Present address: Laura Oberhaus, Paris-Universitas, Environmental Research and Teaching Institute (CERES-ERTI), Ecole Normale Supérieure, Paris, France.
Editor: Riks Laanbroek
Correspondence: Jean-François Humbert, Unité des Cyanobactéries, Institut Pasteur-CNRS URA2172, 28 rue du Dr Roux, 75724 Paris, Cedex 15, France. Tel.: +33 14 438 9316; fax: +331 4061 3042; e-mail: firstname.lastname@example.org
Blooms of freshwater cyanobacteria are typically accompanied by an important decrease in phytoplankton biodiversity in the water bodies where they occur. This study examines the potential production of growth-inhibiting substances by the toxic, bloom-forming cyanobacterium Planktothrix rubescens, following the observation of physical segregation between this and another cyanobacterium during previously performed mixed-culture competition experiments. Inhibition assays examining the growth of target strains exposed to donor culture filtrates showed that the growth of Planktothrix agardhii TCC 83-2, P. agardhii PMC 75.02 and Mougeotia gracillima TCC 50-2 was significantly inhibited in the presence of culture filtrate from P. rubescens TCC 29-1, isolated from Lake Bourget, France. Filtrates from P. rubescens TCC 69-6 and P. rubescens TCC 69-7, isolated from Lakes Nantua and Paladru (France), respectively, did not, however, inhibit the growth of P. agardhii TCC 83-2. This brief exploration of the allelopathic activity of P. rubescens suggests that it may potentially inhibit coexisting competitors as well as phytoplankton isolated from other freshwater ecosystems, and that this capacity may vary among different strains of Planktothrix. The potential importance of this phenomenon in pelagic competition dynamics is discussed.
Studies on competition between aquatic microorganisms have traditionally focused on the availability and partitioning of resources. Among freshwater phytoplankton, many of these studies have examined the role of abiotic conditions such as light, nutrients, and temperature, as well as biotic factors like adaptations to resist parasitism or predation (Reynolds, 1998). However, there are a limited number of papers dealing with the production of growth-inhibiting substances by members of a phytoplankton community, otherwise known as allelopathy, even though this process could provide a direct mechanism for allowing certain species to outcompete others.
Among the available literature on this subject, the production of allelopathic substances that inhibit the growth of other phytoplankton has been observed in dinoflagellates, diatoms, green algae, and cyanobacteria (Gross, 2003; Legrand et al., 2003; Schatz et al., 2005; Leflaive & Ten-Hage, 2007). While cyanobacteria tested for this ability have often occupied benthic or marine habitats, studies on pelagic freshwater cyanobacteria, such as Microcystis sp. (Sukenik et al., 2002), Microcystis aeruginosa (Ikawa et al., 1996), Anabaena flos-aquae (Kearns & Hunter, 2000), and Cylindrospermopsis raciborskii (Figueredo et al., 2007) have shown their capacity for producing substances that inhibit the growth of other phytoplankton. However, studies focusing on allelopathic inhibition by bloom-forming cyanobacteria remain few in number, despite the fact that these microorganisms proliferate in numerous freshwater bodies and that their blooms are typically accompanied by a strong decrease in phytoplanktonic biodiversity.
Given this context, the aim of this work was to determine if allelopathic interactions may occur between Planktothrix strains and between Planktothrix and a potential competitor species, M. gracillima. With this goal in mind, we tested the effects of adding the filtered culture media of five Planktothrix strains, isolated from several ecosystems, on the cellular growth of each of these strains and also on that of an M. gracillima strain.
Materials and methods
The general characteristics of all phytoplankton strains used in this study are summarized in Table 1. Continuous precultures of P. rubescens TCC 29-1 were maintained using a 2-L turbidostat device (Feuillade & Feuillade, 1979), as were those of P. agardhii (TCC 83-2) isolated from Lake Nantua (France). Precultures of P. agardhii PMC 75.02, isolated from the Viry-Châtillon recreational basin near Paris, of P. rubescens TCC 69-6, isolated from Lake Nantua, and of P. rubescens TCC 69-7, isolated from Lake Paladru (France), were raised in batch, as was a second culture of P. rubescens TCC 29-1. Z culture medium (Zehnder, in Staub, 1961) was used for all Planktothrix strains. Mougeotia gracillima TCC 50-2 was maintained in a batch culture using a modified Z medium with silica and vitamins added (ingredients for Z medium, plus Na2SiO3·5H2O at a total concentration of 0.058 g L−1; vitamin B12, 1 × 10−6 g L−1; biotin, 1 × 10−6 g L−1; and thiamin, 2 × 10−4 g L−1), with a final adjusted pH of 7.4. All precultures were nonaxenic and were grown under nonlimiting nutrient conditions, at 20 °C and on a 16-h light : 8-h dark cycle. Batch cultures were near the end of the exponential growth phase at the time of filtering and inoculation into experimental vessels, in order to guarantee sufficient cell density for observing allelopathic effects and to avoid effects of nutrient limitation.
Table 1. Strains used in the growth inhibition assays, their isolation origins, their known capacity to produce microcystins, and the preculture methods and culture media used
Potential production of microcystins
P. rubescens TCC 29-1
P. agardhii TCC 83-2
P. agardhii PMC 75.02
Viry-Châtillon Rec. Basin
P. rubescens TCC 69-6
P. rubescens TCC 69-7
M. gracillima TCC 50-2
Several of the Planktothrix strains tested have been confirmed to produce microcystins: [D-Asp3]- and [D-Asp3, Dhb7]-microcystin-RR, [D-Asp3]-microcystin-LR, and microcystin-YR in the case of P. rubescens TCC 29-1 (Briand et al., 2005), two unidentified variants of microcystin in that of P. agardhii TCC 83-2, and microcystin-YR and –dMe-LR in the case of P. agardhii PMC 75.02 (C. Bernard, pers. commun.).
Donor precultures were filtered through a 2 μm, then through a 1 μm, Millipore filter. In an experiment exploring the possible nature of allelopathic substances, filtrates were allowed to boil gently for 5 min, then to completely cool before their use. Growth inhibition assays were performed on monocultures in 24-well microplates, with three replicates for each treatment, and reciprocal effects were tested in each assay. Wells contained a total volume of 1650 μL, which consisted of 660 μL of donor filtrate (or sterilized Milli-Q water for controls) and 660 μL of the target culture. One-fifth of the total volume of each well (330 μL) consisted of a 5 × concentrated culture medium, to prevent inhibition of growth by nutrient limitation. Calibration curves based on microscope enumeration determined initial culture concentrations to be around 1 200 000 cells mL−1. Microplates wrapped with parafilm were incubated at 20 °C, with an overhead fluorescent white light (Oshram Lumilux de luxe, ‘daylight’) level of 18–30 μmol photons m−2 s−1 on a 16-h:8-h light : dark cycle. This temperature was chosen as an intermediate between those found to favor P. rubescens (15 °C) and P. agardhii (25 °C) in mixed-culture competition experiments (Oberhaus et al., 2007). Light conditions correspond to below-saturation levels, in order to avoid any possible effects of photoinhibition, and also to those used to maintain precultures. Growth was measured by fluorescence, using a Fluoroskan Ascent microplate reader, with excitation and emission set, respectively, at 550 and 650 nm, every 2–3 days for c. 30 days. F and permutation t-tests were performed to highlight significant differences between filtrate treatments and control cultures having received no filtrate (past software).
Results and discussion
Allelopathic inhibition of other phytoplankton by P. rubescens
Experiments testing growth inhibition between the two strains of Planktothrix previously observed to physically segregate in mixed-culture experiments (Fig. 1) showed that culture filtrate from P. rubescens TCC 29-1 strongly inhibited the growth of P. agardhii TCC 83-2 as early as the third day of this assay, and near-complete inhibition of P. agardhii biomass was observed by experiment end for all replicates (Fig. 2a). Further examination of this phenomenon showed that P. rubescens TCC 29-1 filtrate also inhibited the growth of P. agardhii PMC 75.02 (Fig. 2b); in this case, inhibition of growth was evident 6 days after experiment onset. Differences between P. agardhii cultures receiving filtrate from P. rubescens TCC 29-1 and those receiving no filtrate were significant over the entire length of the experiment in both cases (P=0). However, no reciprocal effects were observed, as growth of P. rubescens appeared to be uninhibited in the presence of filtrates from both P. agardhii TCC 83-2 and P. agardhii PMC 75.02 (P=0.736 and 0.272, respectively; data not shown).
Inhibition potential of P. rubescens TCC 69-6, isolated from Lake Nantua, and P. rubescens TCC 69-7, isolated from Lake Paladru, was also tested by growth inhibition assay, using only P. agardhii TCC 83-2 as a target strain. Filtrate from P. rubescens TCC 69-6 did not appear to inhibit the growth of P. agardhii (P=0.363; Fig 2c), nor vice versa (P=0.847, data not shown); this was also the case when P. rubescens TCC 69-7 served as a donor (P=0.783, Fig. 2c) and as a target (P=0.947, data not shown). Differences in batch and turbidostat culture methods were first considered as a possible cause for the noninhibition of P. agardhii by these two strains of P. rubescens. However, verification showed that P. rubescens TCC 29-1, when raised under the same batch culture conditions as these two strains, once again inhibited the growth of P. agardhii TCC 83-2 (P=0.001, data not shown). Among the three P. rubescens strains tested, culture filtrate from P. rubescens TCC 29-1 from Lake Bourget was thus the only one observed to inhibit the growth of P. agardhii TCC 83-2.
The potential inhibitory effect of P. rubescens TCC 29-1 culture filtrate was next tested on M. gracillima TCC 50-2, likewise isolated from Lake Bourget. Growth of M. gracillima was generally slow, and it only showed inhibition starting on the eighth day after exposure to the filtrate of P. rubescens TCC 29-1. Due to delayed inhibition, differences between cultures exposed to P. rubescens TCC 29-1 filtrate and controls were significant from the eighth day to experiment end (P=0.024), but not when examined over the entire duration of the experiment (P=0.109). By experiment end, the replicate-average biomass of M. gracillima was inhibited by a maximum of 57% relative to the control (Fig. 2d). Inhibitory effects upon M. gracillima TCC 50-2 may thus be evidence of a mechanism for P. rubescens TCC 29-1 to inhibit the growth of an important local competitor.
These results seem to contradict the hypothesis of Reigosa et al. (1999), which suggests that coexisting competitors are most likely to have evolved resistance to inhibitory allelochemicals, while strains from other ecosystems should show greater sensitivity. On the other hand, the examination of allelopathic interactions between P. rubescens TCC 69-6 and P. agardhii TCC 83-2, two coexisting cyanobacteria isolated from Lake Nantua, showed no growth inhibition for either strain. This case does in fact seem to agree with the hypothesis of Reigosa et al. (1999) and may be indicative of adaptive resistance by P. agardhii TCC 83-2 to specific allelochemicals produced by P. rubescens TCC 69-6, rather than the nonproduction of such substances. It is however also possible that our assay simply did not detect the more subtle effects of allelochemicals produced by this strain of P. rubescens (Gross et al., 2007). In any case, the strong inhibition capacity exhibited by P. rubescens TCC 29-1 was not observed with P. rubescens strains TCC 69-6 and 69-7.
Our results thus suggest that allelopathic inhibition may occur among sympatric and thus potential local competitors, as with P. rubescens TCC 29-1 and M. gracillima TCC 50-2, as well as between those originating from different ecosystems, as shown by the inhibition of P. agardhii TCC 83-2 (isolated from Lake Nantua) and P. agardhii PMC 75.02 (isolated from the Viry-Châtillon recreational basin) by P. rubescens TCC 29-1. These results further suggest that (1) the inhibitory substance produced by P. rubescens TCC 29-1 may be wide acting, and (2) production of allelopathic inhibitory substances, and perhaps sensitivity to them, could vary on an intraspecific level.
Preliminary information on the inhibitory substance
Due to the confirmed toxicity of P. rubescens TCC 29-1, the question arose as to whether the substance responsible for its inhibition of P. agardhii and M. gracillima could in fact be the microcystins that it produces. Boiling of the P. rubescens TCC 29-1 filtrate before repeating the inhibition assay with P. agardhii TCC 83-2 eliminated its inhibitory effects (P=0.224), suggesting that the active substance can be denatured by this process (Fig. 2a). Microcystins can remain potent even after boiling (Sivonen & Jones, 1999); thus the growth inhibition caused by the P. rubescens TCC 29-1 filtrate does not seem to be caused by these cyanotoxins. Planktothrix rubescens TCC 29-1 and the two P. agardhii cultures used have been confirmed to produce microcystins, also suggesting action by another type of substance, produced solely by P. rubescens TCC 29-1. This is in agreement with other studies showing allelopathic inhibition by compounds other than known cyanotoxins (von Elert & Jüttner, 1997; Legrand et al., 2003; Suikkanen et al., 2004; Schatz et al., 2005; Gantar et al., 2008).
Prior studies have established the existence of bacteria capable of lysing cyanobacteria cells (Rashidan & Bird, 2001; Nakamura et al., 2002) and producing substances inhibitory to cyanobacterial growth (Yoshikowa et al., 2000). Given our unavoidable use of nonaxenic cultures, it was necessary to take into account the presence of such bacteria as a possible cause of inhibition. The bacterial communities of the two Planktothrix turbidostat cultures (P. rubescens TCC 29-1 and P. agardhii TCC 83-2) were compared using PCR-denaturing gradient gel electrophoresis (DGGE), following the method of Dorigo et al. (2006). This confirmed the two communities to be of identical composition, thus probably ruling out the possibility of the presence of lytic bacteria species only in the P. rubescens TCC 29-1 filtrate.
These results strongly suggest that the substance responsible for growth inhibition by P. rubescens TCC 29-1 is indeed an allelochemical other than microcystin. As an important next step, future studies should include isolation and identification of this substance. This information could then allow the precise application of known concentrations of the substance in laboratory experiments, and might pave the way for its isolation from in situ study sites.
Outcomes in the natural environment
The connection between allelopathic interactions observed among phytoplankton in laboratory experiments and those that may exist in natural environments is assumed but has not yet been securely established, because allelopathic substances other than cyanotoxins have not yet been isolated from the natural environment. Moreover, the distance between the cells of donor and target organisms represents a major hindrance to in situ study in pelagic environments especially, where the diffusion and dilution of allelochemicals is a capital issue. Hulot & Huisman (2004) developed a model examining the theoretical fate of phytoplankton allelochemicals in a natural pelagic environment, subjected to degradation by heterotrophic bacteria and vertical mixing. It was found that mixing and the presence of associated bacteria could be important factors in determining the influence of these substances, the first in its ability to carry away or further dilute the substances, or to displace bacteria, and the second in their potential ability to degrade allelopathic compounds, thus influencing arrival of these substances to target strains. The influence of hydrodynamics on this phenomenon is an interesting question with regard to P. rubescens, which has been shown to proliferate in deep lakes with stable metalimnia (Sas, 1989; Jacquet et al., 2005). The role of associated heterotrophic bacteria in allelopathic interactions and the influence of abiotic factors and physiological states on the production of allelochemicals also remain interesting, relatively unexamined, questions.
Our brief exploration of the allelopathic inhibitory potential of a P. rubescens strain found to form blooms in Lake Bourget has shown that freshwater, bloom-forming cyanobacteria may potentially produce allelochemicals, probably other than microcystins, that can inhibit the growth of local competitors or even strains isolated from other ecosystems. Moreover, results obtained under the conditions used suggest that different strains of the same species may exhibit diversity in their production of these compounds, making consideration of this phenomenon possible in inter- as well as intraspecific competition dynamics. Previous results from competition experiments between P. rubescens TCC 29-1 from Lake Bourget and P. agardhii TCC 83-2 from Lake Nantua (Oberhaus et al., 2007) showed differing dominance of these two strains that could be related to light and temperature conditions, with P. agardhii often dominating at a higher temperature. Although removing allelopathy in the study of resource competition may be impossible (Inderjit & Del Moral, 1997), these competition results suggest that the production of allelochemicals by P. rubescens that inhibit P. agardhii's growth may not be sufficient to assure its dominance under abiotic conditions that otherwise favor this chemically disadvantaged competitor. Given this observation, it could be that allelopathy has little effect on competition dynamics at the onset of a cyanobacterial bloom. The ability to produce allelopathic substances could however have a greater influence on bloom maintenance, when biomasses are higher and allelochemical concentrations greater, thus influencing competition with other phytoplankton as well as that among different strains of the bloom-forming species.
The authors thank B. Le Berre, E. Menthon, and Dr C. Leboulanger for their help and technical assistance in the laboratory. The authors would also like to thank Dr C. Bernard (MNHN Paris, France) for kindly providing the PMC 75.02 strain, and the anonymous reviewers whose comments contributed to the improvement of the original manuscript. L.O. was funded by a grant from the Ministère de l'Education Nationale, de l'Enseignement Supérieur et de la Recherche. J.-F.B. was funded by the Ministère de l'Ecologie et du Développement Durable (SACYTOX project – RITEAU program).