The impact of urea on toxic diatoms – Potential effects of fertilizer silo breakdown on a Pseudo-nitzschia bloom

Highlights

• In 2016, 2750 t of fertilizer were spilled into an estuary in Denmark and followed by a toxic bloom of P. seriata.

• Mussel meat from the area contained up to 49 mg domoic acid kg⁻¹, more than double the international regulatory limit.

• The phytoplankton community Jan–Apr 2016 differed markedly to previous years (2013–15).

• Pseudo-nitzschia obtusa is shown here for the first time to produce domoic acid in the absence of copepods.

• Pseudo-nitzschia can grow on available nitrogen sources (i.e. urea and nitrate) with species- and strain-specific variation.

Abstract

In spring 2016, two silos containing liquid nitrogen-containing fertilizer collapsed on a harbor in Fredericia, Denmark. More than 2,750 tons of fertilizer spilled into inner Danish waters. A bloom of Pseudo-nitzschia occurred approximately one month after the incident. The bloom caused a 5-week quarantine of numerous mussel-harvesting areas along the eastern coast of Jutland. The levels of domoic acid measured up to 49 mg kg⁻¹ in mussel meat after the bloom. In the months following the event, the species diversity of phytoplankton was low, while the abundance was high comprising few dominant species including Pseudo-nitzschia. The main part of the liquid nitrogen-containing compound was urea, chemically produced for agricultural use.

To investigate the potential impact of urea on Pseudo-nitzschia, four strains, including one strain of P. delicatissima, two of P. seriata and one of P. obtusa, were exposed each to three concentrations of urea in a batch culture experiment: 10 μM, 20 μM and 100 μM N urea, and for comparison one concentration of nitrate (10 μM). Nitrate, ammonium, and urea were metabolized at different rates. Pseudo-nitzschia obtusa produced domoic acid and grew best at low urea concentrations. Both P. seriata strains had a positive correlation between urea concentration and growth rate, and the highest growth rate in the nitrate treatment. One strain of P. seriata produced domoic acid peaking at low N loads (10 µM N urea and 10 µM N nitrate). In conclusion, the ability to adapt to the available nitrogen source and retain a high growth rate was exceedingly varying and not only species-specific but also strain specific.

Introduction

On 3 February 2016, two silos containing a total of 4750 tons of liquid nitrogen-containing fertilizer collapsed on the harbor of Fredericia, close to Lillebælt, Denmark (Fig. 1). The amount of nitrogen-containing fertilizer reaching the adjacent sea was at least 2750 tons (Schjødt, 2018). The vast majority of the fertilizer consisted of urea (Markager, 2016), but ammonium and nitrate were also a part of the spill. The composition of the fertilizer included both N16 and N32 liquid nitrogen-containing fertilizer. N16, which consists of urea solely, was in the magnitude of 750 tons. 2000 tons of the spill was N32, which is a combination of N sources: 50% of the total N is from urea, 25% from nitrate, 25% from ammonium (Markager, 2016). No inhibitor was added to either of the fertilizer types. Inhibitors are chemical components working against urease and other enzymes to decrease the rate the fertilizer is utilized. The average yearly supply of nitrogen from land in the Lillebælt area is 5517 tons N, the collapse thus constituted an abrupt supply amounting over 50% of the annual terrestrial input in N (Markager, 2016). The timing of the collapse was about two weeks prior to the usual spring bloom in the area (Markager, 2016). In Danish waters, phytoplankton growth is light limited from November to January. The spring bloom usually initiates in early February and utilizes the majority of the available nitrogen and phosphate in February and March. Hereafter growth is limited, mainly due to nutrient depletion, until October (Markager, 2016). The additional supply of N hence most likely favored phytoplankton growth, and notably, a bloom of domoic acid-producing Pseudo-nitzschia (PN) occurred approximately one month after the breakdown of the silos.

In Denmark, all mussel-harvesting areas are closed as soon as domoic acid (DA) or any other phycotoxin is detected. The decision on the management action is based upon the on-going risk assessment carried out by the authorities. The risk assessment takes into account the combined information on algal toxins and the quantitative occurrence of toxic algae. The mussel-harvesting areas are always to be closed if the algal toxin concentrations exceed the international regulatory limit, for DA of 20 mg DA kg⁻¹. Levels of DA up to 49 mg kg⁻¹ in mussel meat were measured (Fig. 2). Detection of DA forced closure of numerous mussel-harvesting areas along the eastern coast of Jutland and resulted in an economical deficit of more than 1 million DKK. The added nitrogen was speculated to increase nitrogen availability during the following months and potentially years and lead to other ecosystem responses such as reduced underwater vegetation as a result of light limitation (e.g. eelgrass) (Markager, 2016). The Lillebælt area is characterized by strong currents and rapid water flow, which therefore potentially distributed the added nitrogen, primarily to the north (Kattegat). The area is characterized by estuarine circulation (Nielsen et al., 2017). This phenomenon occurs when surface and bottom currents flow in opposite directions, meaning that even if the nitrogen load was transported northwards with the surface waters, the uptake and sedimentation in the bottom waters may have retained the nitrogen in the area.

Since the 1970s, the global use of nitrogen-containing fertilizer has increased drastically (Galloway and Cowling, 2002). In 2013, the global use of industrial N was of a magnitude of >170 megatons N yr⁻¹ (Glibert et al., 2014). Since the 1970s, the use of urea has increased more than 100-fold, and at the same time, the composition of the fertilizer has changed, comprising a higher amount of urea. In 2005, urea constituted >50% of the nitrogenous fertilizer used worldwide. As a consequence of this change, urea has become a nitrogen source available to marine and coastal phytoplankton (Glibert et al., 2006).

Apart from organic urea, nitrogen in the ocean mainly occurs as inorganic nitrate and ammonium. The different forms of nitrogen may be utilized with varying efficiency by different phytoplankton species, affecting species competition and thus phytoplankton composition (e.g. Antia et al., 1991). In general, diatoms are considered mainly to utilize nitrate whereas most other phytoplankton is better adapted for using ammonium (Glibert et al. 2016), but the majority of studies regarding nitrogen sources confirm the capability of both freshwater and marine phytoplankton species to utilize urea, either partly or as a sole nitrogen source (e.g. Cira et al., 2016; Cochlan et al., 2008; Donald et al., 2013, Glibert et al., 2014).

The impact of the changing nitrogen composition in agricultural fertilizer along with the increasing use of fertilizer is gaining attention. There is ample evidence for the detrimental impact of the increasing nitrogen usage in aquatic ecosystems (Howarth et al., 2002). The increasing use of urea in fertilizer has been coupled with an increase in documented incidences of Paralytic Shellfish Poisoning (PSP) (Glibert et al., 2006) caused by several toxic dinoflagellate Harmful Algal Bloom (HAB) species. No investigation has yet linked this with Amnesic Shellfish Poisoning (ASP) incidents in the field.

Growth and toxin content of HAB species have also been investigated in the laboratory in relation to nitrogen sources. The dinoflagellate Alexandrium catanella showed e.g. no increase in toxin production related to nitrogen type (Griffin et al., 2019). Contrary the same species (reported as A. tamarense) had a higher toxin content when grown on urea than compared to nitrate, but highest on ammonium (Leong et al., 2004). Another dinoflagellate, Karenia brevis, also showed much higher toxin content when supplied with urea than without urea supplement (Shimizu et al., 1993).

In general, the urea concentration in aquatic ecosystems is lower than the concentration of the inorganic N-sources (nitrate and ammonium), but in a few instances, concentrations up to 25–50 µM N have been reported (Glibert et al., 2005). The fact that each molecule of urea contains two atoms of nitrogen contributes to some confusion, as some refer to units of urea (e.g. 10 µM urea) and some to the nitrogen content in urea (e.g. 20 µM N urea), both examples are equivalent to the same concentration of urea. In the current study, nitrogen content in urea is used, as this unit is dominant in the literature.

The diatom genus PN comprises 52 marine species of which 26 are confirmed toxigenic (Bates et al., 2018). Pseudo-nitzschia is a globally distributed pennate diatom genus. Cells are efficient in utilizing nutrients in upwelling systems as well as during spring and autumn mixing of waters (Lelong et al., 2012), and some species of PN can cause major toxic blooms. The toxin is a simple neurotoxic amino acid called domoic acid (DA). The chemical compounds, metabolic precursors, and the penultimate enzymatic steps for the biosynthesis of DA were recently described (Maeno et al., 2018; Brunson et al., 2018; Harðardóttir et al., 2019) but the final isomerization reaction, sources, and compartmentalization are yet to be elucidated.

Toxic blooms may have severe impacts on marine ecosystems as DA can accumulate in primary consumers, e.g. filtrating mussels, zooplankton, and smaller fish. This can cause intoxication and mortality of secondary consumers like marine mammals, seabirds and humans (Teitelbaum et al., 1990; Lefebvre et al., 2016; Nash et al., 2017). Furthermore, HABs have large economic effects on the mussel industry and fisheries, due to closure of the fishery, lost revenue and discard of already harvested mussels (Fehling et al., 2006; McCabe et al., 2016; Bates et al., 2018).

Domoic acid production is affected by various factors, and it has been suggested that dissimilar species are affected differently (Gai et al., 2018). Nutrient composition and levels have a major influence on DA production. Changes in nitrogen have been found to affect DA production, but notably, nitrogen is a component in DA synthesis, and deficiency may, therefore, limit DA production (Trainer et al., 2012 and references therein). Depletion of silicate and phosphate generally induces high cellular DA (Ryan et al., 2017), and species and strain variation in cellular DA content in response to several environmental changes have been observed for iron concentration, salinity, temperature, pH, light, and presence of bacteria and grazers (Bates et al., 2018).

In addition to the variety of inducing environmental factors, the impact of different growth stages on DA production, is still unclear. Some studies have revealed how PN only produces DA in the stationary growth phase, while other studies have shown that DA production is active in the exponential growth phase (Lelong et al., 2012 and references therein).

From field studies, it is well established that PN benefits from environmental nitrogen loading, both with regard to growth and toxicity (e.g. Downes-Tettmar et al., 2013; Husson et al., 2016; Thorel et al., 2017). Studies assessing the impact of inorganic nitrogen on PN are increasing in number (Auro and Cochlan, 2013; Bates et al., 1993; Bates and Trainer, 2006; Thessen et al., 2009), as are those focusing on organic nitrogen sources such as urea and amino-acids (Auro and Cochlan, 2013; Hillebrand and Sommer, 1996; Howard et al., 2007; Kudela et al., 2008; Martin-Jézéquel et al., 2015.; Radan and Cochlan, 2018). Different nitrogen sources seem to affect PN differently and presently no general pattern in growth, biomass and DA production has been established.

The effect of urea on DA production is, like other DA-inducing factors, unclear in the few studies performed. Urea has been shown to result in higher or lower cellular DA content when compared with other N-sources. When exposing P. cuspidata to different N-sources, an increase in DA content was found when grown with nitrate or ammonium as N-source compared to urea (Auro and Cochlan, 2013). On the contrary, higher DA content was found when P. australis was grown on urea rather than on inorganic N substrates (Howard et al., 2007). Similarly, P. multiseries contained more DA when growing on urea compared to inorganic N-sources (Radan and Cochlan, 2018).

Among PN species, there also seems to be a considerable variation in growth rates in response to N-sources (Radan and Cochlan, 2018). In some, although not all, strains of P. multiseries, P. calliantha and P. fraudulenta, a significant decrease in growth rate was found when growing on urea as the sole nitrogen source compared to ammonium or nitrate (Thessen et al., 2009). On the contrary in another study, P. multiseries and P. australis had similar growth rates when growing on nitrate, ammonium, and urea (Martin-Jézéquel et al., 2015)

The effects of different nitrogen sources have been compared, whereas the effects of different concentrations of urea have presently not been assessed in PN. As urea contains nitrogen, crucial for cell growth and DA production, variations in urea concentration are, however, expected to have an impact on both cell growth and DA content.

A wide variety of PN species are confirmed present in Danish waters and other parts of the North Atlantic area (Lundholm et al., 2010; recently reviewed in Bates et al., 2018), and several have been confirmed as toxigenic: P. seriata (Lundholm et al., 1994), P. pungens (Rhodes et al., 1996), P. fraudulenta (Rhodes et al., 1998), P. australis (Fehling et al., 2004) and P. granii (Trick et al., 2010). The presence of toxigenic PN-species in Danish and neighboring waters poses a risk for toxic events. In 2005, a bloom of P. seriata caused a closure of mussel harvesting areas in Denmark, close to the areas affected in 2016 (Lundholm et al., 2005), and in other North Atlantic areas, several bivalve fishery areas have previously been quarantined e.g. in western Scotland (Fehling et al., 2006, and references therein).

Data were gathered on the phytoplankton community, DA in mussels and the environmental conditions in the months before and following the collapse of the fertilizer silo. The aim was to explore whether urea can induce growth and affect DA content in PN species and hence explain the development of the observed toxic bloom. Did the fertilizer spill have any quantitative or qualitative impact on the phytoplankton community? Can urea induce growth and hence initiate a P. seriata bloom? Can urea induce toxin production in P. seriata and other PN species? To answer these questions, batch culture experiments with strains of P. seriata, P. delicatissima, and P. obtusa were performed. Pseudo-nitzschia seriata (Fig. S2) have previously been shown toxic in Danish, Scottish and Greenlandic waters (Lundholm et al., 1994; Fehling et al., 2004; Hansen et al., 2011) and recently strains from Iceland were tested positive for DA (Lundholm et al., 2018). Pseudo-nitzschia delicatissima has not been found toxic in Danish-, and adjacent waters, but isolates from other geographical regions have been recorded toxic (Bates et al., 2018). Pseudo-nitzschia obtusa strains from Greenland are known to produce DA in the presence of copepods (Harðardóttir et al., 2015), but no other inducing factor is confirmed for this species.