Relationship between dissolved organic carbon and bacterial community in the coastal waters of Incheon, Korea

Surface water samples were collected at three sampling sites in Yeongheung-do, Incheon, Korea, from April 2012 to October 2014 (Fig. 1). Sampling was conducted at St. 1 and St. 2 in 2012-2013. St. 3 was added as another sampling site in 2014. To analyze bacterial composition, both surface and bottom samples were collected from August 2013 to October 2014. Water depth data are shown in Table S1.

Figure 1

Temperature, salinity, dissolved oxygen (DO) and pH were determined using YSI 556 MPS (YSI, Yellow Springs, OH, USA). Water samples for inorganic nutrients were collected from the filtrate through Whatman GF/F filters and stored at -20°C until measured using an autoanalyzer. Chlorophyll a retained on the GF/F filters was determined fluorometrically (Strickland, Parsons 1972) and measured using a fluorometer (10-AU-005; Turner Inc., CA, USA). Dissolved organic carbon (DOC) samples were collected by passing ~30 ml of the filtrate through a pre-combusted (400°C for 8 hours) Whatman GF/F filter in a pre-combusted (400°C for 8 hours) TOC vial (Shimadzu, Kyoto, Japan) and fixed with 2% H₂PO₄ (final concentration). The samples were then stored at 4°C in a fridge until measured using a TOC Analyzer (Shimadzu, Kyoto, Japan).

Samples were fixed with 2% glutaraldehyde (Sigma-Aldrich, St. Louis, MO, USA; final concentration) for 10 min in the dark at room temperature and then stored at 4°C in the dark storage room of the Department of Life Science, Hanyang University. Bacterial abundance was determined via DAPI (4’-6-diamidino-2-phenylindole, Sigma-Aldrich) staining of a sample filtered using a 0.2 μm GTTP Millipore filter membrane (Millipore Filter Corporation, Cork, Ireland). The abundance was then determined using an Olympus epifluorescence microscope (XC10, Olympus, Tokyo, Japan) (Porter, Feig 1980) under 1000× magnification.

Samples were fixed using Lugol’s solution (Throndsen 1978) (final concentration of approximately 2%) and stored at 4°C. A Zeiss stereomicroscope (Axioplan microscope, Zeiss, Oberkochen, Germany) or Olympus stereomicroscope (XC10, Olympus, Tokyo, Japan) were then used to count phytoplankton under 200× magnification. Both microscopes were equipped with Nomarski differential interference contrast (DIC) optics.

Fifty milliliter subsamples were collected and filtered through a 0.2 μm GTTP Millipore filter membrane (Millipore Filter Corporation) to analyze the free-living bacterial community. Filters were transferred to a 2-ml Eppendorf tube containing 0.8 ml of extraction buffer (100 mM of Tris-HCl with pH 8, 100 mM of Na₂-EDTA, 100 mM of sodium phosphate with pH 8, 1.5 M NaCl and 1% CTAB) and stored at -80°C until DNA extraction (Wang et al. 2014). DNA extraction was performed following the EX DNA extraction protocol (Park et al. 2014). Eppendorf tubes (2-ml) containing the membrane filters were immersed in liquid N₂ until completely frozen and then thawed in 65°C water bath. This freeze-thaw process was repeated three times. After adding 8 μl of proteinase K (10 mg ml⁻¹ in TE buffer), the samples were incubated at 37°C for 30 min. Following the addition of 80 μl of 20% sodium dodecyl sulfate (SDS), which was prepared using double-distilled water, the samples were incubated at 65°C for 2 h with occasional stirring. After adding an equal volume of chloroform-isoamyl alcohol (24:1) and shaking the samples, they were centrifuged at 10,000 × g for 5 min. Then aqueous phase of the mixture was transferred to a fresh Eppendorf tube where 3 M sodium acetate (88.8 μl; pH 5.2), prepared using double-distilled water, was added. Next, 586.08 μl of isopropanol (≥99%) was added into the tube. Following centrifugation at 14,000 × g for 20 min, the supernatant was decanted, and 1 ml of cold 70% ethanol was added and the Eppendorf tube was vortexed. Samples were then centrifuged at 14,000 × g for 15 min. Pellets were air-dried before being dissolved in 100 μl of TE buffer (10 mM Tris-HCl, 1 mM EDTA; pH 8). DNA samples were stored at -20°C.

The 341F (5’-CCTACGGGAGGCAGCAG-3’) primer with a GC-clamp (5’-CGCCCGCCGCGCCCCGCGCCCGTCC CGCCGCCCCCGCCCG-3’) attached to the 5’ end and the 518R (5’-ATTACCGCGGCTGCTGG-3’) primer were used to amplify the nuclear SSU rDNA gene (v3 rDNA) (Ferrari, Hollibaugh 1999). PCR reactions were performed in 50-μl reaction mixtures using TaKaRa EX Taq™ (TaKaRa, Shiga, Japan). A touchdown polymerase chain reaction was performed using the Bio-Rad iCycler (Bio-Rad, Hercules, CA, USA) under the following conditions: initial denaturation at 94°C for 5 min, 30 cycles of 94°C for 1 min, annealing for 1 min at 65-55°C (stepping down 1 degree every cycle) for 10 cycles and 1 min at 55°C for 20 cycles, and extension at 72°C for 1 min, and then a final elongation step of 72°C for 15 min. Bands corresponding to the amplified product were excised following gel electrophoresis.

DGGE was performed using the Bio-Rad Dcode universal mutation detection system (Bio-Rad Laboratories); 30 μl aliquots of PCR products were applied to 8% polyacrylamide gels with gradients that contained 40 to 55% urea and 40% deionized formamide. Gels were run for 15 hours at 50 V in 1×TAE buffer. After electrophoresis, gels were stained with ethidium bromide and photographed via UV-transillumination. The banding patterns for each of the DGGE profiles were processed, normalized, and statistically analyzed using the Quantity One software (version 4.6.2; Bio-Rad Laboratories, Inc.). The relationships among the samples were determined by cluster analysis (UPGMA) of banding patterns using the Phylogenetic Tree Quick Guide to the program.

Correlation analysis was performed using the Statistical Package for Social Sciences (SPSS, v. 20.0) based on abiotic and biotic data collected from 2012 to 2014 (St. 1 and 2: 2012-2014; St. 3: 2014). The relationships between bacteria/phytoplankton abundance and environmental parameters were assessed via redundancy analysis (RDA) using CANOCO 5.0 based on the same data set used for correlation analysis from 2012 to 2014.

Temperature, salinity, dissolved oxygen (DO), pH, chlorophyll a (Chl-a), DOC and nutrient were measured at St. 1, 2 and 3 during 2012-2014 (St. 1 and 2: 2012-2014; St. 3: 2014) (Fig. 2–4). These factors exhibited no significant differences among the sampling sites during the study period. The temperature ranged from 5.22 (April 2012) to 26.18°C (August 2013), displaying fluctuations in different months (Fig. 2A). In general, salinity and pH did not change significantly over the study period (Figs 2B-C). However, the pH increased abruptly at St. 1 in May 2013 (Fig. 2C). The average DO concentration was higher in 2013 than in 2012 and 2014. The DO concentration primarily decreased in 2013 (April to October) and 2014 (August to October), and steadily increased in 2012 (April to December) (Fig. 2D). The Chl-a concentration gradually decreased from August to October in both 2013 and 2014 (Fig. 2E) but did not exhibit a significant change in 2012. Interestingly, the notable differences in Chl-a were observed in April 2012 (0.0365 μg l⁻¹) and April 2013 (20.4541 μg l⁻¹), even though these samples were isolated during the same month of both years. DOC concentrations at St. 2 increased from May (598.0 μg C l⁻¹) to December (1164.0 μg C l¹) in 2012, but were generally lower than those at St. 1. However, no significant difference in DOC was observed between the sites in 2013 and 2014 (Fig. 4B). DOC concentration decreased from August (St. 1: 2646.0 μg C l⁻¹; St. 2: 2712.0 μg C l⁻¹) to October (St. 1: 1397.5 μg C l⁻¹; St. 2: 1339.5 μg C l⁻¹) in 2013; however, the opposite trend was observed in 2014.

Figure 2

Nutrient (nitrite, nitrate, ammonium, phosphate, and silicate) concentrations were measured at St. 1, 2, and 3 (Fig. 3). Levels of these nutrients showed similar patterns of fluctuation among the studied sites. As a result, most nutrient concentrations, with the exception of phosphate, were similar among the sampling sites (Fig. 3). Phosphate concentrations were different between St. 1 and St. 2 in December 2012 and May 2013 (Fig. 3A). The phosphate concentration increased in the fall (September 2013 and 2014) and early winter (December 2012) during the study period, however, the increase in 2014 was not as high as the increase in 2012 and 2013. The ammonium concentration was primarily lower than 2 μM, but peaked in August 2013 with maximum concentrations of 3.9641 μM (St. 1) and 3.9047 μM (St. 2) (Fig. 3B). Nitrite did not vary significantly (0.0456-0.4397 μM) in 2012, but increased abruptly in September 2013 (St. 1: 5.992 μM and St. 2: 7.146 μM) and 2014 (St. 1: 3.040 μM and St. 2: 2.800 μM) (Fig. 3C). Nitrate and silicate concentrations displayed similar patterns of fluctuation in 2012 and 2013, but not in 2014 (Figs 3D-E). These concentrations increased continuously from May to September 2013 and then slightly decreased in October, except for the nitrate concentration at St. 1. Nitrate and silicate concentrations exhibited opposite patterns of fluctuation in 2014 (Figs 3D-E). In September 2014, the nitrate concentration slightly decreased, while the silicate concentration increased sharply at the sampling sites. The N:P ratio was generally below 16 for 2012 and 2013, and above 16 in 2014 (Fig. 3F). Altogether these results imply that environmental conditions were significantly different between 2012-2013 and 2014: i) nitrogen was the limiting factor in 2012-2013 and ii) phosphate was the limiting factor in 2014.

Figure 3

Bacterial abundance values did not differ among the sampling sites (St. 1, 2 and 3), but they could display dynamic patterns of monthly differences (Fig. 4A). Bacterial abundance ranged from 1.96 × 10⁵ cells ml⁻¹ to 1.41 × 10⁶ cells ml⁻¹ over the study period. Bacterial abundance peaks were observed in April and June 2012, while the only peaks observed in 2013 and 2014 occurred in August (Fig. 4A). Notably, DOC and bacteria exhibited extremely similar patterns in 2012 and 2013 (Figs 4A-B). DOC and bacteria peaked in April and June 2012 at St. 1. In addition, simultaneous DOC and bacteria peaks were observed at all sampling sites in August 2013 (St. 1 and 2) (Figs 4A-B). These results imply that bacterial abundance is closely related to DOC.

DGGE analysis was conducted to determine the effects of DOC on bacterial composition. Samples from August-October 2013 and August-October 2014 were analyzed. For sampling sites at low depth (Table S1) and in the tidal zone, we expected bottom samples to have a similar bacterial composition and therefore we analyzed these samples at the same time. Additionally, we used DGGE to analyze the bacterial composition of samples collected from August to October 2013 as these samples showed the closest relationship between DOC concentration and bacterial abundance (Figs 4A-B), as well as apparent fluctuations from August to October 2013 at St. 1 and St. 2 (see Figs 5 and 6). Because DOC concentrations were similar in August-September 2014, and the bacterial composition in the samples does not form clades strictly related to particular months (Fig. S1), we did not focus on the samples from 2014 in this study. Bacterial composition analysis and DGGE clustering results for the 2013 samples revealed two clades: one clade formed by samples from August 2013 and the other clade formed by samples from September and October 2013 (Figs 5 and 6). No significant differences were observed between the surface and bottom water samples, except for the surface sample at St. 2 in September 2013. The September samples were closer to the October samples than the August 2013 samples (Fig. 6), indicating that bacterial composition in September and October were more similar. Therefore, DOC may affect not only the bacterial abundance, but also the bacterial composition.

Figure 4

Figure 5

Figure 6

Phytoplankton displayed similar fluctuations at all sampling sites during the study period (Fig. 4C). Phytoplankton abundance ranged from 17 to 3890 cells ml⁻¹ and exhibited a gradually decreasing pattern each year. In addition, Chl-a and phytoplankton displayed similar fluctuation patterns, as shown in Figs 2E and 4C.

Redundancy analysis (RDA) was performed to analyze the relationships between abiotic and biotic factors (Fig. 7). DOC was found to be positively associated with bacterial abundance. Temperature, pH, NO₂^-, and NO₃^- also displayed weak positive correlation with bacterial abundance. These results were confirmed via correlation analysis between abiotic and biotic factors (Table 1). The Pearson correlation between DOC and bacteria was 0.625 (p<0.01), while the Pearson correlation between pH and bacteria was 0.291 (p<0.05). However, none of the other factors was significantly correlated with DOC. RDA and Pearson correlation analyses confirmed a strong relationship between DOC concentration and bacterial abundance.

Figure 7

Table 1

Correlations between physicochemical and biotic factors measured during field monitoring