Microbial control of live / dead zooplankton ratio in Sevastopol Bay

To explain higher fraction of live zooplankton in heavily polluted and eutrophic Sevastopol Bay comparing with cleaner adjacent waters, a hypothesis has been proposed and tested experimentally that more intensive bacteria-driven decomposition of dead organisms in the bay reduced their pool and, as a result, increased the live-to-dead zooplankton ratio. In the experiment, a heat-killed batch culture of the copepod Calanipeda aquaedulcis was used as a substrate for decomposition by natural microbial communities from the waters of different pollution status. Bacterioplankton abundance and in situ decomposition rate of copepod carcasses were shown to be about 3-fold higher in the bay (1.3 × 10 6 cells ml -1 and 0.13 day -1 , respectively) while an approximation of zooplankton non-predatory mortality rates gave equal values for both the sites (about 0.046 day -1 ). These findings call for revising the ways of interpreting the results of zooplankton viability assays in their relation to water pollution status.

Recent studies of live and dead planktonic copepods in Sevastopol bay and adjacent coastal waters (Litvinyuk et al., 2011) have revealed a surprising pattern: average annual fraction of live organisms (FLO) in highly polluted and eutrophic waters of the bay was significantly higher than one in the cleaner offshore area (Fig. 1).This result contradicted the conclusions made earlier by other authors (e. g.Pavlova et al., 2001;Pavlova, Melnikova, 2005) who explained higher necrozooplankton abundances and low FLO by adverse environmental conditions.In this study, bacterial degradation of copepod carcasses was considered a key factor controlling the 'dead pool' and, hence, the value of FLO in the studied area.Higher microbial activity in the polluted bay was hypothesized to accelerate copepod decomposition and, as a result, shift the ratio of live-to-dead organisms to higher values.The hypothesis was tested experimentally and a simple model was used to understand how the pool of dead organisms forms in the water column and how the balance between their inflow (via non-consumptive mortality) and outflow (via decomposition) is maintained in time.

Material and Methods
To estimate microbial abundances and conduct decomposition experiments in situ, seawater samples were collected at two stations in Sevastopol Bay and adjacent, open waters (Fig. 2).The water area was characterized by pronounced gradients of water salinity, contamination and eutrophic state from the mouth of the Black River in the apex of the bay (close to St. 2) to the bay mouth and the open sea outside the bay (St.1).The concentration of biogenic elements in the river outflow was 2 to 10 times higher than in the bay (Ovsyanyi et al., 2007), resulting in formation of stable concentration gradients of nitrite-N (from 0.9 µg L -1 at St. 1 to 2.3 µg L -1 at St. 2), nitrate-N (3.2 and 13.3 µg L -1 , respectively), ammonium-N (17.4-23.1 µg L -1 ), phosphate-P (7.9-8.8 µg L -1 ), and silica (57-81 µg L -1 ) (V. Gubanov, unpublished data).The average E-TRIX eutrophication indices (Vollenveider, 1998) were 4.7 (St. 1) and 5.5 (St.2) (Gubanov et al., 2015).The degree of chronic contamination by petroleum products and heavy metals increased from the open sea outside the bay to the bay center.Thus, the content of oil hydrocarbons in the bottom sediments of St. 1 was ~30 mg 100 g -1 , increasing to 180 mg 100 g -1 at St. 2 (Osadchaya et al., 2004).The rates of degradation and decomposition of dead 'model' organisms (the copepod Calanipeda aquaedulcis Kritchagin, 1873) by natural microbial communities from marine waters of different pollution status (St. 1 vs St. 2) were measured experimentally and compared.A batch culture of the copepods fed microalgal mixture (Bacillariophyceae: Phaeodactylum tricornutum; Chlorophyceae: Chlorella vulgaris, Dunaliella salina; Dinophyceae: Prorocentrum cordata, Prorocentrum micans; Prymnesiophyceae: Isochrysis galbana).The copepods were killed by heat (85 °С for 10 min.),trapped on 100-μm nylon mesh, rinsed and resuspended in native seawater freshly collected at the two stations (June 2012) and prefiltered through 100-µm nylon mesh to remove mesozooplankton.The final concentration of the dead organisms in the experimental flasks was about 300 ind.L -1 .The flasks were prepared in triplicates for each of the locations, submersed to the sea (down to about 2-m depth) and incubated under in situ temperature (22 °С) and light conditions for 4 days that was an expected time period for decomposing the bulk of the carcasses (Tang et al., 2006).Then, decomposition of the copepods was examined by light microscopy.
130 carcasses of nauplii, copepodites and adults were analyzed visually and photographed in every replicate for identifying visual signs of disintegration and decomposition of the carcasses and classifying the extent of their degradation to the five categories: Ino visual signs of any body injuries or internal decomposition, IIpartial postmortal decomposition of the body with light 'spots' and cavities, IIIdistinct visual signs of internal decomposition of up to 50% of the body volume, IVbody injuries and more than 50% internal decomposition, cephalothorax deformation and degradation, Vfragmented and/or empty carcasses with no tissue inside (Fig. 3).The carcasses at the stage V were defined as completely decomposed.To compare average annual bacterioplankton abundances at St. 1 and St. 2, seawater samples were collected in the surface layer bimonthly over 2010-2011.These data and the data on zooplankton FLO in Sevastopol Bay published earlier (Litvinyuk et al., 2011) were both obtained simultaneously at the same stations (as a part of Sevastopol Bay Time Series), thus allowing their inter-comparison.
Bacterial abundances in the samples and in the experimental flasks were estimated by flow cytometry.Counts were performed using a Beckman Coulter flow cytometer (Cytomics FC 500) equipped with an air-cooled blue laser (15 mW, 488 nm) and the standard filter setup.Aliquots of 1ml water samples previously fixed with formaldehyde (2% final conc.) were stained with SYBR Green I (Molecular Probes Inc.), following the procedures described by (Marie et al., 1997;Gasol, Giorgio, 2000).SYBR-Green I fluorescence at the green channel FL1 (525 nm) was considered proportional to intracellular nucleic acid content and interpreted as a measure of bacterial cell-specific metabolic activity according to (Servais et al., 2003).Consequently, the integral metabolic activity of the community was estimated as FL1 × N (in relative units), where N is the total bacterial abundance in the sample.
To analyze the field and experimental results, a simplified model (ignoring sedimentation) has been used, describing dynamics of live and dead organisms and the balance between their pools: where N L and N D are abundances of live and dead organisms, respectively; µ is specific growth rate; m is non-consumptive mortality rate; d is decomposition rate of zooplankton.Under steady-state conditions (µ = m, constant N L ) in the model population, FLO depends on N D and, hence, on the ratio between the nonconsumptive mortality rate (m) of live zooplankton (inflow to the pool of dead organisms) and the decomposition rate (d) of copepod carcasses (outflow from the pool).In the decomposition experiment, the equation 2 was used to calculate d at N L = 0, µ = 0, m = 0, and FLO = 0. Statistical analyses were performed using software STATISTICA ver. 10 (StatSoft, Inc., USA).

Results and Discussion
By fourth day of the exposition, about 20% (St. 1) and 50% (St.2) of the copepods were decomposed completely, reaching the stage V (Fig. 4).Frequency distributions of the copepod decomposition stages differed significantly in the locations, indicating faster process in the bay.The rates of decomposition (d) approximated from these data amounted 0.05 day -1 in the bay (St. 1) and 0.13 day -1 outside it (St.2), thus supporting the hypothesis of microbial control of live/dead zooplankton ratio in Sevastopol bay.Disintegration and decomposition of zooplankton carcasses was earlier shown to be driven by external, not internal, bacteria (Harding, 1973;Lee and Fisher, 1992;Reinfelder et al., 1993), hence, indicating a direct link between FLO and microbial activity in the water column.The initial bacterial abundances in the experimental flasks differed significantly, 0.4 × 10 6 cells ml -1 at St. 1 versus 1.3 × 10 6 cells ml -1 at St. 2 (Fig. 4).Approximately the same difference between the locations maintained over the yearaverage annual abundance of bacterioplankton proved to be about 2.5-fold higher in Sevastopol Bay as against the adjacent waters (Fig. 5).According to the seasonal microbiological data obtained in 2010-2011, no significant difference between the stations has been revealed in terms of bacterial cell volumes and intracellular nucleic acid contents (FL1 signal).However, the integral metabolic activity measured as FL1×N was 2.5-to 3-fold higher in the bay community (Fig. 5).
Thus, both the experimental results and microbiological data have provided an evidence of a linkage between FLO and bacterioplankton.There was an increase in FLO at higher bacterial abundances and high FLO variability at fewer bacteria (Fig. 6).The latter seemed to be a result of contribution of other important factors to the FLO control (like sedimentation or elimination by detritivores).Given constant N L and equality of µ and m (losses are compensated for by reproduction) under stationary conditions, the value of FLO depends only on the d/m ratio irrespective of the initial N L , m and d (Fig. 7).The dependence has allowed a calculation of the d/m ratios from the field data on the FLO annual averages (54% and 77%, Litvinyuk et al., 2011), which amounted about 1.2 and 3.0 at St. 1 and 2, respectively (Fig. 6).Thus, according to the model, the rates of copepod mortality and decomposition were roughly equal in the open waters while in the bay, decomposition occurred three times faster.Combining these calculations with the experimental results gives nearly equivalent rates of zooplankton nonconsumptive mortality at the St. 1 and 2 (0.045 and 0.047 day -1 , respectively) which are within the range (0.01 to 0.1 day -1 ) commonly reported for marine zooplankton (see review in Tang et al., 2014).
In summary, it can be concluded from the present findings that higher FLO values observed in more polluted and eutrophic marine coastal waters were not necessarily a result of higher zooplankton mortality but rather a consequence of higher abundance of heterotrophic bacteria and more intensive bacteria-driven decomposition of dead organisms in the detrital food web.Unexpectedly, the rates of non-consumptive mortality of copepods in polluted Sevastopol Bay were approximated to not differ significantly from those in the less polluted adjacent waters.Thus, results of zooplankton viability assays should be interpreted carefully in their relation to water pollution status.Higher pollution status of an area does not necessarily imply higher zooplankton mortality and/or lower FLO values.

Figure 1 .
Figure 1.Fluorescein diacetate-(FDA) and neutral red (NR) -based estimates of the average annual FLO in the open coastal waters (St. 1 in this study) and the polluted bay (St. 2 in this study) in 2010 -2011.Calculated from the data presented in Litvinyuk et al. (2011).

Figure 2 .
Figure 2. Sampling sites in Sevastopol Bay and adjacent coastal waters.

Figure 3 .
Figure 3. Stages of decomposition of copepod carcasses.Explanations are in the text.

Figure 4 .
Figure 4. Initial bacterial abundances in the experiment (N o , left plot) and frequency distribution of the copepod decomposition stages on the fourth day of exposition at St. 1 and 2 (right plot).Means and standard deviations are presented.

Figure 7 .
Figure 7. Fraction of live organisms (FLO) as a function of the decomposition-to-mortality ratio (d/m) in the model under steady-state conditions (mortality and specific growth rates are balanced, µ = m) and projections of natural zooplankton communities (St. 1 and 2) onto the model curve.