Assimilation of Polysaccharides and Glucose by Major Bacterial Groups in the Delaware Estuary

Preparation of [³H]EPS.

EPS was prepared using an axenic culture of the heterotrophic diatom Nitzschia leucosigma (CCMP 2197) grown on [³H]glucose as the sole carbon source. The diatom culture was grown on 10 μM [³H]glucose (33 Ci/mmol; Amersham) overnight in the dark in filter-sterilized (0.2-μm-pore-size polycarbonate) and autoclaved Sargasso seawater with no additional nutrients. Cells were separated from the culture medium by centrifugation at 2,500 × g (20°C, 20 min), and the supernatant containing the EPS was transferred to a clean tube. The soluble EPS was separated from the culture medium by precipitation with ethanol (70% final concentration) overnight at −20°C (15). The precipitated EPS was resuspended in 70% ethanol (4°C) and precipitated again under the same conditions. This step, which was essential to remove residual monosaccharides, was repeated three times. The EPS was resuspended in 3% NaCl, radioassayed, and tested for carbohydrate concentrations (in glucose equivalents) by using the phenol-sulfuric acid assay (14).

Sample collection and preparation.

Surface water was collected from four locations along the Delaware Estuary in July 2003 (salinities of 0, 13, 21, and 26 practical salinity units [PSU]) and June 2004 (28 PSU). Water samples (30 ml) were incubated with 2 nM [³H]glucose (33 Ci/mmol; Amersham) for 2 h or with 1.5 μM [³H]EPS (concentration in glucose equivalents) for 14 h. Both treatments were incubated at the in situ temperature in the dark. Paraformaldehyde (2% final concentration) was added to killed controls 15 min prior to the addition of the ³H-labeled compounds. At the end of the incubation, samples were fixed with paraformaldehyde (2% final concentration), and all samples were stored at 4°C for 24 h. Samples were then filtered onto 0.2-μm-pore-size polycarbonate filters, which were kept at −20°C until analysis.

FISH and microautoradiography analysis.

FISH analysis was done using the following Cy3-labeled probes: Eub338 for Eubacteria (3); Alf968 for Alphaproteobacteria (20); Bet42a and Gam42a for Beta- and Gammaproteobacteria, respectively (33); CF319a for Cytophaga-like bacteria (32); HGC96a for Actinobacteria (37); and a suite of four probes for SAR11 bacteria (34). Unlabeled competitor probes were used for Beta- and Gammaproteobacteria (33). In addition, a negative probe was used to determine nonspecific binding (24).

Following FISH analysis, samples were subjected to microautoradiography as described by Cottrell and Kirchman (10). Glucose samples were exposed for 12 to 24 h, while EPS samples were exposed for 3 to 6 days. These exposure times were set so that the percentage of total cells associated with silver grains was in the range of 15 to 20%. This range was chosen because that was the maximum percentage of cells assimilating EPS that could be detected. At the end of the exposure time, the slides were developed as described previously (10). The samples were mounted with a 4:1 mixture of Citifluor (Ted Pella) and Vectashield (Vector Labs) containing 0.5 ng/μl of 4′,6′-diamidino-2-phenylindole (DAPI) stain and were covered with coverslips. Slides were stored at −20°C until microscopic analysis.

Total cells (DAPI stained), cells affiliated with a specific bacterial group (Cy3 labeled), and cells that assimilated the radiolabeled compound (with silver grains developed during microautoradiography) were counted using a semiautomatic microscope and image analysis as described previously (10). Data were collected from 30 fields of view, and the numbers of cells counted are indicated in Tables 1 and 2. To calculate the percentage of total silver grain area, the total silver grain area around the probe-positive cells was summed and divided by the sum of total silver grain area associated with DAPI-stained cells. Cell volumes were calculated using the algorithm described by Sieracki et al. (39).

Data expressed as percentages were arcsine transformed before analysis. Model II regressions were performed to examine the relationship between abundance and silver grain area. Cell volumes were log transformed before analysis.

Dominant bacterioplankton groups assimilating glucose and EPS.

The assimilation of glucose was dominated by a few groups that changed along the salinity gradient of the estuary. Actinobacteria accounted for nearly all of glucose assimilation (70%) in the freshwater (Fig. 2A). The contribution of this group decreased dramatically as salinity increased. The opposite trend was observed for Alphaproteobacteria. While this group contributed only 20% to glucose assimilation in the freshwater, Alphaproteobacteria accounted for 60% of assimilation at the highest salinity. The rest of the phylogenetic groups examined in this study assimilated less glucose. Beta- and Gammaproteobacteria and Cytophaga-like bacteria accounted for 15 to 20% of glucose assimilation, depending on location.

While a single group dominated glucose assimilation in the freshwater and saline ends of the estuary, more than one group assimilated EPS at each location. Both Betaproteobacteria and Actinobacteria accounted for 30% of EPS assimilation in the freshwater and were the dominant EPS users in this part of the estuary (Fig. 2B). Assimilation by both phylogenetic groups decreased to 5 to 20% as salinity increased. In the lower part of the estuary (21 to 28 PSU), different groups dominated EPS assimilation. In 21-PSU water, Alpha- and Gammaproteobacteria each accounted for 20% of total silver grain area. Cytophaga-like bacteria were the primary EPS users in 26-PSU water, followed by Alphaproteobacteria. Alphaproteobacteria were the main EPS consumers in the highest salinity, comprising 55% of the total silver grain area (Fig. 2B). Unlike the case for glucose assimilation, where the alphaproteobacterial contribution increased gradually along the estuary, this group used EPS quite similarly in most parts of the estuary (20 to 25% of total assimilation).

In addition to broad phylogenetic groups, we also examined the contribution of the SAR11 clade, a subgroup of the Alphaproteobacteria (19), to assimilation of glucose and EPS. The SAR11 clade was abundant (20 to 33% of all prokaryotes) in the saline part of the estuary (Fig. 3) but was near detection limits in freshwater (27). Glucose assimilation by the SAR11 clade was significantly higher than that of EPS (25% and 15%, respectively) in two of three stations (t test, P < 0.05). However, SAR11 bacteria assimilated these compounds less than expected based on abundance (Fig. 3).

Relationships between abundance and DOM assimilation.

We next examined whether assimilation by a phylogenetic group could be explained by its relative abundance in the community. About 55% of the variation in the assimilation of glucose was explained by the abundance of the bacterial groups (Fig. 4A). A smaller fraction of the variation in EPS assimilation (35%) was explained by abundance (Fig. 4B). However, when excluding the data points that were more than 20% off the 1:1 line (21% and 12% of the data points for glucose and EPS uptake, respectively), the slopes between the silver grain area and the relative abundance of each phylogenetic group were not different from 1 (0.890 ± 0.17 and 0.94 ± 0.36 for EPS and glucose, respectively).

A few bacterial groups assimilated glucose and EPS more than expected based on their abundance. For glucose assimilation, these groups include the Alphaproteobacteria in the lower part of the estuary (salinity of 21 to 28 PSU) and Actinobacteria in the freshwater (Fig. 4A). Although Betaproteobacteria, Gammaproteobacteria, and Cytophaga-like bacteria were abundant in some sites, these groups used glucose as expected or less than expected according to their abundance (Fig. 4A). Alphaproteobacteria assimilated more glucose than expected from their abundances in three locations (Fig. 4A), but this group assimilated more EPS than expected (by 2.5-fold) at only one location (Fig. 4B). Betaproteobacteria (by fourfold) and Cytophaga-like bacteria (by twofold) also assimilated EPS more than expected from their abundances in 28- and 26-PSU waters, respectively. The bacterial groups that assimilated EPS more than expected in the lower part of the estuary were also the groups contributing the most to EPS assimilation (Fig. 2B).

Importance of EPS and glucose assimilation to a phylogenetic group.

The contribution of a specific phylogenetic group to DOM fluxes does not necessarily reflect the importance of a compound to that group. To explore this question, we calculated the fraction of bacteria in each group that assimilated glucose or EPS, i.e., probe-positive cells with silver grains compared with probe-positive cells in that particular group (with and without silver grains). This is different from the case for the previous sections, which present the contribution of each bacterial group to EPS or glucose assimilation, i.e., probe-positive cells with silver grains divided by all cells with silver grains. For example, Alphaproteobacteria accounted for 35% of EPS assimilation in 26-PSU water, but only 12% of Alphaproteobacteria used EPS in this region of the estuary.

Overall, the fraction of bacteria in each group assimilating glucose (15 to 30%) was significantly higher than the fraction assimilating EPS (5 to 20%) (Fig. 5) (paired t test, P < 0.05). More Actinobacteria assimilated glucose than EPS at all locations. Similar fractions of cells affiliated with Alphaproteobacteria assimilated glucose and EPS in 28-PSU water. However, a higher cell fraction of this group assimilated glucose at the other sites. The fractions for the other phylogenetic groups alternated with salinity. For example, more Betaproteobacteria assimilated EPS in 28-PSU water, while more cells of this group used glucose in 21-PSU water.

The fraction of bacteria in groups assimilating either compound did not explain the abundance of these groups (r = 0.28 for EPS; r = 0.15 for glucose) (Fig. 6). Groups that had a high fraction of cells assimilating these compounds were not necessarily abundant. Similarly, groups that were abundant had unexpectedly low fractions of cells assimilating both compounds (Fig. 6). Although Actinobacteria were more abundant in the freshwater part of the estuary, the fractions of this group assimilating glucose were similar throughout the estuary (Fig. 6A). However, relatively more Actinobacteria assimilated EPS in the freshwater section (Fig. 6B). The abundance of Alphaproteobacteria and SAR11 bacteria assimilating both EPS and glucose was usually lower than expected based on the fraction of active cells. A higher fraction of Alphaproteobacteria assimilated both compounds only at 28 PSU.

Cell volumes of active and inactive bacteria.

Cell volumes may change during preparation of samples for the micro-FISH analysis. For that reason, we chose to examine the ratio of the volumes for active versus inactive cells. We assume that any changes in cell volume are cancelled out in the ratio.

The ratios between the volumes of substrate-assimilating and nonassimilating cells indicate that bacteria taking up glucose and EPS were larger than nonassimilating bacteria (t test, P < 0.05). EPS-assimilating bacteria were larger than nonassimilating bacteria by 20 to 44% (Table 1). Cytophaga-like bacteria had the highest ratio of assimilating to nonassimilating cell volumes in the EPS treatment; assimilating cells were 62% larger than nonassimilating ones. This was also the case for Cytophaga-like bacteria cells assimilating glucose (Table 2), in which the assimilating Cytophaga-like bacteria were 57% larger than the nonassimilating cells affiliated with this group. The other phylogenetic groups had smaller differences (range of 10 to 35%) in the glucose treatment. Overall, EPS-assimilating bacteria were larger than glucose-assimilating cells by 10 to 30%.