Changes in bacterioplankton composition under different phytoplankton regimens

Abstract

The results of empirical studies have revealed links between phytoplankton and bacterioplankton, such as the frequent correlation between chlorophyll a and bulk bacterial abundance and production. Nevertheless, little is known about possible links at the level of specific taxonomic groups. To investigate this issue, seawater microcosm experiments were performed in the northwestern Mediterranean Sea. Turbulence was used as a noninvasive means to induce phytoplankton blooms dominated by different algae. Microcosms exposed to turbulence became dominated by diatoms, while small phytoflagellates gained importance under still conditions. Denaturing gradient gel electrophoresis (DGGE) of 16S rRNA gene fragments showed that changes in phytoplankton community composition were followed by shifts in bacterioplankton community composition, both as changes in the presence or absence of distinct bacterial phylotypes and as differences in the relative abundance of ubiquitous phylotypes. Sequencing of DGGE bands showed that four Roseobacter phylotypes were present in all microcosms. The microcosms with a higher proportion of phytoflagellates were characterized by four phylotypes of the Bacteroidetes phylum: two affiliated with the family Cryomorphaceae and two with the family Flavobacteriaceae. Two other Flavobacteriaceae phylotypes were characteristic of the diatom-dominated microcosms, together with one Alphaproteobacteria phylotype (Roseobacter) and one Gammaproteobacteria phylotype (Methylophaga). Phylogenetic analyses of published Bacteroidetes 16S rRNA gene sequences confirmed that members of the Flavobacteriaceae are remarkably responsive to phytoplankton blooms, indicating these bacteria could be particularly important in the processing of organic matter during such events. Our data suggest that quantitative and qualitative differences in phytoplankton species composition may lead to pronounced differences in bacterioplankton species composition.

FIG. 1.

Bacterial abundance (A) and DGGE profiles of bacterial assemblages (B) in the control experiment without phytoplankton. Seawater cultures with bacteria only (fraction < 0.8 μm) were maintained under turbulent (T) and still (S) conditions, with two replicates of each treatment (a and b). Values are the means ± standard deviations (error bars) of duplicate cultures. Community DNA was collected on day 0 (initial) and day 4. The sampling time is shown in days (d).

FIG. 2.

Dynamics in microbial growth variables during phytoplankton bloom experiment 1. (A) Chl a concentrations. (B) Bacterial production estimated by leucine incorporation. (C) Bacterial abundance. Data were from seawater microcosms including the microbial community (fraction < 150 μm) incubated under turbulent (T) and still (S) conditions. Microcosms received nutrients as a batch addition on day 0 (T_B and S_B) or as daily additions (T_D and S_D). Values are the means ± standard deviations (error bars) for samples from duplicate microcosms maintained until day 6. Chl a concentrations are shown in micrograms per liter, bacterial production is shown in micrograms of carbon per liter per day, and the sampling time is shown in days (d).

FIG. 3.

Activities of the four hydrolytic ectoenzymes measured during phytoplankton bloom experiment 1 (left panels) and cell-specific hydrolysis rates obtained by normalizing enzyme activities to bacterial abundance (right panels). Enzyme activity is shown in micromoles per liter per hour, and the hydrolysis rate is shown in femtomoles per cell per hour. The insert graph shows alkaline phosphatase activities (APA) normalized to total bacterial and algal biomass (expressed as micromoles per microgram of C per liter). Microcosms are labeled as described in the legend to Fig. 2. Values are the means ± standard deviations (error bars) for samples from duplicate subsamples. AMA, aminopeptidase activity; d, days.

FIG. 4.

Dendrogram comparing DGGE fingerprints of the bacterial assemblages in the turbulent (T) and still (S) microcosms during phytoplankton bloom experiment 1. The subscript letters indicate batch (T_B and S_B) or daily (T_D and S_D) mode of nutrient addition. The numbers after the microcosm are the day the sample was taken. expt., experiment.

FIG. 5.

Bacterioplankton composition during phytoplankton bloom experiment 1 visualized by DGGE of PCR-amplified partial 16S rRNA genes. Bands that were cut from the gel and sequenced are indicated by their codes at the sides of the gel. Bands with S or T prefix, followed by band number, denote phylotypes mainly found in the still or turbulent microcosms, respectively. Bands with ST prefix, followed by band number, indicate ubiquitous phylotypes. All bands in Table 2 were given the prefix MED. The sampling time is shown in days (d).

FIG. 6.

Phylogenetic tree depicting relationships among partial 16S rRNA gene sequences of Bacteroidetes phylotypes detected during phytoplankton bloom experiment 1 (shown in boldface type) compared to type species of representative genera in the phylum. The approximate position of phylotype MED-S4 is indicated by the short boldface broken line; the sequence was not included in the alignment due to its shorter length. Phylotypes found in natural and experimental algal blooms in previous articles are also included. Phylotype codes are as follows: S and T (this paper), AY and BY (61), AWS (72), BB (5), and DI (J. Pinhassi, unpublished data). Phylotype and isolate names are followed by their GenBank accession number. The scale bar depicts 0.1 substitution per nucleotide position.

FIG. 7.

Time course of Chl a concentration (A) and bacterial abundance (B) in phytoplankton bloom experiment 2. DGGE profiles of the bacterial assemblages (C). Seawater microcosms including the microbial community (fraction < 150 μm) were incubated under turbulent (T) or still (S) conditions. Microcosms without nutrient additions (T₀ and S₀) and enriched with N, P, and Si (T_{16, 24}, and S_{16, 24}). Values are the means ± standard deviations (error bars) of pooled data from the enriched microcosms. Community DNA was collected on day 0 (initial) and on day 4. Chl a concentration is shown in micrograms per liter. The sampling time is shown in days (d).