The role of iron in phytoplankton photosynthesis, and the potential for iron-limitation of primary productivity in the sea

Photosynthesis Research 39: 275-301, 1994. (~) 1994 Kluwer Academic Publishers. Printed in the Netherlands.

Minireview

The role of iron in phytoplankton photosynthesis, and the potential for iron-limitation of primary productivity in the sea

Richard J. Geider 1 & Julie La Roche 2 1College of Marine Studies, University of Delaware, Lewes, DE 19958-1298, USA; 2Oceanic and Atmospheric Sciences Division, Brookhaven National Laboratory, Upton, NY 11973, USA

Received 7 July 1993; accepted in revised form 29 September 1993 °

Key words: chlorosis, diagnostic, flavodoxin, iron-limitation, photosynthesis, phytoplankton

Abstract

Iron supply has been suggested to influence phytoplankton biomass, growth rate and species composition, as well as primary productivity in both high and low NO 3 surface waters. Recent investigations in the equatorial Pacific suggest that no single factor regulates primary productivity. Rather, an interplay of bottom-up (i.e., ecophysiological) and top-down (i.e., ecological) factors appear to control species composition and growth rates. One goal of biological oceanography is to isolate the effects of single factors from this multiplicity of interactions, and to identify the factors with a disproportionate impact. Unfortunately, our tools, with several notable exceptions, have been largely inadequate to the task. In particular, the standard technique of nutrient addition bioassays cannot be undertaken without introducing artifacts. These so-called 'bottle effects' include reducing turbulence, isolating the enclosed sample from nutrient resupply and grazing, trapping the isolated sample at a fixed position within the water column and thus removing it from vertical movement through a light gradient, and exposing the sample to potentially stimulatory or inhibitory substances on the enclosure walls. The problem faced by all users of enrichment experiments is to separate the effects of controlled nutrient additions from uncontrolled changes in other environmental and ecological factors. To overcome these limitations, oceanographers have sought physiological or molecular indices to diagnose nutrient limitation in natural samples. These indices are often based on reductions in the abundance of photosynthetic and other catalysts, or on changes in the efficiency of these catalysts. Reductions in photosynthetic efficiency often accompany nutrient limitation either because of accumulation of damage, or impairment of the ability to synthesize fully functional macromolecular assemblages. Many catalysts involved in electron transfer and reductive biosyntheses contain iron, and the abundances of most of these catalysts decline under iron-limited conditions. Reductions of ferredoxin or cytochrome f content, nitrate assimilation rates, and dinitrogen fixation rates are amongst the diagnostics that have been used to infer iron limitation in some marine systems. An alternative approach to diagnosing iron-limitation uses molecules whose abundance increases in response to iron-limitation. These include cell surface iron-transport proteins, and the electron transfer protein flavodoxin which replaces the Fe-S protein ferredoxin in many Fe-deficient algae and cyanobacteria.

Introduction

In 1988, Martin and Fitzwater published a paper in Nature that provided compelling evidence for

iron limitation of phytoplankton growth in the subarctic North Pacific. Although oceanographers had previously suggested a role for iron in limiting phytoplankton growth in the sea

276

(Hart 1934), all of the early bioassay experiments (Menzel and Ryther 1961, Tranter and Newell 1963, Menzel et al. 1963) have been largely discredited because of inadequate atten- tion to sources of contamination. Martin and Fitzwater (1988) were the first investigators to employ trace metal clean techniques in the study of nutrient limitation of natural phytoplankton assemblages. Subsequent observations in the Southern Ocean (Martin et al. 1990), equatorial Pacific (Martin et al. 1991) and North Atlantic (Martin et al. 1993), indicated that an effect of iron on the phytoplankton community may be a general phenomenon in high nitrate, low chlorophyll open ocean regions. This research spawned a short-lived proposal to ecologically mediate the effects of atmospheric CO 2 buildup from fossil fuel burning by increasing the oceanic sink for CO 2 through iron fertilization of selected regions of the open ocean (Keir 1991). Martin's iron- limitation hypothesis has been the subject of continuing discussion during the past five years, including a 1991 Symposium of the American Society of Limnology and Oceanography published as Volume 36, Number 8, pp. 1507-1970 of the society journal Limnology and Oceanog- raphy.

The controversy over the role of iron in regulating phytoplankton biomass production and growth rate (Martin and Fitzwater 1988, Martin et al. 1990, 1991, Banse 1990, Dugdale and Wilkerson 1990) has served to highlight the limited extent to which we truly understand environmental regulation of phytoplankton photosynthesis and growth in the sea (Cullen 1991, Falkowski et al. 1992). Despite the role of phytoplankton in biogeochemical cycles of the major (C, N, P, S) and minor (Fe, Zn, Mn, Cu) nutrients (Bruland et al. 1991), the relationship between phytoplankton photosynthesis, growth rate and the concentrations of these nutrients is poorly established. Physiological limitation of phytoplankton growth rate in nature by light, temperature or the availability of inorganic nutrients has rarely been directly determined. Rather, the relationships between growth rate and environmental variables have been inferred from bioassay experiments (Hecky and Kilham 1988, Martin et al. 1991), theoretical models

(Riley 1947, Wroblewski 1977, Fasham et al. 1990, Frost 1991) and correlative studies (Riley 1947, Kolber et al. 1990, Platt et al. 1992).

Oceanographers often consider the adjustments of phytoplankton physiology (i.e., phenotypic responses) to be of secondary importance in determining marine primary productivity, whereas phycologists have long recognized the phenotypic plasticity of microalgae and cyanobacteria (see Yentsch 1980 for a consideration of this problem). With the exception of investigations of photosynthesis-irradiance (PI) response curves, the relative neglect of phytoplankton ecophysiology in oceanographic investigations has arisen in part from inadequacy of our tools for investigating physiological responses in nature, and in part from the difficulty of distin- guishing changes in the species composition of the phytoplankton community from changes in the physiological state of the component species. The situation may improve with more wide- spread use of flow cytometry to characterize the taxonomic structure and physiological condition of phytoplankton assemblages.

Phytoplankton show a large plasticity in ultra- structure, bulk biochemical composition, pigment content, nutrient assimilation and photosynthesis (Myers 1980). The carbon:chlorophyll a weight ratio, for example, can vary by over an

- 1 order of magnitude from < 20 to > 200 g C g chl a (Geider 1993). Limnologists and oceanographers have sought to use this plasticity in defining 'physiological indices' or 'diagnostic markers' of nutrient limitation (Healey 1978, Goldman 1980, Zevenboom et al. 1982, Vincent et al. 1984, Falkowski et al. 1992). Although many species show qualitatively similar physiological responses to limiting factors (Goldman 1980), interspecific (genotypic) variation (Donaghay et al. 1978) hinders attempts to apply these indices to mixed assemblages in nature. Physiological indices can be applied with more confidence when the laboratory 'model' species dominates the natural assemblage under investigation (Zevenboom et al. 1982).

Phytoplankton grow rapidly, typically dividing 2 to 7 times per week (specific growth rates 1/N[dN/dt] = 0.2-0.7 d -1) with recent estimates weighted towards the high end of this range

(Laws et al. 1989, Chavez et al. 1991, Welschmeyer et al. 1991). If unchecked by mortality and export, these growth rates would lead to a 4-130 fold increase in abundance during one week. Although temperate phytoplankton communities undergo order of magnitude changes in abundance seasonally, there are locations in the ocean and times of the year when phytoplankton biomass is more or less constant over a period of weeks to months. These include the high nitrate regions in the subarctic and equatorial Pacific. Mortality due to protozoan and metazoan zooplankton grazing, and other losses due to sinking and physical exchange, can crop the phytoplankton at rates equaling or exceeding the cell division rate (Frost 1991, Banse 1992). Most remineralization of organic matter typically occurs within the euphotic zone, and excretion of ammonium, phosphate and other recycled nutrients by zooplankton can be the major source of these nutrients to the phytoplankton. The interplay of phytoplankton with their grazers has led some investigators to suggest that the cell division rate of phytoplankton is rarely nutrient limited in the sea (Goldman 1980). Although it is necessary that mortality plus export balance cell division when phytoplankton biomass remains constant through time (Banse 1992), this condition can be achieved at both low and high phytoplankton growth rates.

In addition to the interplay of physical factors on growth rate and biomass, it is necessary to consider how resource availability affects species composition. Picoplankton dominate in oligotrophic ocean regions, but rarely achieve den- sities in excess of 0.5 mg chlorophyll a m -3 (Chisholm 1992). Although small size increases the competitive ability to assimilate nutrients at low, diffusion-limited ambient concentrations (Raven 1986), it also increases susceptibility to grazing by microzooplankton. Microzooplankton growth, which increases in efficiency as the prey abundance increases, may place an upper bound on picoplankton biomass in the sea. At saturating prey abundance, microzooplankton growth rates can greatly exceed those of their phytoplankton prey (Banse 1982). Increases in chlorophyll concentration above 0.5 mg m -3 are appar-

277

ently achieved through addition of species of larger size (Chisholm 1992). Although they possess a high capacity for nutrient assimilation due to their high surface to volume ratio (Raven 1986), the picoplankton are characterized by an intrinsically lower resource unlimited growth rate than the diatoms. In addition, the dominantly crustacean predators of diatoms have low growth rates, limiting the degree of coupling between diatom growth and zooplankton grazing (Walsh 1976).

Ultimately, the phytoplankton community composition will be determined by competition for limiting resources and by differential susceptibility to death and export (Lehman 1991). The basis for this competition must lie in features of the physiology of competing species that are amenable to study. New approaches borrowed from biophysics, protein chemistry and molecular biology when coupled with epifluorescence microscopy and flow cytometry may allow physiological limitation of phytoplankton photosynthesis and growth to be unambiguously examined in the near future (Falkowski et al. 1992, La Roche et al. 1993).

Although the recent focus of debate on iron- limitation of phytoplankton productivity has cen- tered on the high nitrate, low chlorophyll blue- waters of the open ocean, there is a historical interest in the role of iron in controlling phytoplankton abundance and species composition in coastal waters. Red tides in Florida have been correlated with high concentrations of iron in freshwater runoff (Ingle and Martin 1971, Kim and Martin 1974). Iron was implicated as a factor in a Gymnodinium bloom in the Gulf of Maine (Glover 1978). Iron-limitation has been inferred for flee-living cyanobacteria and symbiotic di- noflageUates on the Great Barrier Reef (Entsch et al. 1983). In addition, recent evidence suggests that iron additions may also stimulate primary productivity in a low nitrate, low chlorophyll oligotrophic ocean region (Young et al. 1991, DiTullio et al. 1993).

Clearly, we can only give a limited account of the iron-limitation hypothesis in this review. We will consider the physiological responses of phytoplankton to iron-limitation, and the evidence for and against an effect of iron-limitation

278

on phytoplankton species composition, growth rates and productivity in the sea. Finally, we will use this opportunity to make the case for increased use of molecular diagnostics in studies of phytoplankton ecophysiology.

Physiological responses of phytoplankton to iron-limitation

Iron containing proteins are essential for photosynthetic and respiratory electron transport, and are directly involved in nitrate and nitrite reduction, N 2 fixation, chlorophyll synthesis, and a number of other biosynthetic or degradative reactions (Table 1). In addition to roles in photosynthetic and respiratory electron transfer, iron is involved in oxygen cycling as a component of catalase, peroxidase, and some superoxide dismutases. It also plays a role in the tricarboxylic acid cycle in the enzyme aconitase which catalyzes the isomerization of citrate to isocitrate. The iron-sulfur protein ferredoxin is the electron donor employed in SO4 reduction, NO 3 reduction and N 2 fixation. Iron-containing catalysts also play important indirect roles in cellular metabolism by regulating enzyme activity. For example, ferredoxin donates electrons to thioredoxin, which in turn provides reductant to enzymes that activate several chloroplast enzymes (Scheibe 1991).

Much of the research on Fe-limitation in the

photosynthetic apparatus and plant metabolism has been conducted with vascular plants (Marschner 1986, Terry and Abadfa 1986). This literature provides considerable background for comparison with the smaller data base on the impact of Fe-limitation on algae and cyanobacteria. There is universal agreement that a reduction in photosynthetic pigment content (chlorosis) accompanies Fe-limitation. Here the agreement ends, and there are conflicting reports on other aspects of the response of photosynthetic physiology to Fe-limitation.

Chlorosis

A decrease in cell chlorophyll content (chlorosis) is one of the most noticeable symptoms of Fe- deficiency in cyanobacteria (Oquist 1974, Glover 1977,~ Guikema and Sherman 1983, Rueter and Ades 1987), algae (Doucette and Harrison 1990, Greene et al. 1991, 1992) and in vascular plants (Terry 1980, Vale t al. 1987). Iron may directly control chlorophyll synthesis by influencing the formation of protochlorophyllide from Mg- protoporphyrin which is catalyzed by the Fe- containing enzyme coproporphyrinogen oxidase (Spiller et al. 1982). Earlier in the biosynthetic reaction sequence, iron availability may regulate the formation of ~-aminolevulinic acid (ALA) (Chereskin and Castelfranco 1982, Miller et al. 1982). Synthesis of ALA requires NADPH and

Table 1. Roles of Fe in plant metabolism from Galliard and Chan (1980), Sandmann and B6ger (1983) Miller et al. (1984), Marschner (1986), Da Silva and Williams (1991)

Catalyst Reaction

cytochromes cytochrome oxidase Fe-superoxide dismutase catalase peroxidase ferredoxin other Fe-S centers succinate dehydrogenase aconitase coproporphyrinogen oxidase nitrate reductase nitrite reductase nitrogenase lipoxygenase glutamate synthetase xanthine oxidase

photosynthetic and respiratory e- transfer O 2+4e +4H +--~2H20 O 5 + 2H + - - . H 2 0 2 + O 2

2H2G 2 -'~ 2 H20 + 02 R(HO)2 + H202 ~ mo 2 + 2H20 e to NADP ÷, NO 3 , SO~, N 2, thioredoxin photosynthetic and respiratory e- transfer FAD + succinate -o fumarate + FADH 2 isomerization of citrate to isocitrate oxidative decarboxylation of Mg-protoporphyrin NO~ + 2e- --~ NO 2 N O ~ + 6 e +3H ÷ - o N H 4 N 2 + 8 H ÷ --~ 2NH 4 fatty acid oxidation, carotenoid degradation glutamine + a-keto glutarate --, 2 glutamate Xanthine + H20 + 02 "-> uric acid + 02

organic acids derived from the Krebs's cycle, which may be Fe-regulated by the activity of the Fe-S protein aconitase (Yu and Miller 1982). Before assigning control to one of these steps in the chlorophyll biosynthetic pathway, it is worth noting that chlorosis is a general mechanism associated with the down-regulation of the photosynthetic apparatus that occurs in response to nutrient limitation of growth rate by nitrogen, phosphorus and sulfur, in addition to limitation by iron (Sakshaug and Holm-Hansen 1977, Col- lier and Grossman 1992, Geider et al. 1993). Thus, it may not be necessary for iron to directly influence the rate of chlorophyll synthesis. Fe- limitation induced chlorosis is accompanied by reductions in the number of chloroplasts per unit cross-sectional area (Doucette and Harrison 1990). Additional ultrastructural changes accom- panying Fe-starvation may include degeneration of lameUar organization in Gymnodinium sanguineum (Doucette and Harrison 1990), large reductions in granal regions in vascular plant chloroplasts (Vesk et al. 1966, Stocking 1975, Spiller and Terry 1980, Pushnik and Miller 1982), and marked reductions in the number of concentric membranes and carboxysome abundance in the cyanobacterium Anacystis nidulans (Guikema and Sherman 1983).

Accessory pigments and light harvesting

The relative abundances of chlorophylls and carotenoids involved in light harvesting are largely independent of iron status. The ratios of chlorophyll b, 0-carotene, neoxanthin and lutein to chlorophyll a were independent of the extent of chlorosis in Fe-starved sugar beet (Beta vulgaris) (Morales et al. 1990). Similarly, the /3- carotene:chlorophyll a and neoxanthin:chlorophyll a ratios were independent of Fe-stress in leaves of apricot and pear, but the chlorophyll b/a ratio decreased and the lutein:chlorophyll a ratio increased (Vale t al. 1987). As in vascular plants, the ratios of light harvesting pigments (chlorophyll c, fucoxanthin and /3-carotene) to chlorophyll a were independent of the degree of Fe-starvation in batch cultures of the diatom Phaeodactylum tricornutum (Geider et al. 1993). In contrast, there appears to be a greater loss of phycocyanin relative to chlorophyll a in Fe-

279

stressed cyanobacteria (Guikema and Sherman 1983).

The ratio of photoprotective xanthophyll cycle pigments (Demmig-Adams 1990) to chlorophyll a appears to increase in iron-starved plants. For example, zeaxanthin increased at low leaf chlorophyll levels and antheraxanthin was independent of leaf chlorophyll a, although violaxanthin changed in proportion to chlorophyll a in iron- stressed sugar beet (Morales et al. 1990). The violaxanthin:chlorophyll a ratio increased in Fe- stressed apricot, but was independent of Fe- stress in pear (Valet al. 1987). Xanthophyll cycle pigments (diatoxanthin and diadinoxanthin) increased in Fe-starved P. tricornutum (Geider et al. 1993).

Photosynthetic electron transfer

Thylakoid proteins decline to a greater extent than stromal proteins and total cell proteins in Fe-limited cells and leaves. Chlorosis involves a reduction in pigment binding proteins as well as photosynthetic pigments. In addition, there appears to be a reduction in the ratio of reaction center and electron transport chain proteins to chlorophyll in Fe-stressed plants. Reductions in the quantities of cytochromes were observed in Fe-deficient cultures of the cyanobacterium Anacystis nidulans (Guikema and Sherman 1983), and decreases in the ratio of cytochrome f:chlorophyU a were observed in Fe-limited continuous cultures of the diatom Phaeodac- tylum tricornutum and the haptophyte Isochrysis galbana (Glover 1977), and in sugar beet (Terry 1983). The P700:chlorophyll a ratio decreased in Fe-stressed sugar beet (Terry 1983), Anacystis nidulans (Oquist 1974) and Phaeodactylum tricornutum (Greene et al. 1991). Loss of the Photosystem II reaction center protein D1 has been observed in Fe-starved Dunaliella tertiolecta and P. tricornutum (Greene et al. 1992).

Changes in the protein composition of the thylakoid membranes are often accompanied by changes in light absorption, fluorescence excita- tion and fluorescence emission spectra. A shift in red chlorophyll a absorption peak from 679 to 673 nm and an associated change in the peak of light absorption by PS I from 683 to 675 nm was observed in Fe-stressed Anacystis nidulans

280

(Oquist 1974, Guikema and Sherman 1983). Accumulation of a chlorophyll binding protein with molecular weight of 36 kDa accompanied these changes in PSI in Anacystis nidulans (Guikema and Sherman 1983). A loss in 77 K chlorophyll fluorescence at 696 and 716 nm was observed in iron-stressed Synechococcus cedrorum (Pakrasi et al. 1985).

The preceding results apply primarily to the abundance of chloroplast proteins, but do not consider the efficiency of partial photosynthetic reactions. A decline in the efficiency of photochemistry in Photosystem II in response to Fe- limitation is indicated by declines in the ratio of variable to maximum fluorescence (Fv/F~) under Fe-starvation in the cyanobacterium Synechococcus cedrorum (Guikema and Sher- man 1983), the chlorophytes DunalieUa tertiolecta and Scenedesmus quadricauda (Rueter and Ades 1987, Greene et al. 1992), the diatom Phaeodactylum tricornutum (Greene et al. 1992), the dinoflagellate Gymnodinium sanguineum (Doucette and Harrison 1990) as well as in higher plants (Val et al. 1987, Morales et al. 1991). There often is a graded response to Fe- limitation. For example, the response of sugar beet to Fe-deficiency occurred in two phases (Morales et al. 1991). In the first phase, a 75% reduction in leaf chlorophyll occurred without reduction in the photochemical efficiency of PS II as judged by invariant and maximal Fv/F m. In the second phase, further reductions of leaf chlorophyll were accompanied by dramatic declines in Fv/F m. Interestingly, the maximum quantum efficiency for 0 2 evolution declined during the first phase while Fv/F m was invariant. This suggests limitation of whole chain electron flow by catalysts other than PS II under moder- ate degrees of Fe-limitation and is consistent with the observation that PSI and the cytochrome b6/f complex may be more susceptible to Fe-limitation than is PS II. Reductions of Fv/F m despite the accumulation of xanthophyll cycle pigments in Fe-stressed P. tricornutum (Geider et al. 1993) and pear (Vale t al. 1987) indicate that this photoprotective mechanism (Demmig- Adams 1990) was insufficient to protect Photo- system II from damage.

Consistent with the declines of reaction center and electron transport chain proteins are ob-

servations that partial and whole chain electron transport activities decline in Fe-starved cells. The uncoupled rates of PS I, PS II and whole chain electron transport per unit chlorophyll decreased in parallel (Nishio et al. 1985b), suggesting that all photosynthetic electron transport chain components declined to an equal extent under Fe-deficiency. PS I activity appeared to be more sensitive than PS II activity to Fe-limitation in tobacco (Pushnik and Miller 1989). The molecular structure of the PS I has been examined in Fe-stressed tobacco, providing at least a partial explanation for the decreased PSI activity. Although six peptides were found in PSI particles from Fe-replete tobacco, only five were found in Fe-starved plants (Pushnik and Miller 1989). The missing polypeptide in the Fe- stressed cells had an apparent MW of 15 kilodal- tons and may have been a nonheme Fe protein (possibly a bound ferredoxin).

The recovery of photosynthetic energy conversion from iron-starvation in the marine chloro- phyte Dunaliella tertiolecta occurred in three stages (Greene et al. 1992). First, there was a rapid (3-5 h) increase in electron transfer on the acceptor side of PS II associated with synthesis of the cytochrome b6/f complex. Second, there was a slower (10-25 h) increase in the quantum efficiency of PSII photochemistry associated with accumulation of the 32 kDa reaction center II protein D1. Third, there was a slow (> 18 h) increase in cell chlorophyll a levels accompanied by an increase in the efficiency of exciton transfer from the light-harvesting complex to reaction center II. Similar sequences of events have been observed during the recovery of chlorotic cyanobacteria and vascular plant chloroplasts following iron resupply. Guikema (1987) found that Fe appeared in cytochrome f prior to being incorporated into PSI reaction center proteins following supply of 59Fe to Fe-starved Anacystis nidulans. Alleviation of Fe-limitation in tobacco was accompanied by a rapid increase in the uncoupled rates of PSI mediated photosynthetic electron flow, while electron flow through PS II increased only after a long lag period (Pushnik and Miller 1989). Consistent with these observations, the PSI and PSI1 reaction center proteins and cytochrome f were resynthesized immediately following Fe resupply to Fe-stressed sugar beet,

whereas synthesis of chlorophylls and light-harvesting complex proteins occurred only after a lag (Spiller and Terry 1980, Nishio and Terry 1983, Nishio et al. 1985a,b).

A decline in the efficiency of photosynthesis is not an inevitable response to iron-limitation. Although cell-specific and biomass-specific rates of photosynthesis decline under Fe-limitation, most of the decline in these rates can be attributed to chlorosis. In many cases chlorophyll a- specific rates of light-saturated photosynthesis are not affected by Fe-limitation suggesting that damage to the photosynthetic apparatus is not a universal feature of the response to iron-limitation. For example, Fe-deficiency does not depre- ss p~l in many vascular plants (Oquist 1974, Terry 1983). In contrast to the reports of im- paired photosynthetic efficiency described in the preceding paragraphs, Terry (1980) observed that the maximum quantum efficiency of light- limited photosynthesis was independent of the degree of iron-limitation in sugar beet. Finally, Guikema and Sherman (1983) did not observe any changes in PS I or PS II activity (per unit chlorophyll) in Fe-starved Synechococcus ced- r o r u m .

Photochemical oxygen reduction

Superoxide produced in illuminated chloroplasts by photooxidation of light-harvesting pigments or the Mehler reaction can damage proteins and nucleic acids (Salin 1987) unless the free radical is consumed by other scavenging mechanisms. The first step in the destruction of active oxygen species is the dismutation of O into H20 2 and 0 2

by superoxide dismutase. Superoxide dismutase is present as three isoenzymes that differ in the nature of the metal prosthetic group: these are Mn-, Cu/Zn- and Fe-superoxide dismutase (Canini et al. 1992). Fe-superoxide dismutase has been found in the cyanobacteria (Canini et al. 1992 and references cited therein). Catalase and peroxidase provide means of disposing of the n 2 0 x generated by the dismutation of O. Catalase is found predominantly in the peroxi- some with H20 2 consumption in the chloroplast catalyzed largely by peroxidases (Salin 1987). The concentrations of peroxidase and catalase are reduced in Fe-deficient plants (Bouma 1983).

281

This may leave Fe-deficient cells more susceptible to photochemical damage.

Iron assimilation

Microorganisms living in freshwater and soil often release specific iron-binding chelators (siderophores) to the environment (Crichton and Charloteaux-Wauters 1987, Reid and Butler 1991). These siderophores complex Fe(III), and the Fe-siderophore complex is assimilated via specific transport systems. Siderophore production is regulated by iron availability, with synthesis suppressed in high iron environments. Siderophores are produced by marine bacteria (Goyne and Carpenter 1974, Reid and Butler 1991, Haygood et al. 1993) and cyanobacteria (Armstrong and Van Baalen 1979). Low molecular weight, hydroxymate-type siderophores are also released during early stationary phase in Fe-limited batch cultures of some eukaryotic phytoplankters (Prorocentrum minutum, P. mariae-lebouriae, Thalassiosira pseudonana and Dunaliella tertiolecta) (Trick et al. 1983). Despite the detection of siderophores in coastal phytoplankton cultures, the role of siderophores in iron acquisition by marine microbes in the open sea has been largely discounted (Morel et al. 1991) because the efficiency of siderophore mediated iron uptake will be reduced by the inability of cells to recover siderophores excreted into the dilute, low biomass open ocean (Hutchins et al. 1991).

Phytoplankton show both genotypic and phenotypic variability in iron requirements. The concentration of dissolved iron required to satu- rate growth of coastal isolates is over 100 times greater than that required by oceanic isolates (Brand et al. 1983, Sunda et al. 1991). Two strategies can be employed to enhance growth at low dissolved iron availability. One is to increase the number of iron transport sites on the cell surface to exploit encounters with rare Fe(III) species. A second strategy is to decrease the cellular iron quota. Both strategies appear to be employed by open ocean phytoplankton. The remainder of this section considers Fe-uptake. Variations in the cell iron quota are considered in the following section.

Phytoplankton possess a high affinity iron

282

transport system (Anderson and Morel 1982, Harrison and Morel 1986, Hudson and Morel 1990, 1993). Large numbers of low molecular weight Fe-binding ligands located on the plasmalemma of phytoplankton have been inferred from pulse chase labelling experiments, with rapid binding of tracer 59Fe to the cell surface and slower internalization of the labelled iron into the cell observed in the diatom Thalassiosira weissflogii and the haptophyte Pleurochrysis car- terae (Hudson and Morel 1990). Using the sub- strate saturated Fe-uptake rate as a measure of the number of uptake sites, Hudson and Morel (1993) found that the concentration of cell surface iron transport ligands increased about 30 fold between iron replete and iron limited Thalassiosira weissflogii. Saturated rates of iron assimilation by phytoplankton in the equatorial Pacific were of the same order as the maximum uptake rate observed in Fe-starved T. weissflogii, consistent with adaptation to low iron availability in situ (Price et al. 1991). Reduction of the high rates of iron assimilation within two days of iron resupply provided further support for the assertion that phytoplankton were iron limited in situ (Price et al. 1994).

Small size can be considered as an adaptation to low nutrient availability. The diffusion limited uptake rate scales as r, membrane area as r 2 and cell volume (and hence nutrient requirement) as r 3. Thus, small size increases the diffusion limited uptake rate relative to the nutrient requirement by r z and maximum transporter limited uptake rate to nutrient requirement by r. Morel et al. (1991) calculated that cells greater than 10 ~zm in diameter were likely to be diffusion limited at dissolved iron concentration less than 0.05 nM typical of the oligotrophic open sea. Significantly, phytoplankton in oligotrophic (low NO 3 low chlorophyll and high N O ; low chlorophyll) regions are dominated by small (2/xm diameter) cells, but large cells bloom following nutrient enrichment.

A unique mechanism has been hypothesized for iron uptake by Trichodesmium colonies (Rueter et al. 1992). These colonies can actively trap iron-enriched 'dust' particles by trichome movement, and natural samples of Trichodes- mium contain iron-rich inorganic particles. Al- though Trichodesmium does not produce sidero-

phores, it does excrete soluble organic material that can solubilize particulate iron, and enrichment experiments have shown a rapid increase in the ratio of phycoerythrin to chlorophyll upon dust addition (Rueter et al. 1992). A variable proportion of the phytoplankton community is associated with aggregates of non-living organic material and its associated microbiota. These aggregates have been hypothesized to be centers of a microbial food web in open ocean waters (Goldman 1984). Do macroscopic aggregates play a similar role to Trichodesmium colonies by trapping dust particles and microbes within a physical matrix? One may speculate that plasmalemma diaphorases in marine phytoplankton such as the diatom Thalassiosira weissflogii (Jones and Morel 1988) could facilitate the reduction of Fe(III) solid phases to Fe(II), with subsequent uptake of the Fe(II) by plasmalemma iron-binding ligands.

Elemental composition and iron-use efficiency

The iron content of phytoplankton is extremely plastic. Iron is required for a large number of catalysts (Table 1), and minimum iron requirements for algal growth have been calculated by Raven (1988, 1990). Under iron replete conditions, much of the iron may be present as the iron storage product ferritin, most likely local- ized in the chloroplasts (Lane and Skopp 1986). Ferritin is found in prokaryotes, plants, fungi and animals (Crichton and Charloteaux-Wauters 1987). Iron makes up a variable proportion of ferritin mass (up to 20%) and is stored as an inorganic core of hydrated ferric oxide resem- bling ferrihydrite (5 Fe20 3 • 9H20) surrounded by a multi-subunit protein shell. The protein shell (apoferritin) consists of 15-28 kDa molecular weight subunits (the low molecular weight forms have been found in bacteria, intermediate molecular weight forms in animals and the high molecular weight forms in plants) (Crichton and Charloteaux-Wauters 1987, Laulh6re et al. 1989).

Minimum iron contents for stationary phase (non-growing Fe-limited) cultures of algae and cyanobacteria are given in Table 2 and observations for plankton are given in Table 3. These are expressed as the C:Fe, N:Fe and P:Fe ratios

283

Table 2. B u l k e l e m e n t a l c o m p o s i t i o n ( m o l a r ra t ios ) o f s t a t i o n a r y p h a s e i r o n - s t a r v e d a l g a e a n d c y a n o b a c t e r i a

S p e c i e s C / F e N / F e P / F e C / N N - s o u r c e

Gymnodinium sanguineum Gymnodinium sanguineuml Phaeodactylum tricornuturn 2 Thalassiosira weissflogii 3 Eukaryotic phytoplankton 4

c o a s t a l i so la tes

o c e a n i c i so l a t e s

C o a s t a l a n d o c e a n i c c y a n o b a c t e r i a 4 Synechococcus W H 8 0 1 8 5

Synechococcus W H 6 5 0 1 5 Anabaena 6

9400 1490 - 6 .3 N O 3

8580 2240 - 3 .8 N H ~

78 000 12 000 760 6 .5 N O 3

330 000 - - - N O 3

- 1 0 0 - 1 3 0 0 - N O 3

- > 10 000 - N O 3

- 2 5 - 5 0 0 - N O 3

3000 - - N O 3

2750 - - N O 3

250 - - N 2

S o u r c e s f o r t h e i n f o r m a t i o n in th i s t a b l e a r e as fo l lows: 1 D o u c e t t e a n d H a r r i s o n (1991) , 2 G r e e n e e t ai . ( 1991 ) , 3 t h e e a r l y

e s t i m a t e o f 100 000 C : F e b a s e d o n r e p o r t s o f A n d e r s o n a n d M o r e l (1982) a n d H u d s o n a n d M o r e l ( 1990 ) h a s r e c e n t l y b e e n r e v i s e d

u p w a r d to 3 3 0 0 0 0 as r e p o r t e d b y M o r e l e t al . ( 1991) , 4 B r a n d (1991) , s R u e t e r a n d U n s w o r t h (1991 ) , 6 H u t c h i n s e t al . ( 1991 ) .

Table 3. B u l k e l e m e n t a l c o m p o s i t i o n o f p l a n k t o n a n d o c e a n i c p a r t i c u l a t e m a t t e r

L o c a t i o n C / F e N / F e P / F e C / N N - s o u r c e

C a r i b b e a n S e a

Trichodesmium 1 - 465 - -

P a r t i c u l a t e m a t t e r 2 - - 2 0 0 - 2 5 0 -

P l a n k t o n 3 20 000 - 200 -

S u b a r c t i c Pac i f ic 4 26 000 - - -

37 000 E q u a t o r i a l Pac i f ic 5 26 000 3000 200 4 .5

( B i o a s s a y s ) 137 000 27 000 2500 9 .0

m

m

S o u r c e s f o r t he i n f o r m a t i o n in th i s t a b l e a r e as fo l lows: ~ R u e t e r e t al . (1992) 2 B r u l a n d e t al . ( 1991 ) , 3 M a r t i n e t al . ( 1 9 8 9 ) , 4 M a r t i n e t al . ( 1991) .

to allow comparison amongst organisms that may vary in volume by several orders of magnitude. Iron-limited phytoplankton show little plasticity in the C:N:P elemental composition (Sakshaug and Holm-Hansen 1977, Greene et al. 1991) allowing use of the Redfield C:N:P ratio of 106:15:1 to compare iron requirements amongst species.

Organic carbon to iron ratios vary by a factor of 500 f rom<2000 C:Fe in the freshwater N 2 fixing cyanobacterium Anabaena sp. to > 1000000 C:Fe in the diatom Thalassiosira oceanica. Excluding the N 2 fixers, which have a very high Fe-requirement (Raven 1988), the low end of the range is still < 10 000 C:Fe in the red tide dinoflagellate Gymnodinium sanguineum and < 20 000 C:Fe in some coastal cyanobacter- ial strains. Despite the advantage for nutrient uptake conferred by the small size of cyanobacteria, this group of prokaryotes has, in general, a higher cellular iron requirement than eu-

karyotic algae (Brand 1991). The high iron requirements of cyanobacteria is to some extent dependent on their higher ratio of P S I / P S I I (Raven 1988, 1990) compared with eukaryotic algae. Therefore, the dominance of cyanobacteria in oligotrophic regions of high nutrient and low chlorophyll is paradoxical. Finally, the large iron requirements of some coastal isolates suggests the potential for iron-limitation in coastal waters, as suggested for Gymnodinium in the Gulf of Maine (Glover 1978).

The elemental composition of stationary phase cultures (Table 2) gives a lower bound on the iron requirements of growing phytoplankton. Measurements of the carbon:iron ratio for cultures in balanced nutrient-replete or iron-limited growth can be considerably lower than in stationary phase cultures. Calculations of iron use efficiency (i.e., [mol CO 2 fixed][mol cellular Fe]-I s- l ) provide a better basis for comparing the competitive abilities of different organisms

284

growing under resource-saturated and resource- l imited conditions. Raven (1988, 1990) calculated theoretical maximum iron-use efficiencies (and corresponding minimum C:Fe ratios) for phytop lankton based on the Fe content and max imum observed catalytic efficiencies of iron containing components . The theoretical calculations include the Fe required for photosynthetic and respiratory catalysts, and in the cases of N O 3 and N 2 assimilation, the Fe in nitrite reductase and nitrogenase, respectively.

Raven ' s (1988, 1990) calculations demonstra te the high cost of N 2 fixation and the modera te cost of N O 3 assimilation relative to growth on a m m o n i u m (Table 4). Nitrate assimilation and reduction increases the iron requirement for growth by about 60% and N 2 fixation increases the requi rement by 100 fold. In addition, these calculations show the strong impact of low light in multiplying cell iron requirements (Table 4). Cells growing at a light-limiting irradiance of 1 ~ m o l photons m -2 s -1 require about 50 times as much iron as do cells growing at light-saturation. The high iron requirement in low light arises because there is an upper bound on the ratio of light harvesting pigments to photosynthetic reaction centers (Raven 1984); the high intracellular chlorophyll content achieved in shade-adapted cells is accompanied by increases in Fe-containing reaction center and electron transport chain components . In addition, the operat ing efficiency of these components is greatly reduced at limit-

ing irradiances, and maintenance metabol ism further taxes net carbon gain. Given the low iron-use efficiency in light-limited conditions, the implication that Antarct ic phytoplankton are unlikely to be iron-limited because they are light- limited (Mitchell et al. 1991, Nelson and Smith 1991) should be treated with caution. We anticipate that phytoplankton can be effectively t rapped in a high-light state if iron-limitation prevents photoadapta t ion to low light in Fe-poor waters. Consistent with this suggestion are limited observations that the chlorophyll a-specific,

chl light-saturated photosynthesis rate (Pro) can be independent of depth in the upper 100 m of the water column in stratified offshore, i ron-poor waters in the Antarctic (Sakshaug and Holm- Hansen 1986). Sakshaug and Ho lm-Hansen (1986) noted the divergence f rom the typical pat tern in which PCmhl declines in deep water due to shade adaptat ion, but were unable to explain the anomaly. Further research examining the interaction of light- and iron-limitation in both controlled laboratory conditions and in natural systems is clearly warranted.

There are very few observations of Fe-use efficiency for phytoplankton (Table 4). How- ever , the achieved Fe-use efficiencies of at least two diatoms exceed the theoretical opt ima calculated by Raven (1988). This discrepancy indi- cates either errors in the measurements of cellular C:Fe ratios and growth rates, or errors in the assumptions of the theoretical calculations.

Table 4. Observed and theoretical iron requirements and iron use efficiencies for photolithotrophic and heterotrophic growth. The iron-use efficiency is calculated as the product of the growth rate and the C:Fe ratio. The iron use efficiency has units of mol C [mol Fe]-t s-1

Species Growth C: Fe Iron- use Nitrogen Limiting rate (d -1) (mmol/mol) efficiency source factor

Thalassiosira oceanica Thalassiosira pseudonana l

Thalassiosira weissflogii 2 Trichodesmium sp? Theoretical optima 4

1.0 500 5.8 NO 3 Iron-limited 2.1 100 2.4 NO3 Resource-saturated 0.6 110 0.76 NO 3 Iron-limited 1.5 330 5.7 NO~ Resource-saturated 0.14 3.0 0.005 N 2 Not specified 2.6 41 1.28 NH 4 ÷ Resource-saturated 2.6 25 0.80 NO~ Resource-saturated 2.6 0.40 0.012 N 2 Resource-saturated 0.04 0.88 0.025 NH 4 ÷ Light-limited

Sources for the information in this table are as follows: 1 Sunda et al. (1991), 2 the earlier estimates of Anderson and Morel (1982) and Hudson and Morel (1990) have recently been revised upward to 330 000 by Morel et al. (1991), 3 growth rate and Fe:N estimated by Rueter et al. (1992) and assuming a C:N ratio of 6.6 by atoms, 4 theoretical optima calculated from Raven (1988, 1990) assuming 0.45 g C g-1 dry weight.

285

One major limitation of the theoretical analysis was an assumption of 1:1:1:1:1 stoichiometry of PS II:cyt b6/f:cyt c:PS I:ferredoxin in the photosynthetic electron transfer chain (Raven 1988). By relaxing this assumption, and considering the observed variability in the stoichiometry of electron transfer chain catalysts, Raven (1990) obtained maximum iron-use efficiencies for growth on ammonium that ranged from 0.62-2.4 [mol CO 2 fixed][mol cell Fe] -1 s -1. Raven (1988) noted that the theoretical maximum use-efficiencies could be increased by about 30% if ATP for biosynthesis is generated directly by photophos- phorylation (e.g., as opposed to CO 2 fixation into carbohydrate followed by ATP synthesis during mitochondrial oxidation of carbohydrate). The maxima could be further increased by about an additional 10% if flavodoxin and plastocyanin substitute for ferredoxin and cytochrome c in the photosynthetic electron transfer chain (Raven 1988). Even making the most liberal assumptions, the discrepancies between observed and theoretical iron-use efficiencies given in Table 4 cannot be completely resolved. However, the known plasticity in the photosynthetic apparatus suggests that the effect of iron availability on the molecular architecture of the photosynthetic apparatus should be examined in a wider range of oceanic and coastal phytoplankton isolates.

Regulation of cellular free iron levels

Regulation of the intracellular free iron concentration is essential both for controlled synthesis of iron containing catalysts and for preventing damage attendant upon free radical formation from the interaction of 02 with iron. The intracellular free iron pool is augmented by iron uptake from the external medium, is buf- fered by exchange with iron-storage proteins (e.g., phytoferritin), and declines when iron is incorporated into proteins (Fig. 1). The free iron pool directly regulates the activity of some enzymes such as aconitase, and indirectly regulates the synthesis of other proteins, including those involved in iron transport and storage. Simi- larities in the structures of aconitase and iron- regulatory proteins suggest a common mecha- msm for sensing intracellular free iron as dis- cussed below (Rouault et al. 1991).

The molecular mechanisms underlying metabolic responses to iron availability have been well studied in mammals where an iron regulatory protein controls the synthesis of ferritin and transferrin. The mammalian protein binds mRNAs containing iron responsive elements (IREs), and is thus referred to as an IRE-binding protein (IRE-BP). Under iron-replete conditions, these IRE-BPs increase the translation of mRNA coding for the iron storage protein ferritin and decrease the stability of the mRNA for a component of the iron uptake system, i.e., transferrin (Theil 1990). Synthesis of ferritin is translationally controlled by an IRE on the upstream 5' end of ferritin mRNA that tightly binds to the Fe3S 4 form of the IRE-BP, thus limiting synthesis of ferritin under low iron conditions. Turnover of transferrin mRNA is controlled by an IRE on its 3' downstream side that binds to the Fe3S 4 form of the IRE-BP, limiting degradation under low iron conditions. Thus, intracellular free iron exerts a coordinated control over the synthesis of an iron storage protein and an iron-transport protein. That the results from research on mammals may be of interest to oceanographers arises from the observation that there is a high degree of sequence and structural homology between IRE-BP and aconitase (Rouault et al. 1991). Rouault et al. (1991) speculate that aconitase and the IRE-BPs may be members of a family of proteins that sense cellular iron status using a dynamic Fe-S center. Given the central role of aconitase in the citric acid cycle, it is possible that similar proteins play a role in iron-regulation in other organisms and that iron responsive elements play a role in the expression of other iron regulated proteins.

The oceanographic evidence for and against iron-limitation

Advection and diffusion of nutrient rich deep waters into the euphotic zone often stimulates phytoplankton growth. Nitrate and phosphate are typically depleted within days to weeks following nutrient enrichment, depending on the stability of the water column and the incident irradiance. However, there are regions where nitrate and phosphate are never depleted from

286

Colloidal & Complexed Fe

f Dissolved Fe(In)

~ Plasma Membrane

Fe/S and non haeml._ -i-- I ~ Iron Proteins ~ Fe(II) - - Phytoferritin k

ae~n e Proteins

\, oa ims ,:,.vodo.in

Fig. 1. Conceptual model of iron regulation in phytoplankton. Dissolved inorganic iron (predominantly Fe(III)) is taken up from solution by cell surface iron transport ligands. Phytoferritin is an intracellular iron storage protein that buffers the intracellular free iron concentration, allowing cellular iron status to become temporarily uncoupled from the availability of dissolved inorganic iron in the environment. Molecular controls refer to the transcriptional and post transcriptional control of m R N A stability and translational control of protein synthesis. These controls are likely to directly affect a wide range of macromolecules including the cell surface transport ligands, phytoferritin, flavodoxin, porphyrins and Fe-containing enzymes. Flavodoxin may serve as a molecular marker of cellular iron status in those organisms showing iron regulation of flavodoxin synthesis. We anticipate that the synthesis of other Fe-containing molecules (see Table 1) will be inversely related to flavodoxin synthesis. Under conditions of balanced growth, the intracellular flavodoxin content should reflect the external dissolved inorganic iron pool. This opens the possibility of developing a biologically relevant assay of iron availability. Not shown is the possible modification of extracellular particulate and dissolved iron speciation by cellular processes. Modified from Da Silva and Williams (1991).

surface waters and the chlorophyll concentration is always low. Three noteworthy examples are the Southern Ocean, the western equatorial Pacific and the subarctic north Pacific. Various hypotheses have been proposed to explain the high nitrate-low chlorophyll regions (Cullen 1991). The iron-limitation hypothesis of Martin and co-workers (Martin and Fitzwater 1988, Martin et al. 1990, 1991) is only one of the possible explanations. Others include grazer control of a phytoplankton population growing at near its maximum potential in the subarctic Pacific (Miller et al. 1991), and light limitation of growth rate and biomass accumulation in the Southern Ocean (Mitchell et al. 1991, Nelson and Smith 1991). It is unlikely that there is a single limiting factor that regulates phyto-

plankton growth and biomass accumulation in any ocean region at any time. Rather, an interaction of physiological limitations (bottom up) and ecological interactions (top-down) are likely to operate in concert. One goal of scientific enquiry is to identify and quantify the factors with a disproportionate impact. With this goal in mind, we proceed to a consideration of the evidence for and against iron-limitation of phytoplankton growth rate and yield, starting with a consideration of iron availability.

Fe availability

Despite the recognition that iron is a bioactive element (Bruland et al. 1991) and has been suggested to limit primary productivity in some

ocean regions (Martin et al. 1990, 1991, 1993), little is known about iron speciation and availability in the sea. In the 1970s, it was recognized that special precautions need to be taken in order to obtain uncontaminated trace element samples from open ocean waters. As a consequence, the accepted estimates of dissolved and particulate iron concentrations in the open oceans have decreased by several orders of magnitude from the micromolar (Tranter and Newell 1963) to the nanomolar range (Martin and Gordon 1988, Martin et al. 1989, Wu and Luther 1994). Most of iron in seawater is present as particulate matter (operationally defined as Fe retained by a 0.4/zm pore size sieve) (Gordon et al. 1982, Landing and Bruland 1987, Martin and Gordon 1988, Martin et al. 1990, Wu and Luther 1994), and is not directly available to phytoplankton. Dissolved iron in oxic seawater is assumed to be present primarily as Fe(OH)3 and Fe(OH)2 at very low concentrations (Hudson and Morel 1990). Any Fe(II) that may be present in rain-water or Fe-rich anoxic sedimentary pore waters is rapidly oxidized to Fe(III) and precipitated as (oxy)hydroxides in oxic water at p H = 8 . 1 (Stumm and Morgan 1981). As a consequence, dissolved inorganic Fe(II) and Fe(III) species are present at very low concentrations in oxic seawater.

It is generally agreed that Fe transport to the ocean occurs primarily as a solid phase via eolian dust (Moore et al. 1984, Duce and Tindale 1991), and that enrichment of natural assemblages with atmospheric dust can stimulate biomass accumulation (Martin et al. 1991). How- ever, the iron on dust particles must be released into solution before it becomes available to the phytoplankton. Iron uptake rates appear to be controlled by the kinetics of complexation of iron-hydroxide species by cell surface transport ligands (Hudson and Morel 1990, 1993). It is thought that only dissolved inorganic iron species are taken up by phytoplankton, and that dissolution of solid-phase iron is required before iron becomes available to phytoplankton, even when the solid phase consists of colloids < 2 nm in diameter (Rich and Morel 1990, Morel et al. 1991). Although much of the aeolian Fe can be rapidly mobilized into a form that passes a 0.4 /xm pore filter (Zhuang et al. 1990, 1992), this

287

Fe is not necessarily dissolved. The commonly employed operational definition of dissolved iron (0.4/xm) is likely to include complexed and colloidal iron (Wells and Goldberg 1991, Wu and Luther 1994) that may not be available to the phytoplankton.

Crystalline iron oxides are less available to phytoplankton than recently precipitated amorphous iron oxides as reflected by differences in dissolution kinetics (Wells 1988/89, Rich and Morel 1990, Wells et al. 1991). However, photochemical processes may solubilize Fe-phases that would otherwise be unavailable. For example, the highly crystalline iron oxyhydroxide goethite, FeOOH, can be made available to phytoplankton by photochemical processes at p H = 8 (Sulzberger 1990, Wells and Mayer 1991, Wells et al. 1991). Organic matter may enhance the photochemical lability of Fe by facilitating the reduction of solid phase Fe(III) to soluble Fe(II) during photolysis (Finden et al. 1984). The photochemically generated Fe(II) may in turn reoxidize to amorphous, labile Fe(III) compounds or be complexed by organic matter. In this regard, it is of interest that Wells and Goldberg (1991) documented Fe-containing organic colloids (120 nm) in oceanic waters.

Wells and coworkers have attempted to esti- mate the proportion of iron in seawater that is 'available' to phytoplankton. The organic chelate 8-hydroxyquinoline was used to complex dissolved and exchangeable iron. The Fe-hydroxyquinoline complex was extracted from seawater to determine the concentration of 'available' iron. We prefer to designate this measurement as 'chelateable' rather than 'available' iron, and note that chemical determinations of chelateable iron do not provide a measure of the dissolved inorganic iron that can be directly taken up by phytoplankton cell surface iron transport ligands. Rather, chelateable iron measures the pool of dissolved, particulate and complexed iron that is potentially available to a chelate with a given stability constant.

Increases in our understanding of iron chemistry in seawater may come from the application of new electrochemical techniques that allow measurements of low level dissolved and total iron concentrations using small volumes and reduced sample handling. Van den Berg et al. (1991)

288

described an electrochemical method for measuring chelateable iron using the chelate 1-nitroso-2- naphthol (1-N-2-N). Wu and Luther (1994) used cathodic stripping square wave voltammetry to increase the sensitivity of Van den Berg's method allowing measurements of 'chelateable' iron in operationally defined dissolved (0.4/ ,m) and particulate fractions. However, this technique also suffers from the inability to measure truly dissolved iron.

Bioassay experiments

Results of bioassay experiments have played a critical role in lending support to the Fe-limitation hypothesis. Bioassay experiments are decep- tively simple manipulations of the natural phytoplankton assemblage. They involve dis- pensing replicate subsamples of a homogeneous sample to incubation bottles, amending the subsamples with different combinations of suspected limiting nutrients, and measuring the change in biomass, photosynthesis or nutrient uptake after incubation under identical conditions for periods of one day to one week. In the subtropical gyres, it may be necessary to add macro-(NO;, PO43, SiO2) and micro- (FeC13) nutrients, alone and in combination (DiTullio et al. 1993) to identify the limiting nutrient(s). The experimental design is considerably simpler in the high NO3-regions where it is sufficient to employ unamended controls and iron amended treatments (but see Coale 1991 for effects due to other trace elements).

A physiological response to nutrient addition is expected to occur within a relatively short interval (1 day) of the initiation of the experiment. Physiological adjustments involve changes in enzyme concentration or activity in the exist- ing assemblage, and should be observed on time scales appropriate to protein synthesis, degradation and activation/inactivation. Ecological responses involve changes in the community structure, and will be observed only on the longer time scales (> 1 day) associated with cell division and net population growth. Most of the bioassay experiments conducted to date have measured ecological responses of chlorophyll a concentration or phytoplankton abundance.

Even in those bioassay experiments where rates of nutrient uptake or carbon assimilation have been measured, it is not easy to differentiate between the ecological effects of changing biomass or community structure, and physiological effects on enzyme activities.

Bioassay experiments conducted in high nitrate surface waters in the subarctic North Pacific (Martin et al. 1989, Coale 1991), equatorial Pacific (Martin et al. 1991, Price et al. 1991), Southern Ocean (De Baar et al. 1990, Helbling et al. 1991, Martin et al. 1991) and North Atlantic (Martin et al. 1993) provide support for the assertion that the phytoplankton community is affected by iron availability. There appears to be general agreement that the final biomass yield is higher in samples enriched with 1-10 nM iron than in unamended samples, whether biomass is measured by chlorophyll, cell abundance or consumption of macronutrients (i.e., CO 2, NO 3 and PO43). However, the interpretation of these findings differs widely. Martin maintains that if iron affects phytoplankton growth and yield in bottles then it must also affect growth and yield in unconfined natural waters. Detractors of the Fe-limitation hypothesis point to the marked increases of biomass often observed in the unamended 'control' samples. The chlorophyll concentrations achieved in unamended bottles often exceed those observed in unconfined natural conditions (Banse 1990, De Baar et al. 1990, Helbling et al. 1991), indicating that there are no true controls in these bioassay experiments. In addition, the initial rates of chlorophyll accumulation in control and iron-enriched samples are often similar (De Baar et al. 1990, Price et al. 1991), suggesting that biomass accumulation in natural assemblages does not become iron-limited until these assemblages are confined in a bottle for more than 24 hours. If the control treatment does not mimic nature, then infer- ences about natural processes that are based on the experimental treatments must be treated with caution.

The phytoplankton are only one component of a dynamic microbial community that includes microzooplankton and bacteria. Placing this community in bottles removes it from the grazing pressure of larger zooplankton. This can affect the balance between phytoplankton cell division

and mortality rates, with consequences for community structure. Microzooplankton grazing on the smallest phytoplankters may be enhanced due to decreased mortality of the microzooplankton associated with elimination of their predators. Alternatively, bioassay experiments may stimulate net population growth of those phytoplankton species that are immune to microzooplankton grazing, but susceptible to grazing by the rare, larger zooplankton that are excluded from the incubation bottles. Bottle experiments also isolate the community at a fixed optical depth, and prevent losses due to sedimentation and mixing. Changes in phytoplankton community structure have been documented during bottle experiments (Martin et al. 1989, Buma et al. 1991, Chavez et al. 1991, Coale 1991, Hel- bling et al. 1991, DiTullio et al. 1993), with the cell division rates of diatoms often elevated in Fe-enriched samples. Red tide dinoflagellates also appear to be selectively stimulated by additions of Fe in bioassay experiments in coastal waters (Graneli et al. 1986), consistent with their high iron requirements (Doucette and Harrison 1990).

The disagreement over the significance of the Fe-addition experiments hinges on the relative roles of iron in limiting the cell division rates, and grazing in limiting rates of net population increase (Donaghay et al. 1991). There appears to be a growing consensus that grazing is the proximate control on the abundance of the dominant picoplankton in the high NO 3 regions (Walsh 1976, Frost 1991, Cullen et al. 1992), but that Fe-availability ultimately limits the growth of diatoms and other nano- and microalgae and hence system productivity (Price et al. 1994). The main piece of evidence for grazer control is that chlorophyll concentration shows little day to day variability in the high nitrate regions indicating that cell division rates are closely balanced by mortality, of which grazing is the dominant component, and export (Frost 1991, Miller et al. 1991, Banse 1992, Cullen et al. 1992). However, it is possible for cell division and grazing to be in balance at both low nutrient-limited or high nutrient-saturated growth rates, and stability of in situ chlorophyll concentrations is not sufficient to assert grazer control. A second piece of evidence for grazer control is that the chloro-

289

phyll increases are dominated by large eukaryotic phytoplankton in both unamended control treatments and Fe-amended treatments during bioassay experiments (Martin et al. 1989, Chavez et al. 1991, Helbling et al. 1991, Coale 1991, Price et al. 1994), presumably due to interference with grazing by copepods and other large zooplankton grazers that are excluded from the bottles (Frost 1991, Welschmeyer et al. 1991, Banse 1992). The final piece of evidence for grazer control comes from estimates that cell division rates measured by a variety of techniques converge on values of 0.5-1.0 d -1 in both the eastern equatorial Pacific (Chavez et al. 1991, Cullen et al. 1992, Price et al. 1993) and sub-Arctic North Pacific (Welschmeyer et al. 1991). These values are considered to be typical of nutrient saturated phytoplankton at the ambient temperatures (Banse 1992). As Cullen et al. (1992) stated 'These estimated growth rates are not consistent with the notion that the supply of trace elements, severely (emphasis added) restricted the specific growth rates of phytoplankton... ' If 'mild' Fe-limitation is allowed, then the unanswered question is 'How far below the resource saturated rate is the in situ phytoplankton cell division rate?' This question leads us to consider the use of physiological and molecular techniques to diagnose (Falkowski et al. 1992) the relative degree of nutrient limitation of phytoplankton division rates.

Physiological and molecular diagnostics

The difficulties in interpretation of nutrient addition bioassays have prompted oceanographers and limnologists to seek other ways to assess nutrient limitation of phytoplankton growth rate. Some indices, such as photosynthate partitioning into macromolecular classes (Laws et al. 1989), are subject to many of the same bottle effects that complicate the interpretation of nutrient addition bioassays. Other indices, such as the carbon:nitrogen ratio (Goldman 1980), are am- biguous when used to characterize a mixed microbial assemblage that may be dominated by bacterioplankton and detritus (Hobson et al. 1973, Fuhrman et al. 1989, Malone et al. 1993). In addition, indices of bulk elemental composition that have been used to assess nitrogen and/

290

or phosphorus limitation of phytoplankton growth (Goldman 1980, Laws et al. 1989), do not apply to iron-limited cells (Sakshaug and Holm-Hanson 1977, Greene et al. 1991, La Roche et al. 1993). Specifically, carbohydrate and lipid storage products accumulate in nitrogen- and phosphorus-limited phytoplankton (Goldman 1980), but they do not accumulate in iron-limited cells (Sakshaug and Holm-Hansen 1977, Greene et al. 1991, La Roche et al. 1993).

Chlorophyll a-specific rates of CO 2 (Cullen 1991, Cullen et al. 1992, Price et al. 1994), NO 3 (Price et al. 1991, 1994) and N 2 assimilation (Rueter et al. 1990) have been used to examine physiological limitation by nutrients. Typically, the initial (presumed in situ rates) are compared with rates measured in control and iron-amended samples in order to compensate for the possibility of pronounced interspecific variability. Thus, the approach still relies on nutrient addition bioassays. However, the emphasis is shifted from a consideration of changes in biomass to a consideration of changes in physiological state. The approach also allows measurement of the physiological state in situ for comparison with experimental treatments. Based on this approach, Price and coworkers found that iron- addition stimulated the rates of NO 3 reduction and CO 2 assimilation in the high nitrate region of the equatorial Pacific. Significantly, the stimulation of chlorophyll-specific NO 3 assimilation rates was limited to cells retained on a 3 /zm sieve. The picoplankton (with a nominal cell diameter of < 2 txm) did not increase their capacity to reduce NO 3 or assimilate CO 2 in response to iron additions. This physiological response provided evidence that the dominant assemblage was not iron-limited in situ, but that rare species which bloomed following iron addition were in fact iron-limited in situ. Further- more, interference with microzooplankton grazing by adding inert microspheres stimulated chlorophyll accumulation without Fe-enrichment (Price et al. 1994). Price et al. (1994) invoked a combination of physiological and ecological factors to account for their observations. They concluded that large species could not take advantage of the high ambient NO~ levels because low iron availability limited the capacity for NO 3 assimilation. Recall that the iron-use

efficiency for growth on nitrate is about 60% greater than for growth on ammonium (Raven 1988, 1990). The small size of the picoplankton allowed them to outcompete the larger phytoplankton for both iron and ammonium, however, net picoplankton growth was controlled by crop- ping by microzooplankton grazers.

Greene et al. (1994) used fast repetition rate fluorometry to assess the quantum yield of Photosystem II photochemistry in the high NO~ equatorial Pacific. Based on a low ratio of variable to maximum fluorescence (Fv/Fm) and recovery of Fv/F m following iron addition, they concluded that the dominant picoplankton assemblage, and not just the rare larger species, were iron-limited. This contrasts with the conclusion of Price et al. (1994). Fv/F m is determined by the functional organization of the photosynthetic apparatus, which can be damaged or altered in response to iron (Greene et al. 1992) and nitrogen (Kolber et al. 1988) limitation. Unlike physiological diagnostics which show pronounced interspecific variability in absolute mag- nitudes, Fv/F m appears to have an absolute upper bound that is independent of the taxon under consideration (Falkowski et al. 1992, Greene et al. 1994). Thus, submaximal values of Fv/F m are taken as strong evidence of physiological stress. The case for iron-limitation of the in situ assemblage comes from observations of low Fv/F m in situ, with recovery in Fe-amended samples, but not in unamended controls (Greene et al. 1994). However, unlike the laboratory situation where recovery of Fv/F m following nutrient resupply precedes the net increase of chlorophyll (Greene et al. 1992, Geider et al. 1993), the increases of Fv/F m and chlorophyll were coincident in the Fe-amended equatorial Pacific assemblage. The biophysical diagnostic Fv/F m provides information on the state of the photosynthetic apparatus that is not available from measurements of chlorophyll a, but it fails to resolve species-specific responses. If some members of the phytoplankton assemblage are more responsive to iron addition than others, the possibility arises that the observed biophysical signal (Fv/Fm) will be 'diluted' by a background of unresponsive taxa until the responsive taxa achieve sufficient biomass.

The contrasting conclusions of Price et al.

(1991, 1994) and Greene et al. (1994) regarding the iron status of the equatorial Pacific point to the difficulties in interpretations which rely on bioassays. In particular, unless the effects due to changes in species composition following nutrient addition can be ruled out, even these physiological and biophysical assays do not provide unambiguous interpretations.

The colonial cyanobacterium Trichodesmium provides a more tractable system for examining the physiological effects of nutrient limitation because large quantities of the organisms can be readily separated from other phytoplankton using nets and screens. Increases in CO2 and N 2 fixation rates, chlorophyll content and in the ratio of variable to maximum fluorescence following addition of iron to natural populations of Trichodesmium provided evidence for in situ physiological limitation by iron in nitrogen-depleted subtropical waters (Rueter 1988, Rueter et al. 1992).

Plant physiologists faced with the problem of assessing physiological limitations under field conditions have devised assays based on the premise that absence of an essential mineral element will lead to a reduction in the concentration of enzymes requiring that nutrient. As a consequence, agricultural researchers have used changes in the activities of specific enzymes to assess mineral nutrition (Brown and Hen- dricks 1952, Bouma 1983). The responses of a number of iron-containing enzymes have been examined. For example, catalase activity (Leidi et al. 1986) and lipoxygenase activity (Boyer and Van der Ploeg 1986) are directly correlated with iron supply. Following a similar tactic, Glover (1977) reported a large decline in the ratio of cytochrome f:chlorophyll a in iron-limited cultures of two marine algae (P. tricornuturn and I. galbana). She also used variations of this ratio to infer iron-limitation of dinoflagellate production in the Gulf of Maine (Glover 1978). Similarly, iron-limitation has been inferred for zooxanthel- lae on Davies Reef, Australia (Entsch et al. 1983), and Trichodesmium colonies from the Caribbean (Rueter et al. 1992) from variations in ferredoxin content.

In the approaches outlined above, a decline in enzyme activity is observed in response to Fe- limitation. These indicators are susceptible to

291

interferences by factors other than iron-limitation such as toxicity or limitation by other elements (Bouma 1983). For example, use of peroxidase activity as an indicator of Fe-status in soya bean was complicated by an interactive effect with Mn (Leidi et al. 1986). It is prefer- able to obtain a 'positive' diagnostic indicator that is turned-on in response to Fe-limitation rather than to use 'negative' indicators which are turned-off. Two categories of catalysts are likely to be fruitful as positive diagnostic indicators. These are (1) proteins associated with iron transport and (2) substitutable catalysts that perform the same function as Fe-containing catalysts but do not contain iron. Changes in iron-uptake ligands were briefly considered in regard to the kinetics of iron transport in a previous section. Here we consider the possibility that substitutable catalysts can serve as diagnostic markers of iron-limitation.

A prime candidate for a diagnostic of iron- limitation is flavodoxin. Flavodoxin is a small electron transfer protein containing ravin mono- nucleotide as the prosthetic group. It can replace ferredoxin on the downstream side of Photo- system I. Flavodoxin and ferredoxin have about equal catalytic activities on a molar basis (Smillie 1965), but flavodoxin is about twice as large as ferredoxin (22-24 kDa versus 10-12 kDa). Flavodoxin has three attributes that make a potentially good diagnostic indicator of iron-limitation in phytoplankton. First, in many organisms, it accumulates only under conditions of iron-limitation (Table 5). Second, it is a stable molecule that can be extracted and quantified. Third, it shows a strong signal in response to iron-limitation and can be detected in a crude soluble protein fraction separated by SDS- PAGE (La Roche et al. 1993a). Given these attributes, it is not surprising that use of flavodoxin to assess Fe-limitation has been suggested before. Entsch et al. (1983) used the presence of flavodoxin in the cyanobacterium Phormidinium sp. to infer iron-limitation in this organism on the Davies Reef, Australia.

Flavodoxin has been shown to be induced only under iron-limiting conditions in representatives from diverse taxa including obligate aerobic, facultative anaerobic, non-oxygenic photosynthetic bacteria, and cyanobacteria (Fox 1976,

292

Table 5. Distr ibution and characteristics of flavodoxins: aa = amino acid sequence, nt = nucleotide sequence

Species Molecular Sequence Iron weight (kDa) availability regulat ion

Prokaryotes A nacystis nidulans l o .23 18.8-20.3 aa / nt induced 23

Anabaena PCC711911 18.9-21.5 aa /n t induced 11 Anabaena variabilis 712024 19.0 nt induced 24 Azotobacter chroococcum 2"3 21 partial not de te rmined Azotobacter vinelandii 3'26"33 19.5-23.0 aa /n t constitutive 2s Clostridium acetabutylicum 3° 15.3 aa /n t not de te rmined Clostridum MP 32'33 15.8 aa induced 27 Clostridium pasteurianum 13.31 14.6 aa induced18 Desulfovibrio vulgaris 6'9'21 15.8-16.3 aa /n t constitutive 8 Desulfovibrio desulfuricans 15 15.6 aa /n t not de te rmined Desulfovibrio gigas s 16.0 not de termined induced 14 Desulfovibrio salexigens TM 15.8 aa /n t not de te rmined Enterobacter agglomeraus20 not determined aa /n t not de te rmined Escherichia coli 29 14.5 aa /n t constitutive 29 Klebsiella pneumoniae 1.7 18.9-19.1 aa / nt not de termine d Megasphaera elsdenii LC132,33 14.5 aa induced32.33 Peptostreptococcus elsdenii 32'33 15.0 aa induced 24a9 Rhodobacter capsulatus 17 17.2 aa /n t not de te rmined Rhodospirillum rubrum 2~ 22.8 partial induced 26 Synechococcus PCC700225 not de termined aa /n t induced 25 Synechococcus lividus 4 17 not de termined constitutive 5 Eukaryo tes Chlorella fusca 36 22.0 not de termined induced 36 Chondrus crispus 12'35 22.5 aa constitutive 12 Dunaliella tertiolecta 22 22 partial induced 22 Phaeodactylum tricornutum 22 23 partial induced 22 Pycnococcus provasoli 22 22 not de termined induced 22

Sources of informart ion for this table are as follows: i Arnold et al. (1988), 2 Bagby et al. 1991, 3 Bennet t et al. (1988), 4 Bothe et al. (1971), 5 Crespi et al. (1972), 6 Curley and Voordouw (1988), 7 D r u m m o n d (1985), 8 Dubourd ieu and Le Gall (1970), 9 Dubourd i eu et al. (1973), 10 Enstch and Smillie (1972), 11 Fillat et al. (1991), 12 Fitzgerald et al. (1978), 13 Fox et al. (1972), 14 Hatchik ian et al. (1972), 15 Helms and Swenson (1991), 16 Helms et al. (1990), 17 Jouanneau et al. (1990), 18 Knight and Hardy (1966), 19 Knight and Hardy (1967), 20 Kreutzer et al. (1991), 21 Krey et al. (1988), 22 La Roche et al. (1993b), 23 Laudenbach et al. (1988), 24 Leonhard t and Straus (1989), 2s Leonhard t and Straus (1992), 26 MacKnight et al. (1974), 27 Mayhew (1971), 28 Mayhew and Ludwig (1975), 29 Osborne et al. (1991), 30 Santangelo et al. (1991), al Tanaka et al. (1971), 32 Tanaka et al. (1973), 33 Tanaka et al. (1974), 34 Tanaka et al. (1977), 35 Wakabayashi et al. (1989), 36 Zumf t and Spiller (1972). O the r dia toms in which flavodoxin is i ron-regulated are Thalassiosira weissflogii, T. oceanica (13-1) , T. pseudonana (3H), Chaetoceros gracilis, C. muelleri, Chaetoceros sp., Nitzschia closterium, Cylindrotheca fusiformis, Phaeodactylum tricornutum, Skeletonema costatum (J. La Roche, unpubl ished results).

Sandmann and Malkin 1983). In addition, flavodoxin is present only under iron-limited conditions in the chlorophytes and diatoms that have been examined to date (La Roche et al. 1993b). However, flavodoxin has been reported to be constitutive in the red alga Chondrus crispus (Fitzgerald et al. 1978). An essential component of the photosynthetic electron transfer chain in the heterocysts of cyanobacteria (Sandmann et al. 1990), flavodoxin synthesis is usually re- pressed under iron replete conditions in other

cyanobacteria (Table 5). Flavodoxin has not been found in higher plants, but substitution of other reducing agents such as flavin mononu- cleotides and riboflavin for ferredoxin has been suggested by Miller et al. (1984). In contrast to the diatoms, chlorophytes and cyanobacteria which produce flavodoxin in iron-limited cultures, Doucette (personal communication) reports that flavodoxin is not induced by iron limitation in a coastal dinoflageUate and two coastal coccolithophorids that he has examined.

293

These findings await corroboration, but indicate the need to examine flavodoxin synthesis in a wide range of oceanic and coastal taxa under appropriate experimental conditions.

Growth conditions are not always well described in early experiments and it is difficult to assess the degree to which iron affects flavodoxin levels. This is clearly acknowledged by Crespi et al. (1972). Although these authors detected flavodoxin in S. lividus under their 'normal' growth conditions, they reported variability in the level of flavodoxin between algae grown in different batches of media. They tentatively attributed these variations in flavodoxin levels to variations in iron levels in the media. In addition, Chondrus crispus is thought to express flavodoxin constitutively but to our knowledge, there are no published accounts of attempts to grow this macroalgae under nutrient replete and iron-deficient conditions (Fitzgerald et al. 1978). Variations in the tolerance range of each species to iron-limitation also prevents the use of a uniform media to study iron regulation. This is especially marked in heterotrophic bacteria which have different requirements because of the diversity of metabolic pathways. For example, D. gigas cannot grow under very low iron concentrations because the respiratory metabolism of this sulfate reducing bacteria requires the synthesis of cytochrome (Hatchikian et al. 1972). What constitutes a low iron media is also am- biguous in much of the work on flavodoxin expression. For example, the ' low iron medium' of Cusanovich and Edmonson (1971) contained 6 /zM iron. There is a clear need for a better description of iron requirements if we are to understand iron regulation of gene expression.

In at least two species of cyanobacteria (Synechococcus PCC7002, formerly Agmanellurn quadruplicatum and Synechococcus PCC7942, formerly Anacystis nidulans R2) the flavodoxin gene isiB is co-transcribed with a gene encoding for a light-harvesting protein of PSII, isiA (Laudenbach et al. 1988, Leonhardt and Straus 1992). The genes are arranged as a dicistronic operon which is transcriptionally regulated by iron. Under iron-limitation or stress, this operon yields two messages, a long one containing both isiA and isiB and a smaller more abundant one (approximately 10 fold) containing only the isiA

gene. In Synechococcus PCC7002, the isiA gene has 40% similarity with the CP 43 gene at the amino acid level. The 36 kDa protein encoded by this gene has been detected during iron stress in Synechococcus PCC7942 (Riethman and Sher- man 1988). Like flavodoxin, the transcript and the 36 kDa protein both disappear during the recovery from iron stress (Pakrasi et al. 1985). Moreover, inactivation of the isiAB operon in Synechococcus PCC7974 produces mutants which do not show the typical symptoms of iron- limitation under iron-stressed conditions (Bur- nap et al. in press). In contrast, the psbC and the petF genes encoding for CP 43 and ferredoxin, respectively, are constitutively expressed under all regimes of iron nutrition. Translational control of these two proteins has been postulated by Leonhardt and Straus (1992). In addition, these authors have found that upstream regulatory sequences of the isiAB operon are similar to the Fur-binding site in the aerobactin promoter, another iron-regulated bacterial gene.

Resource regulation of growth rate and productivity

Increased yield of phytoplankton biomass in bioassay samples amended with iron suggests that iron supply may limit phytoplankton growth rate or regulate the species composition. In- creases of chlorophyll-specific NO 3 uptake rates (Price et al. 1993), decreases of chlorophyll- specific iron incorporation rates (Price et al. 1993) and increases of Fv/F m (Greene et al. 1994) in iron-amended samples suggest that part or most of the in situ assemblage of phytoplankton in the high nitrate equatorial Pacific is iron-limited. However, without corroborating evidence at larger, ecologically relevant scales (Hecky and Kilham 1988), bioassay experiments provide only presumptive evidence for Fe-limitation in the sea. We now turn to an examination of the oceanographic evidence for iron-limitation in the equatorial Pacific.

We can consider the ecological control strength of iron on phytoplankton growth rate in terms of the increase in growth rate of the natural assemblage that could be achieved if iron could be supplied in excess of the maximum

294

potential demand, all other factors remaining equal. However, unlike the laboratory, all other factors cannot be independently controlled in the real ocean, and an extensive and appropriate data base may be required to judge the control strength of iron relative to the control strengths of other potentially limiting factors such as nitrate or light. Such a data set was exploited by Barber and Chavez (1991) to suggest that Fe- limitation was responsible for a 15-20% reduction in phytoplankton growth rate and a 20-25% reduction in integrated euphotic zone primary productivity in the eastern equatorial Pacific. The data set consisted of measurements of primary productivity, chlorophyll and nitrate col- lected at 337 stations obtained on 16 cruises to the equatorial Pacific conducted between 1983 and 1990. Barber and Chavez (1991) concluded that primary productivity was limited by nitrate and/or phosphate injection into the euphotic zone in the western equatorial Pacific. In those waters where the nitrate concentration at 60 m depth was < 12/xM the areal primary productivity rate increased linearly with subsurface nitrate concentration. This positive relation between growth rate and subsurface nitrate in the western equatorial Pacific was due largely to a shift in the depth of the peak chlorophyll concentration toward the surface (and hence to higher irradiances) as the nitrate concentration at 60 m increased (Barber and Chavez 1991). There may also have been a component associated with alleviation of nitrate limitation of growth rate in the light saturated layer. However, higher subsurface nitrate concentrations in the eastern equatorial Pacific did not lead to enhanced primary productivity. Thus, integral primary productivity was a saturating function of subsurface NO3, suggesting a switch to a second limiting factor in the high NO 3 waters. An exception to this pattern was observed in the immediate vicinity of the Galapagos Islands. The Galapagos data fell on the trend expected from the western equatorial Pacific data and provided circumstantial evidence for iron limitation. The higher primary productivity and growth rates near the Galapagos, relative to other waters with similar subsurface nitrate concentrations, was interpreted to result from Fe enrichment of entrained waters by iron released from shallow

sediments. This sedimentary supply of iron near the Galapagos contrasts with the dominance of the eolian source of iron over most of the open oceans. Thus, the data summarized by Barber and Chavez (1991) provides evidence that the control strengths of light, nitrate and iron on phytoplankton growth rate and productivity vary systematically across the equatorial Pacific.

Conclusion and future research

In the five years since Martin and Fitzwater (1988) provided the first compelling evidence in support of the Fe-limitation hypothesis, and despite initial skepticism, a consensus appears to be emerging that Fe does affect phytoplankton growth in many ocean regions. The hypothesis that has gained favor is that the availability of iron differentially affects the physiological state of the picoplankton and the larger phytoplankton. Specifically, iron availability appears to limit the growth rates of rare species, but that most of the phytoplankton assemblage is growing under Fe-replete conditions (Chavez et al. 1991, Price et al. 1993, DiTullio et al. 1993). The rare, Fe-limited species may bloom upon addition of iron to natural systems, and these blooms may have important biogeochemical implications. This hypothesis is in accord with observations that the mean growth rate of the phytoplankton assemblage in the high nutrient regions is high (about 0.7 d -1) (Chavez et al. 1991, Cullen et al. 1992, Price et al. 1994), biomass accumulation in bioassay experiments can be stimulated by iron addition (Martin et al. 1991), but Fe-addition results in a consistent shift in phytoplankton community structure away from picoplankton and towards diatoms (Buma et al. 1991, Coale 1991). Thus, it appears that the addition of Fe may allow rare Fe-limited clones to bloom in an otherwise Fe-sufficient assemblage. However, this consensus is not universal, with recent evidence suggesting that iron availability limits the growth rate of the entire phytoplankton assemblage (Greene et al. 1994).

Unfortunately, the consensus rests on the very shaky foundation of nutrient addition bioassay experiments interpreted in an ecological (as opposed to a physiological) context. Results of a

bioassay are most secure when used to infer (or refute) potential nutrient limitation of physiological processes, such as photosynthetic energy conversion, carbon fixation or nutrient uptake. Use of bioassays to examine ecological responses is suspect because the act of placing the microbial community in a bottle isolates it from important (perhaps the dominant) ecological processes. Examination of the response of phytoplankton within an intact open ocean ecosystem to a mesoscale iron-addition experiment may help to resolve the conflicting evidence (Martin 1992). However, interpretation will be complicated by lack of replication, inability to run appropriate controls and uncertainties about iron chemistry and availability.

An alternative approach is to directly address the fundamental question of whether Fe availability affects the physiological condition of the rare, but potentially bloom-forming phytoplankters in the natural assemblage. This will require species-specific resolution of quantitative differences in the abundance of key (diagnostic) enzymes or other macromolecules. Preliminary results suggest that one such marker is the electron transfer protein flavodoxin. Fluores- cence immunoassays, with single cell resolution, may allow a species-specific determination of Fe-status.

Acknowledgments

We dedicate this paper to the memory of John H. Martin, who more than any other person is responsible for making iron limitation of phytoplankton growth in the sea a legitimate research endeavor. We profited from discussions with our colleagues Richard Greene, Paul Falkowski and George Luther. We thank Louis Sherman and Nell Price for providing prepublication copies of manuscripts describing their recent research, Gregory Doucette for providing unpublished information, and Ric Greene for providing a recent manuscript. The manuscript benefitted from the suggestions of two anonymous review- ers. We thank Peggy Conlon for assistance in preparing the manuscript. Geider's experimental research on iron-limitation in phytoplankton has been supported by the National Science Founda-

295

tion (OCE-8915084 and OCE-9300491), the De- laware Sea Grant Program (R/B-34), and aided by an equipment grant from the University of Delaware Research Fund. La Roche's work on iron-limitation has been supported by the De- partment of Energy under BNL contract DE- ACO2-76CH000016.

References

Anderson MA and Morel FMM (1982) The influence of aqueous iron chemistry on the uptake of iron in the coastal diatom Thalassiosira weissflogii. Limnol Oceanogr 27: 789- 813

Armstrong JE and Van Baalen C (1979) Iron transport in microalgae: the isolation and biological activity of a hydro- xamate siderophore from the blue-green alga Agmenellura quadruplicatum. J Gen Microbiol 111:253-262

Arnold W, Rump A, Klipp W, Priefer UB and Puehler A (1988) Nucleotide sequence of a 24,206-base-pair DNA fragment carrying the entire nitrogen gene duster of Klebsiella pneumoniae. J Mol Biol 203:715-738

Bagby S, Barker PD, Hill HAO, Sanghera GS, Dunbar B, Ashby GA, Eady RR and Thorneley RNF (1991) Direct electrochemistry of two genetically distinct flavodoxins isolated from Azotobacter chroococcurn grown under nitrogen-fixing conditions. Biochem J 277:313-319

Banse K (1982) Cell volumes, maximal growth rates of unicellular algae and ciliates, and the role of ciliates in the marine pelagial. Limnol Oceanogr 27:1059-1071

Banse K (1990) Does iron really limit phytoplankton production in the offshore subarctic Pacific? Limnol Oceanogr 35:772-775

Banse K (1992) Grazing, temporal changes of phytoplankton concentrations, and the microbial loop in the open sea. In: Falkowski PG and Woodhead AD (eds) Primary Prod- uctivity and Biogeochemical Cycles in the Sea, pp 409-40. Plenum Press, New York

Barber RT and Chavez FP (1991) Regulation of primary productivity rate in the equatorial Pacific. Limnol Oceanogr 36:1803-1815

Bennett L, Jacobson M and Dean DR (1988) Isolation, sequencing, and mutagenesis of the nifF gene encoding flavodoxin from Azotobacter vinelandii. J Biol Chem 263: 1364-1369

Bothe H, Hemmerich P and Sund H (1971) Properties of phytoflavin isolated from the blue-green alga Anacystis nidulans. In: Karmin H (ed) Flavins and Flavoproteins, pp 211-226. University Press, Baltimore

Bouma D (1983) Diagnosis of mineral deficiencies using plant tests, ln: L/iuchli A and Bieleski RL (eds) Inorganic Plant Nutrition, pp 120-146. Springer-Verlag, Berlin

Boyer RF and Van der Ploeg JR (1986) Iron metabolism in higher plants. The influence of iron on bean leaf lipoxygenase. J Plant Nutr 9:1585-1600

Brand LE (1991) Minimum iron requirements of marine phytoplankton and the implications for the biogeochemical

296

control of new production. Limnol Oceanogr 36: 1756- 1771

Brand LE, Sunda WG and Guillard RRL (1983) Limitation of marine phytoplankton reproductive rates by zinc, manganese and iron. Limnol Oceanogr 28:1182-1198

Brown JC and Hendricks S (1952) Enzymatic activities as indicators of copper and iron deficiencies in plants. Plant Physiol 27:651-660

Bruland KW, Donat JR and Hutchins DA (1991) Interactive influences of bioactive metals on biological production in oceanic waters. Limnol Oceanogr 36:1555-1577

Buma AGJ, De Baar HJW, Nolting RF and Van Bennekom AJ (1991) Metal enrichment experiments in the Weddell- Scotia Seas: Effects of Fe and Mn on various phytoplankton communities. Limnol Oceanogr 36:1865-1878

Burnap RL, Troyan T and Sherman LA (1993) The highly abundant Chl-protein complex of iron-deficient Synech- ococcus sp PCC 7942 (CP 43') is encoded by the isiA gene. Plant Physiol 103:893-902

Canini A, Civitareale P, Marini S, Grilli Caiola A and Rotilio G (1992) Purification of iron superoxide dismutase from the cyanobacterium Anabaena cylindrica Lemm. and localization of the enzyme in heterocysts by immunogold labeling. Planta 187:438-444

Chavez FP, Buck KR, Coale KH, Martin JH, DiTullio GR, Welschmeyer NA, Jacobson C and Barber RT (1991) Growth rates, grazing, sinking and iron-limitation of equatorial Pacific phytoplankton. Limnol Oceanogr 36: 1816- 18331

Chereskin B and Castelfranco P (1982) Effects of iron and oxygen on the biosynthetic pathway in etiochloroplasts. II. Observations on isolated etiochloroplasts. Plant Physiol 69: 112-116

Chisholm SW (1992) Phytoplankton size. In: Falkowski PG and Woodhead AD (eds) Primary Productivity and Biogeochemical Cycles in the Sea, pp 213-37. Plenum Press, New York

Coale KH (1991) Effects of iron, manganese, copper, and zinc enrichments on productivity and biomass in the subarctic Pacific. Limnol Oceanogr 36:1851-1864

Collier JL and Grossman AR (1992) Chlorosis induced by nutrient deprivation in Synechococcus sp. strain PCC 7942: Not all bleaching is the same. J Bacteriol 174:4718-4726

Crespi HL, Smith U, Gajda L, Tisue T and Ammeraal RM (1972) Extraction and purification of 1H, 2H, and isotope hybrid algal cytochrome, ferredoxin, and flavodoxin. Bio- chim Biophys Acta 256:611-618

Crichton RR and Charloteaux-Wauters M (1987) Iron transport and storage. Eur J Biochem 164:485-506

Cullen JJ (1991) Hypotheses to explain high-nutrient conditions in the open sea. Limnol Oceanogr 36:1578-1599

Cullen JJ, Lewis MR, Davia CO and Barber RT (1992) Phtosynthetic characteristics and estimated growth rates indicate that grazing is the proximate control of primary production in the equatorial Pacific. J Geophys Res 97: 639-654

Curley GP and Voordouw G (1988) Cloning and sequencing of the gene encoding flavodoxin from Desulfovibrio vulgaris Hildenborough. FEMS Microbiol Lett 49:295-299

Cusanovich MA and Edmondson DE (1971) The isolation

and characterization of Rhodospirillum rubrum flavodoxin. Biochem Biophys Res Commun 45:327-336

Da Silva FJJR and Williams RJP (1991) The Biological Chemistry of the Elements. Clarendon Press, Oxford

De Baar HJW, Buma AGJ, Nolting RF, Cadre GC, Jacques G and Trrguer PJ (1990) On iron limitation of the Southern Ocean: Experimental observations in the Wed- dell and Scotia Sea. Mar Ecol Prog Ser 65:105-122

Demmig-Adams B (1990) Carotenoids and photoprotection in plants: A role for the xanthophyll zeaxanthin. Biochim Biophys Acta 1020:1-24

DiTullio GR, Hutchins DA and Bruland KW (1993) Inter- action of iron and major nutrients controls phytoplankton growth and species composition in the tropical North Pacific Ocean. Limnol Oceanogr 38:495-508

Donaghay PL, DeManche JM and Small LF (1978) On predicting phytoplankton growth rates from carbon:nitrogen ratios. Limnol Oceanogr 23:359-362

Donaghay PL, Liss PS, Duce RA, Kester DR, Hanson AK, Villareal T, Tindale NW and Gifford DJ (1991) The role of episodic atmospheric nutrient inputs in the chemical and biological dynamics of oceanic ecosystems. Oceanography 4:62-70

Doucette GJ and Harrison PJ (1990) Some effects of iron and nitrogen stress on the red tide dinoflagellate Gymnodinium sanguineum. Mar Ecol Prog Ser 62: 293- 306

Doucette GJ and Harrison PJ (1991) Aspects of iron and nitrogen nutrition of the red tide dinoflagellate Gymnodinium sanguineurn. I. Effects of iron depletion and nitrogen storage on biochemical composition. Mar Biol 110:165-173

Drummond MH (1985) The base sequence of the nifF gene of Klebsiella pneumoniae and homology of the predicted amino acid sequence of its protein produce to other flavodoxins. Biochem J 232:891-896

Dubourdieu M and Le Gall J (1970) Chemical study of two flavodoxins extracted from sulfate reducing bacteria. Bio- chem Biophys Res Commun 30:965-972

Dubourdieu M, MacKnight ML and Stollin G (1973) Tem- perature-jump studies of Desulfovibrio vulgaris flavodoxin: Kinetics of FMN binding and reduction of semiquinone by methyl viologen. Biochem Biophys Res Commun 60: 649- 655

Duce RA and Tindale NW (1991) Atmospheric transport of iron and its depletion in the ocean. Limnol Oceanogr 36: 1715-1726

Dugdale RC and Wilkerson FP (1990) Iron addition experiments in the Antarctic: A reanalysis. Global Biogeochem Cycles 4:13-19

Entsch B and Smillie RM (1972) Oxidation-reduction properties of phytoflavin, a flavoprotein from blue-green algae. Arch Biochem Biophys 151:378-386

Entsch B, Sim RG and Hatcher BG (1983) Indications from photosynthetic components that iron is a limiting nutrient in primary producers on coral reefs. Mar Biol 73:17-30

Falkowski PG, Greene RM and Geider ILl (1992) Physiolog- ical limitations on phytoplankton productivity in the ocean. Oceanography 5:84-91

Fasham MJR, Dicklow HW and McKelvie SM (1990) A

nitrogen-based model of plankton dynamics in the oceanic mixed layer. J Mar Res 48:591-639

Fillat MF, Borrias WE and Weisbeek PJ (1991) Isolation and overexpression in Escherichia coli of the flavodoxin gene from Anabaena PCC 7119. Biochem J 280:187-191

Finden DAS, Tipping E, Jaworski GHM and Reynolds CS (1984) Light-induced reduction of natural iron(III) oxide and its relevance to phytoplankton. Nature 309:783-784

Fitzgerald MP, Husain A and Rogers Id (1978) A constitutive flavodoxin from a eukaryotic alga. Biochem Biophys Res Commun 81:630-635

Fox JL (1976) Evolution of flavodoxins. In: Singer TP (ed) Flavins and Flavoproteins, pp 439-446. Elsevier Science Publishers, Amsterdam

Fox JL, Smith SS and Brown JR (1972) Amino acid sequences of Clostridium pasteurianurn flavodoxin. Z Natur- forsch 27b: 1096-1100

Frost BW (1991) The role of grazing in nutrient rich areas of the open sea. Limnol Oceanogr 36:1616-1630

Fuhrman JA, Sleeter TD, Carlson CA and Proctor LM (1989) Dominance of bacterial biomass in the Sargasso Sea and its ecological implications. Mar Ecol Prog Ser 57: 207-217

Galliard T and Chan H (1980) Lipoxygenases. In: Stumpf P (ed) The Biochemistry of Plants, Vol 4, pp 131-161. Academic Press, New York

Geider ILI (1993) Quantitative phytoplankton ecophysiology: Implications for primary production and phytoplankton growth. ICES Mar Sci Symp 197:52-62

Geider ILl, La Roche J, Greene, RM and Olaizola M (1993) Response of the photosynthetic apparatus of Phaeodac- tylum tricornutum (Bacillariophyceae) to nitrate, phosphate or iron starvation. J Phycol 29:755-766

Glover HE (1977) Effects of iron deficiency on lsochrysis galbana (Chrysophyceae) and Phaeodactylum tricornutum (Bacillariophyceae). J Phycol 13:208-212

Glover HE (1978) Iron in coastal waters: seasonal variation and its apparent correlation with a dinoflagellate bloom. Limnol Oceanogr 23:534-537

Goldman JC (1980) Physiological processes, nutrient availability and the concept of relative growth rate in marine phytoplankton ecology. Brookhaven Symp Biol 31: 179- 194

Goldman JC (1984) Oceanic nutrient cycle. In: Fasham MJ (ed) Flows of Energy and Materials in Marine Ecosystems: Theory and Practice, pp 137-170. Plenum Press, New York

Gordon RM, Martin JH and Knauer GA (1982) Iron in northeast Pacific waters. Nature 299:611-612

Goyne ER and Carpenter EJ (1974) Production of iron binding compounds by marine microorganisms. Limnol Oceanogr 19:841-842

Graneli E, Perrson H and Edler L (1986) Connection between trace elements, chelators and red tide blooms in the Laholm Bay, SE Kattegat - an experimental study. Mar Environ Res 18:61-78

Greene RM, Geider RJ and Falkowski PG (1991) Effect of iron-limitation on photosynthesis in a marine diatom. Limnol Oceanogr 36:1772-1782

Greene RM, Geider ILl, Kolber Z and Falkowski PG (1992) Iron-induced changes in light harvesting and photochemi-

297

cal energy conversion processes in eukaryotic algae. Plant Physiol 100:565-575

Greene RM, Kolber ZS, Swift DG, Tindale NW and Falkow- ski PG (1994) Physiological limitation of phytoplankton photosynthesis in the eastern Equatorial Pacific determined from natural variablity in the quantum yield of fluorescence. Limnol Oceanogr (in press)

Guikema JA (1987) Differential effects of iron and chlorophyll in the control of membrane synthesis in Anacysfis nidulans. In: Biggens J (ed) Progress in Photosynthesis Research, Vol IV, pp 785-788. Martinus Nijhoff Publishers, Dordrecht, The Netherlands

Guikema JA and Sherman LA (1983) Organization and function of chlorophyll in membranes of cyanobacteria during iron starvation. Plant Physiol 73:250-256

Harrison GI and Morel FMM (1986) Response of the marine diatom Thalassiosira weissflogii to iron stress. Limnol Oceanogr 31:989-997

Hart TJ (1934) On the phytoplankton of the southwest Atlantic and Bellingshausen Sea, 1929-31. Discovery Rep 8:1-268

Hatchikian EC, Le Gall J, Brusch M and Dubourdieu M (1972) Regulation of the reduction of sulfite and thiosul- fate by ferredoxin, flavodoxin and cytochrome cc' 3 in extracts of the sulfate reducer Desulfovibrio gigas. Biochim Biophys Acta 258:701-708

Haygood MG, Holt PD and Butler A (1993) Aerobactin production by a planktonic marine Vibrio sp. Limnol Oceanogr 38:1091-1097

Healey FP (1978) Physiological indicators of nutrient deficiency in algae. Mitt Int Ver Theor Angew Limnol 21: 34-41

Hecky RE and Kilham P (1988) Nutrient limitation of phytoplankton in freshwater and marine environments: A review of recent evidence on the effects of enrichment. Limnol Oceanogr 33:796-822

Helbling EW, Villafane V and Holm-Hansen O (1991) Effect of iron on productivity and size distribution of Antarctic phytoplankton. Limnol Oceanogr 36:1879-1885

Helms LR and Swenson RP (1991) Cloning and characterization of the flavodoxin gene from Desulfovibrio desulfuricans. Biochim Biophys Acta 1089:417-419

Helms LR, Krey GD and Swenson RP (1990) Identification, sequence determination, and expression of the flavodoxin gene from Desulfovibrio salexigens. Biochem Biophys Res Commun 168:809-817

Hobson LA, Menzel DW and Barber RT (1973) Primary productivity and the sizes of pools of organic carbon in the mixed layer of the ocean. Mar Biol 19:298-306

Hudson 1LIM and Morel FMM (1990) Iron transport in marine phytoplankton: Kinetics of cellular and medium coordination reactions. Limnol Oceanogr 35:1002-1020

Hudson RJM and Morel FMM (1993) Trace metal transport by marine microorganisms: Implications of metal coordination kinetics. Deep-Sea Res 40:129-150

Hutchins DA, Rueter JG and Fish W (1991) Siderophore production and nitrogen fixation are mutually exclusive strategies in Anabaena 7120. Limnol Oceanogr 36:1-12

Ingle RM and Martin DF (1971) Prediction of the Florida red tide by means of the iron index. Environ Lett 1:69-74

298

Jones GJ and Morel FMM (1988) Plasmalemma redox activity in the diatom Thalassiosira. Plant Physiol 87: 143- 147

Jouanneau Y, Richaud P and Grabau C (1990) The nucleotide sequence of a fiavodoxin-like gene which precedes two ferredoxin genes in Rhodobacter capsulatus. Nucleic Acids Res 18:5284-5284

Keir RS (1991) Ironing out greenhouse effects. Nature 349: 198

Kim YS and Martin DF (1974) Interrelationship of Peace River parameters as a basis of the iron index: Predictive guide to the Florida red tide. Water Res 8:607-16

Knight E Jr and Hardy RWF (1966) Isolation and characteristics of flavodoxin from nitrogen-fixing clostridium Clostridium pasteurianum. J Biol Chem 241:2751-2756

Knight E Jr and Hardy RWF (1967) Flavodoxin: Chemical and biological properties. J Biol Chem 242:1370-1374

Kolber Z, Wyman KD and Falkowski PG (1990) Natural variability in photosynthetic energy conversion efficiency: A field study in the Gulf of Maine. Limnol Oceanogr 35: 72-79

Kreutzer R, Dayananda S and Klingmuller W (1991) Co- transcription of the electron transport protein genes nifJ and nifF in Enterobacter agglomerans 333. J Bacteriol 173: 3252-3256

Krey GD, Vanin EF and Swenson RP (1988) Cloning, nucleotide sequence, and expression of the flavodoxin gene from Desulfovibrio vulgaris (Hildenbourough). J Biol Chem 263:15436-15443

Landing WM and Bruland KW (1987) The contrasting biogeochemistry of iron and manganese in the Pacific Ocean. Geochim Cosmochim Acta 51:29-43

Lane LC and Skopp RN (1986) A simple method for purifying phytoferritin. J Plant Nut 9:661-669

La Roche J, Geider RJ and Falkowski PG (1993a) Molecular biology in studies of oceanic primary production. ICES Mar Sci Syrup 197:42-51

La Roche J, Geider RJ, Graziano LM, Murray H and Lewis K (1993b) Induction of specific proteins in eukaryotic algae grown under iron, phosphorus and nitrogen deficient conditions. J Phycol 29:767-777

Laudenbach DE, Reith ME and Straus NA (1988) Isolation, sequence analysis, and transcriptional studies of the fiavodoxin gene from Anacystis nidulans R2. J Bacteriol 170: 258-265

Laulhdre JP, Laboure A-M, and Briat J-F (1989) Mechanism of the transition from plant ferritin to phytosiderin. J Biol Chem 264:3629-3635

Laws EA, DiTullio GR, Betzer PB, Karl DM and Carder KL (1989) Autotrophic production and elemental fluxes at 26 ° N, 155 ° W in the North Pacific central gyre. Deep-Sea Res 36:103-120

Lehman JT (1991) Interacting growth and loss rates: The balance of top-down and bottom-up controls in plankton communities. Limnol Oceanogr 36 1546-1554

Leidi EO, G6mez M and de la Guardia MD (1986) Evalua- tion of catalase and peroxidase as indicators of Fe and Mn nutrition for soybean. J Plant Nutr 9:1239-1249

Leonhardt KG and Straus NA (1989) Sequence of the fiavodoxin gene from Anabaena variabilis 7120. Nucleic Acids Res 17:4384-4384

Leonhardt K and Straus NA (1992) An iron stress operon involved in photosynthetic electron transport in the marine cyanobacterium Synechococcus sp. PCC 7002. J Gen Microbiol 138:1613-1621

MacKnight ML, Gray WR and Tollin G (1974) N-terminal sequences of Azotobacter vinelandii and Rhodospirillum rubrum flavodoxins. Biochem Biophys Res Commun 59: 630-637

Malone TC, Pike SE and Conley DJ (1993) Transient variations in phytoplankton productivity at the JGOFS Bermuda time series station. Deep-Sea Res 40:903-924

Marschner H (1986) Mineral Nutrition of Higher Plants. Academic Press, Orlando, FL

Martin JH (1992) Iron as a limiting factor in oceanic productivity. In: Falkowski PG and Woodhead AD (eds) Primary Productivity and Biogeochemical Cycles in the Sea, pp 123-137. Plenum Press, New York

Martin JH and Fitzwater SE (1988) Iron deficiency limits phytoplankton growth in the north-east Pacific subarctic. Nature 331:341-343

Martin JH and Gordon RM (1988) Northeast Pacific iron distribution in relation to phytoplankton productivity. Deep-Sea Res 35:177-196

Martin JH, Gordon RM, Fitzwater SE and Broenkow WW (1989) VERTEX: phytoplankton/iron studies in the Gulf of Alaska. Deep-Sea Res 36:649-680

Martin JH, Fitzwater SE and Gordon RM (1990) Iron deficiency limits phytoplankton growth in Antarctic waters. Global Biogeochem Cycles 4 : 5 - 1 2

Martin JH, Gordon RM and Fitzwater SE (1991) The case for iron. Limnol Oceanogr 36:1793-1802

Martin JH, Fitzwater SE, Gordon RM, Hunter CN and Tanner SJ (1993) Iron, primary productivity and carbon- nitrogen flux studies during the JGOFS North Atlantic Bloom Experiment. Deep-Sea Res 40:115-134

Mayhew SG (1971) Flavin binding in fiavodoxins. Biochim Biophys Acta 235:289-302

Mayhew SG and Ludwig ML (1975) Flavodoxins and electron-transferring flavoproteins. In: Boyer PD (ed) The Enzymes, Vol 12B, pp 57-118. Academic Press Inc, New York

Menzel DW and Ryther JH (1961) Nutrients limiting the production of phytoplankton in the Sargasso Sea, with special reference to iron. Deep-Sea Res 7:276-281

Menzel DW, Hulburt EM and Ryther JH (1963) The effects of enriching Sargasso Sea water on the production and species composition of the phytoplankton. Deep-Sea Res 10:209-219

Miller GW, Denney A, Pushnik J and Yu M-H (1982) The formation of delta-aminolevulinate, a precursor of chlorophyll, in barley and the role of iron. J Plant Nutr 5: 289-300

Miller GW, Pushnik JC and Welkie GW (1984) Iron chlorosis, a world wide problem, the relation of chlorophyll biosynthesis to iron. J Plant Nutr 7 : 1 - 2 2

Miller CB, Frost BW, Wheeler PA, Landry MR, Welschmeyer N and Powell TM (1991) Ecological dynamics in the subarctic Pacific, a possibly iron-limited ecosystem. Limnol Oceanogr 36:1600-1615

Mitchell BG, Brody EA, Holm-Hansen O, McClain C and Bishop J (1991) Light limitation of phytoplankton biomass

and macronutrient utilization in the Southern Ocean. Limnol Oceanogr 36:1662-1677

Moore RM, Milley JE and Chatt A (1984) The potential for biological mobilization of trace elements from aeolian dust in the ocean and its importance in the case of iron. Oceanolog Acta 7:221-228

Morales F, Abadia A and Abadia J (1990) Characterization of the xanthophyll cycle and other photosynthetic pigment changes induced by iron deficiency in sugar beet (Beta vulgaris, L.). Plant Physiol 94:607-613

Morales F, Abadfa A and Abadfa J (1991) Chlorophyll fluorescence and photon yield of oxygen evolution in iron- deficient sugar beet (Beta vulgaris, L.) leaves. Plant Physiol 97:866-893

Morel FMM, Rueter JG and Price NM (1991) Iron nutrition of phytoplankton and its possible importance in the ecology of ocean regions with high nutrient and low biomass. Oceanography 4:56-61

Myers J (1980) On the algae: thoughts about physiology and measurements of efficiency. In: Falkowski PG (ed) Pri- mary Productivity in the Sea, pp 1-15. Plenum Press, New York

Nelson DM and Smith WO Jr (1991) Sverdrup revisited: Critical depths, maximum chlorophyll levels, and the control of Southern Ocean productivity by the irradiance- mixing regime. Limnol Oceanogr 36:1650-1661

Nishio JN and Terry N (1983) Iron nutrition-mediated chloroplast development. Plant Physiol 77:105-111

Nishio JN, Tayler SE and Terry N (1985a) Changes in thylakoid galactolipids and proteins during nutrition-mediated chloroplast development. Plant Physiol 77:705-711

Nishio JN, Abadfa J and Terry N (1985b) Chlorophyll- proteins and electron transport during iron-mediated chloroplast development. Plant Physiol 78:296-299

Oquist G (1974) Iron deficiency in the blue-green alga Anacystis nidulans: Changes in pigmentation and photosynthesis. Physiol Plant 30:30-37

Osborne C, Chen L-M and Matthews RG (1991) Isolation, cloning, mapping, and nucleotide sequencing of the gene encoding flavodoxin in Escherichia coli. J Bacteriol 173: 1729-1737

Pakrasi HB, Goidenberg A and Sherman LA (1985) Mem- brane development in the cyanobacterium Anacystis nidulans during recovery from iron starvation. Plant Phy- siol 79:290-295

Platt T, Sathyendranath S, Ulloa O, Harrison WG, Hoepffner N and Goes J (1992) Nutrient control of phytoplankton photosynthesis in the Western North Atlan- tic. Nature 356:229-231

Price NM, Andersen LF and Morel FMM (1991) Iron and nitrogen nutrition of equatorial Pacific phytoplankton. Deep-Sea Res 37:295-315

Price NM, Ahner, BA and Morel FMM (1994) The equatorial Pacific: Grazer controlled phytoplankton populations in an iron-limited ecosystem. Limnol Oceanogr (in press)

Pushnik JC and Miller GW (1982) The effects of iron and light treatments on chloroplast composition and ultrastruc- ture in iron deficient barley leaves. J Plant Nutr 4:311-321

Pushnik JC and Miller GW (1989) Iron regulation of chloroplast photosynthetic function: Mediation of PS I development. J Plant Nutr 12:407-412

299

Raven JA (1984) A cost-benefit analysis of photon absorption by photosynthetic unicells. New Phytol 98:593-625

Raven JA (1986) Physiological consequences of extremely small size for autotrophic organisms in the sea. Can Bull Fish Aquat Sci 214:1-70

Raven JA (1988) The iron and molybdenum use efficiencies of plant growth with different energy, carbon and nitrogen sources. New Phytol 109:279-287

Raven JA (1990) Predictions of Mn and Fe use efficiencies of phototrophic growth as a function of light availability for growth and of C assimilation pathway. New Phytol 116: 1-18

Reid RT and Butler A (1991) Investigation of the mechanism of iron acquisition by the marine bacterium Alteromonas luteoviolaceus: Characterization of siderophore production. Limnol Oceanogr 36:1783-1792

Rich HW and Morel FMM (1990) Availability of well defined iron colloids to the marine diatom Thalassiosira weissflogii. Limnol Oceanogr 35:652-662

Riethman HC and Sherman LA (1988) Purification and characterization of an iron stress-induced Chl-protein from the cyanobacterium Anacystis nidulans R2. Biochim Bio- phys Acta 935:141-151

Riley, GA (1947) Factors controlling phytoplankton populations on Georges Bank. J Mar Res 6:54-73

Rouault TA, Stout CD, Kaptain S, Harford JB and Klauzner RD (1991) Structural relationship between an iron-regulated RNA-binding protein (IRE-BP) and aconitase: functional implications. Cell 64:881-883

Rueter JG (1988) Iron stimulation of photosynthesis and nitrogen fixation in Anabaena 7120 and Trichodesmium (cyanophyceae). J Phycol 24:249-254

Rueter JG and Ades DR (1987) The role of iron nutrition in photosynthesis and nitrogen assimilation in Scenedesmus quadricauda (Chlorophyceae). J Phycol 23:452-457

Rueter JG, Ohki K and Fujita Y (1990) The effect of iron nutrition on photosynthesis and nitrogen fixation in cultures of Trichodesmiurn (Cyanophyceae). J Phycol 26: 30-5

Rueter JG, Hutchins DA, Smith RW and Unsworth NL (1992) Iron nutrition of Trichodesmium. In: Carpenter EJ (ed) Marine Pelagic Cyanobacteria: Trichodesmium and Other Diazotrophs, pp 289-306. Kluwer Academic Pub- lishers, Dordrecht, The Netherlands

Rueter JG and Unsworth NL (1991) Response of marine Synechococcus (Cyanophyceae) cultures to iron nutrition. J Phycol 27:173-178

Sakshaug E and Holm-Hansen O (1977) Chemical composition of Skeletonema costaturn (Grev,) Cleve and Pavlova (Monochrysis) lutheri (Droop) Green as a function of nitrate-, phosphate-, and iron-limited growth. J Exp Mar Biol Ecol 29:1-34

Sakshaug E and Holm-Hansen O (1986) Photoadaptation in Antarctic phytoplankton: variations in growth rate, chemical composition and P versus I curves. J Plankton Res 8: 459-473

Salin ML (1987) Toxic oxygen species and protective systems of the chloroplast. Physiol Plant 72:681-689

Sandmann G and BSger P (1983) The enzymological function of heavy metals and their role in electron transfer processes in plants. In: L~iuchli A and Bieleski RL (eds)

300

Inorganic Plant Nutrition, pp 563-596. Springer-Verlag, Berlin

Sandmann G and Malkin R (1983) Iron-sulfur centers and activities of the photosynthetic electron transport chain in iron-deficient cultures of the blue-green alga Aphanocapsa. Plant Physiol 73:724-728

Sandmann G, Peleato ML, Fillat MF, Lfizaro C and G6mez- Moreno C (1990) Consequences of the iron-dependent formation of ferredoxin and flavodoxin on photosynthesis and nitrogen fixation on Anabaena strains. Photosynth Res 26:119-125

Santangelo JD, Jones DT and Woods DR (1991) Met- ronidazole activation and isolation of Clostridium acetobutylicum electron transport genes. J Bacteriol 173: 1088-1095

Scheibe R (1991) Redox-modulation of chloroplast enzymes. Plant Physiol 96:1-3

Smillie RM (1965) Isolation of phytoflavin, a flavoprotein with chloroplast ferredoxin activity. Plant Physiol 40: 1124-1128

Spiller SC and Terry N (1980) Limiting factors in photosynthesis. II. Iron stress diminishes photochemical activity by reducing the number of photosynthetic units. Plant Physiol 65:211-215

Spiller SC, Castelfranco AM and Castelfranco PA (1982) Effects of iron and oxygen on chlorophyll biosynthesis I. In vivo observations on iron and oxygen-deficient plants. Plant Physiol 69:107-111

Stocking RC (1975) Iron deficiency and the structure and physiology of maize chloroplasts. Plant Physiol 55:626-631

Stumm W and Morgan JJ (1981) Aquatic Chemistry, 2nd edn. John Wiley and Sons, New York

Suizberger B (1990) Photoredox processes at hydrous metal oxide surfaces: A surface coordination chemistry approach. In: Stumm W (ed) Aquatic Chemical Kinetics, pp 401-429. John Wiley and Sons, New York

Sunda WG, Swift DG and Huntsman SA (1991) Low iron requirement for growth in oceanic phytoplankton. Nature 351:55-57

Tanaka M, Haniu M, Yasunobu KT, Mayhew S and Massey V (1971) Amino acid sequence of the Peptostreptococcus elsdenii flavodoxin. Biochem Biophys Res Commun 44: 886-892

Tanaka M, Haniu M, Yasunobu KT, Mayhew S and Massey V (1973) The primary structure of Peptostreptococcus elsdenii flavodoxin. J Biol Chem 248:4354-4366

Tanaka M, Haniu M, Yasunobu KT and Mayhew S (1974) The amino acid sequence of the Clostridium MP flavodoxin. J Biol Chem 249:4393-4396

Tanaka M, Haniu M, Yasunobu KT and Yoch DC (1977) Complete amino acid sequence of azotoflavin, a flavodoxin from Azotobacter vinelandii. Biochem 16:3525-3537

Terry N (1980) Limiting factors in photosynthesis I. Use of iron stress to control the photosynthetic capacity in vivo. Plant Physiol 65:114-120

Terry N (1983) Limiting factors in photosynthesis IV. Iron stress-mediated changes in light-harvesting and electron transport capacity and its effects on photosynthesis in vivo. Plant Physiol 71:855-860

Terry N and Abadia J (1986) Function of iron in chloroplasts. J Plant Nutri 9:609-46

Theil EC (1990) Regulation of ferritin and transferrin re- ceptor mRNAs. J Biol Chem 25:4771-4774

Tranter DJ and Newell BS (1963) Enrichment experiments in the Indian Ocean. Deep-Sea Res 10:1-9

Trick CG, Andersen ILl, Price NM, Gillam A and Harrison PJ (1983) Examination of hydroxymate-siderophore production by neritic eukaryotic marine phytoplankton. Mar Biol 75:9-17

Val J, Monge E, Heras L and Abadfa J (1987) Changes in photosynthetic pigment composition in higher plants as affected by iron nutrition status. J Plant Nutr 10:995-1001

Van den Berg CMG, Nimmo N, Abollino O and Mentusti E (1991) The determination of trace levels of iron in seawater using adsorptive cathodic stripping voltammetry. Electronanalysis 3:477-84

Vesk M, Possingham JV and Mercer FV (1966) The effect of mineral nutrient deficiencies on the leaf cells of tomato, spinach and maize. Aust J Bot 14:1-18

Vincent WF, Wurtsbaugh W, Vincent CL and Richerson PJ (1984) Seasonal dynamics of nutrient limitation in a tropical high-altitude lake (Lake Titicaca, Peru-Bolivia): Application of physiological bioassays. Limnol Oceanogr 29:540-552

Wakabayashi S, Kimura T, Fukuyama K, Matsubara H and Rogers LJ (1989) The amino acid sequence of a flavodoxin from the eukaryotic red alga Chondrus crispus. Biochem J 263:981-984

Walsh JJ (1976) Herbivory as a factor in patterns of nutrient utilization in the sea. Limnol Oceanogr 21:1-13

Wells ML (1988/89) The availability of iron in seawater: A perspective. Biol Oceanogr 6:463-476

Wells ML and Goldberg ED (1991) Occurrence of small colloids in sea water. Nature 353:342-344

Wells ML and Mayer LM (1991) The photoconversion of colloidal iron oxyhydroxides in seawater. Deep-Sea Res 38:1379-95

Wells ML, Mayer LM and Guillard RRL (1991) Evaluation of iron as a triggering factor for red tide blooms. Mar Ecol Prog Ser 69:93-101

Welschmeyer N, Goericke R, Strom S and Peterson W (1991) Phytoplankton growth and herbivory in the subarctic Pacific: A chemotaxonomic analysis. Limnol Oceanogr 36: 1631-1649

Wroblewski JS (1977) A model of phytoplankton plume formation during variable Oregon upwelling. J Mar Res 35:357-394

Wu J and Luther GW III (1994) Dissolved, colloidal and particulate iron in the water column of the northwest Atlantic Ocean. Limnol Oceanogr (in press)

Yentsch CS (1980) Phytoplankton growth in the sea: A coalescence of disciplines. In: Falkowski PG (ed) Primary Productivity in the Sea, pp 17-32. Plenum Press, New York

Young RW, Carder KL, Betzer PR, Costello DK, Duce RA, Ditullio GR, Tindale NW, Laws EA, Uematsu M, Merrill JT and Feely RA (1991) Atmospheric iron inputs and primary productivity: Phytoplankton responses in the North Pacific. Global Biogeochem Cycles 5:119-134

Yu MH and Miller GW (1982) Formation of 8-aminolevulinic acid in etiolated and iron-stressed barley. J Plant Nutr 5: 1259-1271

Zevenboom W, Bij de Vaate A and Mur L (1982) Assessment of factors limiting growth of Oscillatoria agardhii in hy- pertrophic Lake Wolderwijd, 1978, by use of physiological indicators. Limnol Oceanogr 27:39-52

Zhuang G, Duce RA and Kester DR (1990) The dissolution of atmospheric iron in surface seawater of the open ocean. J Geophys Res 95:16.207-16.216

301

Zhuang G, Yi Z and Duce RA (1992) Chemistry of iron in marine aerosols. Global Biogeochemical Cycles 6:161-173

Zumft WG and Spiller H (1972) Characterization of a flavodoxin from the green alga Chlorella. Biochem Biophys Res Commun 45:112-118

The role of iron in phytoplankton photosynthesis, and the potential for iron-limitation of primary productivity in the sea

Documents