Conceived and designed the experiments: JP KP. Performed the experiments: KP. Analyzed the data: KP JP. Contributed reagents/materials/analysis tools: JP. Wrote the paper: JP KP.
The authors have declared that no competing interests exist.
Free-living nano-sized flagellates are important bacterivores in aquatic habitats. However, some slightly larger forms can also be omnivorous, i.e., forage upon both bacterial and eukaryotic resources. This hitherto largely ignored feeding mode may have pronounced implications for the interpretation of experiments about protistan bacterivory. We followed the response of an uncultured group of omnivorous cercozoan nanoflagellates from the Novel Clade 2 (Cerc_BAL02) to experimental food web manipulation in samples from the Gulf of Gdańsk (Southern Baltic Sea). Seawater was either prefiltered through 5 µm filters to exclude larger predators of nanoflagellates (F-treatment), or prefiltered and subsequently 1∶10 diluted with sterile seawater (F+D-treatment) to stimulate the growth of both, flagellates and bacteria. Initially, Cerc_BAL02 were rapidly enriched under both conditions. They foraged on both, eukaryotic prey and bacteria, and were highly competitive at low concentrations of food. However, these omnivores were later only successful in the F+D treatment, where they eventually represented almost one fifth of all aplastidic nanoflagellates. By contrast, their numbers stagnated in the F-treatment, possibly due to top-down control by a concomitant bloom of other, unidentified flagellates. In analogy with observations about the enrichment of opportunistically growing bacteria in comparable experimental setups we suggest that the low numbers of omnivorous Cerc_Bal02 flagellates in waters of the Gulf of Gdańsk might also be related to their vulnerability to grazing pressure.
Nanoplanktonic flagellates (NF) are important grazers within aquatic microbial food webs. However, they do not represent a homogenous functional guild of predators: members of the smallest size class (<5 µm) are typically responsible for the major part of picoplankton (bacterial and picocyanobacterial) mortality, while larger species may also forage on algae or other NF
Some groups of NF may even explore both, bacteria and eukaryotes as a food source
In aquatic ecosystems, many ciliated protists have been found capable of feeding on both, bacterial and eukaryotic prey. By contrast, such data are scarce for NF, and the current models and experimental approaches to study aquatic microbial food webs do not consider omnivory within this group of organisms
The Baltic Sea is a semi-closed basin with narrow and shallow connection to the North Sea. High riverine run-off and reduced water exchange with the oceanic waters results in vertical and horizontal salinity gradients, from 30 PSU in Kattegat to <1 PSU in the northern reaches of the Bothnian Bay. Anthropogenic pressure on the Baltic Sea is very high, and the ecosystem suffers from pollution and eutrophication. Communities of microorganisms present is the Baltic Sea are a mixture of marine, brackish and freshwater species
We performed food web manipulation experiments to investigate the response to perturbation of a group of omnivorous NF that were also present in surface waters of the coastal Baltic Sea (an uncultured, heterotrophic group of cercozoans from the Novel Clade 2;
The bacterial numbers in the F+D-treatment doubled from t24 to t48. Before and after this period, the total numbers of bacteria remained relatively constant (
Numbers of bacteria (A & C); and composition of bacterial community (B & D) in the F+D- (5 µm prefiltration and 1∶10 dilution) and F- (5 µm prefiltration) treatments. Error bars show standard deviation, based on the triplicate samples. Eub – bacteria targeted by the general probe Eub I-II-III, Alf968 – Alphaproteobacteria (probe Alf 968), Bet42a – Betaproteobacteria (probe Bet42a), Gam42a – Gammaproteobacteria (probe Gam42a), CFB319a – Cytophaga–Flavobacteria (probe CFB319a), HGC69a – Actinobacteria (probe HGC69a). Note that the Y-axis scale differs between the panels.
The total number of bacterial cells in the F-treatment decreased to approximately half within the first 12 h, and remained more or less constant thereafter (
Initially, plastidic NF were slightly more numerous than aplastidic ones in both treatments, but were overgrown within 12 h (
Changes of the numbers of plastidic and aplastidic nanoflagellates (A–B), of cercozoan cells targeted by the probe Cerc_Bal01 (C–D); and of cercozoan cells targeted by the probe Cerc_Bal02 (E–F) in the F+D- (5 µm prefiltration and 1∶10 dilution) and F- (5 µm prefiltration) treatments. Error bars show standard deviation, based on the triplicate samples. Note that the Y-axis scale differs between the panels.
Changes in the numbers of cells targeted by the probe Cerc_Bal01 were similar in both treatments (
Cells targeted by the newly designed probe Cerc_Bal02 were of minor importance at t0 in both treatments (
Assuming that the initial distribution of cell-length was similar in both treatments, a shift in size of Cerc_Bal02 cercozoans towards smaller cells in both treatments was observed (
Changes in the cell length distribution of the cercozoan targeted by the probe Cerc_Bal02 in the F+D- (5 µm prefiltration and 1∶10 dilution) and F- (5 µm prefiltration) treatments. The data from the triplicates were pooled and are shown together on a single graph. Numbers of analysed cells (N) are given in parentheses. The data from the t0 point for the 1∶10 dilution treatment are not available because of low numbers of cells that could be found and measured. The black, vertical lines represent the mean cercozoan cell length.
Both treatments likely resulted in elevated grazing pressure on bacteria, as estimated from the decreased ratio of bacteria to total aplastidic NF (
F+D-treatment | ||||||
t0 | t12 | t24 | t48 | t72 | t96 | |
Bac : aNF | 753±122 | 182±36 | 103±22 | 196±37 | 132±15 | 133±29 |
NF : Cerc_Bal02 | 267±159 | — | — | 52±10 | — | 6.6±0.6 |
Bac : Cerc_Bal02 | 7.4±3.9×104 | — | — | 7.6±2.0×103 | — | 732±194 |
F-treatment | ||||||
t0 | t12 | t24 | t48 | t72 | t96 | |
Bac : aNF | 753±122 | 85±31 | 50±16 | 61±7 | 65±18 | 93±23 |
NF : Cerc_Bal02 | 267±159 | — | — | 58±12 | — | 69±17 |
Bac : Cerc_Bal02 | 7.4±3.9×104 | — | — | 2.4±0.4×103 | — | 4.4±0.9×103 |
Bac, Bacteria, (a)NF: (aplastidic) nanoflagellates. The uncertainty of these values at t0 in the F+D-treatment is relatively higher due to low number of Cerc_Bal02 cells that were counted.
Nanoflagellates were considered bottom-up controlled when the ratio value fell below 1∶100 (based on Gasol
We found both prey types (bacteria and eukaryotes) in food vacuoles of cells targeted by probe Cerc_Bal02 (
Changes in preferred prey of the cercozoan Cerc_Bal02, shown as percentage of the flagellate cells with none, 1 or >1 of eukaryotic or bacterial cells in food vacuoles in (A) the F+D- (5 µm prefiltration and 1∶10 dilution) and (B) F- (5 µm prefiltration) treatments. The data from the t0 timepoint for the F+D-treatment are not available because of low numbers of cells that could be found and analysed. Only significantly different timepoints (p<0.05, χ2-test) are indicated by the corresponding letters.
At the end of the experiment the food ingestion patterns of Cerc_Bal02 cells clearly differed between the two treatments. Eukaryotic food items prevailed in the food vacuoles of Cerc_Bal02 cells in the F-treatment (χ2 = 151.3, P<0.0001), while ingestion of bacteria was more commonly encountered in the F+D-treatment (χ2 = 19.89138, P<0.0001).
We also attempted to specifically investigate the influence of Cerc_Bal02 on the numbers of bacterivorous Cerc_Bal01 NF
(A) Photomicrograph of an unknown protist (olive-green) in the experimental enrichments with an ingested cell detected by probe Cerc_Bal02 (red); (B) Cercozoan cell detected with probe Cerc_Bal02 (red) ingesting a cell detected by probe Cerc_Bal01 (green). Blue objects in both panels: DAPI stained nuclei. Scale bars are 10 µm in panel A and 5 µm in panel B. Depictions are true-colour images from the same microscopic fields obtained by simultaneously exciting with several wave lengths.
The interactions between flagellate species (e.g. grazing, competition) may influence the composition of the pro- and eukaryotic microbial communities, and hence, the functioning of ecosystems
A fractionation of microbial assemblages through filters with a pore size of 5 µm is typically applied to assess the role of NF in controlling the composition of bacterial communities
Dilution experiments have been originally introduced as a means of simultaneous estimation of growth and mortality rates of phytoplankton
We additionally modified the classical dilution treatment by first removing larger protists and metazoans via filtration. This allowed us to directly assess the additional effect of dilution on the studied microbes. For example, there was a clear initial shift of the bacterial assemblage towards
In theory, there ought to be different mechanisms of NF control between the treatments, namely in the F+D-treatment expected bloom of bacteria would later relieve NF from the bottom-up control
The continuous exponential growth and increased contribution of the omnivorous Cerc_Bal02 at low prey to predator ratios in the F+D-treatment (
The observed success of the Cerc_Bal02 population at low prey concentration raises the question about the possible reasons for their low numbers in the studied environment (
The presumably distinct grazing pressure on Cerc_Bal02 cells between the treatments might have also contributed to the observed differences in their size distributions (
It should be noted that the exposure to different regimes of grazing pressure and food availability might have also caused the rise of different genotypes within the diverse group of cercozoans targeted by the probe Cerc_Bal02 (
In summary, our experimental food web manipulation created different scenarios of bottom-up vs. top-down stress for Cerc_Bal02 cercozoans, resulting in contrasting patterns of growth and grazing behaviour. These flagellates were found to be omnivores capable of successful reproduction at low food availability, but seemed to be vulnerable to predation by other NF. This suggests that Cerc_Bal02 cercozoans follow an ‘opportunistic’ life strategy and are possibly controlled by grazers in the environment. Further studies, e.g. of the
Coastal surface water (approx. 400 m off-shore) for the experiments was collected with a clean bucket from the Gulf of Gdańsk (Baltic Sea) on August 20, 2007, prefiltered through a 10 µm plankton net and transported to the laboratory within 15 minutes. Temperature was measured
The collected water was further prefiltered through a 5 µm membrane filter (diameter 47 mm, Isopore, Millipore, USA) at low pressure (0.26 bar). Part of the so pre-treated water was directly used for the experiment (F-treatment), while the rest was ten-fold diluted with sterile seawater (filtration through 5, 1.2 and finally 0.22 µm membrane filters, diameter 47 mm, Isopore, Millipore, USA) (F+D-treatment). 2 litres of the pre-treated water were incubated in triplicates in 5 L glass Erlenmayer flasks in the dark at in situ temperature (19.6°C) and salinity (7.2 PSU) for 96 h in SANYO incubators.
Samples for total numbers of heterotrophic bacteria and nanoflagellates (NF) were taken after 12, 24 h and every 24 h thereafter for a total of 96 h. Bacteria (4–50 ml of water) were fixed with buffered paraformaldehyde solution (pH 7.6, final conc. 1%), and NF (15–100 ml) with alkaline Lugol's solution followed by addition of formaldehyde solution (final conc. 2%) and decolorization with 3% sodium thiosulphate
Samples for CARD-FISH analysis of bacterial groups were collected together with those for the total counts but the tripled volume was filtered on white polycarbonate filters (47 mm diameter, Isopore, Millipore, pore size 0.22 µm). After enzymatic digestion with lysozyme (10 mg ml−1, 1 h) and proteinase K (75 nl ml−1, 30 mins), bacterial cells were hybridized with horseradish peroxidase labelled oligonucleotide probes
Samples for the determination of the abundance of NF affiliated with two groups of Cercozoa were collected three times: at t0, t48 and t96. They were fixed and filtered as described for the total counts. The adjustment of hybridization condition and staining procedure for the flagellates by CARD-FISH is described in Piwosz and Pernthaler
Approx. 100 cells from each triplicate hybridized with the Cerc_Bal02 probe were photographed after visualisation by epifluorescence microscopy (AxioImager.Z1, Carl Zeiss, Jena, Germany, 640× magnification) with an AxioCam MR3 camera (Carl Zeiss), and their cell size was determined using the length tool of the AxioVision software (Carl Zeiss).
The presence of bacteria and eukaryotic prey in food vacuoles was assessed by epifluorescence microscopy based on their DAPI staining (AxioImager.M1, Carl Zeiss) at blue/UV excitation. Prey items were counted only (i) if they were inside a food vacuole, visible as a dark area within a hybridized flagellate cell (
For the analysis of differences in frequency of ingested prey items, a χ2 test was performed. To fulfil the assumption of the χ2 test of expected value for each group to be >10, the cells were divided into three groups: i) no foot items, ii) a single food item, and iii) >1 food items inside food vacuoles. Differences in the control mode (bottom-up vs. top-down) between the treatments were analysed by the Mann-Whitney U-test.
(TIF)
(TIF)
(TIF)
We thank Michael Zeder for assistance with microscopy and Mariusz Zalewski for help with incubators.