Contribution of bacteria in the mucilage of Microcystis spp. (Cyanobacteria) to benthic and pelagic bacterial production in a hypereutrophic lake Free

Abstract

The mucilage of cyanobacteria represents a unique habitat for both water column and sediment bacteria. In Lake Vallentunasjön, Sweden, the pelagic Microcystis-associated bacteria constituted 19–40% of the total bacterial abundance, and their contribution to the total bacterial production was 7–30%. In the sediment, the mucilage bacteria constituted only 1–5% of the total bacterial abundance, but contributed with 8–13% to the total bacterial production during the summer. Microcystis-associated bacteria thus were less active (bacterial production/cell) than ambient water column bacteria, while in the sediments the Microcystis colonies were ‘hot spots’ with enhanced bacterial activity as compared to other sediment bacteria.

1 Introduction

The attachment of bacteria to particles is an important process in aquatic ecosystems. Bacteria may attach, either loosely or more permanently, to various types of suspended particles ranging in size from submicrometer colloids [1], algae and faecal pellets [2], to the large aggregates that ecologists term marine or lake ‘snow’[3,4]. Attachment is normally considered adaptive, e.g. by enhancing uptake of organic compounds and nutrients [5] or providing a refuge from grazing [6], but since almost any particle will be of a size suitable for some grazer, attachment to particles also has certain risks as an adaptive strategy for survival [7,8]. Whether attachment to particles is advantageous in a given environment depends primarily on the types of particles available for colonisation.

Normally, healthy phytoplankton are not colonised by bacteria, but there are exceptions, especially among the cyanobacteria, where attachment of bacteria to heterocytes is a well-known example [9]. Microcystis spp., common non-heterocytic colonial cyanobacteria in eutrophic lakes, are commonly found with numerous bacteria embedded in the mucilage [9,10]. Mucilage bacteria in old, moribund colonies probably benefit from using substrate and nutrients from decaying Microcystis cells, but the bacteria colonising young and healthy colonies may not necessarily be harmful to the host. Even symbiotic interactions may occur that favour the Microcystis cells, e.g. by supplying nutrients [11].

In contrast to the situation in lake water, bacteria in sediments have numerous particles in their environment, and a large part of the bacteria are more or less tightly associated with these. Studies of bacteria associated to various types of sediment particles have, to my knowledge, not been performed. Research is usually focused on different bacterial processes carried out by functional groups of bacteria (e.g. denitrification, sulphate reduction, methanogenesis), and their relative contribution to the degradation of organic material in sediments of various composition. For example, effects of bioturbation and sedimentation on microbial processes are frequently examined [12,13].

Investigations in Lake Vallentunasjön, Sweden, have shown that living Microcystis colonies in the sediment are a significant part of the benthic microbial biomass, and a coupling has been shown between Microcystis biomass and bacterial production in the sediment [14]. The purpose of this study was to assess the production of bacteria embedded in the mucilage of Microcystis colonies and to compare this with the production of other bacteria in the sediment and in the water column. In most studies of attached and free-living bacteria, the separation has been made with filtration techniques. The partitioning of particles into size-classes is never perfect [8] and it is not possible to separate different types of particles within the same size fraction. By removing individual Microcystis colonies from the water column and sediments, I was able to study an important category of attached bacteria in Lake Vallentunasjön.

2 Materials and methods

2.1 Lake description and sampling methods

Lake Vallentunasjön is a shallow (mean depth 2.7 m), hypereutrophic lake near Stockholm, Sweden. Until 1970, the lake was polluted by municipal sewage water during a period of 20 years, and substantial cyanobacterial blooms still occur in late summer yearly, frequently dominated by various Microcystis species (mainly M. wesenbergii and M. viridis). For a detailed lake description, see Brunberg and Boström [14]. The present study was conducted from June 1989 to May 1990. During this period, the surface sediment was sampled on five occasions, using a Willner core sampler. The sediment cores were subsequently sectioned at the lake (0–1 cm layer used in this study). The water column was studied on three dates, including a diurnal study on 2 August. Water samples were taken from the 0–2-m layer of the lake, using a tube sampler.

2.2 Abundance of bacteria and cyanobacteria

Samples for determination of microbial abundance were preserved with 4% (v/v) formaldehyde. Cell numbers of Microcystis (cyanobacterial cells >3 μm in diameter) were determined from autofluorescence with fluorescence microscopy, after sonication and filtration onto 0.2 μm polycarbonate filters. Bacterial abundance was determined after sonication, staining with acridine orange, filtration on 0.2-μm polycarbonate filters prestained with Sudan black and counting of cells using fluorescence microscope. Subtraction of autofluorescing cells (small cyanobacteria, <2 μm in diameter, counted with the same technique as the larger cyanobacteria) from total bacterial counts was made in the appropriate bacterial cell size classes. To determine the colony size of the Microcystis colonies and the number of attached bacteria, 100 individual Microcystis colonies were removed from the samples, sonicated, and the Microcystis cells were counted using the Utermöhl technique with Lugol's solution and an inverted microscope. The bacteria attached to these colonies were enumerated using epifluorescence microscope as described above for the total bacterial counts. For a detailed description of the abundance determinations, see Brunberg and Boström [14].

2.3 Bacterial production

All incubations were started within an hour after sampling. Water samples (15 ml) were incubated in 20-ml glass scintillation vials with 30 nM [methyl-³H]thymidine for 2 h in situ, then formaldehyde was added to a final concentration of 2% (v/v). The fixed samples were transported to the laboratory and 5-ml subsamples from three parallels and one blank were extracted in 0.6 N NaOH containing 0.1% (w/v) SDS and 25 mM EDTA, acidified with cold 100% trichloroacetic acid (TCA), filtered onto 0.45-μm cellulose acetate filters, and subsequently rinsed with 5×1-ml portions of cold 5% (w/v) TCA and 5×1-ml portions of ice-cold 80% (v/v) ethanol [15]. From the remaining three water samples and blank, 100–500 Microcystis colonies were removed by micropipette and processed according to the same protocol. Filters were placed in plastic scintillation vials, scintillation solution was added, and the samples assayed for radioactivity using an LKB Rack-Beta liquid scintillation counter.

Sediment bacterial production was estimated via [³H]thymidine incorporation into DNA following the protocol of Bell and Ahlgren [16]. In short, six parallel samples (0.5 g wet sediment) and two formaldehyde-killed blanks were incubated at in situ temperature with 60 μCi [methyl-³H]thymidine (40–60 Ci mmol⁻¹; Amersham) for 2 h. The incubation was stopped by addition of formaldehyde to a final concentration of 2% (v/v). Three samples and one blank were subsequently assayed according to Bell and Ahlgren [16]. From the remaining samples and blank, 100–500 Microcystis colonies were removed with a micropipette, washed three times in 0.2-μm filtered water (distilled and tap water, 1+1) and transferred to 5 ml of filtered (0.2 μm) lake water containing 4% (v/v) formaldehyde. These samples were then filtered as described for the lake water samples.

The NaOH extraction was used in processing sediment samples [16]. For water samples, radioactivity retained on filters after this procedure was less than that retained after extractions in 5% (w/v) cold TCA extraction (total macromolecules) and slightly more (up to 10%) than the radioactivity incorporated into a DNA fraction using the chloroform–phenol procedure of Wicks and Robarts [17]. Because ≥70% of the thymidine incorporated into macromolecules in this lake was in DNA (Bell 1990, unpublished), the NaOH extraction procedure in this study was considered to be roughly comparable to the DNA fraction. Here it was considered advantageous to have a similar procedure for sediment and water samples. Incubation times are usually kept ≤60 minutes [18]. However, because low total counts were expected, and to ensure diffusion of thymidine throughout the colonies, a longer incubation time was chosen. Normally in this lake, although 5 nM gave maximal incorporation rates, 15–20 nM was the thymidine concentration used to measure production of pelagic bacteria (Bell, unpublished). The higher concentration was used to ensure that the bacteria in the mucilage received a high effective concentration. This was tested in October by measuring the incorporation of [³H]thymidine at concentrations varying between 15 and 75 nM into material retained by a 12-μm pore-sized cellulose nitrate filter. During this period, Microcystis biomass in the lake water was maximal (Chl a >100 μg l⁻¹) and Microcystis colonies were the dominant particles of the >12 μm fraction. Concentrations >30 nM increased the incorporation rate by about 20%, but using higher concentrations increases the risk that macromolecules other than DNA are labelled [19]. Microcystis cells lack the ability to incorporate [³H]thymidine [15,20]. Autoradiographic studies have shown that the [³H]thymidine is incorporated in mucilage bacteria, but not in Microcystis cells [20].

2.4 Statistical treatment

Differences in specific [³H]thymidine incorporation rate between mucilage bacteria and total bacteria were tested statistically with Mann–Whitney test. The differences were considered significant at a level of P<0.05.

3 Results

3.1 Abundance of bacteria and cyanobacteria

The abundance of Microcystis-associated bacteria in the lake water closely followed the abundance of Microcystis cells (Table 1). The maximum abundance of Microcystis cells and Microcystis-associated bacteria, 110×10³ cells ml⁻¹ and 4.07×10⁶ cell ml⁻¹, respectively, were observed on 26 September. The Microcystis-associated bacteria constituted 40% of the total pelagic bacterial abundance on this occasion. From late autumn through spring, pelagic Microcystis colonies were not present in the amounts large enough to allow an investigation of this type. In the sediment, the abundance of Microcystis was highest in December, after the sedimentation of colonies from the lake water. The quotient between bacteria and Microcystis cells varied between 10 and 50 (Table 1) and did not differ significantly from the quotient found for pelagic colonies (Student's t-test). The Microcystis-associated bacteria constituted between 1 and 5% of the total benthic bacteria.

3.2 Bacterial production in pelagic samples

The specific incorporation rate of rate of [³H]thymidine by the pelagic Microcystis-associated bacteria was always lower compared to bacteria which were free-living or attached to other particles in the water column (Fig. 1). This was especially pronounced during the diurnal study. The largest difference was in the morning sampling (10 a.m.), when the total bacterial community had a specific thymidine incorporation rate that was c. 9 times higher than for the mucilage bacteria. In the middle of the day (2 p.m.) the mucilage bacteria showed the highest measured contribution to the total [³H]thymidine incorporation; about 30%. On 26 September, the water temperature had decreased (15°C, compared to 19°C on 2 August). The total [³H]thymidine incorporation was lower than in August, with a minor contribution from mucilage bacteria. In November, when the water temperature was 7°C, no [³H]thymidine incorporation was detectable in the Microcystis-associated bacteria.

3.3 Bacterial production in benthic samples

The rates of [³H]thymidine incorporation in sediment bacteria had a pattern that partly contrasted with the pelagic bacteria (Fig. 2). On two sampling dates, in June and August, the specific [³H]thymidine incorporation rate of the mucilage bacteria was 8 and 3 times higher, respectively, than for the total benthic bacterial population. The production of benthic mucilage bacteria increased dramatically at temperatures >15°C (Fig. 3). When temperatures were lower, the specific production was low in all samples, and the differences between the two bacterial fractions were small. On 26 September, at a temperature of 15°C, the mucilage bacteria had a significantly lower specific incorporation rate of [³H]thymidine than the other sediment bacteria. The results from 6 December were close to the detection limit of the thymidine method, even when the sample size was 500 Microcystis colonies, and there was no significant difference between incorporation rates for mucilage bacteria and total bacteria on this date. Also, in May, when the temperature was 14°C, there was no significant difference in bacterial production between mucilage bacteria and total bacteria.

3.4 Pelagic vs. benthic bacterial production

A comparison between thymidine incorporation rates in pelagic and benthic Microcystis colonies has to be done with caution. It is more complicated to determine thymidine incorporation rates in sediment bacteria than in pelagic bacteria [21]. Thymidine readily adsorbs onto particles in the sediment, and to be able to calculate a correct value of the incorporation rate the isotope dilution has to be determined, i.e. the degree of participation for the labelled thymidine in the thymine synthesis [22]. Fig. 4 shows the thymidine incorporation rate for pelagic and benthic Microcystis colonies, calculated both with and without the isotope dilution estimated for Lake Vallentunasjön sediments. The data suggest that there is a difference in thymidine incorporation rates between pelagic and benthic colonies, with higher rates in the benthic colonies.

In conclusion, this study showed that the specific incorporation rate of [³H]thymidine by the pelagic Microcystis-associated bacteria was always lower compared to bacteria which were free-living or attached to other particles in the water column, while in the sediment, Microcystis-associated bacteria were more active than other bacteria on two of the five sampling occasions (Figs. 1 and 2).

4 Discussion

The thymidine method has been widely used to measure heterotrophic bacterial production in the water column and surface sediments of both marine and freshwater environments [18,22,23], but the analysis protocols may not be applicable for measuring production of bacteria embedded within the mucilage of Microcystis colonies. If the diffusion of thymidine into the mucilage is slow, the bacteria in the Microcystis colonies may be exposed to a lower concentration of thymidine than other bacteria in the surrounding environment. In this study, we tried to optimise the [³H]thymidine concentration and the incubation time to minimise this risk. The results suggest that the rate of thymidine incorporation in mucilage bacteria is not largely underestimated. In the lake water, the incorporation rate in total bacteria was nine times higher than in mucilage bacteria. However, the results were opposite in the sediment, where the mucilage bacteria incorporated up to eight times more [³H]thymidine per cell than the total bacterial community.

There are several possible explanations for the differences in [³H]thymidine incorporation rates between mucilage bacteria and total bacteria. First, there may be different taxonomic composition of the bacterial community in the mucilage of Microcystis colonies, compared to other bacteria in the water column and sediments. The microenvironment of the mucilage may be different regarding e.g. oxygen concentration, redox potential and pH [24] as well as nutrient concentrations. The mucilage even may host bacteria unable to grow outside the colony [25,26].

Besides taxonomic differences in growth rate, there may also be different abilities to take up and incorporate the thymidine into DNA among the bacteria [22,27–30]. This might explain the lower production measured in pelagic Microcystis colonies compared to other pelagic samples. However, autoradiograms [20] indicated that a large part of the mucilage bacteria in Microcystis colonies from Lake Vallentunasjön had incorporated [³H]thymidine. Although thymidine incorporation in aerobic waters is now considered a measure of ‘heterotrophic’ bacterial production (e.g. [23]), this interpretation should be applied with caution to sediment environments in general.

In the pelagic zone, attached bacteria generally incorporate radiolabelled organic compounds at higher rates than free-living bacteria [31], but for [³H]thymidine the rates are generally lower [5,32,33]. Kirchman [34] suggested that attached bacteria have a lower growth efficiency, which may be due to the larger release of extracellular polymers by these bacteria, both for attachment [31,34] and to degrade the particle or the macromolecular compounds bound to the particle [35]. This view is in accordance with the observation that attached bacteria are generally larger than free-living bacteria, although the reverse has also been reported [5,8].

Another cause of low bacterial production in pelagic Microcystis colonies may be slow diffusion within the mucilage. The chemical composition of the mucilage may differ substantially with the time of the year and the condition of the Microcystis cells, but the main constituent is carbohydrate [36]. Low diffusion rates of nutrients from the lake water might cause C:N:P quotients that do not meet the requirements of the growing bacteria. Furthermore, slow diffusion out of the colonies of inhibiting or toxic exudates from the cell metabolism of the Microcystis cells may restrain the growth of attached bacteria.

Finally, the lower production of pelagic Microcystis-attached bacteria may be due to low mortality. Bacterial production depends on both the growth rate and abundance of bacteria. Pelagic environments where bacterial grazers are suppressed develop high abundances of bacteria that are growing slowly, and a major fraction may be inactive. On the other hand, intensive grazing on bacteria keeps the biomass low, but the bacteria are growing rapidly [6,37]. Microcystis is generally considered to be grazed only to a limited extent [38,39]. Bern [40] did not find any grazing on [³H]thymidine-labelled bacteria attached to large colonies of Microcystis wesenbergii in Lake Norrviken, adjacent to Lake Vallentunasjön. Consequently, the biomass of mucilage bacteria in many colonies could be near ‘carrying capacity’, having lower specific growth rates than the other pelagic bacteria.

Contrary to this study, Worm and Sondergaard [41] found a higher specific growth rate of Microcystis-attached bacteria than for ambient pelagic bacteria. The Microcystis colonies were characterised as ‘bacterial incubators’ which supply the ambient lake water with bacterial biomass through shedding. This contrasting result might be due to differences in natural environmental conditions, but may also be explained by differences in methodology. Worm and Sondergaard [41] used filters to separate Microcystis colonies from the lake water. Filtration collects also other lake water particles, which may host numerous bacteria and have a high microbial activity. The method used in the present study, on the other hand, may have a slight bias towards fresh and healthy Microcystis colonies and thereby possibly lower microbial activity. Decaying Microcystis colonies are sometimes difficult to distinguish from other organic debris in a stereomicroscope and may thus be under-represented when picking out samples with micropipette.

In autumn, Microcystis colonies settle to the bottom of Lake Vallentunasjön. Their survival in the sediments is very long, probably several years in some cases, but ultimately the majority of the colonies will be decomposed. Decay and lysis of Microcystis cells will drastically change the extracellular release of inorganic nutrients and organic compounds, but also a switch to a resting stage with changed metabolism may affect the leakage from the cells. Empty Microcystis colonies are easy to identify by epifluorescence microscopy after acridine orange staining. The mucilage keeps the typical shape of a Microcystis colony, especially the mucilage from M. wesenbergii. The mucilage is empty of Microcystis cells, but certainly not of bacteria. In a detailed seasonal study of benthic Microcystis colonies, Brunberg (unpublished) found that 66% (S.D.=12, n=36) of the colonies in the surface sediment (0–10 cm) of Lake Vallentunasjön were devoid of living Microcystis cells. Degradation of moribund Microcystis cells would possibly enhance the bacterial growth in the mucilage, and thus a higher [³H]thymidine incorporation rate would be expected in the benthic mucilage bacteria compared to the pelagic mucilage bacteria (cf. Fig. 4).

The results from 26 September indicate that there may be other reasons, in addition to temperature dependence, for the declining specific production of benthic mucilage bacteria (Fig. 2). The lower value for mucilage bacteria than for total bacteria suggest that the remaining population in the sediment on this occasion was more resistant to decomposition than earlier during the summer. The autumn sedimentation of the water population had not yet started at this time, and the benthic population probably to a large extent was composed of Microcystis colonies which had survived for 1 year or more in the sediment. Results from experimental laboratory studies with Microcystis from Lake Vallentunasjön [42] also support the conclusion that colonies with different age and pre-history have different resistance to decomposition and bacterial attack.

Generally, a large percentage of sediment bacteria are inactive [43,44]. Although a small fraction may be growing rapidly, the specific activity and growth rates calculated for the whole sediment population are low. The activity of sediment bacteria is considered to be governed by the availability of substrate and electron acceptors, and by the temperature. Boström et al. [10] concluded that the bacterial activity in the sediments of Lake Vallentunasjön was primarily governed by temperature. This is further implied by the data in Fig. 3. The results are also in accordance with Kirchman [34], who found that attached bacteria were larger than free-living bacteria in the water of a freshwater pond during July and August, but not in February and May. These findings indicated that attached bacteria were relatively more active than the free-living bacteria in the summer, but during the winter the activities of the two groups of bacteria were equal.

A comparison with data from 1985, when an extensive study of the benthic microbial community was made in Lake Vallentunasjön [10], demonstrates the importance of Microcystis colonies as sites of enhanced bacterial activity in the sediments (Table 2). The 1–5% of Microcystis-attached bacteria in the present study (1989), were responsible for 8–13% of the total bacterial production in the sediment. In 1985, the biomass of Microcystis in the sediment was about five times higher than in 1989, and up to 40% of the sediment bacteria were attached to Microcystis colonies. No corresponding data on the bacterial production of mucilage bacteria are available for 1985, but they may have been responsible for a large part of the high total bacterial production in the sediment that was measured that year. Moreover, the data from a 5-year study of microbial biomass and activity in Lake Vallentunasjön showed a co-variation of Microcystis biomass and bacterial production in the sediment [14].

The immediate advantages and/or disadvantages for bacteria attached to Microcystis colonies thus seem to vary, depending on the environment in which they are situated, but also due to the condition of the Microcystis cells as indicated by the data. However, there is also a long-term aspect on the advantages for mucilage bacteria. The Microcystis colonies may provide microenvironments for bacteria that would not survive in the prevailing environmental conditions of the lake water (cf [45]). When the right conditions occur, the mucilage bacteria may serve as an inoculum which facilitates renewed growth. A low bacterial activity and growth rate in the pelagic Microcystis colonies in the short-term may thus be well paid back in the long-term, when the situation changes and the Microcystis cells start to decay and provide potentially high-quality substrates for the bacteria.

Acknowledgements

This work was performed in close cooperation with the late Dr. Russel T. Bell. Originally, he was a co-author, and participated in the preparation of the first draft of this paper. Russ's expertise on thymidine incorporation methods was important for the planning of the study and in the methodological modifications that we made. Thanks are extended to Jan Johansson for expert technical assistance, and to Dr. L. Tranvik and Dr. J.J. Cole for critical reading of the manuscript. This work was supported by grants from the Swedish Natural Science Research Council (NFR).

References