Introduction

Selenium (Se) is an essential element that is required for the biosynthesis of selenocysteine (Sec), the 21st naturally occurring amino acid in the genetic code (Stadtman, 1996). As a trace element, Se is the 66th most abundant element in the earth’s crust, but can be anthropogenically enriched in aquatic environments by numerous agricultural, industrial and urban activities including fossil fuel (coal) combustion, refinery activity, irrigation of Se-rich soils, fertilizer use and sewage discharge (Bodek et al., 1988; Cappon, 1991; Cutter and Cutter, 2004; US EPA, 2009). The essential nature of Se is related to its occurrence in proteins in the form of Sec residues. Sec is inserted co-translationally in response to the UGA codon, which is recoded from its normal stop codon function if a control element (called SECIS element) is encountered in the 3′-UTR of eukaryotic and archaeal mRNAs, or immediately downstream of UGA in the coding regions of bacterial selenoprotein genes (Berry et al., 1991; Hatfield and Gladyshev, 2002). Selenoproteins can be a thousand times more effective as catalysts than their sulfur-based cysteine homologs (Kim et al., 2006), a fact that likely accounts for the cellular investment in Se-dependent pathways and the specialized machinery used for Se insertion into proteins (Lobanov et al., 2009). Eukaryotic selenoproteomes vary significantly in size but are generally larger in aquatic organisms compared with terrestrial organisms and are particularly large in microalgae, which may have a more consistent supply of bioavailable Se in seawater compared with terrestrial organisms (Lobanov et al., 2009).

Se has been shown to be required by many phytoplankton with examples of absolute Se requirements reported within Bacillariophyceae, Chlorophyceae, Phaeophyceae, Raphidophyceae, Rhodophyceae, Prymnesiophyeae and Prasinophyceae (Fries, 1982; Wheeler et al., 1982; Keller et al., 1984; Harrison et al., 1988; Imai et al., 1996). Se is also required by many dinoflagellates (Dinophyceae), particularly those known to form harmful algal blooms (HABs) including Scrippsiella trochoidea (Harrison et al., 1988), Katodinium rotundatum (Harrison et al., 1988), Gymnodinium catenatum (Doblin et al.,1999, 2000; Band-Schmidt et al., 2004), Alexandrium minutum (Doblin et al.,1999), Protoceratium reticulatum (Mitrovic et al., 2004), Karenia selliformis (Mountfort et al., 2006), Cochlodinium polykrikoides (Lee, 2008), Prorocentrum minimumBaines and Fisher, 2001), Chattonella verruculosa (Imai et al., 1996), and Peridinium gatunense (Lindstrom, 1991). The pelagophyte Aureococcus anophagefferens forms HABs (referred to as ‘brown tides’) that have destroyed shellfisheries and coastal habitats in the United States, South Africa and China (Gobler and Sunda, 2012) and recent sequencing and annotation of its genome (Gobler et al., 2011) revealed an unusually large number of predicted Sec-containing proteins, even when compared with large selenoproteomes in other protistan taxa (Lobanov et al., 2007; Palenik et al., 2007). Despite this recent finding, little is known regarding the organization, composition and properties specific to the selenoproteome as well as the importance of Se in the ecology of A. anophagefferens. Moreover, although the role of macro-nutrients (nitrogen, phosphorus) in the occurrence of HABs has been well-studied (Anderson et al., 2008; Heisler et al., 2008), the importance of trace elements, including Se, for these events has rarely been considered, despite the widespread requirement of Se among harmful algae (Harrison et al., 1988; Lindstrom, 1991; Doblin et al.,1999, 2000; Baines and Fisher, 2001; Band-Schmidt et al., 2004; Mitrovic et al., 2004; Mountfort et al., 2006; Lee, 2008; Gobler et al., 2011).

Here, we examined the role of Se and selenoproteins in the biology and ecology of the harmful alga, A. anophagefferens. The Se requirements of A. anophagefferens were established in culture. The selenoproteome, its components and SECIS (Sec insertion sequence) elements of this species were characterized via informatics and metabolic labeling. The dynamics of naturally occurring blooms of this species and dissolved Se concentrations were quantified and, finally, Se-enrichment experiments were performed with field populations.

Materials and methods

Culture experiments

Se concentrations required to achieve maximal growth rates in axenic A. anophagefferens (clone CCMP1984) were determined using G-medium made from artificial seawater (Doblin et al.,1999) supplemented with differing concentrations of Se added as sodium selenite. Cultures were grown at 21 °C in an incubator with a 14:10 h light:dark cycle, illuminated by a bank of fluorescent lights that provided a light intensity of ∼100 μmol photons per m2 per sec to cultures. These conditions approximated temperature and light exposures found in Long Island estuaries during summer months when A. anophagefferens blooms (Gobler and Sunda, 2012). Cultures were maintained for a minimum of four transfers at each concentration before the collection of final growth rate data to ensure that cells were fully acclimated to treatment conditions and that the carryover of Se from the initial, full strength media was eliminated. Cellular growth rates were calculated for cultures based on cell densities (determined microscopically) in exponential growth phases, using the formula μ=[ln (Bt/B0)]/t, where B0 and Bt are the initial and final abundances, and t is the incubation duration in days. Growth rates were averaged over the entire exponential phase, which typically persisted for 3–6 days, depending on the concentration of Se in the media. Se-limitation of cultures was confirmed by the stimulation of growth of cultures in stationary phase at concentrations below 5 nM following the addition of 10 nM Se. The Monod kinetic terms μmax (maximum growth rate) and KS (half-saturation constant) were derived using Michaelis–Menten curve fitting functions within Kaleidagraph version 4.1.1 (Synergy Software, Reading, PA, USA).

Identification of genes coding for selenoproteins in A. anophagefferens

Identification of selenoprotein genes in the Aureococcus genome was performed using SECISearch (Novoselov et al., 2002; Fomenko et al., 2007), which analyzes primary sequences and secondary structures and then calculates the free energy for various parts of potential SECIS elements. The search was carried out using both default and loose patterns of SECISearch to accommodate identification of unusual SECIS structures. The searches also were then extended to identify organism-specific structures by a modified SECISearch (Novoselov et al., 2002; Fomenko et al., 2007). For identified candidate SECIS elements, open reading frames (ORFs) were predicted in the regions upstream of the SECIS elements. The presence of at least one homologous protein in the NCBI non-redundant database was used as an additional requirement. At the final step, a manual sequence and homology analysis of predicted selenoprotein ORFs located upstream of candidate SECIS elements was carried out.

In addition to the SECIS-based search procedure, the genome was analyzed with TBLAST against all known selenoprotein sequences to identify homologs of previously described selenoprotein genes. A third procedure included a search for Sec/Cys pairs in homologous sequences (Novoselov et al., 2002; Fomenko et al., 2007). We extracted all ORFs containing in-frame UGA codons, and homologs containing Cys in place of Sec were found by identifying the Sec/Cys pairs flanked by homologous sequences. For this procedure, we used TBLASTX to examine all potential ORFs with in-frame UGA codons against NCBI non-redundant protein database. All candidate ORFs were then tested with SECISearch for the presence of SECIS elements. Finally, PSI-BLAST was used to identify additional more distant homologs. All data sets obtained with the three independent methods were combined and the proteins were classified as homologs of previously known selenoproteins, novel selenoproteins and candidate selenoprotein genes.

Metabolic labeling of the A. anophagefferens selenoproteome

To experimentally verify the occurrence of selenoproteins, A. anophagefferens cells were subjected to metabolic labeling with 75Se. The green alga, Chlamydomonas reinhardtii, was labeled in parallel and used as a control. The metabolic labeling was performed as described previously (Novoselov et al., 2002). Briefly, 100 ml of each culture was metabolically labeled with 100 μCi of 75Se ([75Se]selenous acid, specific activity 1000 Ci mmol−1, Research Reactor Facility, University of Missouri, Columbia, MO, USA) for 48 h. Cells were collected, resuspended in PBS and sonicated. Approximately 20 and 50 μg of total soluble protein from A. anophagefferens and C. reinhardtii, respectively, were resolved by SDS–PAGE and transferred onto a PVDF membrane (Invitrogen, Carlsbad, CA, USA). Selenoprotein patterns were visualized with a PhosphorImager.

Ecosystem study

During the spring and summer of 1998, seawater samples were collected twice weekly to fortnightly from West Neck Bay, Stony Brook, NY, USA, a small (1 km2), shallow (2–4 m), enclosed embayment on Shelter Island, within the Peconic Estuary of eastern Long Island. The absence of tributaries and point source anthropogenic inputs make this bay a relatively simple system, where chemical inputs can be largely attributed to groundwater seepage or internal recycling (Gobler and Sañudo-Wilhelmy, 2001). All sampling material used in this study was prepared using trace metal clean techniques (Gobler et al., 2002). Water samples were collected with a peristaltic pump connected to acid-washed Teflon tubing on a bamboo pole extended 4 m upwind and lowered to a depth of 1 m. Dissolved seawater samples were obtained by filtration through trace metal clean, polypropylene capsule filters (0.2 μm). All analyses and sample manipulation performed after collection were conducted within vertical flow clean benches supplied with HEPA-filtered air. Triplicate chlorophyll a samples were collected on GF/F glass fiber filters (nominal pore size=0.7 μm) and analyzed by standard fluorometric methods (Parsons et al., 1984). A. anophagefferens densities were determined on field samples fixed with glutaraldehyde (1% final concentration). Samples were enumerated by direct count methods employing a Zeiss model D-7082 epi-fluorescent microscope (Jena, Germany) and an immunofluorescent label as described by Anderson et al., 1989. The amount of chlorophyll a in West Neck Bay attributable to A. anophagefferens was estimated by assuming a constant chl a per cell value (0.035±0.003 pg per cell for nutrient replete cultures (Gobler and Sañudo-Wilhelmy, 2001)) and multiplying this value by A. anophagefferens densities. Such approximations have been used successfully in the past to compare A. anophagefferens biomass with that of the total algal community (Gobler and Sañudo-Wilhelmy, 2001; Caron et al., 2004).

Groundwater entering West Neck Bay through the intertidal zone was sampled from 1 m, Teflon-lined, PVC piezometers with 2.5 cm horizontal screened slits along the lower 25 cm at six locations across the Bay on multiple occasions. High groundwater seepage rates at West Neck Bay allowed piezometers to fill with fresh groundwater (salinity <0.1 ppt) when sampled during low tide. Groundwater was sampled using a peristaltic pump equipped with acid-washed Teflon tubing. To ensure representative groundwater was sampled, piezometers were purged at <100 ml min−1 and samples were not obtained until the conductivity, dissolved oxygen, temperature and pH of the pumped groundwater stabilized (Puls and Paul, 1995). Filtered groundwater or seawater samples were acidified to pH 1.6 with HCl, stored under dim light at room temperature and then analyzed for total dissolved Se as described by Cutter, 1982, 1983. Briefly, total dissolved Se was determined by boiling a 4 M HCl acidified sample for 1 h with the addition of potassium persulfate. Next, within a glass stripping vessel, selenite was quantitatively converted to hydrogen selenide by adding sodium borohydride to samples along with sulfanilamide (to eliminate interference due to nitrite). The evolved hydrogen selenide was stripped from solution using helium and trapped in a borosilicate U-tube packed with silanized glass wool and immersed in liquid nitrogen. After the trap was removed from the LN2, an atomic absorption spectrometer fitted with an open quartz tube furnace burning an air–hydrogen flame was used to detect the hydride. Instrument response (as peak area) was recorded on a chromatographic integrator. To ensure accuracy, all determinations utilized the standard additions method of calibration, and all samples were analyzed in triplicate to quantify precision (found to be 4%). Detection limits for dissolved Se were 0.02 nM.

Field experiments were conducted on 4 June, 22 June, 8 July and 21 July 1998 using water from West Neck Bay. Within 2 h of collection, 50 ml of seawater was transferred to trace metal clean, 60 ml polycarbonate flasks in a HEPA laminar flow hood to prevent trace metal contamination. Triplicate flasks were amended with sodium selenite (10 nM) or were left unamended as a control treatment. The selenite solution was cleaned of trace metals with Chelex-100 ion exchange resin (Bruland, 1980), filter sterilized (0.2 μm) and frozen before use. Amended and unamended flasks were incubated at the same temperature found in West Neck Bay during water collection under 125 μmol photons per m2 per sec of light on a light:dark cycle which mimicked summer conditions in West Neck Bay (14 h: 10 h) and the light found at ∼1.4 m in the bay (Gobler and Sañudo-Wilhelmy, 2001). After 48 h, samples from each flask were preserved to a final concentration of 1% glutaraldehyde for enumeration of A. anophagefferens as described above. Net specific growth rates of A. anophagefferens were determined as described for culture experiments. Differences between control and treatment flasks were resolved via Student’s t-test.

Results and discussion

Requirement of Se for growth of A. anophagefferens

A. anophagefferens (CCMP9184) cultures displayed a sigmoidal increase in growth rate in response to increasing selenite concentrations (Figure 1) and fit the Michaelis–Menten equation (R=0.97). Cultures did not grow below 10 pM selenite, but displayed a gradual increase in growth above 10 pM and near maximal growth rates (∼0.6 per day) at concentrations of selenite at or above 5 nM (Figure 1). A. anophagefferens had a half-saturation constant (Ks) of 0.27±0.08 nM for selenite and a μmax of 0.60±0.03 per day.

Figure 1
figure 1

Growth rates of A. anophagefferens in media with differing concentrations of selenite.

Full size image

The A. anophagefferens selenoproteome

Computational searches for selenoprotein genes using a combined SECISearch, SECISearch adapted for A. anophagefferens, and the Sec/Cys approaches extended the number of detected selenoproteins from the originally predicted 56 selenoprotein genes (26) to a total of 59. This represents the largest currently known selenoproteome of any organism, being two times larger than the one present in Ostreococcus lucimarinus, previously considered the largest eukaryotic selenoproteome (27,28), and four-fold larger than those of diatoms that compete with A. anophagefferens and green algae (Figure 2). The A. anophagefferens selenoproteome is characterized by the occurrence of almost all known eukaryotic selenoproteins, for example, 20 out of 25 human selenoproteins have selenoprotein homologs in A. anophagefferens and 12 out of 12 Chlamydomonas selenoproteins. Three selenoproteins have previously been detected only in bacteria (outside of A. anophagefferens; Table 1) and most selenoproteins had closest homologs from completely different groups of organisms. In addition, several selenoproteins occurred in multiple copies (Table 1, Supplementary Table 1). For example, methionine sulfoxide reductases were represented by four MsrA isozymes and four MsrB isozymes (Table 1). Several novel selenoproteins with unknown functions were also detected (Table 1). Further analysis of the selenoproteome revealed dominance of the thioredoxin fold, present in more than half of detected selenoproteins. In all of these proteins, Sec was located in place of the catalytic Cys found in thioredoxin. This observation and the additional presence of many other oxidoreductases (MsrAs, MsrBs, GILT, TRs and so on) point to the redox function of Sec in selenoproteins and the role of Se in regulating redox homeostasis in A. anophagefferens.

Figure 2
figure 2

Size of selenoproteomes of seven eukaryotic phytoplankton for which complete genomes are available.

Full size image
Table 1 Selenocysteine-containing proteins encoded in the A. anophagefferens genome
Full size table

A. anophagefferens selenoproteins have a Sec residue encoded by UGA codon, which is inserted with the help of unusual SECIS elements (Figure 3, Supplementary Table 2), which are RNA structures present in the 3′-UTRs of selenoprotein genes. One unique feature of A. anophagefferens SECIS elements is their unusually high diversity (Figure 3). Although several SECIS elements found in the A. anophagefferens genome correspond to the canonical SECIS model, a surprisingly large number of SECIS elements did not (Figure 3). A unique feature of A. anophagefferens SECIS elements is its unusually long stem. A typical length of the stem (distance from the SECIS core to the apical loop) is 10–13 nucleotides, while in Aureococcus SECIS elements it can be nearly twice as long (Figure 3). Another characteristic feature is the extremely small, virtually absent apical loop. The size of the apical loop in the canonical model varies from 7–27 nucleotides, while many Aureococcus SECIS elements have apical loops consisting of only 3 nucleotides and often lacking a conserved unpaired AA sequence (Figure 3). It remains to be seen if there are any extra constrains that disrupt nucleotide pairing near the end of the stem to increase the size of the apical loop.

Figure 3
figure 3

(a) A typical eukaryotic SECIS element that is present in the 3′-UTR of DUF1000. Such structure occurred in only some Aureococcus selenoprotein genes. (b, c) Typical Aureococcus SECIS structures, in 3′-UTRs of SelW and PDI, respectively, featuring a long stem and a small apical loop that lacks a conserved AA sequence and also lacks a mini-stem. An SBP2-binding site at the bottom of the main step is shown in bold. (d) A. anophagefferens Sec tRNA. The anticodon is shown in bold.

Full size image

Owing to the highly unique and diverse nature of SECIS elements in A. anophagefferens our informatics search (SECISearch) to identify selenoprotein genes had to be adjusted to recognize the A. anophagefferens consensus structure developed based on the analyses of SECIS elements in known selenoprotein genes, allowing the selenoproteome to be fully characterized. Although the genome of A. anophagefferens encodes the largest selenoproteome, the set of selenoproteins reported here represents a conservative estimate: the true number of selenoproteins is likely larger. The 59 predicted selenoprotein genes did not include less reliable predictions that could not be unambiguously confirmed owing to lack of homologs in other organisms, identifiable SECIS elements in the 3′-UTR (mostly due to insufficient length, or unusual form of the potential SECIS elements) and possibly incorrect gene models. In addition, proteins that were highly similar to each other were not always reported as separate entities, as additional verification was required to avoid over reporting caused by mis-assembly, splicing variants and sequencing errors. We detected components of Sec insertion machinery, including Sec-specific elongation factor EFsec, a SECIS-binding protein and Sec tRNA (Figure 3d). The latter was not detected by tRNAscan using standard parameters, but could be identified in a maximum sensitivity mode of this program. Overall, it is very likely, that the number of selenoproteins encoded in the A. anophagefferens genome exceeds 60.

Metabolic labeling of A. anophagefferens with 75 Se

Metabolic labeling of A. anophagefferens along with another microalga, Chlamydomonas reinhardtii, which has 12 selenoprotein genes (Novoselov et al., 2007; Palenik et al, 2007; Figure 4), revealed both a large number of expressed selenoproteins in A. anophagefferens and a significantly greater abundance of these proteins compared with C. reinhardtii. Abundant selenoprotein bands in the 8–28 kDa region corresponded to the predicted masses of the majority of Aureococcus selenoproteins. These observations provide further evidence of the enhanced use of Sec and selenoproteins by A. anophagefferens compared with other organisms.

Figure 4
figure 4

Metabolic labeling of Aureococcus anophagefferens (A.a.) and Chlamydomonas reinhardtii (C.r.) with 75Se using 20 μg of A.a. and 50 μg of C.r. proteins separated by SDS–PAGE (left panel) and Coomassie blue staining to show protein loading (right panel). Migration of size markers (in kDa) is shown on the left.

Full size image

Promotion of A. anophagefferens brown tides by Se

During a coastal ecosystem study, dissolved Se concentrations in West Neck Bay, averaged 0.71±0.37 nM (Figure 5). A. anophagefferens densities steadily increased from 1.0 × 104 cells per ml in May to >1.0 × 105 cells per ml in June (Figure 5), representing 10–25% of the total algal community. A. anophagefferens reached peak densities in early July, as cell densities grew to 5.6 × 105 cells per ml (Figure 6) representing >95% of total algal biomass. During the bloom peak dissolved Se levels decreased to their lowest levels of the study (0.05 nM; Figure 5). Following the bloom’s peak, dissolved Se concentrations rose to >1 nM and remained elevated as A. anophagefferens densities steadily declined to nearly zero (Figure 5). Groundwater entering West Neck Bay contained, on average, 3.4±0.3 nM dissolved Se. During incubation experiments conducted in June and late July, the addition of selenite did not significantly alter the growth of A. anophagefferens (Figure 7). During the peak of the A. anophagefferens bloom (8 July), however, the addition of selenite significantly increased A. anophagefferens growth rates compared with unamended controls (Figure 7; P<0.001; t-test). Clearly, Se availability was not the primary driver of this A. anophagefferens blooms, but its importance in supporting brown tide proliferation is evident from its drawdown during the bloom and its stimulatory effect under Se-depleted conditions.

Figure 5
figure 5

Temporal dynamics of dissolved Se and A. anophagefferens abundances during a brown tide bloom in West Neck Bay, NY, USA.

Full size image
Figure 6
figure 6

Nucleotide alignment of Aureococcus SECIS elements. Conserved nucleotides are highlighted. The SBP2-binding site is shown in red above the sequences.

Full size image
Figure 7
figure 7

Net growth rates of A. anophagefferens during incubation experiments conducted using water from in West Neck Bay, NY, USA, during a brown tide event. The asterisk denotes that the addition of Se significantly enhanced the net growth rates of A. anophagefferens during the 8 July experiment (P<0.001).

Full size image

A. anophagefferens possess the largest selenoproteome of any living organism, consisting of at least 59 selenoproteins, and this study demonstrates that its ability to form HABs can be dependent on the bioavailability of Se. A. anophagefferens’ SECIS elements are diverse and, in some cases, unique. Organism-specific features of SECIS elements are known for several species. For example, the majority of Ostreococcus SECIS elements possess an extra-long mini-stem (Lobanov et al., 2007). The SECIS elements of evolutionarily unrelated proteins in Dictyostelium discoideum are highly conserved with five nucleotides preceding and two nucleotides following the SECIS core being identical to all identified SECIS elements (Novoselov et al., 2002). This example substantiates convergent evolution in SECIS structure as well as constraints associated with SECIS-binding proteins. In addition, SECIS elements in Neospora and Toxoplasma show a non-canonical GGGA sequence in the SECIS core (Novoselov et al., 2002; Fomenko et al., 2007). Thus, deviations from the eukaryotic SECIS model, while rare, may be expected, and A. anophagefferens joins the group of organisms with lineage-specific SECIS functions. The diversity of SECIS elements found in the A. anophagefferens genome also suggests a higher-than-normal flexibility in the function of Sec insertion machinery. Once additional genomes with similar properties become available, it will be of interest to identify the elements responsible for recognition of such variable SECIS element structures.

The A. anophagefferens selenoproteome contains both the majority of eukaryotic selenoproteins (for example, 80% of human and 100% of Chlamydomonas selenoprotein families) and several selenoproteins that have previously been detected only in bacteria, suggesting they were acquired through lateral gene transfer from bacteria. In addition, the closest homologs of A. anophagefferens selenoproteins are scattered across diverse eukaryotes, again pointing to lateral transfers of some of these genes. Finally, many completely new selenoproteins were discovered, mostly with unknown functions. Location of Sec in these proteins, however, identifies the catalytic Cys residues in homologs of these proteins in other organisms (Fomenko et al., 2007; Novoselov et al., 2007).

Having the largest and most diverse selenoproteome identified to date implies these selenoproteins have an important role in the fitness and functioning of A. anophagefferens. More than half of A. anophagefferens selenoproteins possess a thioredoxin fold, and many additional proteins are homologs of known thiol oxidoreductases. In addition, A. anophagefferens selenoproteins represent nearly all known redox regulatory systems (thioredoxin, glutaredoxin, glutathione peroxidase, peroxiredoxin, methionine sulfoxide reductase and so on) as well as systems involved in disulfide bond formation and isomerization (protein disulfide isomerase, GILT and so on). The functions of several selenoproteins are unknown, but it is clear that at least some (and possibly most) of them are oxidoreductases. These findings indicate that A. anophagefferens uses Se to fine-tune its cellular redox homeostasis (Stadtman, 1996; Kim et al., 2006; Lobanov et al., 2009), an ability that may impart a competitive advantage to this species over its competitors during algal blooms. The shallow nature and rapid vertical mixing times (∼1 h; Milligan and Cosper, 1997) of estuaries where A. anophagefferens resides chronically expose phytoplankton to extremes in light and temperature, which in turn are likely to create elevated levels of free radicals and intracellular oxidation. For example, exposure to excessive light causes algae to produce reactive oxygen species that result in high levels of oxidative stress and damage unless quickly detoxified (Waring et al., 2010). The large number and multiple redundancies of selenoproteins in A. anophagefferens are likely to help protect these cells against such oxidation through the removal of hydroperoxides and the repair of oxidatively damaged proteins (Stadtman, 1996). Hence, the large arsenal of selenoproteins in A. anophagefferens that prevent cellular oxidation likely infers a competitive advantage to this alga, particularly during summer when it forms HABs and when vertical gradients in temperature and irradiance are at their annual extreme (Cosper et al., 1987).

A highly enhanced use of Sec in A. anophagefferens compared with other organisms that co-occur with it should make this organism particularly competitive during HABs when other, non-Se resources such as nitrogen and light can be scarce (Gobler et al., 2011). Sec-containing proteins can be a thousand times more effective as catalysts than their Cys homologs (Kim et al., 2006) and, therefore, selenoproteins may help A. anophagefferens to more efficiently organize and execute its metabolism. Increased reliance on Se to support biosynthesis of proteins also suggests that A. anophagefferens can efficiently sequester Se from its environment, a hypothesis confirmed by direct metabolic labeling of Aureococcus with 75Se as well as by the decline of dissolved Se in estuaries experiencing brown tides.

Although maximal growth rates for A. anophagefferens occurred at Se at or above 5 nM, its half-saturation constant for growth was 0.27 nM. In the environment, dissolved Se concentrations ranged between 0.05 and 1.35 nM, above and below the half-saturation constant, but below the levels that elicited maximal growth of cultures. Importantly, multiple species of Se are present in marine environments (Cutter and Bruland, 1984) and a substantial portion of the total dissolved Se pool may be comprised of forms that are not readily accessible to phytoplankton (Baines and Fisher, 2001; Cutter and Cutter, 2004). As such, the bioavailable pool of Se for A. anophagefferens during the HAB in West Neck Bay was likely a fraction of the total dissolved pool (as low as 5% in some estuaries; Cutter and Bruland, 1984; Cutter and Cutter, 2004). In marine ecosystems, nutrient concentrations below half-saturation constants are often considered limiting to algal growth (Caperon and Meyer, 1972; Fisher et al., 1992). The lowest concentrations of Se (0.05–0.22 nM) observed during the peak of the A. anophagefferens bloom were below its half-saturation constant and levels required for maximal growth, likely due to the large Se demand created by the bloom. These low Se concentrations likely restricted the growth of A. anophagefferens as the addition of 10 nM Se as bioavailable selenite during this period significantly enhanced A. anophagefferens growth rates. Hence, while a lack of Se may have contributed to the demise of the A. anophagefferens bloom in West Neck Bay, these findings further demonstrate that high levels of Se loading can promote the proliferation of HABs caused by this alga.

A. anophagefferens blooms in shallow, enclosed estuaries such as West Neck Bay where Se concentrations averaged ∼1 nM, but never blooms in deeper estuaries or continental shelf regions that are characterized by lower Se concentrations (Cutter and Bruland, 1984; Measures et al., 1984; Cutter and Cutter, 1995, 2001). The formation of blooms exclusively in shallow estuaries ensures that A. anophagefferens has access to a rich supply of Se required to synthesize these ecologically important and catalytically superior enzymes (Stadtman, 1996; Hatfield and Gladyshev, 2002; Kim et al., 2006). As A . anophagefferens relies on selenoproteins for growth and as a scarcity of Se may prohibit bloom formation in off-shore waters, Se is likely to have a key role in shaping the niche space and bloom occurrences of this species. Moreover, as some phytoplankton do not require Se (Harrison et al., 1988), Se availability is likely to shape the succession and composition of phytoplankton communities in general.

Se can be anthropogenically enriched in aquatic environments (Bodek et al., 1988; Cappon, 1991; Cutter and Cutter, 2004; US EPA, 2009). Consistent with this hypothesis, the watershed of West Neck Bay is densely populated with dwellings that are not connected to a sewage treatment plant but rather have cesspools that leach into the groundwater that enters West Neck Bay. Levels of Se measured in groundwater entering the Bay were four-fold greater than concentrations in the bay and other marine surface waters (Cutter and Bruland, 1984; Measures et al., 1984; Cutter and Cutter 1995, 2001, 2004), likely due to wastewater contamination from cesspools and/or fertilizer use on the watershed (Cappon, 1991; McBride and Spiers, 2001; US EPA, 2009). As groundwater is nearly the exclusive source of freshwater to this (Schubert, 1998) and other estuaries which host brown tides (LaRoche et al., 1997), the anthropogenic loading of Se into this and perhaps other coastal ecosystems may lead to the intensification of HABs caused by A. anophagefferens. As Se is required by many other phytoplankton that form HABs, including C. verruculosa (Imai et al., 1996), S. trochoidea (Harrison et al., 1988), K. rotundatum (Harrison et al., 1988), G. catenatum (Doblin et al.,1999, 2000; Band-Schmidt et al., 2004), A. minutum (Doblin et al.,1999), Protoceratium reticulatum (Mitrovic et al., 2004), K. selliformis (Mountfort et al., 2006), C. polykrikoides (Lee 2008), P. minimum (Baines and Fisher, 2001) and P. gatunense (Lindstrom, 1991). Anthropogenic loading of this element may contribute to the risk of these events globally.