Host-specificity and dynamics in bacterial communities associated with Bloom-forming freshwater phytoplankton

Abstract

Many freshwater phytoplankton species have the potential to form transient nuisance blooms that affect water quality and other aquatic biota. Heterotrophic bacteria can influence such blooms via nutrient regeneration but also via antagonism and other biotic interactions. We studied the composition of bacterial communities associated with three bloom-forming freshwater phytoplankton species, the diatom Aulacoseira granulata and the cyanobacteria Microcystis aeruginosa and Cylindrospermopsis raciborskii. Experimental cultures incubated with and without lake bacteria were sampled in three different growth phases and bacterial community composition was assessed by 454-Pyrosequencing of 16S rRNA gene amplicons. Betaproteobacteria were dominant in all cultures inoculated with lake bacteria, but decreased during the experiment. In contrast, Alphaproteobacteria, which made up the second most abundant class of bacteria, increased overall during the course of the experiment. Other bacterial classes responded in contrasting ways to the experimental incubations causing significantly different bacterial communities to develop in response to host phytoplankton species, growth phase and between attached and free-living fractions. Differences in bacterial community composition between cyanobacteria and diatom cultures were greater than between the two cyanobacteria. Despite the significance, major differences between phytoplankton cultures were in the proportion of the OTUs rather than in the absence or presence of specific taxa. Different phytoplankton species favoring different bacterial communities may have important consequences for the fate of organic matter in systems where these bloom forming species occur. The dynamics and development of transient blooms may also be affected as bacterial communities seem to influence phytoplankton species growth in contrasting ways.

Figure 1. Growth curves of Aulacoseira granulata, Cylindrospermopsis raciborskii and Microcystis aeruginosa.

Lines represent absorbance and symbols represent chlorophyll a concentration. Dashed lines and open symbols are axenic* cultures (−), and solid lines and symbols are non-axenic cultures (+). Error bars represent standard deviation of 3 independent replicates. * C. raciborskii was not axenic. Differences in growth curves were significant to Aulacoseira granulata (ANCOVA, p<0.001, F = 31.82) and Cylindrospermopsis raciborskii (ANCOVA, p<0.001, F = 16.34). No significant effect was observed to Microcystis aeruginosa (ANCOVA, p = 0.11, F = 2.62).

Figure 2. Scanning electron microscopy photomicrograph of the microalgae without (left) and with (right) bacterial inoculum.

Aulacoseira granulata (a,b), Microcystis aeruginosa (c,d) and Cylindrospermopsis raciborskii (e,f) cultures without (a, c, e) and with (b,d,f) environmental bacterial inoculum. Black arrows indicate some bacterial (non-phytoplankton) cells.

Figure 3. Heatmap of the 10 most abundant specific OTUs associated with each studied phytoplankton species.

The non-resampled data was used and only the OTUs that occurred in the 3 independent replicates of one phytoplankton species and that had less than 2 reads in the treatments of the other two were considered specific. Bac_In is the bacterial inoculum; ab indicates attached bacteria and fb, free-living bacteria; dxx indicates the day of sampling. Frequencies are given by relativizing OTUs against the total number of reads of the sample, showing their low proportion. Taxonomic affiliation of two classification databases is shown after identification number: rdp/FW.

Figure 4. Non-metric multidimensional scaling (NMDS) plot showing differences among bacterial communities by phytoplankton species, fraction and growth phase.

Solid and open symbols represent, respectively, free-living and attached communities. The crosses represent Cylindrospermopsis raciborskii controls. 1 = lag or beginning of exponential growth phase, 2 * = exponential growth phase, 3* = stationary growth phase. * To C. raciborskii 2 and 3 were stationary and senescent phases, respectively.

Figure 5. Heatmap displaying the 51 most abundant OTUs after resampling.

Taxonomic affiliation of two classification databases is shown after identification number: rdp/FW. A and F indicate attached and free-living communities, respectively, and dxx indicates day of sampling. Letters in brackets indicate the phytoplankton species in which that OTU was within the most abundant: a, Aulacoseira granulata (18 OTUs); c, Cylindrospermopsis raciborskii (20 OTUS); m, Microcystis aeruginosa (20 OTUs). All these OTUs, except #616, were present in triplicates in at least one treatment of the phytoplankton species where they occurred.

Figure 6. Proportion of bacterial Classes in the replicates of each treatment of the 3 phytoplankton species.

Naïve Bayesian classification was used. Bac_In is the bacterial inoculum. Numbers 1–3 represent the replicate, ab and fb indicate attached and free-living communities respectively, Ct is control and dxx is the sampling day.