The cyanobacteria Planktothrix agardhii and Cylindrospermopsis raciborskii are bloom-forming species common in eutrophic freshwaters. These filamentous species share certain physiological traits which imply that they might flourish under similar environmental conditions. We compared the distribution of the two species in a large database (940 samples) covering different climatic regions and the Northern and Southern hemispheres, and carried out laboratory experiments to compare their morphological and physiological responses. The environmental ranges of the two species overlapped with respect to temperature, light and total phosphorus (TP); however, they responded differently to environmental gradients; C. raciborskii biovolume changed gradually while P. agardhii shifted sharply from being highly dominated to a rare component of the phytoplankton. As expected, P. agardhii dominates the phytoplankton with high TP and low light availability conditions. Contrary to predictions, C. raciborskii succeeded in all climates and at temperatures as low as 11 °C. Cylindrospermopsis raciborskii had higher phenotypic plasticity than P. agardhii in terms of pigments, individual size and growth rates. We conclude that the phenotypic plasticity of C. raciborskii could explain its ongoing expansion to temperate latitudes and suggest its future predominance under predicted climate-change scenarios.
The excessive growth of planktonic cyanobacteria is among the main threats endangering the use of water resources in shallow lakes. Temperature increases in the range of 0.2 °C per decade, and their effects on water mixing regimes, are expected to increase the occurrence, frequency and duration of cyanobacterial blooms in several regions of the planet (Doney, 2006; Falkowski & Oliver, 2007; Markensten et al., 2010). These future changes in climate are also predicted to cause shifts in the species composition of cyanobacterial blooms in favour of invasive species (Mehnert et al., 2010). The modern global distributions and environmental preferences of cyanobacterial species result from differences in evolutionary adaptations and phenotypic traits (Whitton & Potts, 2000). Understanding the characteristics that allow cyanobacterial taxa to succeed in disparate environments is crucial for predicting future bloom-forming behaviour in warming climates.
Filamentous cyanobacteria such as Planktothrix and Cylindrospermopsis, as well as colonial genera at Microcystis, are the most successful bloom-forming organisms in shallow lakes (Padisák & Reynolds, 1998; Nixdorf et al., 2003; Paerl et al., 2011; Tomioka et al., 2011). In particular, Planktothrix agardhii (Order Oscillatoriales) and Cylindrospermopsis raciborskii (Order Nostocales) can be used as model species because of the extensive information available about their distributions. Planktothrix agardhii is a resilient, shade-tolerant species that can produce microcystins and is one of the most common bloom-forming species in temperate lakes (Scheffer et al., 1997). Blooms of C. raciborskii are becoming more frequent in tropical (Figueredo & Giani, 2009; Gemelgo et al., 2009), subtropical (Vidal & Kruk, 2008; Everson et al., 2011) and temperate lakes (Hamilton et al., 2005; Stüken et al., 2006) because of the apparently invasive behaviour of the species (Padisák, 1997). The expansion of C. raciborskii has generated widespread concern as a result of its potential for producing two toxin types, cylindrospermopsins and saxitoxins (Chorus & Bartram, 1999). To date no consensus exists regarding the main mechanisms that have permitted the expansion of C. raciborskii into temperate regions. Proposed hypotheses include climate change–associated water temperature increases (Wiedner et al., 2007), an exceptionally good tolerance of transport (Padisák, 1997), the ecophysiological plasticity of the species (Briand et al., 2004) and the existence of ecotypes with different environmental preferences and tolerances (Chonudomkul et al., 2004; Piccini et al., 2011).
Cylindrospermopsis raciborskii and P. agardhii have similar phenotypic traits, including tolerance to continuous mixing of the water column, high phosphorus storage capacity, buoyancy regulation and shade tolerance (Reynolds, 1993; Padisák & Reynolds, 1998; Istvánovics et al., 2000; Padisák, 2003). These similarities are also reflected in their morphology, indicating that they may be functionally equivalent and occupy a similar ecological niche (Kruk et al., 2010). However, some studies show that C. raciborskii has higher light requirements for growth (Ik) than P. agardhii, suggesting differences in some dimensions of their niches (Briand et al., 2004; Kokociński et al., 2010; Mehnert et al., 2010). Moreover, these two species differ in their capacities to incorporate nitrogen. Cylindrospermopsis raciborskii has the capacity to fix atmospheric nitrogen (N2) through heterocytes, as do other Nostocales, conferring a competitive advantage in nitrogen-depleted environments relative to P. agardhii, which cannot fix nitrogen (Whitton & Potts, 2000). The advantages that explain the recent worldwide expansion of Cylindrospermopsis, combined with predicted changes because of global warming, may imply an impending shift from P. agardhii to C. raciborskii at intermediate latitudes.
Although information about P. agardhii is extensive, and data about C. raciborskii are increasingly available, current knowledge derives either from experiments or field sampling. Very few studies simultaneously compare both species (Dokulil & Teubner, 2000; Wiedner et al., 2007; Kokociński et al., 2010). Functional traits, including morphological and physiological features, govern individual ecological performance and summarize organism responses to the environment (McGill et al., 2006; Violle et al., 2007; Kruk et al., 2010). A comparative approach to studying the morphological and physiological traits and distributions of P. agardhii and C. raciborskii can provide insight into the behaviour of these key cyanobacterial species. Moreover, this approach can contribute to more general ecological questions such as microorganism invasions (McGill et al., 2006; MacDougall et al., 2009).
Our aim was to evaluate the global distribution and ecological preferences of C. raciborskii and P. agardhii, and to determine the implications for the geographical expansion of C. raciborskii. We assembled a large database spanning wide latitudinal gradients and different climatic regions and carried out laboratory experiments to characterize the morphological and physiological traits of the two species.
Materials and methods
We constructed a database of 940 samples taken from 28 mesotrophic to hypereutrophic lakes where P. agardhii and/or C. raciborskii were present in at least one sample. In 125 samples neither species was present. Species data were obtained from published (Padisák, 1994; Aubriot et al., 2000, 2011; Kruk et al., 2002, 2010; Marinho & Huszar, 2002; Soares et al., 2009) and unpublished material (kindly provided by F. Bressan, A. Ferreira and S. de Melo). Three climate regions were represented in the lake database: tropical (08°02′–22°33′S), subtropical (34°33′–34°55′S) and temperate (35°30′–38°80′S and 46°50′–52°23′N). Studied lakes in the temperate zone were from Hungary (Balaton Lake), The Netherlands (Deest and Ochten floodplain lakes: D1, D2, D3, D4, D5, O2, O3, O4, O5, O6) and the Argentinean Pampas lakes AR19, AR20, AR29, AR30, AR31 and AR32; in the subtropical zone from Uruguay (Laguna Blanca, Canteras, Chica, Javier, Rodó, Sauce and Ton-Ton); and in the tropical zone from Brazil (Funil, Imboassica, Juturnaíba, Tabocas and Tapacurá). All samples were used to determine the distribution of each species in relation to selected environmental variables, excluding observations with zero biovolume. The environmental variables were lake area (area, ha), maximum depth (Zmax, m), mixing depth (Zmix, m), euphotic/mixing depth ratio (Zeu/Zmix), water temperature (T, °C), pH, conductivity (K, μS cm−1), alkalinity (Alk, mg CaCO3 L−1) and total phosphorus (TP, μg L−1). The Zeu/Zmix ratio was used as a proxy of the light available in the environment for phytoplankton growth (Jensen et al., 1994). The largest number of observations (~ 61%) was from shallow lakes (Zmax < 4 m), although several cases corresponded to deep lakes (93 data points, Zmax > 20 m, maximum: Funil Reservoir, Zmax: 50 m). A wide range of lake areas were included (0.5–7200 ha, plus Balaton Lake which is 59300 ha) but only 19 lakes were smaller than 100 ha. Our data set had an extensive range of TP (12–1653 μg L−1) where few observations (n = 5) indicated mesotrophic status (< 30 μg L−1 TP) and 100 were from hypereutrophic conditions (> 200 μg L−1 TP).
Despite the diversity of lakes and locations, phytoplankton sampling always followed routine protocols, and thus, the samples were representative of lake conditions. Samples were obtained at different depths within the mixed, illuminated zone, or in permanently mixed shallow lakes, the whole water column was integrated using sample bottles or tubes. Phytoplankton samples were fixed with Lugol's solution and settled in counting chambers (Utermöhl, 1958); at least 100 individuals of the most frequent species or 400 individuals in total were counted in random fields in an inverted microscope as described in (Kruk et al., 2010). Individual volume (V, μm3) was calculated for each taxon according to simple volumetric formula, considering the organism as the unit, and biovolume was expressed as mm3 L−1. The surface are volume ratio (S/V, μm−1) and filament maximum linear dimension (MLD, μm) were estimated as detailed in Kruk et al. (2010). The community was analysed in terms of species richness (expressed as the number of taxa per sample, S), the absolute biovolume of P. agardhii and C. raciborskii, the relative contribution of each to total biovolume, and their frequency of occurrence (number of observations). Total biovolume was considered low when < 1 mm3 L−1, and a species was considered dominant when it represented at least 30% of the total biovolume in a particular sample. The frequency of occurrence, the median and the range of the two species in terms of biovolume were analysed with all samples in the data set (including zero data).
The physiological and morphological responses of the two species were compared using Uruguayan isolates: P. agardhii (MVCC11) and C. raciborskii (MVCC14). Isolate MVCC11 was collected from Lago Rodó (34°55′S, 56°10′W), a eutrophic to hypereutrophic shallow lake used for recreation (Area: 1.5 ha, Zmax: 2.5 m, TP: 70–565 μg L−1) (Scasso et al., 2001). Isolate MVCC14 was collected in Laguna Blanca (34°53′ S, 54°20′ W), a eutrophic shallow lagoon (Area: 40.5 ha, Zmax: 2.6 m, TP: 86 μg L−1) used as a drinking water supply (Vidal & Kruk, 2008). Static cultures of the isolates were kept in BG11 medium at 26 °C (±1 °C), as described in the study carried out by Piccini et al. (2011), which is the normal summer water temperature in Uruguayan lakes where the species were isolated (Vidal & Kruk, 2008; Aubriot et al., 2011).
Two set of experiments were performed: light intensity gradient and temperature experiments, and the growth rates (physiological trait) of the two species were determined and compared. In addition, for the light intensity gradient experiments, we evaluated the physiological response of pigment structure change and the morphological trait changes of V, S/V and MLD.
To determine the effect of light intensity on growth rates of P. agardhii, four growth curves under six light intensity levels (from 5 to 180 μmol photons m−2 s−1) were repeated in 5-day experiments at 26 °C (±1 °C). Data for growth curves of C. raciborskii were obtained from Piccini et al. (2011), who performed the experiment under the same conditions. Before beginning the experiments, cultures were acclimated to each light level for 10–15 days and replicated when the biomass was duplicated (three replicates, except for 5 μmol photons m−2 s−1: one). The experiments were run in 100-ml bottles, filled with 80 mL BG11 medium and inoculated with cyanobacterial culture in exponential growth phase. Optical density (OD, absorbance at 750 nm) was used as an indicator of biomass and the initial inoculum for all experiments was 0.1 absorbance units. Absorbance at 440 nm was used to determine the light extinction coefficient (Kirk, 1996) in order to calculate the light intensity inside the bottles. Absorbance measurements were taken in a spectrophotometer (Thermo Evolution 60). The growth rate (μ, d−1) of each isolate was calculated in 24-h intervals during the exponential phase as:
where ODi and ODf are the estimated biomasses at initial (ti) and final (tf) times, respectively. Maximum specific growth rate, μmax, the initial slope, α, and the irradiance at the onset of light saturation, Ik (Ik = μmax/α), were derived from the fitted model of Jassby & Platt, (1976) for photosynthesis.
Samples were taken at the end of two growth experiments (i.e. 20 and 100 μmol photons m−2 s−1) in order to quantify several characteristics of the two taxa. To compare morphological changes, V, S/V and MLD were calculated for each isolate and replicate, based on microscopic measurements of 60 organisms made under an Olympus BX40 optical microscope at 1000× magnification. To compare changes in relative pigment concentration, in vivo relative concentrations of phycocyanin and Chl a fluorescence were measured in a fluorometer (Turner, Aquafluor), with phycocyanin relative concentration standardized against Chl a in vivo fluorescence. Finally, the pigment structure of each taxon was characterized using high-performance liquid chromatography (HPLC). Samples from one replicate of each taxon were filtered onto GF/C glass-fibre filters and kept frozen (−80 °C) until pigment extraction. HPLC methods and protocols followed those described in Bonilla et al. (2005). Carotenoids were detected by diode-array spectroscopy (350–750 nm), chromatograms were obtained at 450 nm (for carotenoids), and Chl a was detected by a fluorescence detector (excitation λ = 440 nm; emission λ = 650 nm). The identification and quantification of the pigments (Chl a, aphanizophyll, β,β-carotene, echinenone and zeaxanthin) was based on commercial standards as detailed in Bonilla et al. (2005). Unknown carotenoids were quantified by applying the calibration curves used for β,β-carotene. The final concentration of each pigment is expressed in nmol L−1, and changes in carotenoids were analysed using ratios to Chl a.
In order to determine the influence of low temperature on growth rate, 4-day experiments were run for both isolates at three temperatures: 15, 20 and 25 °C (±1 °C) at both 60 and 135 μmol photons m−2 s−1, with four replicates for each condition. The experimental setup and growth conditions were the same as described earlier, with cultures allowed to acclimate to each temperature for 15 days (three times, except at 15 °C: once). Biomass and growth rate were calculated as previously described for light intensity experiments. The parameter Q10 (15–25 °C) for each light intensity (60 and 135 μmol photons m−2 s−1) was calculated using the maximum average growth rate (n = 4) obtained at each temperature for each species.
The annual coefficient of variation was calculated to determine the variability of the biomass of the two species in nature, based on temporal data series for 17 lakes in the database, for temperate (11 lakes), subtropical (1) and tropical (5) regions. As the objective of this particular analysis was to determine the amplitude of biomass change, all data, including observations with zero values, were used.
To evaluate the success of the two species in relation to key environmental variables, we examined the maximum relative contribution of each species to total biovolume distribution along gradients of temperature, Zeu/Zmix and TP. For these analyses, data were segregated into groups every one degree Celsius, 0.1 Zeu/Zmix unit, and 10 μg L−1 TP. We then performed linear and nonlinear regressions between species biovolume and each environmental variable. The simple functions with best fit were selected following parsimony criteria of maximum explained variance with the minimum number of parameters and best significance (F-test). Linear relationships with breakpoints, such as those suggested for Planktothrix biovolume to Zeu/Zmix and to TP plots, can indicate ecological thresholds (Toms & Lesperance, 2003). We therefore applied a simple piecewise linear regression to these data and breakpoints were determined after 200 iterations. Data were tested for normality and homogeneity of variance prior to analyses and log10 transformed when necessary (P. agardhii biovolume distribution on Zeu/Zmix gradient).
The three climatic regions were compared by examining data from winter and summer months for each lake and year in the dataset (n = 445) in terms of temperature, Zeu/Zmix, TP, and the biovolume of the two species. Differences between environmental and biotic variables were analysed with nonparametric Kruskal–Wallis (K–W) tests, all pairwise multiple comparison tests (Dunn's Method) and Mann–Whitney tests (when P. agardhii was present in only two regions).
Differences between physiological and morphological experimental responses of the two species to light and temperature gradients were compared using t-test analysis and, when normality failed, with the nonparametric Mann–Whitney (M–W) test. All analyses were performed with the programs statistica 6.0 and sigma plot 11.0.
Species distributions and their relation to environmental factors
Cylindrospermopsis raciborskii was observed in a higher number of samples than P. agardhii (306 and 199 samples, respectively), with the two species co-occurring on 34 occasions (all in Lake Balaton) in a wide range of environmental conditions. We analysed lakes where one of the two species was present on at least one sample. Cylindrospermopsis raciborskii was absent in all samples of subtropical (Lago Rodó) and small Dutch temperate lakes, while P. agardhii was absent in most of the subtropical Uruguayan lakes (except Lago Rodó) and all tropical Brazilian lakes. Each species reached high biovolume and had a high contribution to total biovolume in several samples (Table 1, Fig. 1a and b). Cylindrospermopsis raciborskii was dominant (at least 30% of total biovolume) more frequently than P. agardhii, in samples of both high (> 1 mm3 L−1) and low (< 1 mm3 L−1) total biovolume. In most cases, P. agardhii was a minor component of the phytoplankton, representing < 10% of total biovolume, but in several cases, it was strongly dominant (> 50% of total biovolume) and its maximum absolute biovolume was one order of magnitude higher than C. raciborskii (Table 1). Biovolume variability over time also differed between species. Temporal data series of 17 lakes showed that the annual variation of P. agardhii biovolume was significantly higher (P < 0.05) than C. raciborskii, shifting from low to high values (Fig. 1c and d). Phytoplankton species richness also differed when the dominance of one or the other species occurred; in general, the number of species was higher when C. raciborskii was dominant (Fig. 2a and b).
Table 1. Community and environment characteristics (median, minimum–maximum between brackets cursive numbers indicate the number of cases) for Planktothrix agardhii and Cylindrospermopsis raciborskii, when present
Lakes with P. agardhii
Lakes with C. raciborskii
Significant differences (Mann Whitney,
P < 0.05) between the measured variables for each species are indicated.
ns, not significant; BV, biovolume; S, species number; Zmax, maximum depth; Zmix, mixing zone; Zeu, euphotic zone; K, conductivity; Alk, alkalinity; TP, total phosphorus.
In the field, C. raciborskii and P. agardhii occurrences differed significantly relative to temperature, lake area, maximum depth, mixing depth, conductivity, alkalinity and TP (Table 1). However, no significant differences were found for pH and light availability (Zeu/Zmix) (Table 1). In general, C. raciborskii was dominant at higher temperatures than P. agardhii. Several occurrences of P. agardhii were reported at temperatures below 15 °C, and below 4 °C its contribution varied between 0.2% and 13% of total biovolume (0.1–1.4 mm3 L−1). Almost all data for C. raciborskii appeared at temperatures higher than 20 °C. However, it is notable that we observed C. raciborskii with high biovolume at 11 °C (2.1 mm3 L−1, 95% of total biovolume) in a subtropical lake (Lago Javier, Uruguay). Both species were dominant in eutrophic to hypereutrophic lakes. Cylindrospermopsis raciborskii attained higher biomass under lower TP, and no occurrences of P. agardhii were found in samples with < 50 μg L−1 TP (Table 1).
Planktothrix agardhii biovolume shifted sharply from high to low values across thresholds in the temperature, Zeu/Zmix and TP gradients (as identified by parameter c in the logistic function and breakpoints in piecewise linear regressions in Fig. 3a, c and e). Planktothrix agardhii biovolume decreased abruptly below 11 °C, above 1.62 Zeu/Zmix and above 159 μg L−1 TP. Cylindrospermopsis raciborskii biovolume was inversely related to TP, and no significant relation was found with temperature or Zeu/Zmix (Fig. 3b, d and f). However, maximum C. raciborskii biovolume was observed with Zeu/Zmix values ≤ 1. Also, this species had higher biovolumes than P. agardhii in fully illuminated water columns (Zeu/Zmix: 3–4).
Comparison of the species among climates: temperate, subtropical and tropical
There were significant differences (P < 0.05) between geographical regions in terms of temperature, light and phosphorus. Water temperature was higher in the tropics, more transparent waters were found in tropical systems, and higher trophic states (TP) were found in subtropical lakes (Table 2). Planktothrix agardhii occurred only in temperate and subtropical water bodies, where it had a significantly higher average contribution to total biovolume than C. raciborskii. Cylindrospermopsis raciborskii occurred in the three regions and, notably, had no significant differences in its contribution to total biovolume between tropical and temperate regions (Fig. 4).
Table 2. Water temperature, light availability (Zeu/Zmix) and total phosphorus (TP) of studied lakes grouped by regions (temperate, subtropical and tropical) and based on winter and summer data (median and minimum and maximum between brackets and number of samples in cursive)
Temperate (35°30′–38°80′S, 46°50′–52°23′N)
Significant differences (Mann–Whitney, P < 0.05) between regions are indicated with different letters in the table.
ns, not significant; Zeu, euphotic zone; Zmix, mixing zone.
19.1 (0.50–27.6)a 207
22.8 (10.0–26.4)b 40
26.9 (20.0–31.6)c 107
0.58 (0.17–4.15)a 257
0.84 (0.17–2.7)a 75
0.64 (0.09–3.0)b 56
TP (μg L−1)
105 (50–1652)a 210
158 (46–422)b 56
91.7 (12.4–794)a 4
Morphology, pigment structure and growth rate of P. agardhii and C. raciborskii were compared under different light intensities and temperatures (Tables 3 and 4, Fig. 5). Increments of light intensity from 20 to 100 μmol photons m−2 s−1 induced adaptive morphological responses in C. raciborskii. A significant increase in C. raciborskii MLD and individual V (M–W, P < 0.05) occurred when cultures grew at high light intensity (Table 3).
Table 3. Morphology (average ± standard deviation, n = 60) and pigment structure (molar pigment ratios to Chl a) of Planktothrix agardhii and Cylindrospermopsis raciborskii isolates grown under 20 and 100 μmol photons m−2 s−1
P. agardhii MVC11
C. raciborskii MVCC14
The maximum absorbance peaks are indicated between brackets for the unknown carotenoids (Myxol-like and Car 1).
S / V, surface/volume ratio; Myxol-like, 4-keto-myxol-2′-methylpentoside-like; TCAR, total carotenoids; Phy, phycocyanin; nd, not detected.
Light intensity (μmol photons m−2 s−1)
Size and shape
3573 ± 1771
6169 ± 3968
414 ± 150
1406 ± 523
258 ± 111
318 ± 200
132 ± 48
200 ± 75
0.98 ± 0.09
0.82 ± 0.04
2.02 ± 0.01
1.35 ± 0.03
Carotenoid ratios to Chl a
Car 1 (477/505)
Table 4. Growth parameters under different light intensities (maximum growth rate: μ, slope of the light-limited portion of the curve: α, and subsaturating light: Ik); Q10 based on the maximum growth average (n = 4) at 15 and 25 °C for both isolates
P. agardhii MVCC11
C. raciborskii MVCC14
Mean ± standard deviation for light growth response.
Significant differences between species (t-test, P < 0.05).
Lipid pigment composition and responses to light intensity also differed between species. Typical cyanobacterial carotenoids were detected in both species. Planktothrix agardhii had higher concentrations of zeaxanthin, β,β-carotene, an undetermined glycosidic carotenoid similar to 4-keto-myxol-2′-methylpentoside (myxol-like) and an unknown carotenoid (car 1), while C. raciborskii had high concentrations of aphanizophyll and echinenone (Table 3). In both species, protective and accessory pigment concentrations changed in response to light intensity, with notable differences. The phycocyanin/Chl a ratio decreased with higher light intensity in P. agardhii and increased in C. raciborskii. The magnitude of change in carotenoids/Chl a was also higher in C. raciborskii than in P. agardhii (Table 3). Total carotenoids increased sixfold in C. raciborskii, largely because of aphanizophyll, but also because of echinenone and β,β-carotene. Planktothrix agardhii total carotenoids increased 1.5 times with higher light, mainly because of myxol-like and carotenoid 1, whereas β,β-carotene, echinenone and zeaxanthin decreased.
Growth curve experiments performed along a light intensity gradient (from 5 to 180 μmol photons m−2 s−1) indicated strong similarities between the two species under light-limited conditions (indicated by α and Ik), although C. raciborskii reached significantly higher growth rates (μmax) than P. agardhii (Table 4, Fig. 5a and b). Temperature growth experiments at two light intensities demonstrated the different behaviour of the two species. Planktothrix agardhii growth rates were significantly higher than those of C. raciborskii at 15 and 20 °C at low light intensity (60 μmol photons m−2 s−1) (Fig. 5c), although no differences were found at 25 °C. However, C. raciborskii grew significantly faster than P. agardhii (Fig. 5c) at high light intensity (135 μmol photons m−2 s−1) at 25 °C. Q10 values also showed that C. raciborskii growth rate had a higher response to a temperature increase at 135 μmol photons m−2 s−1 than P. agardhii (Table 4).
Our extensive data set and laboratory experiments indicated that although C. raciborskii and P. agardhii overlap in their distribution relative to temperature, light and trophic status, they differed in their biovolume distributions along these gradients. Our results support the hypothesis that Cylindrospermopsis is tolerant to a wide range of climates, from tropical to temperate (Briand et al., 2004). Although many studies have suggested that the optimum water temperature of the species is from 25 to 35 °C (Saker & Eaglesham, 1999; Briand et al., 2004; Mehnert et al., 2010), high biomass has been observed in subtropical lakes at 19 °C (Everson et al., 2011), and C. raciborskii was equally dominant throughout the year in a tropical lake independent of water temperature variation (17–24 °C) (Figueredo & Giani, 2009). Still other studies found some strains to be capable of sustaining biomass or growing at temperatures as low as 14–17 °C (Chonudomkul et al., 2004; Piccini et al., 2011). Fabre et al. (2010) observed C. raciborskii occurrence during winter in a subtropical lake (Lago Javier, Uruguay), and in our database the biovolume of C. raciborskii in this lake reached 95% of the total (i.e. 2.2 mm3 L−1) in winter (water temperature: 11.2 °C). To our knowledge, this is the lowest temperature at which C. raciborskii has been observed to reach high biovolume and dominate the phytoplankton. The success of C. raciborskii in a wide range of temperatures observed in our data set and other recent studies (Vidal & Kruk, 2008; Kokociński et al., 2010; Everson et al., 2011) suggests that current concepts of C. raciborskii as a tropical species may be due more to a lack of information than to any physiological restriction.
We observed P. agardhii only in temperate and subtropical lakes, but in a wide range of temperature conditions. This species can reach high biomass in a range of temperatures from < 2 °C (Toporowska et al., 2010) to 29 °C in tropical ecosystems (Crossetti & Bicudo, 2008; Gemelgo et al., 2009). In our database, P. agardhii dominated the phytoplankton of Lago Rodó (subtropical) in all seasons at temperatures ranging from 10 to 31 °C, indicating substantial tolerance to temperature variation.
Our Q10 data indicated that C. raciborskii grows faster than P. agardhii when temperatures shift towards warmer conditions and thus may be favoured by climate warming. Experimental studies showed that P. agardhii maximum growth occurred between 20 and 25 °C (Post et al., 1985; Sivonen, 1990), and its growth rate increased significantly between 15 and 25 °C (Oberhaus et al., 2007). Cylindrospermopsis raciborskii has been shown to benefit more than other cyanobacteria from high temperatures (Mehnert et al., 2010) and also to have higher photosynthetic activity and lower light requirements than other cyanobacterial species (Wu et al., 2009). Dominance of this species, however, cannot be predicted from any single factor. In this sense, our Q10 values (at 60 and 135 μmol photons m−2 s−1) suggest that C. raciborskii could have a competitive advantage over P. agardhii at conditions with both high light and high temperature. These differences are attributable to the ability of C. raciborskii to increase light-harvesting capacity through changes in shape and pigment composition and their proportions as we demonstrated in our experiments. Different photoprotective responses were also suggested by pigment changes after light increases. In this sense, the acclimation capacity of C. raciborskii is illustrative of its phenotypic plasticity.
Planktothrix agardhii showed a higher competitive capacity than C. raciborskii under low light and lower temperatures, which agrees with its broad distribution in turbid temperate lakes (Dokulil & Teubner, 2000; Nixdorf et al., 2003). Based on the biovolume distribution and growth of P. agardhii in our study, we suggest that this species has limited plasticity, as its physiological response to temperature increase under high light intensity was less pronounced than that of C. raciborskii.
According to our Ik values, both species are shade-tolerant, implying that they can succeed in turbid, eutrophic lakes (Padisák & Reynolds, 1998), as originally proposed for Oscillatoriales (Scheffer et al., 1997). The Ik values that we obtained for both species were lower than those reported in the literature (~ 20 μmol photons m−2 s−1) (Talbot et al., 1991; Shafik et al., 2001; Briand et al., 2004) which may support the existence of ecotypes suggested by Piccini et al. (2011). The presence of ecotypes with different environmental preferences confers a wide intra-specific variability to C. raciborskii; ecotypes are among the hypotheses advanced to explain the species’ expansion. The pigments we identified in both species were typical for cyanobacteria (Millie et al., 1990), although they were present in markedly different proportions.
Planktothrix agardhii is favoured in continuously mixed, shallow lakes (Scasso et al., 2001; Kruk et al., 2002; Stüken et al., 2006). Similarly, blooms of C. raciborskii are commonly reported in mixed conditions (Bouvy et al., 1999; Huszar et al., 2000; Briand et al., 2002; Figueredo & Giani, 2009) and rarely in stratified deep reservoirs (Padisák et al., 2003). This clearly indicates that both species have a wide tolerance for mixing. While in our study there was no clear relationship between C. raciborskii biovolume and light availability (Zeu/Zmix), P. agardhii biovolume was higher in turbid conditions below the threshold value of 1.62 Zeu/Zmix ratio, suggesting its dependence on turbidity.
The cyanobacterial contribution to total phytoplankton biomass increases markedly above 30 μg L−1 TP in temperate lakes (Watson et al., 1997; Dokulil & Teubner, 2000), and P. agardhii biovolume distribution in our data set reflect this general pattern. However, the sudden changes we observed in biovolume distribution in hypereutrophic conditions suggested a threshold near 160 μg L−1 TP, above which other factors affected biovolume accumulation. In contrast, there was a negative relationship between trophic status and C. raciborskii contribution to total biovolume, with increasing dominance of the phytoplankton below 200 μg L−1 TP. Higher ranges of phosphorus cell quota in C. raciborskii relative to P. agardhii may permit C. raciborskii to better exploit low P environments (Ducobu et al., 1998; Istvánovics et al., 2000). Some studies suggest that P. agardhii growth is greatly dependent on high-frequency phosphate availability (Catherine et al., 2008; Crossetti & Bicudo, 2008; Kokociński et al., 2010; Aubriot et al., 2011), while C. raciborskii is able to dominate with small, low-frequency phosphate inputs (Posselt & Burford, 2009). Phytoplankton from oligotrophic and mesotrophic ecosystems may thus be sensitive to a replacement by C. raciborskii as a dominant species under small nutrient enrichments.
According to our data, C. raciborskii can dominate the phytoplankton at lower overall biovolume than P. agardhii, giving insight into the ability of C. raciborskii to colonize and rapidly succeed in new habitats. While variations of C. raciborskii biovolume during the year were gradual, P. agardhii was either dominant or scarce in phytoplankton. Moreover, the dominance of P. agardhii in the phytoplankton appears to occur in more eutrophic conditions (i.e. higher phytoplankton biovolume). Scheffer et al. (1997) observed similar behaviour in eutrophic temperate shallow lakes and proposed hysteretic mechanisms to explain the distribution and resilience of P. agardhii.
Our data indicated that higher diversity (as taxonomic richness) is supported under dominance of C. raciborskii than that of P. agardhii. This suggests a higher capacity of C. raciborskii for co-existence with other species (Kokociński et al., 2010) and may support the hypothesis of its greater plasticity. Sperfeld et al. (2010) demonstrated experimentally that the invasion and success of C. raciborskii was not affected by the diversity of the host phytoplankton community. The relatively higher diversity associated with C. raciborskii dominance also has implications for food webs, as some studies also found positive correlations between the biomass of C. raciborskii and zooplankton (Bouvy et al., 2001; Soares et al., 2009). Conversely, the lower diversity of communities dominated by P. agardhii may result from a capacity of this species to generate limiting conditions (i.e. high turbidity) for potential phytoplankton competitors.
The greater plasticity of C. raciborskii in response to key environmental factors (temperature and light intensity) may explain its gradual response to changing environments. Conversely, the lower plasticity of P. agardhii fits with its narrower distribution in nature. Aquatic environments are highly variable habitats in terms of light and nutrient resources at the time scale of phytoplankton life spans. Reversible plastic phenotypes represent an advantage for organisms in highly variable environments (Piersma & Drent, 2003), allowing the adjustment of their functional responses and increasing their invasive potential (Litchman, 2010). Ecotypes with differing environmental tolerances such as those shown in C. raciborskii (Piccini et al., 2011) further strengthen its aptitude for invasive behaviour and success in different climates.
In summary, C. raciborskii and P. agardhii behaved differently as a result of contrasting strategies for responding to environmental constraints. Further research is required to determine whether this pattern may represent differing strategies in bloom-forming filamentous cyanobacteria of the orders Oscillatoriales and Nostocales. Differences between P. agardhii and C. raciborskii, as well as between other organisms with comparable strategies, will likely affect the future distribution of these species in projected warming climates where blooms will be enhanced. The high phenotypic plasticity of C. raciborskii, and its wide tolerance ranges to key environmental factors, explains its current expansion to temperate latitudes and forecasts its further increase in the future.
We thank Marie-Josée Martineau for technical assistance and Warwick F. Vincent for kindly providing access to laboratory facilities. This work was financed by ANII FCE2007_353 and PEDECIBA. We also thank two anonymous reviewers for their comments.