Bacterial response to glucose addition: growth and community structure in seawater microcosms from North Pacific Ocean

Environmental parameters

Sampling locations, air temperature, water temperature, water depth, salinity, nutrient concentrations (NO₃-N, NO₂-N, NH₄-N, SiO₄, PO₄-P), and incubation temperatures are shown in Table 1. The air and water temperatures of the studied locations were 11 and 16.6 °C, 3.1 and 3.8 °C, 3.1 and 3.7 °C, 24.5 and 25.9 °C, 24.5 and 18.8 °C, respectively in the Kuroshio Current, SPG surface layer, SPG chlorophyll maximum zone, STG surface layer, and STG chlorophyll maximum zone. At SPG, the values of different parameters were quite similar (p = 0.62, two-tail t-Test; at 5% level of significance) between surface (5 m) and chlorophyll maximum (37 m), indicating the vertical mixing in the upper water column. At STG, the values were relatively different (p = 0.39, two-tail t-Test; at 5% level of significance) between surface (5 m) and chlorophyll maximum (125 m), suggesting the vertical stratification of the water column. The in-situ (water) temperatures (6.4 °C, 0.2 °C, 0.3 °C and 4.2 °C) were lower than the incubation temperatures compared to those of Kuroshio Current, SPG surface layer, SPG chlorophyll maximum zone, and STG chlorophyll maximum zone, while 2.9 °C higher than the incubation temperature of the STG surface layer. Nutrient assays revealed a big difference in nutrient concentrations between SPG and STG; the waters from the station STG were nutrient-poor. The incubation temperatures of the onboard microcosms were 23 ± 1 °C, ~ 4 °C, and 23 ± 1 °C in the case of Kuroshio Current, SPG, and STG, respectively (Table 1).

Bacterial cell densities and cell volumes

At initial incubation periods (12 h to 24 h), the cell densities between the glucose-amended and non-treated microcosms were similar (p = 0.74, two-tail t-Test; at 5% level of significance). Highly significant differences (p < 0.001, two-tail t-Test; at 5% level of significance) was observed later after 36 h of incubation in Kuroshio and STG samples (Table 2). In case of SPG, the density of bacteria increased slowly. In surface seawater, glucose treated microcosms showed higher density than non-treated control at 96 and 120 h. However, in chlorophyll maximum seawater, no difference was noticed between the two. In case of STG, the addition of glucose resulted in marked increase in the density after 36 h in both seawater samples from surface and chlorophyll maximum layer (Table 2).

Bacterial growth as cell volume change (mean ± standard deviation) was also measured for each microcosm (Table 3). In case of Kuroshio seawater, the average cell volume in treated microcosm increased by almost 20-fold after 72 h incubation, while much less increase was observed in the control microcosm. In case of SPG, cell volumes were virtually unchanged even after 120 h of incubation in both surface and chlorophyll maximum seawater microcosms. In the STG seawaters, cell volumes were almost similar up to 24 h and increased primarily in the treated microcosms for both surface and chlorophyll maximum water.

Succession and modification in bacterial community structure

After sequencing all samples, a total of 15,163 operational taxonomic units (OTUs), which were available for molecular phylogenetic analysis, were obtained from 659,424 raw sequence data (supplementary Figs. 1, 2, and 3 showing the rarefaction curves). The community compositions were shown in Fig. 2 at phylum or class levels (e.g., Bacteroidia, Alphaproteobacteria, and so on). The phyla, which contributed less than 1% of the total abundance, were combined and referred to as “Others,” and those with no affiliation were referred to as “Unclassified.” At the initial stages, the Kuroshio Current and SPG seawaters were dominated by Bacteroidota (class Bacteroidia) and Proteobacteria (classes Alphaproteobacteria and Gammaproteobacteria). In contrast, STG waters were dominated by Cyanobacteria, Marinimicrobia, and Proteobacteria (classes Alphaproteobacteria and Gammaproteobacteria) (Fig. 2). Bacterial communities changed in the treated microcosms collected from Kuroshio Current and STG visibly after 36 h (Fig. 2A,C); the class Gammaproteobacteria proliferated in the treated microcosms. Bacteroidia decreased in early incubation periods of the Kuroshio Current but increased after 48 h incubation in both the treated and control microcosms (Fig. 2A). For the microcosms collected from SPG, minor or no community changes were observed for both the surface and chlorophyll maximum seawaters regardless of the nutritional amendment. In addition, linear discriminant analysis effect size (LEfSe) measurement was conducted to detect any taxa with a significant differential abundance between “control” and “treated” samples of the SPG samples (surface and chlorophyll maximum). No significant (p value cutoff p < 0.05; LDA score, log10 > 3) features were identified with the given criteria for both SPG-surface and SPG-Chlorophyll maximum (supplementary Table 2).

The major groups (classes/orders) in the most dominant phylum, Proteobacteria, were separately shown in Fig. 3. Within the phylum, the composition of the bacterial community gradually changed into a simplified form over time; the class Gammaproteobacteria become dominant (Fig. 3). Within the class Gammaproteobacteria, the genus Vibrio (order: Enterobacterales, family: Vibrionaceae) were abundant in Kuroshio, while Alteromonas (order: Enterobacterales, family: Alteromonadaceae) in STG (Fig. 3A,C). Another notable feature of the two major gyre systems (SPG and STG) in the mid-latitudes of the Northern Pacific Ocean is the dominancy of the orders SAR11 (Alphaproteobacteria) and Pseudomonadales (Gammaproteobacteria; mostly SAR86_clade). These two groups were later outnumbered in the case of STG but remained almost unchanged in SPG (Fig. 3B,C). These two groups were also dominant in seawater from Kuroshio; at later periods of incubation, especially the group SAR11 showed a gradual decrease in the treated microcosms (Fig. 3A).

Species richness and diversity indices

The Chao1 index was calculated for species richness (Table 4) and rarefaction curves were considered for diversity indices (supplementary Figs. 1, 2, and 3). In general, the species richness was relatively higher in unfiltered water samples (samples just after collection), which was reduced after filtration, observed at the initial stages (12 h) of incubation and increased again at 24–48 h. The richness was significantly reduced (p = 0.0006 and t = 2.26, two-tail t-Test; at 5% level of significance) from the initial period (12 h) to the final periods of incubation (60–96 h) for different microcosms regardless of treatments. However, the Chao1 index values were relatively higher in the control microcosms for almost all the cases at the end of the incubation periods (Table 4). The rarefaction curves also showed that the glucose treated samples were less diverse compared to the controls (supplementary Figs. 1, 2, and 3).

Influence of environmental/incubation parameters on bacterial community structure

Meta-NMDS and RDA were performed based on the relative abundance data of the obtained OTUs to clarify the influence and/or establish relationship between incubation conditions (time, temperature and glucose treatment) to the bacterial community structure. Meta-NMDS analysis with all the samples showed that they were clustered according to the sampling stations (r² = 0.86 and P = 0.001, based on 1000 permutations), incubation temperature (r² = 0.79 and P = 0.001, based on 1000 permutations), and salinity (r² = 0.82 and P = 0.001, based on 1000 permutations) (Fig. 4A). RDA also showed similar and significant influences of sampling stations (r² = 0.54 and P = 0.001, based on 1000 permutations), glucose treatment (r² = 0.67 and P = 0.001, based on 1000 permutations), incubation temperature (r² = 0.71 and P = 0.001, based on 1000 permutations), and salinity (r² = 0.76 and P = 0.001, based on 1000 permutations) to the bacterial community structure (Fig. 4B).

The incubation periods significantly influenced (r² = 0.7 and P = 0.02, based on 1000 permutations) the clustering of the samples of Kuroshio than that of the glucose treatment (r² = 0.3 and P = 0.1, based on 1000 permutations) (Fig. 5A). Similar analysis with the samples from SPG also showed significant influence of incubation periods (r² = 0.4 and P = 0.01, based on 1000 permutations), and seawater source (surface and chlorophyll maximum zone) (r² = 0.2 and P = 0.02, based on 1000 permutations) with no particular arrangement pattern for the glucose treatment (r² = 0.2 and P = 0.2, based on 1000 permutations) (Fig. 5B). However, the NMDS with the samples of STG showed clear arrangement according to the water source (r² = 0.5 and P = 0.001, based on 1000 permutations) and glucose treatment (r² = 0.7 and P = 0.001, based on 1000 permutations). The incubation periods also significantly influenced (r² = 0.3 and P = 0.05, based on 1000 permutations) the clustering of the samples of the STG (Fig. 5C).

Top twenty most influential OTUs, responsible for differentiating the community structure between the samples, were obtained (based on the Bray–Curtis dissimilarity indices) through similarity percentage (SIMPER) analysis (supplementary Table 3). Their (top 20 most influential OTUs) relative abundance data were applied for redundancy analysis (RDA) to establish relationship between community composition and environmental features. The influential OTUs were mostly from the phylum Proteobacteria such as—OTU01 (SAR11_clade (Alphaproteobacteria)), OTU02 (Alteromonas (Gammaproteobacteria)), OTU07 (Alteromonas (Gammaproteobacteria)), OTU16 (Photobacterium (Gammaproteobacteria)), OTU13(Aurantivirga (Bacteroidota)), OTU05 (Prochlorococcus (Cyanobacteria)), OTU08 (SAR11_clade (Alphaproteobacteria)), OTU09 (SAR86_clade (Gammaproteobacteria)), OTU06 (SAR86_clade (Gammaproteobacteria)), OTU20 (Alcanivorax (Gammaproteobacteria)), OTU17 (Alteromonas (Gammaproteobacteria)), OTU11 (SUP05_cluster (Gammaproteobacteria)), OTU04 (SAR11_clade (Alphaproteobacteria)), OTU14 (NS5_marine_group (Bacteroidota)), OTU19 (SAR11_clade (Alphaproteobacteria)), OTU18 (SAR11_clade (Alphaproteobacteria)), OTU22 (Rhodobacteraceae (Alphaproteobacteria)), OTU03 (Vibrio (Gammaproteobacteria)), OTU23 (SAR11_clade (Alphaproteobacteria)), and OTU35 (unclassified bacteria). Based on the relative abundances of these dominant OTUs, their association to environmental factors was examined by redundancy analysis (RDA) (Fig. 6). RDA showed that the samples are separated by treatments (r² = 0.66 and P = 0.001, based on 1000 permutations) and stations (r² = 0.53 and P = 0.001, based on 1000 permutations). The incubation temperature (r² = 0.76 and P = 0.001, based on 1000 permutations) and salinity (r² = 0.70 and P = 0.001, based on 1000 permutations) also had strong influences on the clustering of the samples, while the incubation periods showed none (r² = 0.02 and P = 0.62, based on 1000 permutations) (Fig. 6).

Source: Ecology - nature.com