1

Long-term allelopathic controlofphytoplankton by the submerged macrophyte Elodeanuttallii

Maarten Vanderstukken1, Steven A.J. Declerck2,3, Ellen Decaestecker1& Koenraad Muylaert1,*

1 Laboratory of Aquatic Biology, KU Leuven campus Kulak, E. Sabbelaan 54, 8500 Kortrijk, Belgium

2 Laboratory of Aquatic Ecology and Evolutionary Biology, KU Leuven, Ch. Deberiotstraat 32, 3000 Leuven, Belgium

3 Department of Aquatic Ecology, Netherlands Institute of Ecology (NIOO-KNAW), P.O. 50, 6700 AB Wageningen, The Netherlands

* corresponding author:

Tel.: +32 56 24 62 83

Fax: +32 56 24 69 99

Short title: Allelopathic control of phytoplankton by Elodea

Keywords: allelochemicals; chemical ecology; shallow lakes; nutrient limitation; competition

Summary

  1. It is well known that submerged macrophytes can suppress phytoplankton blooms in lakes and thus promote water quality and biodiversity. One of the possible mechanisms through which submerged macrophytes control phytoplankton is by producing allelochemicals that suppress phytoplankton growth rates. The in situ importance of allelopathy, however,is often questioned because it is assumed that phytoplankton communities can rapidly evolve resistance to allelochemicals.
  2. Here, we present the results of two mesocosm experiments in which we evaluated whether the submerged macrophyte Elodea nuttallii is capable of controlling phytoplankton biomass over periodsof 4 to 8 weeks. Such a time scale is long relative to the generation time of phytoplankton and is therefore expected to allow the development of resistance through compositional shifts at both population and community levels.
  3. Although the mesocosms were inoculated with a diverse phytoplankton inoculum including species that had previously been exposed to Elodea, phytoplankton biomass remained consistently low during the course of the experiments in the treatments with Elodea. As zooplankton grazing and competition for nutrients and light by macrophytes were excluded in our experiments, this suggests that phytoplankton was controlled by allelopathy.
  4. Dialysis bag assays, performed at the end of each mesocosm experiment, showed that phytoplankton communities from mesocosms with Elodea were equally sensitive to exudates from Elodea than phytoplankton communities from mesocosms without Elodea.
  5. These results suggestthat phytoplankton communities do not evolve resistance to allelochemicals from Elodea. This may allow Elodea to control phytoplankton in natural ecosystems over prolonged time-periods through allelopathy.

Introduction

Phytoplankton compete with submerged macrophytes for light and nutrientsin shallow lakes. This has received much attention because theoutcome of competition has important consequences for the ecological integrity of these ecosystems (Scheffer et al. 1993;Kostenet al. 2009). Macrophyte-dominated systemsgenerally tend to havegood water quality and high biodiversity, while phytoplankton-dominated systemsare more often associated with low biodiversity and poor water quality (Declercket al., 2011). Submerged macrophytes are superior competitors for nutrients because they have access to nutrient pools in the water column and sediment, whereas phytoplankton have only access to water-column nutrients (Sand-Jensen Borum, 1991). Submerged macrophytes, however, may require a period of clear water in spring to achieve a sufficiently high standing crop biomass to become a significant competitor for nutrients. Phytoplankton arein general the superior competitor for light, because theycan effectively block access of submerged macrophytes to light (Scheffer et al. 1993), although dense stands of submerged macrophytes may block light availability for phytoplankton. Submerged macrophytes will outcompete phytoplankton at low nutrient concentrations, while phytoplankton aremore likely to win incompetition at high nutrient concentrations.

Inferior competitors may influence the outcome of resource competition by producing allelochemicals that suppress the superior competitor. Allelopathy may allow submerged macrophytes to win incompetition from phytoplankton even when nutrient concentrations are relatively high (Hilt Gross 2008). Allelopathic interactions are notoriously difficult to study because of the difficulty of separatingallelopathic effects from other biological interactions (Harper 1975; Gross, Hilt, Lombardoet al. 2007). Nevertheless,several lines of evidence indicate that submerged macrophytes are capable of controlling phytoplankton through allelopathy. First, crude extracts or purified compounds obtained from different submerged macrophytes have been shown to display algicidal activity (e.g. Wium-Andersenet al.1982; DellaGrecaet al.2001; Erhard Gross 2006). Second, exudates or culture filtrates of submerged macrophytes have beenshown to suppress phytoplankton growth rates (e.g. KörnerNicklisch 2002; Mulderijet al. 2006; Chang, Eigeman & Hilt 2012), indicating that such compounds may actually be released in the surrounding water at concentrations that can affect phytoplankton productivity. Third, coexistence experiments between submerged macrophytes and phytoplankton and also field studies have reported inhibitory effects of submerged macrophytes on phytoplankton under conditions where alternative control mechanisms such as nutrient competition or zooplankton grazing could be excluded (MjeldeFaafeng 1997; Van Donk Van den Bund 2002; Vanderstukkenet al. 2011).

One of the major criticisms of allelopathy as a key factor determining the outcome of competition among primary producers in natural ecosystems is that the target biota are expected to evolve resistance to allelochemicals (Rabotnov 1974; Reigosa, Sanchez-Moreiras & Gonzales 1999; Fitter 2003). A number of studies in terrestrial systems have indeed shown that target plants can rapidly evolve resistance to allelochemicals from competing plants (Callawayet al. 2005; Jensen Ehlers 2010). Limited evidence suggests that resistance may also evolve in phytoplankton communities exposed to allelochemicals from macrophytes. Different phytoplankton species anddifferent strains ofthe same species display variation in resistance to allelochemicals from submerged macrophytes (Jasser 1995;KörnerNicklisch 2002; Al-Shehri 2010). Therefore, phytoplankton communities may be able to develop resistance to allelochemicals through compositional changes, both at the community level (species composition), as well as at the individual population level (genetic composition) through selection on genotypes. Given the short generation times and high dispersal rates of most phytoplankton strains (Finlay 2002), this may occur over a time-scale of weeks.

Allelopathic effects of submerged macrophytes on phytoplankton are typically studied in short-term experiments, often with a single phytoplankton species or strain. By design, such experimentspreclude the development of resistance to allelochemicals that might happen innatural ecosystems. The goal of this study was to test whether diverse phytoplankton communities are capable of adapting to allelopathic substances from submerged macrophytes when they are exposed to these allelochemicals for a prolonged period of time. Therefore, we carried out two coexistence experiments in cattle tank mesocosmsthatwere inoculated with a diverse phytoplankton assemblage andwere monitored for several weeks. We combined these mesocosm experiments with additional assays to test for the effect of allelopathy on phytoplankton growth and to exclude alternative mechanisms of phytoplankton suppression, such as nutrient or light limitation and grazing by zooplankton.

Methods

General experimental setup

We carried out two mesocosm experiments to study long-term allelopathic control of phytoplankton by Elodea nuttallii (henceforth referred to as Elodea). Elodea is a commonsubmerged macrophyte in Western Europe and is known to possessallelopathic activity towards phytoplankton (Erhard Gross 2006). The experiments were carried out at an experimental mesocosm terrain at KU Leuven in July – August 2007 (experiment I) and August –September 2008 (experiment II). The second experiment was carried out primarily to confirm the results ofthe first experiment, but included a few modifications to demonstrate allelopathic activity of Elodea towards phytoplankton in a more rigorous way.

The experimental mesocosms consisted of 200 L polyethylene tanks (height: 0.5 m; diameter: 0.77m) with a 4 cm layer of fine white sand on the bottom. The tanks were filled with distilled water enriched with nitrogen (2.4 mg N L-1as NaNO3) and phosphorus (0.3 mg P L-1 as K2HPO4). Micronutrients were added in concentrations as used in Wright’s Cryptophytephytoplankton culture medium (GuilardLorenzen 1972. To maximize the potential for the phytoplankton community to adapt to allelochemicals, each mesocosm was inoculated with a diverse phytoplankton inoculum originating from about 20 lakes, several of which contained dense stands ofElodea (cf. Declercket al. 2007).The inoculum was filtered througha 100 µm nylon mesh to remove large zooplankton such as Daphnia. The Elodea plants used in the experiments were collected from a nearby creek. They were thoroughly rinsed with tap water and treated withthe insecticide carbaryl (40 µg L-1) to remove zooplankton and macroinvertebrates. This treatment effectively removed all large cladocera and all macroinvertebrates except for snails (Lymnaeasp.). Therefore, snails were also added to the mesocosms without Elodea or with plastic Elodea. The snails reproduced rapidly in the mesocosms and controlled growth of periphyton and filamentous algae. After carbaryl treatment, the plants were incubated in tanks for at least two weeks during which the water was replaced several times to remove all traces of carbaryl. Each Elodea mesocosm treatment was inoculated with 800 g (fresh weight) plant biomass, resulting in a plant infested volume of about 35 %.Elodea plants were planted in the mesocosm two weeks prior to the start of the experiment to allow the plants to take root and for periphyton to colonize. All mesocosms were covered with 1 mm nylon mesh to avoid colonization by zooplankton or macroinvertebrates.

Experiment I

In experiment I, we compared mesocosm treatments with and withoutElodea. Each mesocosm treatment was replicated 6 times and the experiment was monitored for54 days. Chlorophyll a concentration was measured twice weekly.Elodea plants did not grow strongly during this experiment and pruning was not necessary to keep plant biomass constant. Elodea biomass at the end of the experiment was 5.5% lower than initial biomass.

Nutrient competition between Elodea and phytoplankton was avoided by biweekly additions of N and P to the mesocosms (50%of the initial concentrations were added each time, including micronutrients). At the end of the experiment, nutrient concentrations were measured and compared between the treatments with and without Elodea to evaluate whether Elodea reduced nutrient concentrations relative to the treatment without Elodea. In addition, we carried out nutrient limitation assays in each mesocosm to confirm whether or not phytoplankton communities in the mesocosms were nutrient-limited. Despite the fact that great care was taken to avoid introduction of zooplankton in the mesocosms, colonization of themesocosms by small copepods could not be avoided (they were introduced as naupliar stages with the phytoplankton inoculum). Therefore, we carried out zooplankton grazing assays in each mesocosm to estimate the impact of these small copepods on phytoplankton and torule out that differences in phytoplankton biomass between the mesocosm treatments would be due to herbivory.

At the end of the experiment, we compared phytoplankton species composition between the two treatments to evaluate whether Elodeahad induced species sorting in the phytoplankton community. We used dialysis bag assays (cf. KörnerNicklisch 2002) to investigate whether phytoplankton communities from Elodea mesocosms were better adapted to allelochemicals from Elodea than phytoplankton communities from mesocosms without Elodea. For these dialysis bag experiments, phytoplankton from each mesocosm was transferred to a dialysis bag and incubated in separate 65 L tanks filled with nutrient-enriched water with or without liveElodea plants, and thus with or without exudates of Elodea. After 3 days, chlorophyll a concentration and the maximum quantum yield of photosystem II (Fv/Fm) were compared between dialysis bags incubated in the presence or absence ofElodea exudates.

Experiment II

In experiment II, we orthogonally combined an Elodea treatment(plastic versus real Elodea) with a nutrient addition treatment (non-saturating versus saturating nutrient concentrations) to separate allelopathic control of phytoplanktonby Elodeafrom control by nutrient competition.In this experiment, the real Elodea plants grew strongly during the course of the experiment. The plants were pruned twice to keep biomass constant and both times about 50% of the initial biomass was removed. Final biomass was 8% higher than initial biomass. In the treatments with saturating nutrient concentrations, dissolved N and P concentrations were measured twice weekly and re-adjusted to initial levels. In the treatments with non-saturating nutrient concentrations, 50% of initial nutrient concentrations were added every two weeks, as in experiment I. Each mesocosm treatment was replicated 5 times and the experiment was monitored for 28 days. Chlorophyll a concentration was measured twice weekly asin experiment I.

Asin experiment I, zooplankton grazing assays were carried out in each mesocosm to ensure that differences between treatments were not due to zooplankton grazing. At the end of the experiment, dialysis bag assays were again carried out to test whether phytoplankton communities from mesocosms with Elodea were more resistant to allelochemicals from Elodea than those originating from mesocosms without Elodea. In contrast to experiment I, dialysis bags were incubated in GF/F filtered exudates from Elodea enriched with nutrients rather than in water with live Elodea to completely rule out nutrient competition.

Sampling and analyses

Chlorophyll a concentration was measuredthrough in vivo fluorescence using an AquaFluor fluorometer (Turner Designs, Sunnyvale, USA), calibrated before each experiment by means of spectrophotometric measurements of chlorophyll a extracted from GF/F filters. Samples for total and dissolved (GF/F filtered) nutrients were stored frozen until analysis. Samples for total nitrogen (TN) and total phosphorus (TP) were digested using alkaline persulphateprior to the analysis. Soluble reactive phosphorus (SRP; PO43-) and dissolved inorganic nitrogen (DIN: sum of NO3-, NO2- and NH4+) were measured with automated colourimetric methods using a microflow segmented flow analysis system (QuAAtro, Seal Analytical, Norderstedt, Germany) equipped with standard manifolds for the analysis of NH4+, NO2-+NO3- and PO43-. The sample for phytoplankton community analysis collected at the end of the experimentIwas fixed with 4% formalin and analyzed using inverted microscopy. Two hundred units (cells or colonies) were identified to genus level and enumerated.Phytoplankton biovolume was estimated from cell size measurements.

Nutrient limitation, zooplankton grazing and dialysis bag assays

To estimate zooplankton grazing in each mesocosm, we compared the increase in chlorophyll a concentration over 24 hoursin 600 mL polyethylene bag incubations with and without zooplankton. Zooplankton was removed by filtering water over a 64 µm nylon mesh. The polyethylene bags were incubated in duplicate just below the water surface in each mesocosm and were homogenized every 8 hours to avoid sedimentation of phytoplankton.Chlorophyll aconcentration was measured at the start and end of the incubationusing in vivo fluorescence.

To estimate nutrient limitation, we compared the increase in chlorophyll a concentration in similar incubations with or without the additionof saturating concentrations of N and P (650 µg P L-1 as K2HPO4 and 5.4 mg N L-1 as NaNO3) and micronutrients. The treatment with saturating nutrient concentrations also served as the control treatment for the zooplankton grazing experiment. Zooplankton was removed prior to the incubation to avoid interference by zooplankton grazing. The polyethylene bags were incubated and monitored just as those of the zooplankton grazing assays.

We used dialysis bag assays to estimate the sensitivity of phytoplankton communities from the mesocosms to exudates of Elodea. For each mesocosm, 64 µm filtered water was divided intotwo 100 ml dialysis bags (Wienie-Pak skinless sausage casings; ViskoTeepak, Lommel, Belgium) (KörnerNicklisch 2002). In experimentI, the dialysis bags were incubated in 65 L tanks with and without Elodea (35% plant infested volume). The tanks were filled with water enriched with high concentrations of N and P (1.5 mg P L-1 as K2HPO4 and 12 mg N L-1 as NaNO3) to avoidnutrient competition between phytoplankton in the dialysis bags and Elodea. In experimentII, the dialysis bags were incubated in 2 L aquaria containing GF/F filtered waterwith or without exudates ofElodeainstead of in water containing living Elodea plants. Water with exudates was obtained from mesocosm treatments with plastic or realElodea receiving saturating nutrient concentrations. The water was enriched with N, P and micronutrientsatthe same concentrations as in experiment Ito avoid nutrient limitation of phytoplankton in the dialysis bags. In both experiments, chlorophyll a concentration in the dialysis bags was measured at the start and after 3 days of incubation. The percentage increase in chlorophyll awas compared between incubations with or without exudates to evaluate sensitivity of the phytoplankton communities to allelochemicals from Elodea. We also measured the maximum quantum yield of photosystem II (Fv/Fm) as an indicator of phytoplankton stress response to allelochemicals of Elodea (cf. KörnerNicklisch 2002). Fv/Fm was calculated as (Fm-Fo)/Fm, with Fo being the minimum fluorescence in the dark (measured after 30 min dark acclimation at 20° C) and Fm the maximum fluorescence after a saturating light pulse. Fo and Fm were measured using a WalzPhytoPAMpulse amplitude modulated fluorometer (Walz, Effeltrich, Germany).

Statistical analyses

Differences in chlorophyll a and nutrient concentrations between the mesocosm treatments were evaluated using ANOVA. Analyses were done on time-weighted averages of chlorophyll a, total and dissolved nutrient concentrations (Stephen et al. 2004). The data were log-transformed to meet the ANOVA assumptions, except for the SRP data from experimentI, which were square root transformed.

Nutrient limitation or zooplankton grazing was evaluated by paired t-tests for each mesocosm treatment separately. The t-tests compared the percentage increase in chlorophyll a concentration between control treatments and treatments with nutrient or zooplankton addition. For the dialysis bag assays, a global analysis was done using repeated measures ANOVA in which we simultaneously evaluated the effect of the exudates (within subjects effect) and differences in this effect between phytoplankton communities originating from different mesocosm treatments (interaction between within subjects and between subjects effect). The percentage increase in chlorophyll a concentration and Fv/Fm were dependent variables.All univariate statistical analyses were performed with the software package STATISTICA 8.0 (StatSoft, Tulsa, USA).

We used redundancy analysis (RDA) to test fordifferences in phytoplankton community composition between treatments with and without Elodea in experimentI. Only species that had an abundancegreater than 10% of total phytoplankton community in at least one sample orgreater than 5% in at least five samples were included in the RDA analysis. The RDA analyses were performed on square root transformed data. The statistical significance of the RDA model (effects of treatment and their interactions) was evaluated with 999 random Monte Carlo permutations using CANOCO 4.5 (Microcomputer Power, Ithaca, USA)(ter BraakŠmilauer 2002).

Results

Experiment I

Chlorophyll a concentrations were significantly higher in the mesocosms without Elodea than in those withElodea (Fig. 1a, Table 1). Concentrations of DIN and TP did not differ between mesocosms with and without Elodea, while concentrations of SRP and TN were significantly lower in mesocosms with Elodea (Fig. 2a, d, Table 1). SRP concentrations were particularly low in the mesocosms withElodea (3 µg P l-1)(Fig. 2d). Nutrient limitation assays, however, showed no significant response of phytoplankton to nutrient addition in the mesocosms withElodea (Fig. 3a). This suggests that phytoplankton in mesocosms with Elodea was controlled by another factor than the low nutrient concentration. In contrast, phytoplankton communities in mesocosms without Elodea responded significantly to nutrient additions, indicating nutrient-limitation of phytoplankton in the absence of Elodea.