CELL-SPECIFIC EXTRACELLULAR PHOSPHATASE ACTIVITY OF DINOFLAGELLATE POPULATIONS IN ACIDIFIED MOUNTAIN LAKES¹

Study site and sampling. We studied the three following small mountain lakes in the Bohemian Forest (Šumava Mts., Czech Republic): Čertovo Lake (13°12′ E, 49°10′ N), Prášilské Lake (13°24′ E, 49°05′ N), and Plešné Lake (13°52′ E, 48°47′ N). Their crystalline bedrock is geologically sensitive to acidic deposition; catchments are small (65–88 ha), with steep slopes covered with thin acidic soils; and the vegetation is dominated by Norway spruce (Veselý 1994). The lakes are situated at elevations of 1,027–1,087 m, their surface areas range between 4.2 and 10.7 ha, and the maximum depths are 17–35 m (Nedbalová et al. 2006).

The epilimnion of the lakes was sampled at the 0.5 m depth regularly in 3-week intervals from May to September 2007. The samples were immediately prefiltered with a 200 μm polyamide sieve (Silk & Progress s. r. o., Brněnec, Czech Republic) and transported to the laboratory. For the determination of phytoplankton species composition and biovolume, a subsample of ∼0.5 L was fixed with acid Lugol’s solution and stored until processing.

Chemical and plankton analyses. In the laboratory, samples were filtered with glass-fiber filters (pore size 0.4 μm, Macherey-Nagel GmbH, Düren, Germany), except for samples for pH and alkalinity (acid neutralizing capacity determined by Gran titration), and total concentrations of aluminum (Al), phosphorus (P), and carbon (C), which were not filtered beyond the field prefilter. Dissolved organic C (DOC) was analyzed in the filtrate, and particulate C (PC) was analyzed by combustion of the retained particulate organic matter on the glass-fiber filter (both with TOC 5000A analyzer; Shimadzu, Kyoto, Japan). Soluble reactive P (SRP) was determined by the molybdate method. When the SRP concentration was below the detection limit of 50 nmol · L⁻¹, half of this value was used in subsequent data evaluation. Total and dissolved P (TP and DP) were determined by perchloric acid digestion and the molybdate method; samples were 4-fold concentrated by evaporation (at ∼100°C prior digestion) to obtain a detection limit of 15 nmol · L⁻¹. No TP and DP concentration was below this detection limit. Particulate P (PP) was the difference between TP and DP. Fractionation of Al, that is, total reactive Al (Al_t), dissolved Al (Al_d), and organically bound Al (Al_o), was analyzed in nonfiltered samples, filtered samples, and cation exchange treated samples after their filtration, respectively (Driscoll 1984). Ionic Al (Al_i) was calculated as the difference between Al_d and Al_o, and particulate Al (Al_p) as the difference between Al_t and Al_d. Details on analytical methods are given in Kopáček et al. (2006). All chemical analyses were finished within 24 h after the sampling.

Phytoplankton counting was done in Utermöhl’s sedimentation chambers on an inverted microscope (Nikon Diaphot; Nikon, Tokyo, Japan). A presedimentation of samples (5×) was necessary except for Plešné Lake samples. A minimum of 400 individuals were counted in each sample. Mean cell volume (MCV) of each species was estimated by shape assimilation to known geometric forms (Hillebrand et al. 1999), and total biovolume was expressed in mm³ · L⁻¹. Total biovolume of filamentous cyanobacteria was estimated in sedimentation chambers using the line intercept method (Nedoma et al. 2001). All species’ biovolumes were summed as phytoplankton fresh mass. Concentration of chl a was determined spectrophotometrically on Whatman GF/C filters (Maidstone, Kent, UK) after acetone extraction (Lorenzen 1967); values were not corrected for phaeopigments.

Detection and quantification of PA. Samples were processed immediately after transportation to the laboratory (usually <2 h after sampling). Two distinct assays were used for detection of extracellular PA in the samples: (i) bulk activity and (ii) single-cell activity (FLEA assay).

Bulk PA was measured using common fluorogenic substrate, 4-methylumbelliferyl phosphate (MUFP; Hoppe 1983). Duplicate 4.5 mL subsamples were supplemented with 0.5 mL of MUFP solution (100 μmol · L⁻¹ final concentration; Glycosynth, Warrington, Cheshire, UK), mixed, and incubated at in situ temperature for 6–15 min. Then, the enzymatic hydrolysis was stopped by addition of 100 μL of 2 M NaOH, followed by intensive stirring. Fluorescence was read on a Spekol 11 photometer with a fluorometric device (Carl Zeiss, Jena, Germany).

The protocol for FLEA (Nedoma et al. 2003) was followed for the detection and measurement of cell-specific PA in the phytoplankton. Duplicate 2 mL (Plešné Lake) or 5 mL (other lakes) subsamples were incubated with fluorogenic substrate ELFP (Molecular Probes; Invitrogen, Eugene, OR, USA). The substrate was split into small portions (100 μL, stored at −18°C) used up within one experiment. ELFP was always filtered through a spin filter (0.2 μm pore size, Ultrafree-MC; Millipore Corp., Bedford, MA, USA) before use. The incubation was started by the addition of the ELFP solution (20 μmol · L⁻¹ final concentration) and lasted 60 min at 20°C. Each incubation was terminated by transferring the sample to a filter holder, and the sample was immediately supplemented with HgCl₂ as a fixative (4 mmol · L⁻¹ final concentration). The fixation of incubated samples with HgCl₂ prevents the destruction of fragile algal species, especially flagellates, during filtration and microscope slide preparation (Štrojsová et al. 2003). Then samples were filtered over mild vacuum (<20 kPa) through polycarbonate filter (pore size 2 μm, filter diameter 16 mm; Osmonics, Minnetonka, MN, USA). The filter with retained plankton was placed on a microscopic slide, embedded with the antifading reagent Citifluor AF1 (Citifluor, London, UK), and covered with a coverslip.

The phytoplankton was examined for the presence of ELFA precipitates with an epifluorescence microscope (Nikon Eclipse 90i). For the detection of extracellular phosphatases, we used an image analysis system consisting of a monochromatic digital integrating camera Vosskühler COOL-1300Q (Vosskühler GmbH, Osnabrück, Germany), mounted onto the fluorescence microscope Nikon Eclipse 90i and connected to a PC. The image analysis software used for microscope control, image acquisition, and processing was NIS-Elements 2.30 (Laboratory Imaging, Praha, Czech Republic). Two kinds of images were obtained: (1) ELF-images of the ELFA deposits’ fluorescence for the detection and quantification of PA, using ELFA-specific microscope filter block (excitation/emission: 360–370 nm per 520–540 nm), and (2) CHL-images of chl a autofluorescence for identification of algae, using a chl-specific filter block (excitation/emission: 510–550 nm per >590 nm). Both ELF image and CHL image of the same field were stored in one image file of JPEG2000 image format.

To quantify cell-specific PA of algal species, a set of images was taken using ×200 magnification to measure at least 30 randomly selected cells of each species, both unlabeled or ELFA labeled. The FLEA assay, therefore, quantified an average PA per cell of the entire population (including inactive cells on a filter). The cell-specific PA measurement was done in the program NIS-Elements 2.30. Relative ELFA fluorescence in fluorescence units (FU) on a selected area (object) was calculated from the object area (Area), mean gray of the object (MG_obj), the mean gray of the background (MG_bgr), camera exposure time (ET), and a calibration factor (CF), compensating experiment-to-experiment changes in optical setup (optical alignment, fluorescent lamp aging, etc.) as follows:

(1)

Calibration factor was determined using a standard solution of 9.4% fluorescein (0.1 g · mL⁻¹) as described in Model and Burkhardt (2001). FU was averaged per sample and converted to amount of ELFA precipitate by using a conversion factor of 0.021 fmol · FU⁻¹, determined as described in Nedoma et al. (2003). Species-specific PA was expressed per cell in fmol · cell⁻¹ · h⁻¹. To compare cell-specific PA among the species of different size (i.e., MCV), the values were normalized per unit of cell biovolume and expressed as CSPA_B in amol · μm⁻³ · h⁻¹.

Statistics. Medians of seasonal data were tested using repeated measures nonparametric one-way analysis of variance (ANOVA; Friedman test with Dunns posttest) in the program Prism 4.0 (GraphPad, San Diego, CA, USA).

Multidimensional analysis was done in the program CANOCO 4.5 (ter Braak and Šmilauer 2002). Since the gradients calculated by the detrended correspondence analysis were short, we used linear gradient analysis (redundancy analysis). A step-forward regression was performed to explore the environmental parameters significantly correlated with the species-specific CSPA_B (P < 0.05; Monte Carlo randomization test with 500 permutations).

Complex aluminum effects under acidic stress. A high proportion of dinoflagellates in phytoplankton biomass is characteristic for acidified lakes (Almer et al. 1978, Findlay et al. 1999, Vrba et al. 2003a, Nedbalová et al. 2006), which used to be explained by their efficiency in competition for nutrients and tolerance to high Al concentrations (Hörnström 2002).

In this study, Al concentration together with pH were identified as the most important factors among tested parameters, which control—through P availability—CSPA_B in the three dinoflagellates (Fig. 4). Yet there is, of course, natural autocorrelation of high Al concentrations with low pH. Although the Al_t concentrations have decreased markedly in the study lakes since the peak of acidification in the 1980s, they still remain rather high compared to other lake districts (Vrba et al. 2003a, 2006). Jansson (1981) observed that high Al concentrations had an inhibitory effect on extracellular phosphatases at low pH. The inhibition effect on PA is dependent on Al speciation; Al_i is considered to be a competitive inhibitor of extracellular phosphatases. Bittl et al. (2001) studied the impact of Al on bulk PA in Čertovo, Prášilské, and Plešné lakes and confirmed a direct inhibitory effect of Al_i on extracellular phosphatases at diverse Al_i concentrations (0–1,000 μg · L⁻¹) in a laboratory experiment. They set a threshold value of pH < 5 for direct impact of Al on PA, which agreed well with the decrease in Al_i proportion in Al_t at pH > 5 (Driscoll 1985). Vrba et al. (2006) proposed that Al plays a key role in P availability in the acidified Bohemian Forest lakes, namely, through the inhibition of extracellular phosphatases and P inactivation due to the formation Al oxyhydroxides (Al_p) strongly binding P_i (Kopáček et al. 2001). It was proposed that planktonic organisms produce more extracellular phosphatases to obtain a sufficient amount of P in order to overcome the inhibitory effect of Al_i as well as the P inactivation by Al_p (Vrba et al. 2006). This mechanism might take effect in the bacterioplankon that were largely ELFA labeled (e.g., in mesotrophic Plešné Lake). In fact, the plankton P demand causes such an excessive release of extracellular phosphatases that their natural substrate pool may turn over ∼120 times a day in Plešné Lake, while only eight times a day in Prášilské Lake (cf. turnover time in Table 3).

According to the above-mentioned assumption, one would expect a positive correlation between CSPA_B and Al concentration at low pH, as it was in the case of bulk PA (Table 1). The analysis of particular phytoplankton populations, however, showed a negative correlation between CSPA_B of the three species and both Al_t and Al_i concentration (Fig. 4). Moreover, the species-specific response to the analyzed factors was similar for P. umbonatum and Gymnodinium sp. Neither Al effect can satisfactorily explain the overall discrepancy of the dinoflagellates’ CSPA_B. Vrba et al. (2006) also recognized Al as the key factor determining food web structure and plankton dynamics. Consequently, some indirect effect of Al is very likely superior to its direct interactions with the extracellular phosphatases, and the interlake differences in the plankton composition may better explain residual CSPA_B variability.

Top-down effects of zooplankton. Unlike in Prášilské Lake, cladoceran zooplankton are entirely absent in the other two lakes under study (Vrba et al. 2003a,b), and these lakes further differ in both the composition and grazing efficiency of their zooplankton assemblages (Table 3, Nedbalová et al. 2006). While rotifers hardly can release much P_i in Čertovo Lake, abundant populations of both Cyclops abyssorum (recently reintroduced to the lake, see Kohout and Fott 2006) and indigenous Heterocope saliens likely may control the dinoflagellate populations as well as overall P regeneration in Plešné Lake. Therefore, lake internal P turnover might be higher than all indices implied (Table 3).

In Prášilské Lake, the summer shift toward the remarkable dominance of large G. uberrimum (cell diameter ∼35 μm; note also its largest MCV in this lake, Table 2) was also observed during previous studies and resulted from its resistance to the grazing of Daphnia longispina (Vrba et al. 2003b). Because of the permanent CSPA_B detected in G. uberrimum, ELFA-labeled species formed almost 100% of the total phytoplankton biomass in this lake. In Čertovo Lake, the majority of phytoplankton biomass was formed by G. uberrimum and P. umbonatum, with the exception of two samples, when Dinobryon was dominant, likely taking advantage of its phagotrophy and motility (Vrba et al. 2003b). As a fast-growing colonial mixotrophic flagellate (Bird and Kalff 1987), Dinobryon frequently formed summer peaks in both Čertovo and Prášilské lakes (Vrba et al. 2003b, Nedbalová et al. 2006).

Plešné Lake showed a distinct phytoplankton structure (Nedbalová et al. 2006), with several-fold higher total biomass caused by higher P input from the watershed (cf. Table 3), which likely favored the nonmotile green alga Monoraphidium dybowskii and filamentous cyanobacteria Limnothrix sp. and Pseudanabaena sp. (Vrba et al. 2006). The dominance of M. dybowskii in this lake also can be owing to its tolerance to high Al concentrations (Hörnström et al. 1995). Surprisingly, the production of extracellular phosphatases was detected only rarely in M. dybowskii during this study, although this species was ELFA labeled in all earlier studies (Štrojsová and Vrba 2006, 2009). We assume this fact likely reflects high P regeneration due to recent zooplankton grazing (see above). The same effect (i.e., higher ambient P_i supply) perhaps could partly explain lower CSPA_B of all the dinoflagellate species in Plešné Lake. Their low biomass was certainly influenced by the relatively rare occurrence of large G. uberrimum; this species was never found before in Plešné Lake despite extensive observations (Nedbalová and Vrtiška 2000, Vrba et al. 2003a, Nedbalová et al. 2006). Its low abundance and smaller size in this lake could be caused by a specific predation of Heterocope saliens, dwelling solely in the epilimnion of Plešné Lake.

The effect of light availability. Although we are aware of dinoflagellate motility (e.g., Reynolds 1997), different light conditions could provide another explanation of the controversial CSPA_B results. If we consider the Z_eu:Z_mix ratio as a proxy of light availability for algae, its median <1 clearly indicates common light insufficiency in the epilimnion of mesotrophic Plešné Lake compared to the oligotrophic lakes (Table 3) (cf. Kalff 2002). Consequently, the dinoflagellates living in this mesotrophic lake could be limited by light rather than P. On the other hand, sufficient light conditions in the both oligotrophic lakes likely allowed for close to maximal CSPA_B of all three species (Fig. 3, Table 3). Similarly, the highest CSPA_B of all the dinoflagellates in Plešné Lake occurred in June (Fig. 2) and coincided with the highest Z_eu:Z_mix = 1.6 (Table 3).

Seasonal coincidence of high bulk alkaline PA with good light conditions (high Z_eu:Z_mix ratio) was observed in a dimictic eutrophic reservoir (Nedoma et al. 2006). Our study indicates the possible importance of light for the production of extracellular phosphatases in natural dinoflagellate populations. Light availability may change cell stoichiometry from P to C limitation (Sterner and Elser 2002). Ecological stoichiometry (growth rate hypothesis) also predicts a high need in P during intensive cell growth (Sterner and Elser 2002). Both the high P need and optimal light conditions likely caused parallel peaks of each species-specific PA in all the lakes in June–July, whereas the persistence of ice cover till late April, and thus lower intensity of phytoplankton growth and its lowered P demand, could cause relatively low values of PA in early May (Fig. 2). This new view should be considered in further studies on P-limited phytoplankton.

At the same time, insufficient light conditions favor mixotrophic species in the phytoplankton competition, and dinoflagellates are potential mixotrophs; in particular, bacterivory of G. uberrimum has been reported (Callieri et al. 2006, Niesel et al. 2007). Its grazing on the abundant bacterioplankton might be a rich P source in Plešné Lake. Such a parallel uptake of both P and C could be an efficient alternative strategy to phosphatase expression under light deficiency. If this was the case, the mixotrophy of G. uberrimum also would explain the weaker interlake correlations of its cell-specific PA and other parameters compared to the other two dinoflagellates under study.