Significance of Plankton Community Structure and Nutrient Availability for the Control of Dinoflagellate Blooms by Parasites: A Modeling Approach

Abstract

Dinoflagellate blooms are frequently observed under temporary eutrophication of coastal waters after heavy rains. Growth of these opportunistic microalgae is believed to be promoted by sudden input of nutrients and the absence or inefficiency of their natural enemies, such as grazers and parasites. Here, numerical simulations indicate that increasing nutrient availability not only promotes the formation of dinoflagellate blooms but can also stimulate their control by protozoan parasites. Moreover, high abundance of phytoplankton other than dinoflagellate hosts might have a significant dilution effect on the control of dinoflagellate blooms by parasites, either by resource competition with dinoflagellates (thus limiting the number of hosts available for infection) or by affecting numerical-functional responses of grazers that consume free-living parasite stages. These outcomes indicate that although both dinoflagellates and their protozoan parasites are directly affected by nutrient availability, the efficacy of the parasitic control of dinoflagellate blooms under temporary eutrophication depends strongly on the structure of the plankton community as a whole.

About the Authors:

Catharina Alves-de-Souza

* E-mail: [email protected]

Affiliation: Laboratório de Ficologia, Departamento de Botânica, Museu Nacional/Universidade Federal do Rio de Janeiro, Quinta da Boa Vista S/N, São Cristóvão, Rio de Janeiro, RJ, Brasil

David Pecqueur

Affiliation: UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, Université Montpellier, Place E. Bataillon, CC 093, Montpellier, France

Emilie Le Floc’h

Affiliations UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, Université Montpellier, Place E. Bataillon, CC 093, Montpellier, France, UMS 3282 OSU OREME-MEDIMEER, IRD, UM, CNRS, SMEL, 2 rue des Chantiers, Sète, France, UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, SMEL, Sète, France

Sébastien Mas

Affiliation: UMS 3282 OSU OREME-MEDIMEER, IRD, UM, CNRS, SMEL, 2 rue des Chantiers, Sète, France

Cécile Roques

Affiliation: UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, Université Montpellier, Place E. Bataillon, CC 093, Montpellier, France

Behzad Mostajir

Affiliation: UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, Université Montpellier, Place E. Bataillon, CC 093, Montpellier, France

Franscesca Vidussi

Affiliation: UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, Université Montpellier, Place E. Bataillon, CC 093, Montpellier, France

Lourdes Velo-Suárez

Affiliation: Ifremer, DYNECO PELAGOS, Plouzané, France

Marc Sourisseau

Affiliation: Ifremer, DYNECO PELAGOS, Plouzané, France

Eric Fouilland

Affiliation: UMR 9190 MARBEC Center of Marine Biodiversity, Exploitation & Conservation, CNRS, UM, IRD, IFREMER, SMEL, Sète, France

Laure Guillou

Affiliation: CNRS, UMR 7144 & Université Pierre et Marie Curie, Station Biologique de Roscoff, Place Georges Teissier, Roscoff, France

Introduction

Most harmful algal blooms (HABs) are formed by toxic or otherwise harmful dinoflagellate species that have potentially devastating effects on marine habitats including the closure of shellfish beds, massive fish kills, and death of other marine fauna [1]. They are frequently initiated in semi-confined areas like harbors, lagoons, and estuaries where sudden nutrient pulses from rivers draining urban centers and agricultural watersheds result in temporary eutrophication [2–5]. Such increases in nutrient availability could promote bloom formation not only by boosting dinoflagellate growth rates (bottom-up control) [6], but also by changing the structure and dynamics of plankton communities, including grazers and parasites that are natural enemies of dinoflagellates (top-down control) [7–9]. Amongst the top-down mechanisms influencing dinoflagellate blooms, parasitism stands out as one of the most direct and widespread [10]. Yet, little is known about how shifts in plankton community composition driven by changes in nutrient concentration could affect host-parasite dynamics in plankton systems.

Amoebophrya (Alveolata, Syndiniales) is a genus of obligate parasitic eukaryotes infecting a wide range of marine dinoflagellates [10]. The life-cycle of these parasites is characterized by an alternation between a biflagellated free-living infective stage (the dinospore) and an intracellular stage (the trophont). Once inside the host, maturation takes 2–3 days and eventually culminates with the death of the host and the release of a motile worm-shaped multinucleated and multiflagellated structure (the vermiforme), which fragments within a few hours, releasing hundreds of dinospores, each potentially capable of infecting a novel host [11, 12]. This simple life-cycle, combined with a short generation time, make Amoebophrya spp. excellent models for studying host-parasite dynamics within plankton communities.

Both field observations and model predictions indicate that these highly virulent parasites have the capacity to efficiently control blooms of their dinoflagellate hosts [10, 13–15]. However, parasite and host populations do not exist as isolated pairs, but rather as parts of multispecies systems. In fact, dinoflagellates could be simultaneously affected by other biotic interactions, such as competition for nutrient uptake with other phytoplankton and grazing by microzooplankton [6, 16]. In addition, grazing on free-living parasite stages can significantly decrease parasite infections in plankton systems [9, 17]. Thus, in order to understand the factors affecting parasitic control of dinoflagellate blooms, it is important to consider the structure of the whole plankton community, including both the specific composition and the relative abundance of their members.

Here, the relative importance of nutrient availability and biotic interactions (i.e., parasitism, grazing and competition for nutrients) to dinoflagellate bloom dynamics was evaluated by numerical simulations. The main goal was to assess how changes in the relative abundance of host and non-host species driven by changes in nutrient concentration could affect the dilution effect on parasite infections. Model outcomes indicate that although increasing nutrient concentration promotes both the growth of dinoflagellates and the development of their protozoan parasites, the efficacy of parasitic control of dinoflagellate blooms under temporary eutrophication depends strongly on the structure of the plankton community as a whole.

Materials and Methods

Modeling approach

The model was based upon previous mathematical approaches related to Amoebophrya-host dynamics [15, 18]. In order to avoid the inclusion of too many trophic links (which would lead to an unnecessary complexity), the modeled community included as few components as possible. Thus, we considered a simple plankton community constituted by eukaryotes, only, including dinoflagellates, their parasites, their potential competitors (phytoplankton other than dinoflagellates), as well as their grazers (microzooplankton). Although mesozooplankton can exhibit high grazing rates on blooming dinoflagellates [19], they were not considered because their long generation times reduce the likelihood of exerting immediate control over phytoplankton blooms following nutrient pulses [20]. Nitrate concentration was chosen as the forcing variable because nitrogen is generally the most limiting nutrient in marine systems [5]. Thus, it would be reasonable to consider it as the main nutrient to take into account when considering a situation of temporary eutrophication, in which a phytoplankton community is suddenly exposed to an increased concentration of nutrients. Moreover, this is the nutrient with the largest volume of available information in the literature on nutrient uptake, which facilitated the parameterization of the model regarding phytoplankton growth.

Briefly, the phytoplankton community was represented in the model by nanophytoplankton other than diatoms and dinoflagellates (A), diatoms (D), uninfected (H) and infected (I) dinoflagellates. For modeling purposes, phytoplankton were considered to be divided in nanophytoplankton, diatoms and dinoflagellates with no overlapping between those groups, although the term nanophytoplankton usually refers to phytoplankton smaller than 20 μm that may include small diatoms and dinoflagellates. Rotifers (R) grazed only on prey smaller than 20 μm [21], i.e. nanophytoplankton (A) and dinospores (P) (the free-living parasite stages). Although microciliates (C) can have a wide prey range [22], modeled microciliates grazed only on diatoms and dinoflagellates. By assigning different preys to microciliates and rotifers, we were able to evaluate the individual importance of these two grazers for the host-parasite dynamics.

Parameters/state variables and equations used in the simulations are shown in Tables 1 and 2, respectively. The initial values of the variables reported in Table 1 were based on the specific composition of the eukaryotic plankton community observed at the beginning of a bloom of the dinoflagellate Prorocentrum triestinum (infected by Amoebophrya spp.) following an experimental nutrient pulse obtained during a mesocosm experiment performed in Thau Lagoon (southern France) in autumn 2009. Similarly, the NO3+NO2 values observed before (1 μM) and after (36 μM) the experimental nutrient pulse were used here as the nitrate concentrations for running the simulations under oligotrophic and eutrophic conditions, respectively. The choice of the plankton community from Thau Lagoon as a template for the model is justified because this system is periodically subject to dinoflagellate blooms [23] and temporary eutrophication following river discharge [24]. More details regarding the methods used in the determination of these variables as well as a short description of the composition of the plankton community in Thau Lagoon, are available in the S1 Appendix (Sections S1-1 and S1-2). No specific permissions were required for these locations/activities and these field collections did not involve endangered or protected species.

[Figure omitted. See PDF.]

Table 1. Values of parameters and state variables considered in the numerical simulations.

https://doi.org/10.1371/journal.pone.0127623.t001

[Figure omitted. See PDF.]

Table 2. Differential equations used in the numerical simulations.

https://doi.org/10.1371/journal.pone.0127623.t002

Sensitivity analysis

The Sobol' method [35] was used to conduct the sensitivity analysis of the proposed model. This analysis is a global approach to estimate the effect of one parameter on the model output when all other parameters vary, enabling the identification of interactions in the model [36]. Sobol’s first and total order sensitivity indices were estimated for 30-d simulations using the SBtoolbox [37] using Matlab (MathWorks).

Numerical simulations

Different scenarios representing an increasing degree of community complexity were simulated: i) parasites and dinoflagellates only, ii) parasites, dinoflagellates and grazers, and iii) parasites, dinoflagellates, grazers and other phytoplankton (i.e., nanophytoplankton and diatoms). In each scenario, simulations were run for 30 days under two trophic conditions: oligotrophic (1 μM nitrate) and eutrophic (36 μM nitrate). A 30-d period was chosen because this represents a period in which both the development and the demise of dinoflagellate blooms are expected to occur in Thau Lagoon [38].

Dinoflagellate growth was further modeled, in the presence and absence of parasites, under a series of abundances of phytoplankton other than dinoflagellates (104, 105, 106, 107, 108, 109 cells L–1) and increasing concentrations of nitrate (0.5, 1, 5, 10, 20, 30, 40 and 50 μM). In all simulations, the proportion between nanophytoplankton and diatoms was maintained at 99:1. The abundances of dinoflagellates (250,000 cells L–1), microciliates (3,200 cells L–1) and rotifers (20 ind L–1) were fixed. For simplicity, simulations with the presence of parasites started with no infected dinoflagellates and with a dinospore:host proportion of 2:1 (i.e., dinospores = 125,000 cells L–1).

The effect of each component of the modeled plankton community on the parasite-dinoflagellate dynamics was assessed by performing additional 30-d simulations where the different components were excluded from the model (one at a time). Negative effects were divided into two types: those that acted during the exponential growth phase (i.e., they affected the maximal abundance of dinoflagellates and dinospores); and those that contributed to the elimination of dinoflagellates and dinospores once their maximal cell concentrations were achieved. For that, we evaluated whether the removal of a particular model component resulted in an increase in maximal abundances of dinoflagellates and/or dinospores. At the same time, we estimated the percentage of loss caused by parasites and/or grazers during the demise of the blooms using the loss terms of the differential equations and estimated the relative importance of parasites and grazers for the demise of blooms under eutrophic conditions (36 μM nitrate). In both cases, the intensity (%) of the negative effect was estimated by comparing the maximum abundance of the affected component in the presence and absence of the tested component under a low and a high proportion of other phytoplankton (104 and 108 cells L–1, respectively). In all simulations, the proportion between nanophytoplankton and diatoms was set at 99:1. Initial abundances of the other components were as follows: dinoflagellates = 250,000 cells L–1, microciliates = 3,200 cells L–1, rotifers = 20 ind L–1, dinospores = 125,000 cells L–1.

Finally, grazing on dinospores by rotifers was simulated under increasing nanophytoplankton concentrations (104, 105, 106, 107, 108 and 109cells L–1). To remove possible interference related to the release of dinospores from infected hosts and nanophytoplankton growth, simulations were carried out in the absence of hosts over a period of 24 h. In all simulations, dinospore and rotifer abundances were 125,000 cells L–1 and 20 ind L–1, respectively. At the same time, we assessed how two nanophytoplankton concentrations (105 and 107 cells L–1) affected the consumption of dinospores by rotifers during 30-day simulations under eutrophic condition (36 μM nitrate). Initial abundances of other components of the plankton community were as follows: dinoflagellates = 250,000 cells L–1, diatoms = 29,000 cells L–1, microciliates = 3,200 cells L–1, rotifers = 20 ind L–1, dinospores = 125,000 cells L–1. Simulations started without the presence of infected dinoflagellates.

All the simulations were performed using Matlab (MathWorks). Scripts are available on request from the first author.

Results

The effect of oligotrophic and eutrophic conditions (1 and 36 μM nitrate, respectively) on host-parasite dynamics were simulated under three scenarios representing an increasing degree of plankton community complexity. Without grazers (scenario 1; Fig 1A), simulations resulted in much higher abundance of dinoflagellates and dinospores in eutrophic condition than those observed under low nitrate concentration (Fig 1D and 1E). Although the prevalence (% of infection) reached almost 100% in both nitrate concentrations, a delay was observed in the timing of the parasite-host dynamics under low nitrate concentration (maximal prevalence observed on days 9 and 17 in eutrophic and oligotrophic conditions, respectively). Parasite-host dynamics remained relatively similar when grazers were added to the model (scenario 2; Fig 1B). The only difference was a faster decline of both dinospores and parasite prevalence in scenario 2 than in scenario 1 (Fig 1F and 1G). Addition of phytoplankton other than dinoflagellates (nanophytoplankton and diatoms) (scenario 3; Fig 1C) had a negative effect on dinoflagellate and dinospore abundances in both simulated trophic conditions. In contrast to scenarios 1 and 2, nitrate availability strongly affected parasite prevalence when other phytoplankton were present (maximal parasite prevalence of 99% and 10%, in eutrophic and oligrotrophic conditions, respectively) (Fig 1H and 1I).

[Figure omitted. See PDF.]

Fig 1. Interplay between nutrient concentration and plankton community complexity on parasite-host dynamics.

(A–C) Different scenarios representing an increasing degree of community complexity: parasites and dinoflagellates only (A), parasites, dinoflagellates, and grazers (B), and parasites, dinoflagellates, grazers, and phytoplankton other than dinoflagellates (C); h = uninfected dinoflagellates, i = infected dinoflagellates, a = nanophytoplankton, d = diatoms, p = dinospores (parasite free-living stages), c = microciliates, r = rotifers. (D–I) Temporal dynamics of dinoflagellates (cells L–1106), dinospores (cells L–1109) and prevalence (% of infection) obtained from 30-day simulations under oligotrophic (OLIG) and eutrophic conditions (EUTR) (1 and 36 μM nitrate, respectively).

https://doi.org/10.1371/journal.pone.0127623.g001

Dinoflagellate growth was further modeled, in the presence and absence of parasites, under varying abundance of phytoplankton other than dinoflagellates and increasing concentrations of nitrate (Fig 2A). Dinoflagellate blooms were observed only at nitrate concentrations above 20 μM and low abundance of other phytoplankton (> 107 cells L–1). These conditions also corresponded to the highest mortality caused by parasites (maximal dinoflagellate abundance of 108 and 107 cells L–1 in the presence and absence of parasites, respectively). Under high abundance of other phytoplankton (>108 cells L–1), dinoflagellates were not greatly impacted by an increasing nutrient availability and parasites had no effect on dinoflagellate growth.

[Figure omitted. See PDF.]

Fig 2. Effect of the different components of the modeled plankton community on the parasite-dinoflagellate dynamics.

(A) Results of simulations assessing the effect of the initial abundance (cells L–1) of other phytoplankton (i.e. nanophytoplankton and diatoms) on maximum dinoflagellate abundance (cells L–1), in the presence (colored mesh plot) and absence (red mesh plot) of parasites, under different nitrate concentrations (μM). The composition of the plankton community was the same as in Fig 1C. (B–C) Individual relevance of the different components of the simulated plankton community that negatively affected dinoflagellates and their parasites ineutrophic conditions (36 μM nitrate) under low (B) and high (C) abundance of other phytoplankton (104 and 108 cells L–1, respectively). The proportion between nanophytoplankton and diatoms was 99:1 in all simulations. Components of the plankton community are identified by the same letters as indicated in Fig 1C. Arrow thickness indicates the intensity (%) of negative effects (only values higher than 1% are shown). Negative effects were divided into two types: those that acted during the exponential growth phase (i.e. they affected the maximum number of individuals in the population) (black arrows); and those that contributed to the elimination of dinoflagellates and parasites (yellow arrows). Dashed arrows indicate indirect negative interactions.

https://doi.org/10.1371/journal.pone.0127623.g002

An additional set of simulations showed how the different components of the modeled plankton community (i.e., other phytoplankton, grazers and parasites) can negatively affect the initiation and development of dinoflagellate blooms and/or cause the demise of their populations under eutrophic conditions (36 μM nitrate) by removing the different components of the model (one at a time). Also, these simulations indicated how parasite dinospores were affected by other components of the plankton community besides dinoflagellates. Under low abundance of other phytoplankton (104 cells L–1; Fig 2B), parasites were the main factor controlling not only the initiation and development of dinoflagellate populations but also their extinction. Microciliates had a slight indirect negative effect on maximal dinospore abundance whereas rotifers contributed to 28% of dinospore mortality. Under high abundance of other phytoplankton (108 cells L–1; Fig 2C), the initiation and development of dinoflagellate bloom were avoided mostly by competition with diatoms. Parasites had only a slight effect on the extinction of dinoflagellates which was caused mainly by ciliate grazing. Both diatoms and nanophytoplankton had an indirect negative effect on the maximal dinospore abundance whereas rotifers contributed up to about 84% of dinospore mortality. Increasing abundance of nanophytoplankton resulted in the increase of both growth and grazing rates of rotifers (Fig 3A). When nanophytoplankton abundance was lower than 104 cells L–1, rotifers consumed dinospores mostly after the dinospore burst with no effect on the parasite-host dynamics (Fig 3B and 3C). On the other hand, grazing pressure on dinospores was high from the beginning of the simulations when nanophytoplankton abundance was higher than 107 cells L–1 leading to a decrease in the number of infective dinospores (Fig 3D and 3E).

[Figure omitted. See PDF.]

Fig 3. Influence of nanophytoplankton abundance on rotifer grazing of dinospores.

(A) Results from simulations assessing the effect of increasing nanophytoplankton abundances (cells L–1) on the number of dinospores consumed by rotifers (dinospores rotifer–1 d–1) and rotifer growth rates (d–1). The dashed line indicates the point where rotifer growth = 0. (B–E) Effect of two nanophytoplankton concentrations, 105 and 107 cells L–1 (B–C and D–E, respectively), on the consumption of dinospores by rotifers during 30-day simulations.

https://doi.org/10.1371/journal.pone.0127623.g003

Results from the global sensitivity analysis (Fig 4) indicated that, under nutrient limitation (oligotrophic condition; 1 μM nitrate), dinoflagellates were affected mainly by the half-saturation constant of diatoms. On the other hand, the half saturation constant for ciliate growth was the most important model parameter affecting dinoflagellates under high nutrient availability (eutrophic condition; 36 μM nitrate). The amount of infected dinoflagellates was greatly influenced by the number of dinospores released per infected host under oligotrophic condition whereas the prey threshold for ciliate growth was the key parameter influencing infections under eutrophic conditions. Finally, dinospore density was most affected by the half-saturation constant for rotifer grazing and the maximal ingestion rate of rotifers under oligotrophic and eutrophic conditions, respectively.

[Figure omitted. See PDF.]

Fig 4. Results of Sobol’s sensitivity analysis.

Ranking of first index sensitivities for the most relevant 5 parameters affecting dinoflagellate abundance (H; red bars), infected dinoflagellate cells (I; grey bars) and dinospore abundance (P; black bars) under oligotrophic and eutrophic conditions (1 and 36 μM nitrate, respectively). Parameters are numbered from 1 (most influencing) to 5. Meanings of parameter symbols are the same as indicated in Table 1.

https://doi.org/10.1371/journal.pone.0127623.g004

Discussion

Eutrophication has been indicated as one of the main factors contributing to the increase in frequency and geographic distribution of HABs observed in the past decades [39]. In this context, considerable effort has been made to understand how changing nutrient concentrations and ratios result in the tendency of toxic or otherwise harmful opportunistic algal species to bloom (reviewed by [6]). However, blooms should not be considered to be exclusively regulated by abiotic factors as they are also under biological control [40]. In this study, numerical simulations indicated that parasites had greater impact on the development of dinoflagellate blooms than microciliates (as previously pointed out by [15]). Grazing by microzooplankton has been proposed as an important mechanism controlling dinoflagellate blooms since protistan grazers are capable of growing as fast as their phytoplankton preys [41]. However, the generally high prey threshold of microzooplankton organisms such as ciliates implies a large time delay in their numerical-functional responses (i.e., increase/decrease in growth and grazing rates of predators according to the total amount of available preys) [20]. The sensitivity analysis of our model indicated that dinoflagellate abundance was sensitive to changes in the values of the half saturation constant for ciliate growth (Fig 4). This is in agreement with simulations performed by Montagnes et al. [15] where ciliates were only able to control dinoflagellate blooms when values for their grazing parameters were lower than recorded in the literature. As a consequence, although microzooplankton is capable of eliminating dinoflagellate blooms once they are established [16, 42], they are not likely to be capable of preventing their initiation [20].

Parasites of genus Amoebophrya have been identified as potentially important biotic factors controlling dinoflagellate blooms [13, 14]. Host abundance has been proposed as the main factor affecting infective rates of these parasites [10]. This is related to the density-dependence of parasite dynamics since increasing host density enhances the risk of the host to become infected through increased encounter rates [43]. Given that each Amoebophrya infection produces hundreds of dinospores [12], even a small increase in the number of available hosts would result in a significant increase in the number of propagules released at any given moment of the parasite-host dynamics. Based on that, the stimulation of dinoflagellate growth resulting from increased nutrient availability would always lead to higher parasitic rates. However, the modeling approach considered in this study indicated that the effect of temporary eutrophication on the parasitic control of dinoflagellate blooms should be considered within a community context.

Nutrient availability affected parasitic control of dinoflagellates, but the magnitude of this effect was likely determined by the presence/absence as well as the relative abundance of other members of the plankton community which would exert direct and/or indirect effects on host-parasite dynamics. When host and parasites were considered independently, parasites exerted a complete control on the dinoflagellate host (Fig 1D and 1E), as expected from previous results of both experimental and modeling studies [12, 18]. We expected that the addition of grazers to this simple system would affect host-parasite dynamics, since it has been shown that grazing on the free-living stages of parasites could significantly diminish parasite prevalence in plankton systems [9, 17, 44]. However, while grazers by themselves were not important (Fig 1F and 1G), differences in both parasite prevalence and dinospore density related to nutrient concentrations were greatly magnified by the addition of other phytoplankton (Fig 1H and 1I).

Numerical simulations indicated that although nutrient availability importantly affected dinoflagellate growth (no bloom was formed at nitrate concentrations below 20 μM) (Fig 2A), the predominant biotic mechanism controlling dinoflagellate populations depended strongly on the relative abundance of the different phytoplankton groups at the beginning of the simulations. When diatoms and nanophytoplankton were dominant, the formation of dinoflagellate blooms under high nutrient concentrations was impeded mostly by resource competition for nutrients (mainly with diatoms) (Fig 2B), related to higher growth rates and low half-saturation constant for nitrate uptake usually observed in diatoms [45]. This is in agreement with the results of the sensitivity analysis that indicated the half saturation constant for nitrate uptake of diatoms as the most relevant parameter of the model affecting dinoflagellate abundance (Fig 4). Parasites importantly reduced the size of dinoflagellate populations, but caused its elimination only when dinoflagellates dominated phytoplankton assemblages (Fig 2C).

The indirect negative effect of diatoms and nanophytoplankton on free-living parasite stages was an unexpected result of these simulations. By outcompeting dinoflagellates for nutrient uptake, other phytoplankton limited the number of hosts available for infection, which in turn decreased the total amount of released dinospores. Interestingly, a high abundance of other phytoplankton also determined ciliates (and not parasites) to be the main cause for the elimination of dinoflagellates. Simultaneously, the nanophytoplankton stimulated predation on the parasites by affecting numerical-functional responses of rotifers (Fig 3A). When nanophytoplankton abundance was lower than 107 cells L–1, the total amount of available prey was below the threshold for rotifer growth. As a consequence, rotifers consumed dinospores mostly after their maximal abundance was observed (and most of the hosts were killed) with no effect on the host-parasite dynamics (Fig 3B and 3C). On the other hand, when the abundance of nanophytoplankton was higher than 107 cells L–1, grazing pressure on dinospores was high from the beginning of the simulations leading to a decrease in the number of dinospores able to establish infections (Fig 3D and 3E). This is very relevant considering that the sensitivity analysis indicated that dinospore abundance was mostly affected by parameters related to rotifer grazing (Fig 4). Predation on free-living parasite stages is an important biotic factor affecting parasite transmission [46]. However, previous experimental evidence is contradictory in that the presence of predators could lead either to an increase or decrease in parasite prevalence [47, 48]. Our results suggest that the presence of alternative prey (potentially affecting the numerical-functional response of predators) should also be taken into account to understand the conditions under which predators affect host-parasite dynamics.

Parasite-mediated interactions are widely known to significantly affect population structure, trophic relationships and energy flow within ecological communities [49]. Less frequent is the perception that parasites are in turn affected by interactions with other species besides their focal hosts. The ‘dilution effect hypothesis’ suggests that the net effect of increasing species richness reduces the risk of certain infectious diseases in ecological communities [50]. Although initially focused on how the presence of a diverse assemblage of relatively inefficient hosts would reduce parasite transfer to competent hosts [51], it is now clear that other mechanisms related to non-host species may also act as buffers against parasite transmission [47, 52]. Here, we provide evidence that diverse phytoplankton assemblages that are functionally different from the host (e.g. diatoms and nanophytoplankton) may have an important indirect dilution effect on the parasitic control of dinoflagellate blooms either by competing for resources with dinoflagellates (and thus limiting the number of hosts available for infection) or by stimulating microzooplankton grazing on free-living parasite stages. We therefore speculate that parasites are likely more efficient as controlling agents of dinoflagellates during monospecific blooms. Forthcoming experimental research should consider possible interplay between parasitism and strategies frequently used by dinoflagellates to increase their competitive performance (e.g., mixotrophy and allelopathy [53, 54]).

The assessment of the relevance of eutrophication on HAB dynamics should take into account that plankton communities as a whole and not only HAB species are impacted by nutrient input, resulting in changes of the biotic interactions [40]. The modeling approach proposed in this study indicated that although nutrient availability promotes both the growth of dinoflagellates and the development of their protozoan parasites, the efficacy of parasitic control of dinoflagellate blooms under temporary eutrophication depends strongly on the structure of the plankton community at the moment when the nutrient pulse occurs. The modeling outcomes further underscore the importance of accounting for indirect interactions when assessing the relative importance of bottom-up and top-down factors on HAB dynamics.

Supporting Information

[Figure omitted. See PDF.]

S1 Appendix. Plankton structure used for the formulation of the model.

Section S1–1. Determination of the variables. Section S1–2. Nitrate concentrations and composition of the plankton community.

https://doi.org/10.1371/journal.pone.0127623.s001

(DOC)

S1 Fig. Outputs of the 30-d numerical simulations for phytoplankton other than dinoflagellates (A–B) and microzooplankton (C–D).

https://doi.org/10.1371/journal.pone.0127623.s002

(PNG)

Acknowledgments

We thank A. Trottet, C. Leboulanger, J. Nouguier, R. Pete, C. Vasseur M.-G., Tournoud and C. Salles for their invaluable help during samplings in Thau Lagoon. We thank D. Bonnet for rotifer counting and J. Koester for her English review of the manuscript. Results provided by the national French EC2CO programme RESTHAU (2008–2009) coordinated by E. Fouilland and using MEDIMEER facilities (MEDIterranean platform for Marine Ecosystems Experimental Research) was partly used in the present study. The modeling part was supported by the national French EC2CO program PARALEX (2009–2011) coordinated by L. Guillou. L. Velo-Suárez was supported by a Marie Curie International Outgoing Fellowship (IOF; grant agreement: MOHAB PIOF-GA-252260).

Author Contributions

Conceived and designed the experiments: EF SM ELF. Performed the experiments: EF SM ELF. Analyzed the data: DP CR EF BM FV CAS LG LVS MS. Contributed reagents/materials/analysis tools: EF LG BM FV. Wrote the paper: CAS DP ELF CR SM FV BM LVS MS EF LG. Formulated the mathematical model and performed the numerical simulations: CAS LVS MS.

Citation: Alves-de-Souza C, Pecqueur D, Le Floc’h E, Mas S, Roques C, Mostajir B, et al. (2015) Significance of Plankton Community Structure and Nutrient Availability for the Control of Dinoflagellate Blooms by Parasites: A Modeling Approach. PLoS ONE 10(6): e0127623. https://doi.org/10.1371/journal.pone.0127623

References

1. Anderson DM, Cembella AD, Hallegraeff GM. Progress in understanding harmful algal blooms: paradigm shifts and new technologies for research, monitoring, and management. Annu Rev Mar Sci. 2012;4:143–76. pmid:22457972

2. Walsh JJ, Jolliff J, Darrow B, Lenes J, Milroy S, Remsen A, et al. Red tides in the Gulf of Mexico: where, when, and why? J Geophys Res. 2006;111(C11003): pmid:20411040

3. Warrick JA, Washburn L, Brzezinski MA, Siegel DA. Nutrient contributions to the Santa Barbara Channel, California, from the ephemeral Santa Clara River. Estuar Coast Shelf Sci. 2005;62(4):559–74.

4. Hallegraeff GM. Ocean climate change, phytoplankton community responses, and harmful algal blooms: a formidable predictive challenge. J Phycol. 2010;46(2):220–35.

5. Paerl HW. Assessing and managing nutrient-enhanced eutrophication in estuarine and coastal waters: interactive effects of human and climatic perturbations. Ecol Eng. 2006;26(1):40–54.

6. Heisler J, Glibert PM, Burkholder JM, Anderson DM, Cochlan W, Dennison WC, et al. Eutrophication and harmful algal blooms: a scientific consensus. Harmful Algae. 2008;8(1):3–13.

7. Buskey EJ. How does eutrophication affect the role of grazers in harmful algal bloom dynamics? Harmful Algae. 2008;8(1):152–7.

8. Yih W, Coats DW. Infection of Gymnodinium sanguineum by the dinoflagellate Amoebophrya sp.: effect of nutrient environment on parasite generation time, reproduction, and infectivity. J Eukaryot Microbiol. 2000;47(5):504–10. pmid:11001148

9. Johansson M, Coats DW. Ciliate grazing on the parasite Amoebophrya sp. decreases infection of the red-tide dinoflagellate Akashiwo sanguinea. Aquat Microb Ecol. 2002;28:69–78.

10. Park MG, Yih W, Coats DW. Parasites and phytoplankton, with special emphasis on dinoflagellate infections. J Eukaryot Microbiol. 2004;51(2):145–55. pmid:15134249

11. Cachon J. Contribution à l'étude des péridiniens parasites. Cytologie, cycles évolutifs. Ann Sci Nat B. 1964;6:1–158.

12. Coats DW, Park MG. Parasitism of photosynthetic dinoflagellates by three strains of Amoebophrya (Dinophyta): parasite survival, infectivity, generation time, and host specificity. J Phycol. 2002;38:520–8.

13. Coats DW, Adam EJ, Gallegos CL, Hedrick S. Parasitism of photosynthetic dinoflagellates in a shallow subestuary of Chesapeake Bay, USA. Aquat Microb Ecol. 1996;11:1–9.

14. Chambouvet A, Morin P, Marie D, Guillou L. Control of toxic marine dinoflagellate blooms by serial parasitic killers. Science. 2008;322:1254–7. pmid:19023082

15. Montagnes DJS, Chambouvet A, Guillou L, Fenton A. Responsibility of microzooplankton and parasite pressure for the demise of toxic dinoflagellate blooms. Aquat Microb Ecol. 2008;53:211–25.

16. Calbet A, Vaqué D, Felipe J, Vila M, Sala MM, Alcaraz M, et al. Relative grazing impact of microzooplankton and mesozooplankton on a bloom of the toxic dinoflagellate Alexandrium minutum. Mar Ecol Prog Ser. 2003;259:303–9.

17. Kagami M, Bruin A, Ibelings BW, Donk E. Parasitic chytrids: their effects on phytoplankton communities and foodweb dynamics. Hydrobiologia. 2007;578:113–29.

18. Salomon PS, Stolte W. Predicting the population dynamics in Amoebophrya parasitoids and their dinoflagellate hosts using a mathematical model. Mar Ecol Prog Ser. 2010;419:1–10.

19. Teegarden GJ, Cembella AD. Grazing of toxic dinoflagellates, Alexandrium spp., by adult copepods of coastal Maine: implications for the fate of paralytic shellfish toxins in marine food webs. J Exp Mar Biol Ecol. 1996;196(1):145–76.

20. Sherr EB, Sherr BF. Capacity of herbivorous protists to control initiation and development of mass phytoplankton blooms. Aquat Microb Ecol. 2009;57:253–62.

21. Hansen B, Wernberg-Møller T, Wittrup L. Particle grazing efficiency and specific growth efficiency of the rotifer Brachionus plicatilis (Muller). J Exp Mar Biol Ecol. 1997;215:217–33.

22. Verity PG. Feeding In planktonic protozoans: evidence for non-random acquisition of prey. J Eukaryot Microbiol. 1991;38(1):69–76.

23. Laanaia N, Vaquer A, Fiandrino A, Genovesi B, Pastoureaud A, Cecchi P, et al. Wind and temperature controls on Alexandrium blooms (2000–2007) in Thau lagoon (Western Mediterranean). Harmful Algae. 2013;28:31–6.

24. Tournoud M-G, Payraudeau S, Cernesson F, Salles C. Origins and quantification of nitrogen inputs into a coastal lagoon: Application to the Thau lagoon (France). Ecol Model. 2006;193(1):19–33.

25. Chan LL, Hodgkiss IJ, Wan JMF, Lum JHK, Mak ASC, Sit WH, et al. Proteomic study of a model causative agent of harmful algal blooms, Prorocentrum triestinum II: the use of differentially expressed protein profiles under different growth phases and growth conditions for bloom prediction. Proteomics. 2004;4(10):3214–26. pmid:15378703

26. Seeyave S. Nitrogen nutrition of harmful algal blooms in upwelling systems: University of Southampton; 2009.

27. Zheng-Fang W, Qing Z, Min G. The effects of nitrogen, phosphorus, vitamins and trace metals on the growth of the red tide organism Prorocentrum micans. Chin J Oceanol Limnol. 1995;13(4):338–42.

28. Furnas MJ. Growth Rates of Summer Nanoplankton (< 10/ m) Populations in Lower Narragansett Bay, Rhode Island, USA. Mar Biol. 1982;70:105–15.

29. Montagnes DJS, Berges JA, Taylor FJR, Harrison PJ. Estimating carbon, nitrogen, protein and chlorophyll a from cell volume in marine phytoplankton. Limnol Oceanogr. 1994;39:1044–60.

30. Örnólfsdóttir E, Lumsden S, Pinckney J. Phytoplankton community growth-rate response to nutrient pulses in a shallow turbid estuary, Galveston Bay, Texas. J Plankton Res. 2004;26(3):325–39.

31. Collos Y, Vaquer A, Bibent B, Slawyk G, Garcia N, Souchu P. Variability in nitrate uptake kinetics of phytoplankton communities in a Mediterranean coastal lagoon. Estuar Coast Shelf Sci. 1997;44(3):369–75.

32. Jeong H, Yoon J, Kim J, Yoo Y, Seong K. Growth and grazing rates of the prostomatid ciliate Tiarina fusus on red-tide and toxic algae. Aquat Microb Ecol. 2002;28:289–97.

33. Ciros-Pérez J, Carmona M, Serra M. Resource competition between sympatric sibling rotifer species. Limnol Oceanogr. 2001;46(6):1511–23.

34. Montagnes D, Kimmance S, Tsounis G, Gumbs J. Combined effect of temperature and food concentration on the grazing rate of the rotifer Brachionus plicatilis. Mar Biol. 2001;139(5):975–9.

35. Sobol' IyM. On sensitivity estimation for nonlinear mathematical models. Matematicheskoe Modelirovanie. 1990;2(1):112–8.

36. Cariboni J, Gatelli D, Liska R, Saltelli A. The role of sensitivity analysis in ecological modelling. Ecol Model. 2007;203(1):167–82.

37. Schmidt H, Jirstrand M. Systems Biology Toolbox for MATLAB: a computational platform for research in systems biology. Bioinformatics. 2006;22(4):514–5. pmid:16317076

38. Collos Y, Gagne C, Laabir M, Vaquer A, Cecchi P, Souchu P. Nitrogenous nutrition of Alexandrium catenella (Dinophyceae) in cultures and in Thau lagoon, southern France. J Phycol. 2004;40:96–103.

39. Anderson DM, Glibert PM, Burkholder JM. Harmful algal blooms and eutrophication: nutrient sources, composition, and consequences. Estuaries. 2002;25(4):704–26.

40. Smayda TJ. Complexity in the eutrophication–harmful algal bloom relationship, with comment on the importance of grazing. Harmful Algae. 2008;8(1):140–51.

41. Strom SL, Morello TA. Comparative growth rates and yields of ciliates and heterotrophic dinoflagellates. J Plankton Res. 1998;20(3):571–84.

42. Rosetta CH, McManus GB. Feeding by ciliates on two harmful algal bloom species, Prymnesium parvum and Prorocentrum minimum. Harmful Algae. 2003;2(2):109–26.

43. Anderson RM, May RM. The population dynamics of microparasites and their vertebrate hosts. Phil Trans Royal Soc B. 1981;210:658–61.

44. Bachvaroff TR, Kim S, Guillou L, Delwiche C, Coats DW. Diversity of the Syndinian genus Euduboscquella based on single cell PCR. Appl Environ Microbiol. (in press).

45. Litchman E, Klausmeier CA, Schofield OM, Falkowski PG. The role of functional traits and trade‐offs in structuring phytoplankton communities: scaling from cellular to ecosystem level. Ecol Lett. 2007;10(12):1170–81. pmid:17927770

46. Thieltges D, Jensen K, Poulin R. The role of biotic factors in the transmission of free-living endohelminth stages. Parasitology. 2008;135(4):407. pmid:18208633

47. Johnson PTJ, Dobson A, Lafferty KD, Marcogliese DJ, Memmott J, Orlofske SA, et al. When parasites become prey: ecological and epidemiological significance of eating parasites. Trends Ecol Evol. 2010;25:362–71. pmid:20185202

48. Orlofske SA, Jadin RC, Hoverman JT, Johnson PT. Predation and disease: understanding the effects of predators at several trophic levels on pathogen transmission. Freshwat Biol. 2014;59:1064–75.

49. Hatcher MJ, Dick JTA, Dunn AM. Diverse effects of parasites in ecosystems: linking interdependent processes. Front Ecol Environ. 2012;10:186–94.

50. Keesing F, Holt RD, Ostfeld RS. Effects of species diversity on disease risk. Ecol Lett. 2006;9:485–98. pmid:16623733

51. Ostfeld RS, Keesing F. Biodiversity and disease risk: the case of Lyme disease. Conserv Biol. 2000;14:722–8.

52. Johnson PTJ, Thieltges DW. Diversity, decoys and the dilution effect: how ecological communities affect disease risk. J Exp Biol. 2010;213:961–70. pmid:20190121

53. Granéli E, Weberg M, Salomon PS. Harmful algal blooms of allelopathic microalgal species: the role of eutrophication. Harmful Algae. 2008;8(1):94–102.

54. Burkholder JM, Glibert PM, Skelton HM. Mixotrophy, a major mode of nutrition for harmful algal species in eutrophic waters. Harmful Algae. 2008;8(1):77–93.