In the present study, we investigated bacteria and protists in lakes that were classified as eutrophic according to their trophic state index (TSI) values (56.9 to 62.8). Lake Śniardwy, with a TSI of 51.5 indicating meso/eutrophic properties, was included in the investigation because results of recent studies showed that this lake is undergoing the fast process of water eutrophication; bloom of the filamentous cyanobacteria is regularly observed during the summer [Chróst, unpublished data]. Moreover, bacterial and protistan numbers (similar to or even higher than in some eutrophic lakes) may indicate that this lake was more similar to the eutrophic than mesotrophic lakes. Our previous research [13, 14], as well as this study, indicated that microbial parameters may be useful indicators of lake trophic status in addition to chemical parameters. Despite the similar trophic status, the studied lakes differed markedly in the abundances of bacteria and protists and domination structure of ciliate communities. The most visible differences (4-fold) were observed in nanoflagellates, especially in their autotrophic forms (7-fold differences).
Our studies were conducted at the end of April, which is a period of dynamic changes in the pelagic communities of plankton. These changes occur over a relatively short time and their role depends mostly on abiotic factors such as lake morphometry and climate variation . This period differs from other seasons in many aspects. As shown by Kiersztyn et al. , TSI values estimated for the same lakes differ evidently between spring (April) and summer (July). In the present study, the TSI values of most of the studied lakes were higher than those estimated during the summer stratification period (unpublished data). Studies on the seasonal dynamics of planktonic organisms showed that bacteria e.g. , nanoflagellates [1, 8, 16, 20, 21] and ciliates [8, 10, 22], reach usually maximal abundances and/or biomass during spring period, especially in eutrophic/hypereutrophic lakes. At this time, they are probably controlled by different environmental and biotic factors than in summer and/or autumn. According to Šimek et al. , the bottom-up control may be more important during spring and top-down control during summer.
In the euphotic zone, we found high numbers of nanoflagellates (8.89 to 33.56 × 103 cells ml-1; mean 17.36) and ciliates (32.9 to 100.5 ind. ml-1; mean 59.9) that were 2–3-times higher than those noted by Chróst et al.  during the summer stratification period. In contrast, bacterial abundances were relatively low (2.60 to 7.88 × 106 cells; mean 4.76); 3–4 times lower in comparison to the results of studies by other authors that were conducted in the same lakes [14, 24]. In the euphotic zone, among the investigated groups of organisms, only bacterial numbers increased with the increasing trophic status of the studied lakes. This fact may indicate that not all microorganisms were affected in the same way by eutrophication .
Season strongly affects the relationships between different planktonic groups of organisms [8, 16]. Sanders et al.  found that the top-down control (predation) seems to be more important in regulating prokaryotic abundances in eutrophic systems than in oligotrophic systems in which bottom-up control (substrate and nutrient availability) plays a significant role. It is well known that protists, especially heterotrophic nanoflagellates, are the most effective consumers of bacteria [1, 26]. However, depending on the trophic status of lake ciliates, autotrophic nanoflagellates, rotifers and cladocerans can also be important bacterivores . In addition, the role of protists in controlling bacterial abundances is altered in different seasons and years in the same reservoir [1, 2]. According to several authors [8, 27, 28], the abundance of bacteria and protists often shows a lack of coupling across trophic gradients, especially in more eutrophic environments during spring.
In the present study, we did not find any significant correlations between bacterial and protistan numbers, indicating that both nanoflagellates and ciliates were not a major factor controlling bacterial abundance in eutrophic Mazurian lakes during spring. This conclusion is supported by the results of the CCA analysis, which did not show clear relationships between these groups of organisms, both in the euphotic and bottom water layers (Figure 7). In contrast, studies by Koton-Czarnecka and Chróst , which were conducted in the pelagial of ten Mazurian lakes, showed that grazing on bacteria was more efficient in April than in July and protistan grazing removed more than 31% of the bacterial biomass. Ciliates play a significant role as consumers of small algae, autotrophic picoplankton and nanoflagellates. They may also be active bacterivores, particularly in eutrophic lakes, where bacterial densities are sufficient to maintain ciliate populations [1, 30]. This study showed that the ciliate community was dominated mainly by small prostomatids Balanion planctonicum and Urotricha spp. (mainly U. furcata) feeding on bacteria and algae (including flagellates), small oligotrichs Rimostrombidium humile, feeding mainly on algae (including diatoms), and larger (40–60 μm in size) scuticociliate Histiobalantium bodamicum, feeding mostly on algae (e.g. cryptophyte Rhodomonas sp.) . Significant correlations between oligotrichs ciliates and chlorophyll a concentrations, as well as between nanoflagellates and dominating ciliate orders such as Oligotrichida and Prostomatida or dominating taxa such as Urotricha and Balanion planctonicum, may indicate that ciliates were probably associated with spring phytoplankton bloom and played an important role as consumers of small algae and nanoflagellates. All of these facts and a lack of coupling between ciliates and bacteria show that ciliates were probably more important with respect to algivory than bacterivory. A similar conclusion was found by other authors e.g. [10, 11, 31, 32], who suggested that high abundances of small algivorous ciliate taxa in spring (April) in dimictic lakes are related to the mass development of small algae (including flagellates Rhodomonas, Cryptomonas and diatoms) and may be the most important grazers of the spring phytoplankton bloom.
Bacteria differed not only in cell-size and community composition, but mainly varied in metabolic activities in the studied lakes. Studies performed in both marine and freshwater environments, using different staining techniques, showed that only a small fraction (generally less than 20%) of the total bacterial microscopic counts is metabolically active e.g. [33–36]. Factors regulating the numbers and proportion of metabolically active bacteria (temperature, viruses, nutrients, organic substrates and grazing by flagellates and ciliates) are not fully understood  and vary with a very short time scale . Many studies, mostly experimental, have shown that the low percentage of active cells was the consequence of strong effective and continuous protozoan grazing pressure e.g. [37, 39, 40]. Among protists, HNF plays the crucial role in shaping abundance, biomass and morphology of the active fraction of bacteria [37, 41, 42].
In this study, three staining methods (CTC, EST, and NuCC) were used to determine different aspects of the cellular state or metabolic activity of bacteria. The percentage of bacteria with an active electron transport system (CTC+ cells) was the lowest (1.1–15.4% of the total DAPI counts) among the studied fractions of activity. This confirms the observations of Sherr et al. , who showed that CTC+ cells represent a varied active fraction of the total bacterial abundance. A relatively low percentage of CTC+ cells (0.1 to 9.5% of the total counts) was found by other authors [3, 39]. Higher proportions of these cells (about 20–30%) were recorded by Berman et al. , Søndergaard and Danielsen , and del Giorgio and Scarborough .
There are not many data published on bacteria with cellular esterase activity (EST+). As shown by Schumann et al. , the percentage of these cells ranged from 2 to 24% depending on the water systems and was higher in estuaries (mean 12%) than in freshwater (9%) and the Baltic Sea (5%). In this study, the share of EST+ bacteria was low in all lakes (2.7 to 14.2% of the total counts) and was comparable to the CTC+ bacterial cells.
Comparison of the major methods used to assess active bacterial cells showed that the lowest percentage of active cells in freshwaters were recorded with the CTC and EST assay in contrast to NuCC (nucleoid-containing bacteria) and MEM+ (cells with an intact membrane) methods . According to the literature data, NuCC may constitute 37–90% of total counts , while MEM+ cells 42–90% [3, 35] of the total bacterial counts, with the maximal percentage in spring . In the present study, the share of NuCC+ in the total numbers was relatively high and did not differ markedly between individual lakes (50.5–97.3% of the total counts in the euphotic zone); this is in contrast to CTC+ cells, which showed 10–fold differences.
Several authors indicated that the abundance and proportion of active cells (mostly CTC+) increased with increasing lake productivity during the summer period [38, 44]. However, little information is available on the influence of lake trophic status on other fractions of active bacterial cells (EST+, NuCC+, MEM+). Adamczewski  showed that in Mazurian lakes, the share of active bacteria to the total bacterial numbers and biomass (4–27%) markedly differed among the studied lakes, and did not show a relationship with lake water trophic status, either in the surface or bottom water layers. Moreover, mesotrophic lakes were characterized by a considerably higher contribution of MEM+ than eutrophic lakes. Also, Chróst et al.  indicated that the contribution of MEM+ bacteria to the total bacterial numbers (16–24%) and bacterial biomass (9–14%) did not respond strongly to the trophic status of lakes. In our study, only the percentage of EST+ bacteria correlated negatively with TSI of lakes. Correlations between EST+ and NuCC+ cells and total phosphorus (TP) concentrations were also found; however, EST+ correlated negatively, while NuCC+ correlated positively. Highly significant positive correlations were found between DOC concentrations and the numbers of CTC+ cells and EST+ cells, which may suggest that nutrients and organic substrates were important factors controlling the active metabolism of bacteria. Our results are in agreement with other studies [33–35]. We did not find any correlation between nanoflagellates, both autotrophic and heterotrophic, and numbers of active cells of bacteria. This may imply that the bottom-up control by substrates was more important than top-down grazing control by flagellates in regulating the abundance of metabolically active bacterial cells in eutrophic lakes during the spring period. The major source of labile carbon (i.e. easily accessible to bacteria, e.g. proteins) is phytoplankton primary production. During the spring, a large portion of organic matter in lake waters is of allochthonous origin because of the surface runoff from surrounding watershed . The allochthonic matter is poor in proteins, but rich in aromatic cyclic compounds (e.g. polyphenols) that are resistant to microbial degradation; this is an argument supporting a bottom-up control mechanism. Literature data show that ciliates may also be a factor determining active bacterial cells [34, 45]. However, ciliates are not selective and feed on all available bacterial-food. In addition, as mentioned above, ciliates preferentially consumed algae than bacteria in eutrophic Mazurian lakes.
This study was conducted at the end of April, which is the initial phase of stratification and phytoplankton bloom. We did not observe a vertical uniformity in temperature; however, in most lakes the temperature was only slightly lower at the bottom (2.9–8.3°C) than in the euphotic zone (7.3–13.3°C). Thus, plankton organisms were not evenly distributed between the two layers, euphotic and bottom, as during the spring lake water circulation, and did not show a typical distribution in the entire water column as during the summer stratification period . However, studies performed by Berdjeb et al.  in deep pre-alpine lakes, which differ in trophic status, revealed significant differences in the bacterial community structure with depth, regardless of whether the water column was mixed or stratified. It is known that many species of flagellates and ciliates may migrate diurnally throughout the water columns due to top-down grazing pressure by zooplankton e.g. [48, 49]. Thus, microbial communities may vary more with water column depth than through a season, especially during periods of increasing thermal stratification [47, 50, 51].
In the bottom waters, ciliate (19.2–55.8 ind. ml -1 – values 2–3-times lower than in the euphotic zone) and nanoflagellate numbers (3.30–6.04 × 103 cells ml-1 – values 2–5-times lower than in the euphotic zone) were relatively high. The numbers of HNF were more or less evenly distributed between the two studied zones, whereas ANF occurred in relatively low numbers in comparison to the euphotic zone. Both nanoflagellates and ciliates correlated positively with chlorophyll a and negatively with TP concentrations. In addition, highly significant positive correlations were found between oxygen concentrations and nanoflagellates (both total nanoflagellates, HNF and ANF) and ciliates, as well as between temperature and both groups of protists. All of these correlations may indicate that trophic conditions and environmental physical parameters were important in controlling protozoan communities at the bottom level. The results of CCA analysis provided similar conclusion (Figure 7). Similarly as in the euphotic zone, algivorous ciliates (Tintinnidium sp. and Urotricha spp.) dominated. Small bacterivorous scuticociliates and omnivorous Mesodinium sp. were more abundant in the deeper water and their occurrence was mainly associated with a large amount of detritus-bound bacterial food . Significant positive correlations between ciliates and total numbers of nanoflagellates and both their autotrophic and heterotrophic forms suggest the mutual relationship between these components.
Bacterial numbers, biomass, cell-size, and secondary production were generally higher in the bottom lake water than in the euphotic zone. Only the contributions of metabolically active cells of bacteria to the total bacterial counts were lower in the bottom water layer, suggesting that metabolically active cells were probably under high grazing pressure. Similar results were found by Berman et al. , who showed that the mean percentage contribution of metabolically active bacterial cells was lower at 1 m depth (CTC – 4.1%, NuCC – 7.2%, MEM + − 7.6%) than at 38 m depth (CTC – 6.5%, NuCC – 9.7%, MEM + − 8.8%) in meso-eutrophic Lake Kinneret. In contrast, Adamczewski  showed that the maximal share of MEM+ in total bacterial numbers was higher in the hypolimnion (26.3%) than in the epilimnion (18.8%) in 17 Mazurian lakes of different trophies. The author suggested that the higher proportion of metabolically active cells in anoxic water at the bottom of lake indicated lower grazing pressure by bacterivorous protists in comparison to the surface water layer of the studied lakes where the grazing of heterotrophic nanoflagellates leads to a lower abundance and biomass of metabolically active bacteria. In the present study, bacteria, both their total numbers and metabolically active cells, did not correlate with the studied abiotic (DOC, TP, temperature, oxygen) and biotic (chlorophyll, nanoflagellates, ciliates) parameters. However, besides the above-mentioned factors regulating bacterial communities, “inside-control” by viral lysis exists e.g., [37, 47, 52]. As shown by Berdjeb et al. , the magnitude of this mechanism may vary by depth. The authors indicated that top-down regulation by flagellates together with ciliates or viruses was important in controlling bacterial community structure only in the hypolimnion, in contrast to the epilimnion, where bacteria were dependent mainly on bottom-up factors. In our study, the lack of coupling between bacteria and the studied parameters may suggest that infection by viral lysis probably affected bacteria at the bottom of eutrophic Mazurian lakes during the spring period.