NaOH was added to both fractions to completely cleave any remaining DMSP to DMS. DMS was then measured in both the fungal fraction and the medium as a way of quantitating DMSP.

Chemicals. DMSP was synthesized from DMS and acrylate according to the method of Chambers et al. (7). DMSP was standardized against the pure commercial reagent obtained from Research Plus, Bayonne, N.J. Acrylate and DMS were obtained from Aldrich, Milwaukee, Wis. Glycine betaine was purchased from Sigma.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

DMSP lyase producers. The correlation between fungi which have DMSP lyase activity and their plant origins is seen in Table 1. Seventeen of the 36 strains tested had DMSP lyase activity, and most of these (13 of 17) were residents of DMSP-producing plants. Among decomposers of S. alterniflora tested, only the two isolates of Lachnum spartinae (strains I and II) had no DMSP lyase activity. A principal fungal inhabitant of dead leaves of this grass, the ascomycete P. spartinicola, occupying an average of 89% of the leaf area (36), was able to degrade DMSP with the subsequent release of DMS. None of the oomycotes tested had DMSP lyase activity, and they exhibited no growth on medium containing DMSP alone. All DMS emissions from these fungal suspensions were dependent on added DMSP. Aliquots of the fungal and oomycote suspensions were plated onto nutrient agar after the experiment to ensure that a pure culture was involved. In all cases, only the species of interest was present. DMSP lyase-specific activity for all fungal isolates except two is shown in Table 1. Different fungi had markedly different enzyme activities, with the highest activity being found in a S. alterniflora dweller, Phaeosphaeria neomaritima II, and the lowest activity being found in isolates of Tremateia halophila, which live on a non-DMSP-producing plant. The specific activities for DMSP lyases from Mycosphaerella sp. II strain FSU and D. salina, which were obtained from S. alterniflora and R. mangle (red mangrove), respectively, were not determined because the DMS produced was simultaneously taken up by these fungi (see below).

DMSP lyase induction. Two general patterns of DMSP lyase induction, each represented by a number of the fungal species, are indicated in Table 1. All DMS-producing isolates were induced by DMSP, but some were also induced by glycine betaine. Kinetics representative of all species having DMSP lyase activity following induction are shown in Fig. 1. B. spartinae (and the other species identified in Table 1, footnote c) was induced by both DMSP and glycine betaine. The following additional observations were made (data not shown) concerning the induction of DMSP lyase by various fungal species: (i) Phaeosphaeria halima, a S. alterniflora inhabitant, appeared to express DMSP lyase activity constitutively; (ii) in both strains of Passeriniella obiones, glycine betaine induced DMSP lyase faster than did DMSP; (iii) P. spartinae, Phoma sp., Tremateia halophila, and B. spartinae required a longer (4-h) induction period before enzyme activity was seen; and (iv) preincubation with glycine betaine seemed to delay (by ~1 h) enzyme induction by DMSP in P. spartinicola.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Kinetics of DMSP lyase activity by the marine fungus B. spartinae II. These results are representative of those fungi whose DMSP lyases were inducible by DMSP alone or by either DMSP or glycine betaine. The glycine betaine-inducible strains are identified in Table 1. At time zero, 1 mM DMSP was added to washed cell suspensions that were either uninduced () or induced with 2 mM DMSP () or 2 mM glycine betaine (). This experiment was repeated twice for each fungal strain, and the data shown here are representative of the results.

DMS uptake. During the DMSP lyase induction experiments, it was noted that when DMSP was added to suspensions of Mycosphaerella sp. II strain FSU and D. salina, they produced a small amount of DMS, which rapidly disappeared. It appeared that these strains were utilizing the DMS they had just produced. Thus, the fungi were tested for DMS uptake, and of all the fungal isolates seen in Table 1, only Mycosphaerella sp. II strain FSU and D. salina had this activity. Figure 2 shows the kinetics of the DMS uptake experiments for Mycosphaerella sp. II strain FSU; D. salina, at equivalent protein concentrations, showed similar results. Therefore, these fungi not only degrade DMSP, releasing DMS, but also recycle the DMS produced. The addition of 5 µg of nystatin ml¹ completely inhibited the DMS uptake system, suggesting an active (energy-dependent) process.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   DMS utilization by Mycosphaerella sp. II strain FSU. DMS was added to cell suspensions at time zero, and its disappearance was monitored by gas chromatography. Mycosphaerella suspensions contained 5 µg of nystatin ml1 () or no addition (). A suspension of P. halima, representing those fungi that do not take up DMS, served as a control (). The data presented are representative of two experiments.

Acrylate mineralization. All species of fungi having DMSP lyase activity (Table 1) were tested for their ability to grow on and mineralize acrylic acid (CH₂CH₂COO). All DMSP lyase-containing species, including the two which also utilized DMS, grew on the acrylate and oxidized it to CO₂ (range = 0.5 to 2.6 µmol of CO₂ produced bottle¹ day¹ [mean = 1.6]).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

DMS emission to the atmosphere has the potential to lead to changes in the albedo effect and thereby to influence global climatic patterns (8). Reported rates of emission of DMS from smooth cordgrass (S. alterniflora) salt marshes range up to tens of micromoles m² h¹ (31), greater than rates of emission from noncordgrass marshes or through the air-water interface above suspended phytoplankton (26, 31). The shoots of smooth cordgrass, which, when living, contain high levels of DMSP (up to >100 µmol g¹ [41]), are the principal sources of DMS from the cordgrass marsh (9). Rates of DMS release from smooth cordgrass shoots peak in September and October (45), at the time when the death rate of smooth cordgrass leaves is also at a maximum (10, 38, 42).

The autumnal maxima for DMS release and the rate of appearance of standing dead shoot parts suggest that the standing dead shoot parts are the source of DMS (9). S. alterniflora could possess DMSP lyase activity of its own, so the release may be partly due to activity of a DMSP lyase in senescent shoots, but an alternative or additional explanation is now available: the ascomycetous-fungal decomposers of standing cordgrass shoot parts, strong secondary producers within the standing dead material (37), are responsible for the breakdown of cordgrass DMSP via fungal DMSP lyase activity. Several of the smooth cordgrass ascomycetes in Table 1 (i.e., P. spartinicola, B. spartinae, and P. obiones), with substantial DMSP lyase activity (3 to 16 nmol of DMS min¹ mg of fungal protein¹), are among the established secondary producers (40). The rates of DMSP lyase activity that we observed for mycelium grown in rich, proteinaceous liquid medium could be low compared to rates for mycelium inside the natural solid substrate. The fact that we observed growth rates of fungi on solid DMSP medium that were approximately fourfold higher than those on liquid medium suggests that our measured rates of DMSP lyase activity in liquid may also be low. DMSP would be expected to be present primarily in the cell cytoplasm (16), and early in their digestion of cordgrass tissue, cordgrass ascomycetes completely destroy the cytoplasmic contents of leaf cells (see Fig. 2 of reference 40). The fact that decomposer fungi of DMSP-containing plants have the potential to be involved in release of DMS from dead plant material has implications beyond the salt marsh ecosystem, e.g., for sugarcane fields (41a).

Given the distinct possibility that salt marsh ascomycetes cause the release of a portion of the DMS from cordgrass shoots, it is possible that the methods used to measure DMS release from cordgrass marshes have yielded underestimates of natural release rates. The activity of cordgrass ascomycetes is strongly dependent on the water content of standing decaying cordgrass shoots (39). In order to measure DMS release from marshes, chambers have been placed over cordgrass shoots (9, 31, 45). Dacey et al. (9) pointed out that the enclosure chambers could cause inaccuracies in DMS flux estimates due to CO₂ stripping. Although the chambers would be expected to maintain high internal humidity, they could also prevent water contact (e.g., tides, rain, fog, and dew) with dead shoot material, thereby potentially limiting the rate of fungal DMSP lyase activity. Note that in the work of Morrison and Hines (31), when sediments were flooded, DMS rates increased. This increase was attributed to deterrence of DMS flux to sediments by the water cover (9, 31), but the increase could also have been partially due to tidal wetting of the least-elevated decaying blades, permitting enhanced fungal activity.

We screened only one oomycote (H. vesicula) from the salt marsh ecosystem (the strain from P. serotina leaves [Table 1]). The logic for this limited screening was that oomycotes do not appear to have a substantial presence in decomposing leaves of marsh grasses (33). Neither this salt marsh oomycote nor any of the other, mangrove ecosystem oomycotes that we screened exhibited DMSP lyase activity (Table 1). This finding might be viewed as counterintuitive, since the oomycotes have an evolutionary lineage that associates them with microalgae (15), some species of which are well established producers of DMSP and DMSP lyase (1, 19). However, since (i) the oomycotes are believed to be best adapted to growth in fallen leaves (especially those of mangroves) (33), (ii) at least one mangrove species (R. mangle) has been found not to accumulate DMSP in its leaves (9), and (iii) oomycotes commonly utilize proline as the compatible solute when faced with conditions of low water potential (30), perhaps the absence of DMSP lyase in marine oomycotes is to be expected.

Most of the fungi in which we detected DMSP lyase activity are known to inhabit only the decaying parts of shoots of the DMSP-containing plants listed in Table 1. One exception to the restriction to DMSP-producing hosts may be P. neomaritima, which may also occur in non-DMSP-producing grasses (23). However, in a recent exhaustive examination of a reported non-DMSP-containing host, J. roemerianus, Kohlmeyer et al. (reference 24 and references therein) did not find P. neomaritima. A second exception is D. salina, a common inhabitant of intertidally deposited marine substrates of many kinds, including drift seagrasses and macroalgae (34, 35, 43), which are known producers of DMSP (9, 16). The two additional exceptions are the Phoma sp. and T. halophila that we obtained from the non-DMSP-producing grasses S. patens and J. roemerianus, respectively. The explanation for these two exceptions is not obvious. Since DMSP is a substrate analog of other readily usable carbon sources found in non-DMSP-producing grasses, the possibility exists that this lyase acts upon other substrates.

Only one of our screened fungal species that are restricted to decomposition of standing dead smooth cordgrass was found not to possess DMSP lyase: L. spartinae (Table 1). One potential reason for this might be that L. spartinae is found on older decaying leaf sheaths (4); it may be that by the time that L. spartinae is most active, succeeding other leaf sheath species such as P. neomaritima, all DMSP from the dead leaf sheaths has been lost. Furthermore, L. spartinae is an apothecial ascomycete, a taxon quite distinct from those of the other S. alterniflora and S. foliosa ascomycetes, which are pseudothecial or perithecial ascomycetes (17).

The utilization of DMSP by suspensions of the fungi Mycosphaerella sp. II FSU and D. salina, and the subsequent appearance and disappearance of the DMS from these cultures (Table 2), appears to be the first evidence of DMSP lyase and DMS uptake activities occurring in the same organism. Presumably the DMS is utilized as an additional carbon and energy source by these cells. We have not, however, studied DMS utilization by these fungi in the absence of DMSP catabolism. Both DMSP catabolism to DMS and DMS uptake are common phenomena in natural populations from sea and estuarine waters (21, 47, 51, 53) and salt marsh sediments (22). However, these environments contain complex microbial populations, and it cannot be determined if any of their members are capable of both activities. Furthermore, there is no indication from the literature that any of those microbes whose DMSP lyase has been studied in pure culture can also take up DMS (11, 12, 29, 49). The results of the acrylate mineralization test showed that these two fungi utilize not only DMS but also the other product of DMSP lyase activity, acrylate (data not shown).

The DMSP lyase activity of the salt marsh plant fungal inhabitants ranged from 3 to 41 nmol of DMS min¹ mg of cell protein¹. These values are lower than those seen in bacteria, where they range from 100 to 750 nmol of DMS min¹ mg of cell protein¹ when assayed under comparable conditions (11, 12, 55). Other values for bacterial DMSP lyase exist in the literature, but these rates were calculated on a per-cell basis (28, 29, 54). The fungus F. lateritium, whose DMSP lyase we recently characterized (2), had activities that were generally higher than those of the plant-associated species reported here; values of the former ranged from 20 to 80 nmol of DMS min¹ mg of cell protein¹. We must emphasize again that the relatively lower values we report here for fungal DMSP lyase activity were from cells grown in rich, proteinaceous liquid medium and may be lower than those that are possible on solid natural substrata. S. alterniflora is reported to have 5 to 50 µmol of DMSP g (fresh weight) of tissue¹ (reference 16 and references therein), which, after several assumptions have been made, can be calculated to be approximately 5 to 50 mM DMSP in the cytosol, concentrations far above those required for optimal activity of F. lateritium (2) and presumably of the Spartina decomposers listed in Table 1.

Our finding that DMSP lyase activity in salt marsh plant fungal decomposers was highly correlated with DMSP production by the plants suggests that an evolutionary relationship may exist between these plants and the decomposer fungi they harbor. That is, the fungi may have developed the lyase concomitantly with the salt marsh plants' acquiring the ability to synthesize DMSP. Alternatively, those decomposer species that have the ability to degrade the large supply of cellular DMSP could have an adaptive advantage over non-DMSP lyase-producing fungi for this niche in the Spartina leaf. Consistent with both hypotheses is the fact that the Spartina DMSP lyase producers are also capable of utilizing the resulting acrylate.