1. Paul Falkowski. OCE-9911948. Infrared Fast Repetition Rate Fluorometry for detection and characterization of photosynthetic bacteria in deep sea vents. $25,000 (1 Oct 1999-30 Sept 2000) (1 paper submitted to Science).
2. Kenneth G. Miller EAR97-08664; Global Sea level and Late Cretaceous to Miocene sequences: Completion of the New Jersey Coastal Plain Drilling Project. $503,694 (6/1/1998-5/30/00). Twenty seven papers in Miller and Snyder, eds., (Proc. ODP, Sci. Results, 150X, 388p., 1997), and publications in Rev. Geophy., Science,Geology (6 papers), GSA Bull., and others.
3. Andrew Knoll. EAR-9805032. Paleobiology of a Uniquely Well-Preserved Window on the Early Evolution of Multicellular Organisms (09/01/98 - 08/31/00) $149,739. Seven papers including Xiao, S., Y. Zhang, and A.H. Knoll (1998) Nature 391: 553-558.
4. Oscar Schofield OCE 97-27342 Impact of episodic event on bio-optical characteristics and phytoplankton dynamics of southern Lake Michigan 7/1/97 - 6/30/02. $252,375
5. Costantino Vetriani - no prior NSF support for this new faculty member at IMCS .

"As the present condition of nations is the result of many antecedent changes, some extremely remote and others recent, some gradual, others sudden and violent, so the state of the natural world is the result of a long succession of events, and if we would enlarge our experience of the present economy of nature, we must investigate the effects of her operations in former epochs." Charles Lyell, Principles of Geology, 1830

1. Statement of the problem and primary goal of the proposed research
This proposal is about the origin, radiation, and ecological success of three groups of eucaryotic phytoplankton, namely dinoflagellates, coccolithophorids, and diatoms. In the contemporary oceans, photosynthetic carbon fixation by these three taxa is the primary fuel supporting metazoans and the sinking flux of organic carbon to the ocean interior and seafloor (Goldman, 1988; Rivkin, 1997; Bienfang, 1992). The fossil record indicates that these three groups of phytoplankton rose to taxonomic and ecological prominence in the Mesozoic Era, between 250 and 65 Ma. What happened in that geological period that led to the selection of three major eucaryotic phytoplankton taxa and why have these organisms been so successful ever since? The focus of this multidisciplinary research program is to understand the historical origins and environmental conditions that led to selection and radiation of the modern dominant eucaryotic phytoplankton taxa, and the ecological processes that contribute to their continued success in the contemporary ocean. The proposed research utilizes a combination of geological, molecular biological, ecological, and modeling approaches to address an important and complex, tractable yet unsolved puzzle in Earth system science. Our primary goal is to develop the first quantitative models of eucaryotic phytoplankton community structure in the contemporary oceans based on paleoecological and evolutionary inference.
2. Context of the problem and formulation of central hypotheses
The central question raised in this proposal is: Why have three phylogenetically diverse groups of eucaryotic, unicellular algae been so ecologically successful, and what does their evolutionary history tell us about the history of Earth and the ability of eucaryotic phytoplankton to accommodate to change in the future?
a. The key hypotheses
           The proposed research seeks to test a set of three related hypotheses, from which we will develop a conceptual model for evolution and ecological success (dominance) of key phytoplankton taxa in the contemporary ocean. The central hypotheses are:
1. The three dominant phytoplankton taxa in the contemporary ocean evolved in shallow shelf-seas in the Mesozoic Era in response to such changes in the ocean environment as anoxia, sea level and/or tectonic processes that excluded ecological advantages previously afforded to chlorophytes.
         This hypothesis will be tested by: (a) comparing the micropaleontological record of the three groups to contemporaneous faunal assemblages and geochemical analyses of proxy variables for paleoenvironmenal conditions; (b) examining how molecular biological and biochemical markers comport with plastid and host origins and their "chemical history"; and (c) using microcosm analogs of paleoceanic conditions to examine the behavior of the chromophyte taxa in relation to representativechlorophyta.
2. Once established, these groups radiated rapidly. The rapid tempo of evolution was a consequence of high mutation frequencies relative to reversion and sexual recombination, resulting in high genetic potential and DNA content relative to genetic expression in the three taxa. The rapid tempo of evolution in the three taxa has permitted rapid radiation and adaptation to changing oceanic conditions throughout the Mesozoic. This rapid tempo continues to the present time.
         This hypothesis will be tested by examining the number of fossil genera for each major phytoplanktonic taxa in specific periods throughout the Mesozoic. The apparent rates of radiation will be compared with rates of change in a suite of target genes from representative species within the three major taxa, using homologous genes in chlorophytes and cyanobacteria as outgroups. The primary aim of this research effort will be to develop independent estimates of the tempo of evolution from geological and biological inference.
3. The ecological dominance of the three major eucaryotic phytoplankton taxa is a consequence of pan-division traits that permit individual species within each group to rapidly accommodate large variations in oceanic conditions.These traits include the evolution of cell walls and vacuoles that, respectively, provide protection from predation while simultaneously optimizing the exploitation of pulsed nutrient supplies. A corollary of this hypothesis is that the structure of marine food webs in the contemporary ocean is primarily a consequence of the tempo of evolution of the three major taxa of eucaryotic phytoplankton, which is itself a consequence of continuous changes in oceanic regimes.
           This hypothesis will be tested through a combination of mathematical models and experimental laboratory and microcosm studies, under non-steady state conditions. The mathematical models will examine how eucaryotic phytoplankton community structure is affected by resource acquisition, quantum efficiency and growth optimization, costs and benefits of cell wall synthesis, maintenance of vacuoles, motility, and the co-evolution of phytoplankton and their predators.

b. Evolutionary pattern of eucaryotic phytoplankton
          The evolution of eucaryotic protists can be traced by three independent strategies: examination of the fossil record, investigation of biochemical and morphological homology, and construction of phylogenetic trees using gene sequences. Ideally, these three strategies should converge on common points of origin and common points of radiation. The compilation of information from the three strategies provides clues about selection processes, the tempo of evolution within and between the taxa, and the adaptive potential of the groups in the contemporary ocean.

        Isotopic geochemical evidence suggests that photosynthetic organisms originated more than 3460 Ma (Schopf, 1993), and preserved biomarker molecules indicate the presence of cyanobacteria by 2700 Ma (Brocks et al., 1999). By ca. 2100 Ma, the photosynthetic activity of cyanobacteria had oxidized Earth's atmosphere and surface oceans (Rye and Holland, 1998). Fossils indicate that by 1200 to 1000 Ma, eucaryotic algae had diverged to form at least two major groups: a "green" line, which became the forerunner of all chlorophyte algae and (subsequently) all higher plants, and a "red" line, which ultimately was incorporated into the chromophyte algae and, in at least one instance, into dinoflagellates (Fig 1). Indeed, fossils suggest that the secondary endosymbiosis that gave rise to the photosynthetic stramenopiles (including diatoms, chrysophytes, and brown algae) occurred over a billion years ago (German, 1990; Knoll, 1996; Woods et al., 1998). Despite these early origins, primary production in the ocean appears to have been dominated by cyanobacteria and green algal flagellates until the end-Permian extinction, 251 Ma (Tappan, 1980; Lipps, 1992; Knoll, 1989; Knoll, 1992). The first identifiable thecate (and presumably, photoautotrophic) dinoflagellates are found in the Early Triassic, about 10 m.y. after the end-Permian extinction (Fensome et al. 1996) (Fig. 2). Coincidentally, the early Triassic appears to have witnessed a widespread ocean anoxic event (OAE) (Isozaki, 1997). Coccolithophorids emerged late in the Triassic, under oxic conditions (Young et al., 1999). Both of these groups radiated extensively throughout the Jurassic and into the Cretaceous. The radiations are positively correlated with changes in sea level, suggesting a role of shelf seas (Stover et al., 1996). Although there are reports of fossil diatoms from the Toarcian stage in the Early Jurassic (Rothpletz, 1896, 1900), this phytoplankton group did not appear to rise to prominence until the Cretaceous. The earliest diatoms appear to be neritic, and their radiation appears roughly consistent with regression events.
        The recorded first appearances for these three groups represent minimum estimates of the time of origin. A relatively small fraction (ca. 15%) of modern dinoflagellates produce fossilizable cysts (Head, 1996), and preserved biomarker molecules provide evidence for the clade as early as the Neoproterozoic Era (Walter and Summons, 1990; Moldowan and Talyzina, 1998). Nonetheless, dinoflagellate biomarker concentrations increase significantly in Triassic organic matter, in parallel with the radiation recorded by microfossils. Thus, the radiation of dinoflagellates appears to be faithfully chronicled by the geological record. The same is true for coccolithophorids and diatoms; their radiations in the fossil record are marked by clear changes in the marine carbonate and silica cycles (e.g., Maliva et al., 1989; Siever, 1991).
       Molecular biological analyses are broadly consistent with the fossil record. The photosynthetic apparatus in cyanobacteria reveals clear homology with two groups of anoxygenic photosynthetic bacteria, each of which provided one of the two reaction centers requisite for oxygenic photosynthesis (Barber, 1992; Michel and Deisenhofer, 1988). All photosynthetic eucaryotes are oxygenic; i.e., they contain two types of photosystems. Plastid ultrastructure, rRNA sequences, and the conservation of key photosynthetic proteins suggest a common origin for all plastids, namely a procaryotic oxygen evolving group closely related to extant cyanobacteria (Cavalier-Smith, 1982; Margulis, 1974). The points of divergence and rates of evolution of the eucaryotic photoautotrophic can be inferred by comparing SSU rRNA sequences between and within taxonomic divisions (Delwiche, 1999). Such results provide clues about probable time of origin and rates of evolution of the host and/or the associated plastid, but are often ambiguous about the timing of the symbiotic event and the rates of evolution of the symbiotic association. In contrast to the reaction center proteins, the photosynthetic pigment sequences that comprise antenna complexes are highly variable and the proteins, in which the pigments are contained, are largely encoded in the nucleus (Green and Durnford, 1996). Indeed, the chemistry of the pigments has traditionally provided a basis for taxonomic classification of phytoplankton (Jeffrey 1980). Eucaryotic cells containing chlorophyll b form one major group, or "subkingdom", the Viridiplantae, whereas cells containing chlorophyllide c (and its isomers) form a second group, the polyphyletic Chromophyta (Cavalier-Smith, 1993). Phylogenetic trees and morphological evidence suggest that the former evolved from the incorporation and sequestration of a cyanobacterial endosymbiont, whereas the latter group evolved from secondary and, in some cases, tertiary symbiotic incorporation of a red alga in a suite of heterotrophic protistan host cells (Delwiche, 1999). Genes encoding for the pigment protein complexes and other structural elements potentially provide independent information about rates of evolution of the symbiotic association (i.e., the eucaryotic cell).
        Catastrophe hypotheses in geology (Berggren and van Couvering, 1984), and disturbance hypotheses in ecology (e.g., Paine and Vadas, 1969; Connell, 1978) provide conceptual models that are useful in understanding phytoplankton evolution and succession, respectively (Reynolds, 1997). The essence of these hypotheses is that stochastic as well as periodic environmental variations provide opportunities for some species to temporarily dominate a given assemblage; however, because the environment is continually fluctuating, any competitive advantage afforded to one species or group of species is fleeting. Thus, under ideal conditions, species within a community move through a phase space with a strong attractor, leading to reproducible patterns (Ascioti et al., 1993). We hypothesize that on geological time scales, deep disturbances in oceanic regimes have led to evolutionary succession (selection) by replacement of dominant eucaryotic species within a cohort of contemporaneous species in a division or class; i.e., the attractor continually moves through phase space, providing similar, but never reproducible, patterns of community structure. The selection process is dependent upon the genetic potential within the cohort of species. This hypothesis extends to ecological succession in marine phytoplankton communities in the contemporary ocean.
        Several specific hypotheses can be advanced to account for the initial selection and subsequent radiation of the three eucaryotic phytoplankton groups under consideration. Among these hypotheses are: (1) the initial, independent symbiotic acquisition of photosynthetic capability by these three groups was driven by chance encounters of host cells with prospective plastids; (2) the long-term effects of one or more ocean-scale perturbations (e.g., OAEs, sea level changes, and tectonic activity) changed the balance of the planktonic biota, favoring the selection of the specific taxa in certain ocean regimes; (3) a long-term change in the physical circulation and nutrient status of the oceans favored the Chl a+c clades (a pan-clade advantage); or (4) some combination of these factors, including the co-evolution of major grazing zooplankton. One objective of the proposed research is to examine each of these hypotheses within the context of geological, geochemical, biological and ecological possibilities. We suggest that the relationship between short–term environmental perturbations in phytoplankton community structure (occurring on scales of decades and centuries), will be mirrored in long-term changes in the global oceans.
        Consequences of a basin-scale OAE include: sequestration of organic carbon in the deep ocean and sediments with high levels of CO_2, CH₄, and H₂S; a massive loss of fixed inorganic nitrogen; acidification of the deep ocean and transfer of carbonate alkalinity to the upper ocean; and a change in trace metal inventories and composition. Each of these consequences has potentially been recorded in the geological record and embossed in the genetic and biochemical infrastructure of selected organisms. For example, under oxidizing conditions in the ocean, soluble Fe, which is essential for all organisms, is extremely scarce, whereas Cu is relatively abundant; the converse is true under anoxic conditions. However, even without a change in the ocean redox chemistry, a relatively small decrease in ocean pH would have slowed the oxidation kinetics of iron thereby increasing its bioavailability (Millero et al. 1987). Curiously, all chlorophyte algae possess a Cu-containing protein, plastocyanin that ferries electrons between cytochrome f and PSI reaction centers. Chromophyte algae use an Fe containing complex to accomplish the same result (Raven et al.. 1999; Sandmann et al. 1983). Does the selection of metals in the plastid reflect originating conditions for the symbiotic association? High concentrations of CO₂ and phosphate prevent calcification; consequently, coccolithophorids and other calcifying organisms would be at a disadvantage under hypoxic conditions. Were coccolithophorids able to rapidly exploit the Jurassic ocean as a consequence of restoration of oceanic pH and carbonate alkalinity in the late Triassic? Was the initial radiation of diatoms due in part to the culling of dinoflagellates during an OAE?
       The outcome of the evolution is determined by fitness. Physiological processes in extant organisms partially, albeit imperfectly, reflect the evolutionary selection pressures of their taxonomic predecessors. We hypothesize that four fundamental features, common to all three taxa, have allowed them to dominate eucaryotic phytoplankton communities since the Mesozoic. First, all three taxa are armored; i.e.,, many of the species within each of the taxonomic groups contain cell walls that potentially have helped avoid or escape death by predation or viral attack. However, there are energetic and material costs of forming a cell wall that must be offset by the potential benefits. Second, all species in the three taxa possess either permanent or temporary storage vacuoles. When nutrients are supplied in pulses, a vacuole potentially gives a large, eucaryotic cell a competitive advantage over picoplankton in a dilute nutrient-poor sea. Again, however, there are costs as well as benefits to forming a vacuole (Raven 1997). Third, all three groups of organisms form benthic resting cysts or spores, suggesting that the coupling between benthic and pelagic systems is critical for continuity of the species. Finally, all three divisions have relatively high ratios of genetic material relative to genes expressed (Table 1). Although, taken seperately, these factors cannot account for ecological success, they almost certainly are contributing factors to the continued success of the dominant divisions (Smetacek, 2000). To our knowledge, however, the evolution of these features and their ecological consequences have not been mathematically modeled.
3. Research team and structure
        To test the three core hypotheses, we propose to incorporate information from three groups of investigators, selected not only for their individual expertise, but also for their proven ability to work collaboratively. The three groups possess expertise in geology and geochemistry, molecular biology and biochemistry, and algal physiology and ecological modeling. The fundamental concept is to compare paleoecological data, inferred primarily from geological and geochemical proxies, with molecular biological and biochemical information to test hypotheses 1 and 2; the paleoecological data will serve to help guide physiological experiments and ecological models to test hypotheses 1 and 3. Our proposed research program is described below.

Table 1. Average genome sizes for various microbial procaryotes and eucaryotic algae. Assuming an average protein contains ca. 300 amino acids and there are ca 10⁵ proteins per eucaryotic cell, between 10% and 50% of the chlorophyte genome encodes expressed genes, whereas only 2-10% is expressed in a diatom or coccolithophore, and <0.1% is expressed in a dinoflagellate.

Proposed Research.
1. Disturbance and phytoplankton radiation in the Mesozoic. The basic goals of this effort are to define paleoenvironmental conditions in key periods in the Mesozoic and reconstruct, from new observations and extant data, the radiation rates for the three key groups. Specifically, we propose to: (a) reconstruct the microfossil assemblages and paleoenvironmental regimes for key periods in the Mesozoic, (b) compare this information to the molecular evolution and emergent biochemical properties of the three taxa, and (c) use the geological and molecular results to constrain algal competition microcosm and modeling experiments for simulated ocean chemical regimes.
         There is an extensive literature on geology and evolution during the Mesozoic Era and it is not our intention or desire to redo what has already been done. A fundamental goal of this effort is to integrate information across disciplines and in that light, we will use published information as a guide in our efforts to answer questions that remain critical for understanding how the ecological structure of eucaryotic phytoplankton communities reflects geological and evolutionary history.
        It is unclear whether the end-Permian to Early Triassic period witnessed whole ocean anoxia ("superanoxia" sensu Isozaki, 1997) or maintained an oxygenated surface layer overlying an anoxic interior. The distinction is critical, as it leads to a central question: Is anoxia a selective mechanism for phytoplankton or was some other ecological variable, (e.g., nutrient supply or shelf sea area) more significant? If, for example, the surface ocean became anoxic, trace metal availability would have been a potentially strong selective agent for new phytoplankton taxa. The effects could be direct, in that specific trace metals are critical for biochemical redox and other enzymatically catalyzed reactions; or indirect, in that changes in trace metal inventories could alter the optical properties of the surface ocean. Surface water anoxia would have also been a strong selective force constraining the survival of aerobic, heterotrophic protists; was the endosymbiotic appropriation of a photoautotrophic cell a response to an OAE? That deep ocean anoxia prevailed in the late Permian and early Triassic is reflected by high organic carbon content in the sediments, absence of bioturbation, a high abundance of mineral sulfides, and a paucity of benthic foraminifera. The relatively high abundance of radiolarians suggests productive oxygenated surface waters (Isozaki, 1997); though such evidence is not conclusive, it is supported by records of C isotopes in surface waters suggesting high surface productivity and increased partitioning of carbon into the deep ocean (Gruszczynski et al., 1989; Holser et al., 1989; Magaritz et al., 1992). The corollary of a highly stratified ocean is also consistent with sulfur isotope records exhibiting a "mirror image" pattern in deep and shallow sites (Claypool et al., 1980; Kajiwara et al., 1994) and with shallow waters records of strontium ( ⁸⁷ Sr/⁸⁶ Sr) (Gruszczynski et al., 1992) and neodymium ( ¹⁴³ Nd/¹⁴⁴ Nd) (Gruszczynski et al., 1992; Martin and Macdougall, 1995); the isotopic composition of these two systems in a well mixed ocean reflects the relative contributions from continental and hydrothermal sources, which carry very distinctive isotopic signatures. In a highly stratified ocean, however, the deep ocean chemistry would tend to more closely reflect hydrothermal inputs, whereas the upper ocean chemistry would more closely reflect continental sources.
Disturbance and Paleooceanography of the Mesozoic. Given our interest in the conditions, which set the stage for phytoplankton radiation in the Mesozoic, we propose to focus on four key 5-20 m.y.-long "time slabs" centered across key evolutionary and paleoceangraphic events:
(1) The Permo-Triassic boundary and subsequent Early Triassic aftermath. Unlike the K/T boundary (Hildebrant, 1990), there is no evidence that the mass extinction at the P/Tr boundary is due to an extraterrestrial impact (Erwin, 1993). However, geochemical and mineralogical evidence hints that there were widespread environmental perturbations (Holser and Magaritz, 1987; Magaritz et al., 1992), apparently associated with deep ocean anoxia. Hypotheses for the cause of the P/Tr extinction include a massive overturn of the density-stratified oceans causing poisoning of the shallow waters (Knoll et al., 1996) and expansion of the anoxic conditions into shelf waters (Haltom and Wiggle, 1997; Isosaki, 1997).
(2) The Tr/J boundary provides another interval for evaluating the role of mass extinction in providing ecological opportunities for marine eucaryotic phytoplankton. One of the five largest mass extinction events, the Tr/J boundary has excellent chronostratigraphic control (cyclo- and magneto stratigraphy; Kent and Olsen, 1999). It is only after that event, that two of the three groups, the dinoflagellates and coccolithophores, begin to radiate (e.g., Fig 2). What did this event do to lead to such opportunity in the subsequent Jurassic period?
In the Cretaceous, our efforts will focus on 3 specific time intervals:
(3) The early Aptian OAE1a (the Selli level; ca. 120 Ma) which was an interval of diversification of calcareous nanoplankton (Roth, 1987; 1989; Bralower et al., 1994). Whether the speciation reflects increased nutrient competition (Roth 1987, 1989) or increased ocean fertility perhaps associated with sea level rise (Mutterlose 1992, Bralower et al. 1994) continues to be a subject of debate. This event is also associated with the collapse of the nannoconids, a group of nannoplankton that had dominated assemblages for the previous 20 million years, and is therefore known as the "nannoconid crisis" (Coccioni et al., 1992; Erba, 1994). A possible explanation for the crisis was that a mantle superplume event (e.g., Larson, 1991) directly or indirectly caused a change in the thermal or nutrient structure of oceanic surface waters (Erba, 1994). Clarification of the causal mechanisms for the diversification and the nannoconid crisis will be elucidated using the multiple geochemical proxies. OAE1a is succeeded in the late Aptian and Early Albian sequence by a diverse community of neretic diatoms. Clarification of the causal mechanisms for the diversification will be elucidated using the multiple geochemical proxies.
(4) The Cenomanian/Turonian boundary (OAE2) records a ca. 93 Ma anoxic event that provides the best preserved and most complete records of the development of ocean anoxia. Although the dominant phytoplankton groups had radiated by this time, analysis for OAE2 will allow us the best opportunity to ground-truth proxies for anoxia and for evaluating the cause and effects of ocean anoxia and their effects on phytoplankton and zooplankton lineages.
Proposal field sites: (1) Basal Triassic: (a) Meishan, China. Carbonaceous shales and marls in proposed type section for Permian-Triassic boundary; shelf succession well exposed and accessible. (Yang et al. 1993) (b) Eastern Shikoku and southwestern Honshu, Japan, pelagic siliceous and carbonaceous claystone (Isozaki, 1997; Kajiwara, et al., 1994); (c) Carnic Alps, Austria, predominantly carbonate shelf succession, with basal Triassic black shales. Drill core may be available (Holser, W.T. and H.P. Schönlaub, 1991) (2) Late Triassic (Norian-Rhaetian) (a) Exmouth Plateau, Australia , Pelagic sediments; exceptional material at ODP Sites 761 and 764 (Bralower et al. 1991); (3) Cretaceous OAE Intervals (a) OAE1 (Selli) Excellent material available in drill cores; Cismon core; DSDP site 398 (Eastern N. Atlantic), DSDP 463 (mid Pacific), ODP 641 (Eastern N. Atlantic) (Erba et al., 1999; Bralower et al.,1994) (4) OAE2 (Bonarelli) Bass River, NJ, ODP Leg 174: marl; expanded, only moderately lithified C/T boundary section; Ancora, NJ, ODP Leg 174: marl; relatively unlithified C/T boundary section; Pueblo, CO: Well studied reference section (Miller et al.,1997, 1999; Sugarman et al., 1999).

Geochemical studies: We propose to study the elemental and isotopic chemistry of biogenic and inorganic carbonates from shallow water sites. For the P/Tr transition we will focus on brachiopodes and conodont shells; later, we will use foraminifera and ammonites. The following geochemical studies are designed to address these following questions: (1) what was the extent of anoxia in each of these events? was it limited only to the deep ocean or, at times, expanded also to shallow waters? (2) What were the exchange dynamics between the deep and shallow ocean during such periods of strongly stratified ocean? What were the effects on metal and nutrient inventories in the surface ocean?

Mg/Ca, Li/Ca d⁴¹ Ca, d⁶ Li, d³⁴ S and REEs as indicators for ocean stratification. A deep ocean that has been stratified for >1 My will become depleted in magnesium and enriched in lithium in the interior as a consequence of Mg precipitation and Li input at hydrothermal vents (Von Damm et al., 1985). The shallow ocean should have had an inverse response (i.e., enriched in Mg and depleted in Li), reflecting its riverine inputs. Similarly, the two water masses should have had very different isotopic compositions with respect to d⁶ Li, d³⁴ S, and different rare earth element (REE) patterns than is expected from a well-mixed ocean due to the significantly different signatures of the continental and hydrothermal end-members (Faure, 1986; Chan and Edmond, 1988; Chan et al., 1992; Kajiwara and Kaiho, 1992). Thus, sedimentary records of these proxies should provide additional constraints on the different hypotheses of ocean anoxia and overturn. Calcium isotopes have the potential to provide another new tool to assess these hypotheses as well as changes in biological productivity; the isotopic composition of d⁴¹ Ca shouldhave been affected similarly by ocean stratification. However, one would expect that changes in calcareous productivity of should have left an additional imprint (Zhu and MacDougall, 1998).

Trace-metal/calcium ratios Cd/Ca, Cu/Ca, Zn/Ca, Pb/Ca, Fe/Ca, U/Ca, and V/Ca as indicators of micro nutrient availability. Trace metals in biogenic carbonates have been used for reconstructing Quaternary seawater nutrient concentrations (Boyle, 1988; Lea and Boyle, 1989). However on longer geological timescales the seawater concentrations of these elements are sensitive to the oceanic redox conditions due to their uptake by sulphidic minerals (e.g., Rosenthal et al., 1995). Similarly, U/Ca and V/Ca have been used to study variations in the aerial extent of reducing sediments (Hastings et al., 1996; Russell et al., 1996). Thus, the chemical composition of carbonates should provide a good indication of the trace-metal chemistry of shallow waters during OAE and subsequent radiative periods.

Nitrogen isotopes. We propose to study the nitrogen isotopic composition in sedimentary bulk organic matter. The N isotopic composition in organic matter should reflect the source of nitrate to the surface waters; nitrogen fixation imparts an isotopically light signature (~0‰), whereas nitrate supplied from the deep anoxic ocean (either due to incomplete denitrification or oxidation of NH₄) imparts an isotopically heavy signature. We will augment these analyses by additional measurements of mineral-bound organic matter in carbonate fossils (brachiopods/conodonts); mineral-bound organic matter is considered better preserved from diagenetic processes and should, in principle, provide a more reliable record of the surface water d¹⁵ N at that time. Analysis of organic matter in carbonate fossils can be done after separating the organic matter from the CaCO₃ by dissolution in dialysis bag.

Geochemistry notes. (1) All elemental analyses will use methods developed for our Magnetic Sector ICP-MS at Rutgers for rapid and precise measurements on small samples (Field et al., 1998; Rosenthal et al., 1999); (2) Sulfur isotopes will be measured at Rutgers by light isotope mass spectrometry (Prohaska et al., 1999); (3) Li isotopes will be measured by L-H Chan at Louisiana State university; (4) Ca isotopes will be measured by multi-collector ICP-MS either at Harvard; (5) Oxygen and carbon isotopes in carbonate samples will be determined for chronostratigraphic purposes at Rutgers; (6) Nitrogen isotopes in organic matter will be measured at Rutgers.
           Micropaleontological studies The coccolithophorid record of OAE1a and OAE2 is fairly well established (e.g., Bralower, 1998; Mutterlose, 1992; Bralower et al., 1994). By comparison, the diatom and dinoflagellate record is much less well-known. To fully establish the ecological and oceanographic controls on microplankton evolution, and the possible effects of OAEs on radiation, we must obtain a more accurate record of the exact timing of speciation of all the groups under consideration. Coccolithophorids thrive today under oligotrophic conditions whereas the other two groups are adapted to prosper under more eutrophic conditions. In addition, we might expect coccolithophorids to be less tolerant of dysoxic conditions in the water column. Thus, differences in the evolutionary record of the groups are expected and will provide vital clues as to the exact controls on microplankton radiation.
         To reconstruct the record of diatoms and dinoflagellates around OAE1a and OAE2, we will conduct a general survey of a number of different deep sea and marginal locations, including both DSDP/ODP sites and land sections. We will screen samples for both diatoms and dinoflagellates using routine sample preparation techniques. Sections containing one or both groups will be studies in detail, with a full suite of geochemical analyses (see above). If unavailable, coccolithophorid biostratigraphy will be carried out.
Molecular ecology and biochemistry. The fossil record will be compared to phylogenetic history reconstructed from a suite of functional genes and gene products from the extant dominant marine photosynthetic eucaryotes. The phylogenetic efforts will focus on reconstructing the timing of their evolutionary history and the tempo of their evolution. Biochemical data will examine the correspondence between protein selection and its implication for selective agents.
        There has been considerable analysis of rRNA sequences in the eucaryotic phytoplankton taxa; such analyses have been used to predict the origin and rates of evolution in key groups (e.g., diatoma and haptophytes: Kooistra et al. 1998, Medlin et al. 1997; dinoflagellates: Taylor et al., in progress). These studies are being supplemented with other single gene analyses (e.g., tufA; Medlin, in progress). Our efforts will focus on a suite of functional genes (light-harvesting pigments, RuBisCO, superoxide dismutase, and nitrate reductase) that are ubiquitious in eucaryotic phytoplankton but have undergone significant, although differing (and in some cases, co-varying), environmental selection. In addition, the gene products define a range of physiological capabilities for the algal taxa. Comparison between the functional genes will be used to constrain rRNA estimates of evolution, simultaneously providing independent information about paleoenvironmental features to which the taxa have adapted. Dramatic  changes in the climate and geochemistry of the Mesozoic ocean may have induced huge biological adaptations, the signatures of which are encrypted and still recognizable in the evolutionary history of extant organisms and their molecules. For example, differences in environmental metal concentrations, due to changes in the geochemistry of the Mesozoic ocean, may have selected for organisms and metalloenzymes with different adaptations and metal requirements (Kirschvink et al., 2000). An overarching goal is to estimate relative rather than absolute rates of evolution within the three taxa, and compare those rates with those estimated for chlorophyll b-containing eucaryotic algae (i.e., to estimate the "tempo" of evolution for the major eucaryotic taxa).
       Our first effort involves simply determining the copy number for the basic functional genes. Is there any relationship between the genetic content of the cell and the number of functional genes? We emphasize that the phylogenetic analysis of multifamily proteins reflects the evolutionary history of single molecules and not necessarily the evolution of organisms (as the analysis of homologous genes, such as the ribosomal RNA does). To the extent that the degree of sequence divergence represents elapsed time since a common ancestor, the evolutionary history of these genes could provide information on the relative timing of such major evolutionary events, as the rise and radiation of chromophytes. Amazingly, such basic information for many of these genes or their products is lacking for the major groups of phytoplankton (Table 2); thus, our efforts must begin with acquiring fundamental information about the genes.

Table 2.  A survey of key functional proteins and their distribution in selected taxa of eucaryotic algae.  The number to the left indicates the number of gene and protein sequences reported to public gene banks,  the number in brackets indicates the number of species within the taxonomic group from which the sequences have been obtained.

Light-harvesting complex protein (LHC). The light-harvesting complex proteins (LHCs) bind photosynthetic pigments (chlorophylls and carotenoids). Assuming, based on phylogenetic analysis, that all chlorophyte antennae evolved from a red algal Lhc gene (Durnford et al. 1999), we can develop a molecular clock based on the divergence of the algal Lhc genes using the known red algal protein sequences as a standard reference point from which to calculate divergence and estimate time. As there is evidence indicating that the red and green algal plastid evolved from a common primary endosymbiotic event, we can use the red algal sequences to compare the tempo of evolution in the chromophyte and chlorophyte lineages based on the divergence of Lhc genes. Several red algal LHC-like sequences are in the databases from which we can initiate our calculations (Table 2). Further characterization of the Lhc genes from the dinoflagellates (including the water-soluble peridinin chlorophyll proteins) and coccolithophores is required in an attempt to set the tempo of phytoplankton evolution using several Lhc gene types in the analyses. The divergence of the Lhc genes into functionally specialized members offers a unique opportunity to examine the evolution of the eucaryotic algae by determining which Lhc homologues are possessed by the different divisions of algae.
       Although Lhc protein sequence can be used to examine phylogenetic relationships and, potentially, the rate of evolution within taxa, environmental selection on the chromophores is poorly understood. We propose to analyze this issue in conjunction with the protein sequence analyses; we call this analysis "paleobio-optics". For example, the sequence of the chlorophyll a/b binding protein in the procaryotic Prochlorophyta clearly emerged from an iron stress-induced protein (CP43') in phycobilin containing cyanobacteria. The evolution of the Chl b binding proteins in the procaryotes suggests that a paucity of bioavailable iron (presumably as a consequence of oxidation of the Proterozoic oceans) led to their selection as a scaffold for chlorophyll chromophores. Interestingly, high iron concentrations, which would have been present in the surface waters prior to their oxidation, would have shifted the maximum wavelength of light penetration from blue-green to longer wavelengths; i.e., those matching the absorption spectrum of phycoerythrin and phycocyanin. The oxidation event would have shifted the spectrum to shorter wavelengths, as iron precipitated; i.e., matching the Soret absorption bands of chlorins. Similarly, the water soluble peridinin-chlorophyll proteins, that characterize many thecate dinoflagellates (and presumably have a red algal lineage, see Fig. 1), have high absorption cross-sections in the green, whereas fucoxanthin chlorophyll proteins are more blue shifted. Do these spectral biases in the light harvesting complexes of eucaryotic algae reflect differences in spectral irradiance in the Mesozoic at the time of origin of the complexes? If so, what might have led to those changes in spectral irradiance?
RuBisCO. RuBisCO has undergone significant evolutionary selection pressure, and there is some information on gene sequences for the major eucaryotic taxa. Two basic forms of the enzyme, Form I (L8S8) and form II (L2) are found in procaryotes and, until recently, it was assumed Form II was confined to photosynthetic (anaerobic) bacteria and chemoautotrophs (Tabita, 1999). However, Form II RuBisCO was found in symbiotic dinoflagellates (Morse et al. 1998). Does its presence imply an anaerobic origin of some dinoflagellates through the endosymbiotic association of an anaerobic photosynthetic bacterium, or does it represent a lateral gene transfer somewhere in the evolution of the symbiotic dinoflagellates? We propose to characterize the RuBisCO genes from representative species within the three eucaryotic taxa to address these questions and to construct a cladogram for RuBisCO evolution. In the chromophytes, both subunits of the protein are encoded in the plastid, whereas in chlorophytes the small subunit is nuclear encoded. In the latter group, the SSU gene has undergone considerable evolutionary divergence, while that encoding the LSU is relatively highly conserved. This distinction permits evaluation of the rate of evolution of two plastid-encoded genes compared over long time periods and under different selection pressures. We predict that both the LSU and SSU in diatoms will diverge at comparable rates.
Superoxide dismutase (SOD). SOD catalyses the dismutation of the superoxide ion into oxygen and hydrogen peroxide. There are four known classes of SODs, each characterized by the distinctive catalytic metal at the active site; these are, Cu/Zn-SOD, Mn-SOD, Fe-SOD and Ni-SOD. The distribution and metal selection of SODs in eucaryotic phytoplankton is essentially unknown. However, molecular evolution studies suggested that SOD may be evolving at a relatively constant rate when the constrains of molecular clock operation are taken into account (Fitch and Ayala 1994). That makes SOD a good candidate for a reference/standard molecular clock. It has been suggested that the metal selection comports with the redox conditions in the environment at the time of origin (Kirschvink et al., 2000). We will determine what SOD proteins are expressed in dinoflagellates, coccolithophorids and diatoms, and compare the phylogenetic distribution of SOD with the paleoreconstruction of metal chemistry of the organisms in the Mesozoic.
Glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Glyceraldehyde-3-phosphate dehydrogenase is in both the glycolitic pathway and the Calvin cycle. GAPDHs are nuclear-encoded; they are present in both cytosol and plastid, therefore they represent potential markers of ancestral endosymbiont and host cell components. GAPDH genes have been identified in cryptomonads, diatoms and dinoflagellates, and consensus amino acid motifs conserved among GAPDHs can be used to rapidly isolate GAPDH genes from expression libraries. We propose to obtain GAPDH sequences from selected members of the three divisions in question (Table 2), and use them for phylogenetic analyses in conjunction with selected sequences available from the database.
Nitrate reductase. There are three forms of assimilatory nitrate reductase (NR); all are nuclear encoded. To date, 19 genes from 16 plant and algal species have been sequenced (Zhou and Kleinhofs, 1996). All algal nitrate reductase sequences are from Chlorophyta; no representatives of dinophytes, haptophytes, or diatoms exist in the database. The enzyme contains a conserved heme binding domain and a Mo cofactor-binding domain along with a catalytic domain. While phylogenetic analysis of NR is consistent with information derived from 18S rRNA genes, epitopic analysis suggests rapid evolution of the protein within diatoms and between diatoms and other chromophyte algae (Vagara et al., 1998). We suspect there may be a much more rapid rate of evolution of this protein in algae than previously thought. This gene is important in nutrient acquisition for the major phytoplankton groups and a rapid tempo of evolution in this gene may be in response to highly variable nutrient conditions in the oceans.
Microsatellites. Microsatellites consist of non-coding repetitive sequences of combinations of di-, tri- or tetra-nucleotide repeats (i.e., (GT)n, (GTA)n, (ACTG)n), n > 6) that are highly abundant and scattered through the genome of eucaryotic organisms with frequencies of 10³ -10⁵ copies (Wright and Bentzen, 1993). The importance of microsatellites in an evolutionary context stems from the fact that microsatellites are highly heterozygotic and posses large numbers of alleles resulting from high mutation rates (Schriver et al., 1995). Comparative studies have shown that results obtained using microsatellite analyses are congruent with other genetic markers (Desplanque et al., 1999). Overall, microsatellites can be considered as useful and neutral Mendelian markers that may provide information about the nature of mutation and the role of selection and recombination in evolution within a species (Jarne & Lagoda, 1996). We will use microsatellites as a proxy to elucidate evolutionary steps for each functional group for which there is representation in the geological to: (1) determine genetic relatedness, i.e., genetic distance between groups using Jaccard’s coefficient given by genetic distance matrices calculated from the allelic frequencies for a given microsatellite locus; and (2) determine rates of speciation for representative taxa within each functional group by comparing the variance of allele frequencies between individuals with the estimated origin time based on the geological record.

Competition between eucaryotic algae in Mesozoic. These molecular biological efforts will be complemented with a series of competition experiments to characterize the relative success between eucaryotic algae under conditions that simulate both historical and modern oceans. Isotopic and geochemical profiles characterized by the paleooceanography group will constrain the historical experimental conditions. These historical medias are composed of seawater salts and trace metals at their hypothesized concentrations for both oxic and anoxic oceans and constrained by thermodynamic calculations. This approach has been applied successfully to studies of cyanobacteria (Berman-Frank and Falkowski, in progress). Redox and pH conditions are altered to obtain thermodynamic equilibrium media, simulating historical conditions. Growth rates and physiological profiles will be contrasted between the modern (simulated by standard medias) and historical oceans. The broad suite of physiological measurements and productivity parameters (see below) will provide inputs to community models that will assess productivity in the historical and modern ocean.
2. Maintenance of phytoplankton diversity in the modern ocean. The high genetic potential generated during the Mesozoic has permitted rapid and continual radiation in response to changing oceanic conditions. This provides physiological flexibility for maintaining eucaryotic diversity in the modern ocean. This set of hypotheses will be assessed by (1) examining the behavior of community structure in quantitative models and, (2) developing physiological profiles for eucaryotic phytoplankton over a wide range of abiotic conditions that will provide inputs for mathematical models of ecosystem structure.
         Modeling. A primary goal of EREUPT is to develop quantitative models of eucaryotic phytoplankton community structure in the contemporary ocean (and oceans of the coming century) and use these models to simulate the paleoecological and evolutionary information from Objective 1. To achieve this end, we propose four integrated modeling components: (1) box models that constrain past, present, and future environmental conditions; (2) biogeological models, that examine the rate of evolution and adaptation in response to trended (e.g., sea level change) and abrupt (e.g., OAE, impact) disturbance phenomena; (3) cost-benefit analyses, designed to mathematically quantify the fitness of each of the eucaryotic phytoplankton taxa for resource acquisition, physiological acclimation, and genetic potential; and (4) community structure models designed to predict the outcome of competition between taxa under conditions in which growth and mortality (e.g., grazing) interact. These four groups of models will be coupled in a nested array to examine eucaryotic phytoplankton community structure under historical conditions, contemporary conditions, and future ocean scenarios. The models serve to both synthesize the geological, genetic, and physiological components of the proposed research, and to provide guidance for the continuous development of experimentally testable hypotheses. A brief description of the model efforts follows.
Geobiological models. Steady-state box models are a computationally convenient approach to assessing course scale interactions between processes. Such models are useful for assessing changes in ocean chemistry in response to changes in circulation or oceanic biological processes. Box models, similar to those used to study glacial climates (Siegenthaler and Wenk 1984, Knox and McElroy 1984, Toggweiler and Sarmiento 1985), will be used to assess the potential mechanisms for (e.g.,) anoxia and the corresponding impact on the eucaryotic algal productivity and community structure (Falkowski, 1980). Similar models, focusing on ocean circulation during the Cretaceous anoxic conditions have been used to assess the impact of increased vertical overturning on ocean productivity. In the proposed research, we will include community models (see below) and nutrient concentrations to explore the feedbacks between productivity, ocean circulation, and deep ocean anoxia. For example, higher nutrient fluxes on broad continental shelf-seas stimulate eucaryotic phytoplankton blooms, which lead to anoxic conditions especially during post-bloom stratification events. Such scenarios are observed in the contemporary ocean (Falkowski, 1980) and it is hypothesized that the frequency of these events is increasing globally, due to human-induced change. By incorporating the paleo and contemporary biogeochemical information into the box models, we can explore how ocean chemistry potentially influences the selection and overall productivity of eucaryotic taxa which, in turn, alters ocean chemistry. Model predictions will be independently tested in microcosm laboratory experiments and from the fossil record. Among the properties to be included in the model are phosphate, silicate, alkalinity, the inorganic carbon system, oxygen, and certain trace metals, (e.g., Cu and Fe). We will use carbon, sulfur, and strontium isotopes, where possible, to constrain the model solutions for the different geologic slices. We will first examine homogenous models, which assume the ocean is well mixed, and then extend these to models thataccount for known mixing regimes to explore the role of heterogeneity in diversification.
          We will couple a plankton community model to the geobiological box models. The community model consists of four algal functional groups: chlorophytes, coccolithophorids, diatoms, and dinoflagellates. The phytoplankton species are generally described by their nominal sizes, Lhcs, and nutrient uptake/storage capacities (e.g., Armstrong 1999, Bissett et al. 1999). Grazing is represented in a form whereby total community grazing pressure is distributed proportionally to algal abundance in parts of the algal size spectrum. A delay in the distribution allows phytoplankton to grow fast enough, and become resistant enough to predation, to escape predator control and bloom. To apply the model, we will specify the growth dynamics and predation susceptibilities of phytoplankton "species" (i.e., representative taxa). Growth will be based on the variable internal stores model of Droop (1968; see also Burmaster 1979, Morel 1987). In this model, uptake and growth are separated dynamically in time. Similar models, based on resource competition, successfully simulate phytoplankton dynamics on time scales of decades (Huisman and Weissing, 1999). Uptake of a limiting nutrient causes the "cell quota" of that limiting nutrient (the amount of that nutrient in the cell's cytoplasm) to increase; algal growth rate then responds to the internal store. In addition, for species with vacuoles, we will add a separate compartment that will allow the species to store nutrients, from which it can supplement its (cytoplasmic) cell quota. By running the model over a range of conditions in historical and modern ocean, we will assess the potential importance of vacuoles for cells of similar size. These simulations will all be conducted in the context of a larger community with microbial loop components and other plankton components over month, year, to geological scales. The coupled geobiological model will be used to assess competitive outcomes over a range of scenarios to address several testable hypotheses. What is the biological forcing and corresponding feedbacks of the anoxic events in the ocean? What is the competitive succession that occurs over long timescales in the oceans? What is the relative importance of abiotic and biotic forcing of the phytoplankton community composition over geologic scales? What is the functional role of morphology?
        Whereas the geobiological models can provide insight to the forcing, maintenance, and consequences of phytoplankton community models, they do not necessarily provide insight to the tempo and functional implications of evolution. These processes will be examined in a series of cost-benefit models that will use the tempo of evolution defined for the functional genes in Objective 1 and extrapolated to assess the physiological fitness for the given eucaryotic taxa. The cost-benefit models quantify fitness in terms of the biosynthetic costs of a compound versus that of the environmental constraints (e.g., Raven, 1997). There are two major questions to be addressed by this effort. First, are the major physiological innovations, which provide significant benefit, the result of unique events in which environmental constraints provided the necessary "activation energy" for innovation or, is the process a trended change that occurs over geologic scales? Do the innovations appear as punctuated change as they co-occur with the major disturbances? Model results will be explored through laboratory experiments, defining the physiological profile for monospecific cultures and conducting competition between two or more taxa experiments under fluctuating conditions. The laboratory work will focus on characterizing growth, differential protein expression and bulk chemistry (C/N/P) of exponentially and stationary phase semi-continuous batch laboratory cultures over a wide range of temperature, nutrient, and light conditions. These laboratory datasets will provide the input data to the geobiological model.
Ecophysiological experiments. There is a vast literature on the ecophysiology of numerous representative species within each of the phytoplankton taxa. Our efforts will use that information base in conjunction with novel approaches designed to test key assumptions and inferences drawn from the models and paleoenvironmental and molecular biological components. These new approaches include: quantifying the role of armor in zooplankton grazing experiments (e.g., The AWI group is using computer models and micromanipulators to study the breaking strength of diatom frustules); elucidating the differential expression of specific functional genes (Liang and Pardee, 1992; Jones et al., 1997) under imposed historical and contemporary oceanic conditions; and assessing the utilization efficiency of light under simulated (modeled) radiative transfer conditions calculated for