Our data establish a pathway for DMSP biosynthesis in marine algae. This pathway has no steps in common with that in higher plants, which proceeds via SMM and 3-dimethylsulphoniopropionaldehyde 6,7 . DMSP biosynthesis must therefore have evolved independently at least twice. Our results have two other implications. The first stems from the finding that a transaminase reaction stands at the head of the DMSP pathway; this may help explain why nitrogen deficiency enhances DMSP production 12,23,30 . Depletion of cellular amino acids would favour the transamination reaction, thereby promoting DMSP synthesis when nitrogen is limiting. Second, our results suggest that DMSP may not be the only precursor of the DMS produced by living algae: DMSHB is another potential precursor in vivo. In support of this possibility, we have obtained preliminary evidence for extensive catabolism of supplied DMSHB to DMS in Tetraselmis sp. and E. huxleyi.

Emiliania huxleyi clones CCMP 370 and CCMP 373 produced similar amounts of dimethylsulfoniopropionate (DMSP) during axenic exponential growth, averaging 109 mM internal DMSP. Both clones had detectable DMSP lyase activity, as measured by production of dimethyl sulfide (DMS) during in vitro assays of crude cell preparations, but activities and conditions differed considerably between clones. Clone 373 had high activity; clone 370 had low activity and required chloride. For both strains, enzyme activity per cell was constant during exponential growth, but little DMS was produced by healthy cells. Rather, DMS production was activated when cells were subjected to physical or chemical stresses that caused cell lysis. We propose that DMSP lyase and DMSP are segregated within these cells and reaction only under conditions that result in cell stress or damage. Such activation occurs during microzooplankton grazing. When these clones were grazed by the dinoflagellate Oxyrrhis marina, DMS was produced; ungrazed cells, as well as those exposed to grazer exudates and associated bacteria, generated no DMS. Grazing of clone 373 produced much more DMS than grazing of clone 370, consistent with their relative in vitro DMSP lyase activities. DMS was only generated when cells were actually being grazed, indicating that ingested cells were responsible for the DMS formation. We suggest that even low levels of grazing can greatly accelerate DMS production.

The activities of unicellular microbes dominate the ecology of the marine environment, but the chemical signals that determine behavioral interactions are poorly known. In particular, chemical signals between microbial predators and prey contribute to food selection or avoidance and to defense, factors that probably affect trophic structure and such large-scale features as algal blooms. Using defense as an example, I consider physical constraints on the transmission of chemical information, and strategies and mechanisms that microbes might use to send chemical signals. Chemical signals in a low Re, viscosity-dominated physical environment are transferred by molecular diffusion and laminar advection, and may be perceived at nanomolar levels or lower. Events that occur on small temporal and physical scales in the "near-field" of prey are likely to play a role in cell-cell interactions. On the basis of cost-benefit optimization and the need for rapid activation, I suggest that microbial defense system strategies might be highly dynamic. These strategies include compartmented and activated reactions, utilizing both pulsed release of dissolved signals and contact-activated signals at the cell surface. Bioluminescence and extrusome discharge are two visible manifestations of rapidly activated microbial defenses that may serve as models for other chemical reactions as yet undetected due to the technical problems of measuring transient chemical gradients around single cells. As an example, I detail an algal dimethylsulfoniopropionate (DMSP) cleavage reaction that appears to deter protozoan feeding and explore it as a possible model for a rapidly activated, short-range chemical defense system. Although the exploration of chemical interactions among planktonic microbes is in its infancy, ecological models from macroorganisms provide useful hints of the complexity likely to be found.

Oceanic dimethylsulfide (DMS) emissions to the atmosphere are potentially important to the Earth's radiative balance. Since these emissions are driven by the surface seawater concentration of DMS, it is important to understand the processes controlling the cycling of sulfur in surface seawater. During the third Pacific Sulfur/Stratus Investigation (PSI-3, April 1991) we measured the major sulfur reservoirs (total organic sulfur, total low molecular weight organic sulfur, ester sulfate, protein sulfur, dimethylsulfoniopropionate (DMSP), DMS, dimethylsulfoxide) and quantified many of the processes that cycle sulfur through the upper water column (sulfate assimilation, DMSP consumption, DMS production and consumption, air-sea exchange of DMS, loss of organic sulfur by particulate sinking). Under conditions of low plankton biomass (<0.4 gg/L chlorophyll a) and high nutrient concentrations (>8 tam nitrate), 250 km off the Washington State coast, DMSP and DMS were 22% and 0.9%, respectively, of the total particulate organic sulfur pool. DMS production from the enzymatic cleavage of DMSP accounted for 29% of the total sulfate assimilation. However, only 0.3% of sulfate-S assimilated was released to the atmosphere. From these data it is evident that air-sea exchange is currently only a minor sink in the seawater sulfur cycle and thus there is the potential for much higher DMS emissions under different climatic conditions. Introduction Oceanic dimethylsulfide (DMS) is currently thought to be the major natural source of sulfur to the atmosphere [Bates et al., 1992; Spiro et al., 1992]. Once in the atmosphere, DMS is oxidized to produce aerosol particles which affect the acidbase chemistry of the atmosphere [Charlson and Rodhe, 1982] and the radiative properties of marine stratus clouds [Charlson et al., 1987; Falkowski et al., 1992]. This latter effect is calculated to have a major impact on the Earth's radiative balance and hence its climate [Charlson et al., 1987]. The starting point in the marine atmospheric sulfur cycle is the air-sea exchange of DMS which is a function of the gas i NOAA/Pacific Marine Environmental Laboratory, Seattle, Washington. transfer velocity and surface seawater DMS concentration. The gas transfer velocity is controlled primarily by surface turbulence, seawater temperature and gas diffusivity and can be modeled as a function of wind speed for various trace gases [Liss and Merlivat, 1986; Wanninkhof, 1992]. The different T. S. Bates, NOAA/Pacific Marine

Isochrysis galbana Parke, Emiliania huxleyi (Lohm.) Hay and Mohler, and some related prymnesiophyte algae produce as neutral lipids a set of polyunsaturated long-chain (C 37-39 ) alkenones, alkenoates, and alkenes (PULCA). These biomarkers are widely used for paleothermometry, but the biosynthesis and cellular location of these unique lipids remain largely unknown. By staining with the fluorescent lipophilic dye Nile Red, we found that I. galbana and E. huxleyi, like many other algae, package their neutral lipid into cytoplasmic vesicles or lipid bodies. We found that these lipid bodies increase in abundance under nutrient limitation and disappear under prolonged darkness and show that this pattern correlates well with the concentration of PULCA as measured by TLC. In addition, we show that lipid vesicles purified by sucrose density gradient centrifugation consist predominantly of PULCA. We also found significant pools of neutral lipid associated with chloroplasts, and PULCA component profiles in lipid vesicles and chloroplasts are similar. Examination of cell ultrastructure shows conspicuous cytoplasmic and chloroplast lipid bodies, and we suggest that PULCA may be synthesized in chloroplasts and then exported to cytoplasmic lipid bodies for storage and eventual metabolism. Our results connect and extend prior observations of lipid bodies and membrane-unbound PULCA in I. galbana and E. huxleyi, as well as the behavior of PULCA during nutrient and light stress.

We characterised and conlpared dirnethylsulfoniopropionate (DMSP) lyase isozymes in crude extracts of 6 axenic Emiliania huxleyi cultures (CCMP 370, 373, 374, 379, 1516, and strain L). This enzyme cleaves DMSP to form dimethyl sulfide (DMS), acrylate and a proton, but the function of this reaction in algae is still poorly understood. Most of the cultures produced high concentrations of intracellular DMSP, which was constant over the growth cycle and ranged from 157 to 242 mM, except for 1516 which had 50 mM DMSP cell ' Extracts of all strains produced DMS from exogenous DMSP i n vitro. DMSP lyases appeared constitutive, but enzyme activity and behaviour varied greatly among strains, and did not correlate with intracellular DMSP concentration. Strains 373 and 379 showed high DMSP lyase activities (12.5 and 6.1 fmolDMS cell-' min-', respectively), whereas DMS production was more than 100-fold lower in 370, 374, 1516 and L This difference was intrinsic and the general pattern of high-and low-activity strains remained true over more than a 1 yr cultivation period The cleavage reaction was optimal at pH 6 in the strains with high lyase activity and pH 5 was optimal for 374, 1516 and L. Strain 370 showed increasing activity with increasing pH. Experiments with additions of 0.125 to 2 M NaCl indicated halotolerant DMSP lyases in 373, 379 and 374. However, the halophilic DMSP lyases in 370 and L required 1 M NaCl addition for optimal DMSP cleavage, and 1516 showed optimal activity at 2 M NaCl These results suggest that there are several structurally different DMSP lyase isozymes within E. huxleyj. However, it cannot be ruled out that varying concentrations of DMSP lyase per cell may have contributed to the differences in enzyme activity per cell. Comparison with other algal taxa indicates several families of DMSP lyases, hinting at possibly different cellular locations and functions, and varying DMS production under natural conditions.

[1] We conducted isothermal (15°C) batch culture experiments with the coccolithophorid Emiliania huxleyi (strain NEPCC 55a) to evaluate the extent to which nutrient and light stress contribute to variability in the alkenone unsaturation index U 37 K 0. Alkenone content and composition were constant throughout exponential growth in both experiments when nutrients (nitrate and orthophosphate) were replete. Stationary phase (nutrientstarved) cells continued to produce alkenones, amassing concentrations (AEAlk) ! 3 times higher than those dividing exponentially (1.5-2 pg cell À1 ), and the U 37 K 0 of ''excess'' alkenone dropped by 0.11 units. In contrast, 5 days of continuous darkness resulted in a 75% decrease in cellular AEAlk and a significant U 37 K 0 increase (+0.11 units). Given an established 0.034 unit/°C response for exponentially growing cells of this strain, the observed range of U 37 K 0 variability at 15°C corresponds to an uncertainty of ±3.2°C in predicted growth temperature. This level of variability matches that of the global U 37 K 0 annual mean sea surface temperature calibration for surface marine sediments, begging the question: What is the physiological condition of alkenone-producing cells exported to marine sediments? Comparison of our laboratory results for a strain of E. huxleyi isolated from the subarctic Pacific Ocean with depth profiles for alkenones in surface waters from two contrasting sites in the northeast Pacific Ocean suggests that the answer to this question depends on the ocean regime considered, a possibility with significant bearing on how stratigraphic U 37 K 0 records in marine sediments are to be interpreted paleoceanographically.

Activated defenses against herbivores and predators are defenses whereby a precursor compound is stored in an inactive or mildly active form. Upon damage to the prey, the precursor is enzymatically converted to a more potent toxin or feeding deterrent. In marine systems, activated defenses are only known to exist in a few species of tropical macroalgae. In this study, we examined an activated defense system in temperate marine macroalgae in which the osmolyte dimethylsulfoniopropionate (DMSP) is converted to acrylic acid or acrylate, depending upon the pH, and dimethyl sulfide (DMS) by the enzyme DMSP lyase upon damage to the alga. We surveyed 39 species of red, green, and brown algae from the Washington and Oregon coasts, and found high concentrations of DMSP in the chlorophytes Acrosiphonia coalita, Codium fragile, Enteromorpha intestinalis, E. linza, Ulva californica, U. fenestrata, and U. taeniata, and in the rhodophyte Polysiphonia hendryi. Concentrations of DMSP ranged from 0.04% of the alga's fresh mass (FM) to 1.8% FM. We found significant DMSP lyase activity in 1 green alga, U. fenestrata, and 1 red alga, P. hendryi, with DMSP cleavage rates approaching 300 mmol kg-1 FM min-1. Loss of DMSP and the production of DMS when the tissues of U. californica and P. hendryi were crushed suggested that physical damage results in DMSP cleavage. In laboratory feeding preference experiments, acrylic acid deterred feeding by the sea urchin Strongylocentrotus droebachiensis at concentrations of 0.1 to 2% FM and by S. purpuratus at 0.25 to 2% FM, while the precursor DMSP functioned as a feeding attractant to both sea urchins. In contrast, feeding by the isopod Idotea wosnesenskii was not deterred by acrylic acid even at concentrations as high as 8% FM. Our data suggest that DMSP may function as a precursor in an activated defense system in diverse species of temperate macroalgae and may possibly contribute to the widespread success of the Ulvophyceae. This chemical system is also found in unicellular phytoplankton, and presents an opportunity to compare and contrast the ecological role of chemical defense among micro-and macroorganisms.