Inadequacy of the Eucaryote Inhibitor Cycloheximide in Studies of Protozoan Grazing on Bacteria at the Freshwater-Sediment Interface

Four guilds from a lake sediment-water interface microbial community were isolated and tested for sensitivity to cycloheximide (0.1 to 200 mg liter⁻¹). Field experiments were conducted to compare the inhibition, dilution, and filtration methods for determining grazing rates. Cycloheximide inhibited anaerobic bacteria at 50 mg liter⁻¹, and inhibition of bacterial growth was observed in the grazing experiments. The results show that the assumption of selective inhibition of heterotrophic eucaryotes was violated and preclude the use of cycloheximide in grazing experiments.     

Active Bacterial Populations and Grazing Impact Revealed by an In Situ Experiment in a Shallow Aquifer

To elucidate bacterial population dynamics in an aquifer, we attempted to reveal the impact of protozoan grazing on bacterial productivity and community structure by an in situ incubation experiment using a diffusion chamber. The abundance and vertical distribution of bacteria and protozoa in the aquifer were revealed using wells that were drilled in a sedimentary rock system in Itako, Ibaraki, Japan. The water column in the wells possessed aerobic and anaerobic layers. Active bacterial populations under the grazing pressure of protozoa were revealed through in situ incubation with grazer eliminating experiment by the filtration. On August 19, 2003, the total number of bacteria (TDC) decreased from 1.5 × 10⁶ cells ml^{? 1} at 2.2 m depth to 3.0 × 10⁵ cells ml^{? 1} at 10 m depth. The relative contribution of the domain Bacteria to TDC ranged between 63% and 84%. Protozoa existed at a density of 4.2 × 10⁴ to 1.9 × 10⁵ cells ml^{? 1} in both aerobic and microaerobic conditions. A grazing elimination experiment in situ for 6 days brought about clearly different bacterial community profiles between the 2.2 m and 10 m samples. The bacterial composition of the initial community was predominantly β- and γ -proteobacteria at 2.2 m, while at 10 m β-, α - and γ -proteobacteria represented 56%, 26% and 13% of the community, respectively. The distribution of bacterial abundance, community composition and growth rates in the subsurface were influenced by grazing as well as by geochemical factors (dissolved oxygen and concentrations of organic carbon, methane and sulfate). Results of the in situ incubation experiment suggested that protozoan grazing contributes significantly to bacterial population dynamics.     

Comparison of Rates of Flagellate Bacterivory and Bacterial Production in a Marine Coastal System

Protozoan predation on bacteria and bacterioplankton secondary production were simultaneously determined in La Salvaje Beach water during 1990. Protozoan grazing on bacterioplankton was measured from fluorescently labeled bacterium uptake rates; estimates of bacterial secondary production were obtained from [³H]thymidine incorporation rates. Two different conversion factors were used to transform thymidine incorporation rates into bacterial production rates; both of them were specific for La Salvaje Beach and were calculated by using empirical and semitheoretical approaches. The average flagellate predation rate was 14.0 bacteria flagellate^-1 h^-1; the average population predation rate was 7.35 x 10⁶ bacteria liter^-1 h^-1. The estimates of bacterial production differed greatly depending on the conversion factor used, and so did the percentages of bacterial production consumed by flagellated protozoa (4.6% when the empirical conversion factor for La Salvaje Beach was used and 113% when the semitheoretical conversion factor specific for this system was used). The ecological implications of each of these values are discussed.     

Protozoan Grazing and Bacterial Production in Stratified Lake Vechten Estimated with Fluorescently Labeled Bacteria and by Thymidine Incorporation

In stratified Lake Vechten, The Netherlands, protozoan grazing was estimated on the basis of uptake of fluorescently labeled bacteria and compared with bacterial production estimated on the basis of thymidine incorporation. By using a grazer-free mixed bacterial population from the lake in continuous culture, an empirical relationship between cell production and thymidine incorporation was established. Thymidine incorporation into total cold-trichloroacetic-acid-insoluble macromolecules yielded a relatively constant empirical conversion factor of ca. 10¹⁸ (range, 0.38 × 10¹⁸ to 1.42 × 10¹⁸) bacteria mol of thymidine⁻¹ at specific growth rates (μ) ranging from 0.007 to 0.116 h⁻¹. Although thymidine incorporation has been assumed to measure DNA synthesis thymidine incorporation appeared to underestimate the independently measured bacterial DNA synthesis by at least 1.5- to 13-fold, even if all incorporated label was assumed to be in DNA. However, incorporation into DNA was found to be insignificant as measured by conventional acid-base hydrolysis. Methodological problems of the thymidine technique are discussed. Like the cultures, Lake Vechten bacteria showed considerable thymidine incorporation into total macromolecules, but no significant incorporation into DNA was found by acid-base hydrolysis. This applied not only to the low-oxygen hypo- and metalimnion but also to the aerobic epilimnion. Thus, the established empirical conversion factor for thymidine incorporation into total macromolecules was used to estimate bacterial production. Maximum production rates (141 × 10⁶ bacteria liter⁻¹ h⁻¹; μ, 0.012 h⁻¹) were found in the metalimnion and were 1 order of magnitude higher than in the epi- and hypolimnion. In all three strata, the estimated bacterial production was roughly balanced by the estimated protozoan grazing. Heterotrophic nanoflagellates were the major consumers of the bacterial production and showed maximum numbers (up to 40 × 10⁶ heterotrophic nanoflagellates liter⁻¹) in the microaerobic metalimnion.     

Altering the balance between bacterial production and protistan bacterivory triggers shifts in freshwater bacterial community composition

Bacterivorous protists are known to induce changes in bacterial community composition (BCC). We hypothesized that changes in BCC could be related quantitatively to a measure of grazing: the ratio of bacterial mortality to growth rate. To test this hypothesis, we analyzed time-course changes in BCC, protistan grazing rate, and bacterial production from 3 in situ studies conducted in a freshwater reservoir and three laboratory studies. In the field experiments, samples were manipulated to yield different levels of protistan bacterivory and incubated in dialysis bags. Laboratory investigations were continuous cultivation studies in which different bacterivorous protists were added to bacterial communities. BCC was assessed using 4–6 different rRNA-targeted oligonucleotide probes for community analysis. Change in BCC (Δ BCC) was estimated as the sum of changes in the proportions of the two phylogenetic groups that showed the largest shifts. Analysis of a set of 22 estimates of shifts in the ratio of grazing to production rate over periods of 48–72 h and Δ BCC showed that Δ BCC was positively and tightly correlated (r ² = 0.784) with shifts in the ratio of grazing mortality to cell production. While the nature of a shift in BCC is unpredictable, the magnitude of the change can be related to changes in the balance between bacterial production and protistan grazing. This revised version was published online in August 2006 with corrections to the Cover Date.     

Channeling of bacterioplanktonic production toward phagotrophic flagellates and ciliates under different seasonal conditions in a river

The objective of this study was to analyze the flux of biomass through the communities of bacteria and phagotrophic protists in the cold and warm conditions occurring seasonally in Butrón River. Bacterial and heterotrophic protistan (flagellate and ciliate) abundance was determined by epifluorescence direct counts; protistan grazing on planktonic bacteria was measured from fluorescently labeled bacteria uptake rates; and the estimate of bacterial secondary production was obtained from [³H]thymidine incorporation rates. The abundance of bacterial, flagellate, and ciliate communities was similar during cold and warm situations. However, we observed that estimates of dynamic parameters, i.e., secondary bacterial production and protistan grazing, in both situations were noticeably different. In the warm situation, grazing rates of flagellates and ciliates (bacteria per protist per hour) were, respectively, 7 times and 18 times higher than those determined in the cold situation, and the grazing rates of the protistan communities (bacteria per protists present in 1 ml of water per hour) increased up to 5 times in the case of flagellates and 42 times in the case of ciliates. Estimates of bacterial secondary production were also higher during the warm situation, showing a ninefold increase. The percentage of bacterial production preyed upon by flagellates or ciliates was not significantly different between the two conditions. These results showed that in the different conditions of a system, the flux of biomass between the trophic levels may be quite different although this process may not be reflected in the abundance of each community of bacteria, flagellates, and ciliates. Offprint requests to: J. Iriberri.     

Differential effect of cycloheximide on penicillin and cephalosporin biosynthesis by Cephalosporium acremonium

In resting cells of Cephalosporium acremonium CW19, protein synthesis was inhibited completely by 100 μg cycloheximide per ml. Furthermore, ongoing protein synthesis halted abruptly when cycloheximide was added after 20 min of incubation. Although cycloheximide did not affect the specific rate of penicillin N production, it markedly inhibited the specific rate of cephalosporin C production. The effect of cycloheximide was not influenced by the carbon source used to prepare the cells for the resting cell system.     

Pheopigment dynamics,zooplankton grazing rates and the autumnal ammonium peak in a Mediterranean lagoon

The dilution technique was used to estimate chlorophyll and pheopigment, net and gross production as well as zooplankton grazing over a 12-month period in a coastal lagoon in Southern France. Chlorophyll a (Cha) based gross growth rates of phytoplankton ranged from undetectable in February to 2.6 day⁻¹ in June, corresponding to 3.8 divisions per day. Cha-based grazing rates ranged from undetectable in February to 1.1 d⁻¹ in June. The seasonal growth pattern of picoplankton was similar to that of the whole community, with a peak in July, corresponding to four divisions per day. Grazing processes represented from 20 to 150% of the phytoplankton daily growth, and the grazing pressure was stronger on small phytoplankton cells than on larger cells. Gross growth rates of phytoplankton were related to zooplankton grazing rates, and both were related to water temperature. Mesozooplankton which escaped sampling or oysters had to be also invoked as additional sinks for the primary production. In the fall, pheopigment concentrations greater than chlorophyll concentrations coincided with high ammonium levels in the water column. Pheopigment a production rates were highly correlated to chlorophyll -based microzooplankton grazing rates. The pheopigment a to chlorophyll a ratio was correlated with ammonium concentrations and could be used an index of the balance between ammonium supply (degradation) and demand (uptake by phytoplankton). In addition, pheopigment degradation rates in absence of grazing could be related to irradiance, indicating photo-degradation of these compounds.     

Role of bacteria as food for zooplankton in a eutrophic tropical pond (Ivory Coast)

The aim of the study was to determine whether bacteria could be a substantial source of carbon for zooplankton and whether the grazing pressure of these metazoan filter-feeders could influence the fate of bacterial production. Eight grazing experiments using natural bacteria labelled with³H thymidine were conducted in a tropical pond (Ivory Coast) during various phases of biological colonization (rotifer-dominated and copepod-dominated phases of the colonization). Higher grazing and clearance rates were observed with rotifers (Brachionus plicatilis andHexarthra intermedia), while very low values were obtained when the cyclopoid copepodApocyclops panamensis was dominant. Less than 1% of the bacterial production was harvested when copepods were dominant, whileB. plicatilis consumed up to 36% of this production. However, this consumption of bacteria appeared to contribute only to an insignificant proportion of the daily carbon intake (e.g. 0.9 to 7.1% of body carbon for rotifers). The low contribution of bacteria in the nutrition of zooplankton is discussed in terms of their cell size and their relative abundance in the total amount of seston available.     

Pulse labelling: a method for measuring microbial growth rates in the ocean

Pulse labelling provides a means of generating the growth curve of a microbial population in natural samples. By suitable choice of radioisotopes, a variety of cellular components may be labelled, although adenine to label nucleic acids and ³²P to label phospholipids proved most successful. The labelling of phospholipids also enabled us to measure the growth of algae. When cycloheximide was used as a eukaryote inhibitor to prevent grazing it served to reduce the incorporation of adenine into the bacterial component of the population. Sample dilution as means of controlling predation effects actually increased the sample growth rate until the bacterial component reached the same level as in the undiluted sample, at which time the growth rate fell to that found in the undiluted control. This finding suggests that the bacterial population may be maintained at the carrying capacity of the menstruum, and dilution actually induces an artifact rather than eliminate predation effects.     

Growth Response of Soda Lake Bacterial Communities to Simulated Rainfall

Moderately saline soda lakes harbor extremely abundant and fast growing bacterial communities. An interesting phenomenon of an explosive bacterial growth in shallow soda lakes in Eastern Austria after dilution with rainwater, concomitantly with a significant decrease in temperature was observed in a former study. In the present study, we tried to identify the factors being responsible for this enhanced bacterial growth in laboratory batch cultures. Three experiments were performed with water taken from two different lakes at different seasons. Natural soda lake water was diluted with distilled water, artificial lake water, sterile filtered soda lake water, and grazer-free water to test (1) for the influence of compatible solutes released to the environment and reduced salt stress after osmotic down-shock, (2) for the influence of nutrients, which may be washed in from the dry areas of the lake bottom after rainfall and (3) for the decrease of grazing pressure due to dilution. The potential influence of (4) viruses was indirectly deduced. The response of the bacterial community to the manipulations was measured by changes in bacterial numbers, the incorporation of ³H-leucine and the concomitant determination of the amount of ³H-leucine uptaking bacteria by microautoradiography. The influence of the environmental factors enhancing bacterial growth after a simulated rainfall event showed variations between the lakes and over the seasons. The addition of nutrients was, in all experiments, the main factor triggering bacterial growth. The decrease in grazing pressure and viral lysis after dilution was of significant importance in two of three experiments. In the experiment with the highest salinity, we could show that either compatible solutes released after osmotic down-shock and used as a source of nutrients for the soda lake bacterial populations or reduced salt stress were most probably responsible for the observed marked enhancement of bacterial growth.     

Benthic Bacterial Production and Protozoan Predation in a Silty Freshwater Environment

The interrelation of heterotrophic bacteria with bacterivorous protists has been widely studied in pelagic environments, but data on benthic habitats, especially in freshwater systems, are still scarce. We present a seasonal study focusing on bacterivory by heterotrophic nanoflagellates (HNF) and ciliates in the silty sediment of a temperate macrophyte-dominated oxbow lake. From January 2001 to February 2002 we monitored the standing stock of bacteria and protozoa, bacterial secondary production (BSP, ³H-thymidine, and ¹⁴C-leucine incorporation), and grazing rates of HNF and ciliates on bacteria (FLB uptake) in the oxic sediment of the investigated system. BSP ranged from 470 to 4050 µg C L^–1 wet sediment h^–1. The bacterial compartment turned out to be highly dynamic, indicated by population doubling times (0.6–10.0 d), which were comparable to those in the water column of the investigated system. Yet, the control mechanisms acting upon the bacterial population led to a relative constancy of bacterial standing stock during a year. Ingestion rates of protozoan grazers were 0–20.0 bacteria HNF^–1 h^–1 and 0–97.6 bacteria ciliate^–1 h^–1. HNF and ciliates together cropped 0–14 (mean 4)% of BSP, indicating that they did not significantly contribute to benthic bacterial mortality during any period of the year. The low impact of protozoan grazing was due to the low numbers of HNF and ciliates in relation to bacteria (1.8–3.5 × 10⁴ bacteria HNF^–1, 0.9–3.1 × 10⁶ bacteria ciliate^–1). Thus, grazing by HNF and ciliates could be ruled out as a parameter regulating bacterial standing stock or production in the sediment of the investigated system, but the factors responsible for the limitation of benthic protistan densities and the fate of benthic BSP remained unclear.     

Periphyton responds differentially to nutrients recycled in dissolved or faecal pellet form by the snail grazer Theodoxus fluviatilis

1. We aimed to separate the effects of grazers on periphyton via grazing from that of nutrient recycling from their faecal pellets. 2. We set up three different experimental treatments (snails/no snails/faecal pellets) and sampled over 16 days. The ‘snail’ treatment contained a low density (snail biomass c. 14 g^?2) of the gastropod grazer Theodoxus fluviatilis and the ‘faecal pellet’ treatment received the same amount of faecal pellets as were produced in the ‘snail’ treatment. Whereas the ‘faecal pellet’ treatment provided extra nutrients to periphyton from the faeces, the ‘snail’ treatment provided nutrients in the form of both faeces and in excreta. There was also direct grazing on periphyton in the ‘snail’ treatment. The ‘no snail’ was not grazed and received no nutrients in faeces or excreta. 3. We measured periphyton C, N and P content, chlorophyll‐a (chl‐a), primary production, bacterial biomass, bacterial production and bacterial respiratory activity. In the water column we measured dissolved inorganic N and soluble reactive P. 4. Snails increased the amount of dissolved inorganic N in the water. On day 16, the periphyton N : P ratio in the ‘faecal pellet’ treatment was lower, and periphyton P content was higher, than in the other two treatments. N : P ratios decreased over time in the ‘faecal pellet’ treatment. Primary and bacterial production were positively correlated in all treatments. 5. Algal chl‐a and primary production of periphyton per unit area and periphyton chl‐a : C ratios increased over the 16 day in the ‘snail’ treatment, and thus excretion of dissolved N by snails had a stronger positive effect on the periphyton community than N and P in faecal pellets. 6. Our data show that excretion and egestion can have different effects on periphyton, probably because of the higher proportion of dissolved N in excreta and the higher proportion of P recycled in faecal pellets. The relative effect of nutrients recycled in egesta or in excretions, probably depends on the form of nutrient limitation of the periphyton. Further, the different components of the periphyton matrix could react differently to the different forms of nutrient recycling. 7. We conclude that direct grazing effects are less important than nutrient effects when nutrients are limiting and grazing pressure is low. Further, the spatial separation of different grazing effects can lead to differences in periphyton production and nutrient stoichiometry. This might be an explanation for the patchiness of periphyton in nature.     

Influence of oxygen concentration on the metabolism of Penicillium chrysogenum

In large-scale bioreactors, there is often insufficient mixing and as a consequence, cells experience uneven substrate and oxygen levels that influence product formation. In this study, the influence of dissolved oxygen (DO) gradients on the primary and secondary metabolism of a high producing industrial strain of Penicillium chrysogenum was investigated. Within a wide range of DO concentrations, obtained under chemostat conditions, we observed different responses from P. chrysogenum: (i) no influence on growth or penicillin production (>0.025 mmol L⁻¹); (ii) reduced penicillin production, but no growth limitation (0.013–0.025 mmol L⁻¹); and (iii) growth and penicillin production limitations (<0.013 mmol L⁻¹). In addition, scale down experiments were performed by oscillating the DO concentration in the bioreactor. We found that during DO oscillation, the penicillin production rate decreased below the value observed when a constant DO equal to the average oscillating DO value was used. To understand and predict the influence of oxygen levels on primary metabolism and penicillin production, we developed a black box model that was linked to a detailed kinetic model of the penicillin pathway. The model simulations represented the experimental data during the step experiments; however, during the oscillation experiments the predictions deviated, indicating the involvement of the central metabolism in penicillin production.     

Production and Vertical Flux of Attached Bacteria in the Hudson River Plume of the New York Bight as Studied with Floating Sediment Traps

We investigated the growth and vertical flux of attached bacteria with floating sediment traps in the Hudson River Plume of the New York Bight during the spring diatom blooms. Traps were floated at the base of the mixed layer (ca. 10 m) for 1-day periods. After recovery, we measured bacterial abundance and rates of [methyl-³H]thymidine incorporation in the trap samples. The vertical flux of attached bacteria was estimated with a model formulated to distinguish between bacterial accumulation in traps due to in situ growth and that due to vertical flux. Attached bacterial flux ranged from 0.6 × 10¹¹ to 2.0 × 10¹¹ cells m⁻² day⁻¹, and attached bacterial settling rates of 0.1 to 1.0 m day⁻¹ were observed during periods of vertical particulate organic carbon flux ranging from 254 to 1,267 mg of C m⁻² day⁻¹. In situ growth of bacteria in sediment traps was unimportant as a source of bacterial increase when compared with vertical flux during our study. The vertical flux of attached bacteria removed 3 to 67% of the total daily bacterial production from the water column. Particulate organic carbon is not significantly mineralized by attached bacteria during its descent to the sea floor in the plume area during this period, when water temperature and grazing rates are at their annual minima.     

Pelagic food web processes in an oligotrophic lake

Major pelagic carbon pathways, including primary production, release of extracellular products (EOC), bacterial production and zooplankton grazing were measured in oligotrophic Lake Almind (Denmark) and in enclosures (7 m³) subjected to artificial eutrophication. Simultaneous measurements at three days interval of carbon exchange rates and pools allowed the construction of carbon flow scenarios over a nineteen day experimental period.The flow of organic carbon was dominated by phytoplankton EOC release, which amounted from 44 to 58% of the net fixation of inorganic carbon. Gross bacterial production accounted for 33 to 75% of the primary production. The lower values of EOC release (44%) and bacterial production (33%) were found in the enclosures with added nutrients. The release of recently fixed photosynthetic products was the most important source of organic carbon to the bacterioplankton. Uptake of dissolved free amino acids was responsible for 52 to 62% of the gross bacterial production. Thus, amino acids constituted a significant proportion of the EOC. Zooplankton (< 50 µm) grazing on algae and bacteria accounted only for a minor proportion of the particulate production in May. Circumstantial evidence is presented that suggests the chrysophycean alga Dinobryon was the most important bacterial remover.The results clearly demonstrated EOC release and bacterial metabolism to be key processes in pelagic carbon cycling in this oligotrophic lake.     

Zooplankton feeding on algae and bacteria under ice in Lake Druzhby,East Antarctica

The feeding of the cladoceran Daphniopsis studeri on algae and bacteria was investigated under ice in an ultra-oligotrophic Antarctic lake from late autumn (May) to early spring (October) in 2004. D. studeri fed on both algae and bacteria with estimated filtering rates of 0.048 and 0.061 l ind⁻¹ day⁻¹), respectively. Algae seemed to be the major food resource for the D. studeri population, however at times of low algal densities the bacterioplankton represented an important alternative food resource. The D. studeri grazing impact on the algal and bacterial standing stock was in general low (0.6–4.6% removed per day), but during the winter period this organism can remove up to 34% of the bacterial production (BP). At times D. studeri grazing can temporarily have a significant impact on the BP rates, though their impact was relatively low when compared to viral-induced bacterial mortality in the lake.     

Potential capacity of coprecipitation of dissolved organic carbon (DOC) with iron(III) precipitates

Hypolimnetic anoxic water of Lake Onogawa was subjected to aeration experiments. When the samples were agitated by magnetic stirrers for 24 h, dissolved oxygen increased from 0 to more than 7.6 mg l⁻¹, dissolved iron decreased from 98% to about 5% of the initial total iron, and from 32% to 48% of the dissolved organic carbon (DOC) disappeared. On the other hand, when the anoxic waters were left unstirred, dissolved oxygen increased from 0 to 2.2 mg l⁻¹, dissolved iron decreased from 98% to 31%, and 20% of the DOC disappeared within 48 h. Further 24-h incubation had little effect on the DOC loss, although dissolved oxygen increased to 3.9 mg l⁻¹ and dissolved iron decreased to 5%. These rates of DOC disappearance are too large to be explained by bacterial decomposition. It is quite conceivable that a part of the DOC is coprecipitated with iron(III) precipitates. When Fe(II) in the anoxic hypolimnion is oxidized by autumnal water mixing, probably anoxic water is mixed with aerobic water. The anoxic water must receive oxygen from the aerobic water during this mixing and be simultaneously diluted with the aerobic water. Because the present experimental conditions, especially the stirred one, significantly differ from in situ conditions, the present results are thought to be a potential capacity of DOC coprecipitation.     

Fungal control of nitrous oxide production in semiarid grassland

Fungi are capable of both nitrification and denitrification and dominate the microbial biomass in many soils. Recent work suggests that fungal rather than bacterial pathways dominate N transformation in desert soils. We evaluated this hypothesis by comparing the contributions of bacteria and fungi to N₂O production at control and N fertilized sites within a semiarid grassland in central New Mexico (USA). Soil samples were taken from the rhizosphere of blue grama (B. gracilus) and the microbiotic crusts that grow in open areas between the bunch grasses. Soils incubated at 30% or 70% water holding capacity, were exposed to one of three biocide treatments (control, cycloheximide or streptomycin). After 48 h, N₂O and CO₂ production were quantified along with the activities of several extracellular enzymes. N₂O production from N fertilized soils was higher than that of control soils (165 vs. 41 pmol h⁻¹ g⁻¹), was higher for crust soil than for rhizosphere soil (108 vs. 97 pmol h⁻¹ g⁻¹), and increased with soil water content (146 vs. 60 pmol h⁻¹ g⁻¹). On average, fungicide (cycloheximide) addition reduced N₂O production by 85% while increasing CO₂ production by 69%; bactericide (streptomycin) reduced N₂O by 53% with mixed effects on CO₂ production. N₂O production was significantly correlated with C and N mineralization potential as measured by assays for glycosidic and proteolytic enzymes, and with extractable nitrate and ammonium. Our data indicate that fungal nitrifier denitrification and bacterial autotrophic nitrification dominate N transformation in this ecosystem and that N₂O production is highly sensitive to soil cover, N deposition and moisture.     

Marine bacterial community structure resilience to changes in protist predation under phytoplankton bloom conditions

To test whether protist grazing selectively affects the composition of aquatic bacterial communities, we combined high-throughput sequencing to determine bacterial community composition with analyses of grazing rates, protist and bacterial abundances and bacterial cell sizes and physiological states in a mesocosm experiment in which nutrients were added to stimulate a phytoplankton bloom. A large variability was observed in the abundances of bacteria (from 0.7 to 2.4 × 10⁶ cells per ml), heterotrophic nanoflagellates (from 0.063 to 2.7 × 10⁴ cells per ml) and ciliates (from 100 to 3000 cells per l) during the experiment (∼3-, 45- and 30-fold, respectively), as well as in bulk grazing rates (from 1 to 13 × 10⁶ bacteria per ml per day) and bacterial production (from 3 to 379 μg per C l per day) (1 and 2 orders of magnitude, respectively). However, these strong changes in predation pressure did not induce comparable responses in bacterial community composition, indicating that bacterial community structure was resilient to changes in protist predation pressure. Overall, our results indicate that peaks in protist predation (at least those associated with phytoplankton blooms) do not necessarily trigger substantial changes in the composition of coastal marine bacterioplankton communities.