FIGURE 1 Sampling sites at (a) the Krka River Estuary and (b) the Wenchang River Estuary.
The water samples were collected in the Krka River Estuary from 4 to 9 September 2014 and in the Wenchang River Estuary from 8 to 10 May 2015. The samples were collected from the surface (0 m) using 5-litre Niskin bottles along the salinity gradient.
2.2Basic environmental analysis
The salinity and temperature of the water in the Krka River Estuary were measured on site using an HQ40D portable probe (Hach Lange, Germany), while a portable multiparameter WTW multi 350i probe (Geotech Environmental Equipment, Denver, USA) was used for the Wenchang River Estuary.
The samples (50 mL) for the analysis of ammonium (NH4+) were stabilized by the addition of 2 mL phenol solution (1 mol L-1; 95 % ethanol) (Ivančić and Degobbis, 1984) and stored in the dark at 4 °C. The samples (500 mL) for all other nutrients were stored at -20 °C. The concentrations of dissolved inorganic nitrogen (DIN=nitrate (NO3¯), nitrite (NO2¯), and ammonium (NH4+) and orthophosphate (PO43¯) were determined using spectrophotometric methods according to Strickland and Parsons (1972). Their method accuracies are ±3%, ±3%, ±5% and ±6%, respectively, while detection limits are 0.05 µmol L-1, 0.01 µmol L-1, 0.1 µmol L-1 and 0.05 µmol L-1, respectively.
2.3Pigment analysis
While Chlorophyll a (Chl a ) is used as a suitable proxy for phytoplankton biomass, many other phytoplankton pigments show chemotaxonomic associations that can be used to characterize phytoplankton assemblages (Gibb et al., 2000). Pigment-based measurements can be a good tool to determine phytoplankton diversity (Schlüter et al., 2016 and references therein). Analyses of phytoplankton diversity in estuaries using microscopy can be very difficult due to the complex composition of marine and freshwater phytoplankton, cells degraded by osmotic shock in the halocline layer, and the usual taxonomic challenges of smaller cells and unresolved identification issues. Although the assignment of pigments to different phytoplankton groups has its drawbacks, e.g. if some of the phytoplankton groups have similar pigment profiles, the method can be used as a useful tool to describe the functional diversity of phytoplankton in combination with rapid screening of samples under the microscope to identify the dominant species (Sarmento and Descy, 2008, Schlüter et al., 2006). In the Krka River Estuary and in other saltwedge estuary studies, which have combined both methods, pigment analyses have been proposed as a reliable method for determining phytoplankton diversity (Viličić et al., 2008, Šupraha et al., 2014).
We have detected the following marker pigments of the microphytoplankton: fucoxanthin (Fuco , diatoms),peridinin (Perid , dinophytes); mostly nanophytoplankton marker pigments: 19’–hexanoyloxyfucoxanthin (Hex , prymnesiophytes), chlorophyll b (Chl b , chlorophytes),violaxanthin (Viola , chlorophytes and prasinophytes),alloxanthin (Allo , chrysophytes and cryptophytes),lutein (Lut , chlorophytes and prasinophytes); and the marker pigment of picophytoplankton: zeaxanthin (Zea , cyanobacteria) (Jeffrey and Vesk, 1997).
For pigment determination, 1 L of seawater was filtered through 0.7 µm Whatman GF/F filters, which were pre-burned at 450°C/5h. Extraction was performed in 4 mL of cold 90% acetone by sonication, followed by centrifugation. The pigments were separated by reversed-phase HPLC (Barlow et al., 1997). The composition of the phytoplankton pigments was analyzed by HPLC following Barlow et al. (1997) method. The extracts were mixed 1:1 (v/v) with 1 M ammonium acetate and injected into the HPLC system with the 3-mm Thermo Hypersil-Keystone column MOS2, C-8, 120 A pore size, 150×4.6 mm (Thermo Hypersil-Keystone, Bellefonte, PA, USA). The pigments were separated at a flow rate of 1 mL min–1 using a linear gradient programme lasting 40 min, using solvents A and B. Solvent A consisted of 70:30 (v/v) methanol: 1 M ammonium acetate and solvent B was 100% methanol. Chlorophylls and carotenoids were detected by absorbance at 440 nm (SpectraSYSTEM, Model UV 2000, Thermo Fischer Scientific, USA). The qualitative and quantitative analyses of individual pigments were performed by external standard calibration with authentic pigment standards (VKI, Denmark).
2.4Lipid class analysis
To determine the lipid classes, 3 L of seawater were collected in glass containers and passed through the 200 µm stainless steel screen to remove zooplankton and larger particles. Immediately afterwards, the seawater was filtered through 0.7 µm Whatman GF/F filters pre-burned at 450 °C for 5 h. The filters of Krka River Estuary samples were stored in liquid nitrogen for six months, while the samples from the Wenchang River Estuary were stored at -80 °C for 2-3 weeks before analysis. The particulate lipids were extracted with a modified one–phase solvent mixture of dichloromethane–methanol–water (Bligh and Dyer, 1959). N–hexadecanone was added to each sample as an internal standard to estimate recoveries in the subsequent steps of sample analysis. The extracts were evaporated to dryness under a nitrogen atmosphere, stored at -20 °C for one day and dissolved in 20 µL dichloromethane immediately prior to analysis.
The lipid classes were determined using TLC–FID (Iatroscan MK–VI, Iatron, Japan) (Gašparović et al., 2015). The lipid classes were separated on Chromarods SIII and quantified by an external calibration with a standard lipid mixture at a hydrogen flow of 160 mL/min and an air flow of 2000 mL/min. The standard deviation determined from duplicate runs accounted for 1–14% of the relative abundance of the lipid classes. Eighteen lipid classes were detected by this technique. The separation scheme comprised of successive elution steps in solvent systems of increasing polarity, followed by a subsequent partial combustion of Chromarods. Total lipid concentrations were determined by summing all lipid classes quantified by TLC-FID. Detailed procedures are described in Gašparović et al. (2015). However, in Gašparović et al. (2015) there was an error for the elution of the solvent mixture MGDG and DGDG (chloroform–acetone (72:28, v:v), which should be “chloroform–acetone (28:72, v:v)” (Gašparović et al. 2017). So, we tested the elution of the standards for MGDG, DGDG and SQDG with that incorrect protocol (Gašparović et al. 2015) and found that the MGDG standard mixture splits into two peaks (termed first and second MGDG, fMGDG and sMGDG, respectively), while DGDG and SQDG co-elute in one peak in the next elution step.
2.5MGDG fatty acid composition
The separation of the MGDG lipids present in the sample mixture was carried out using the UltiMate 3000 Rapid Separation HPLC (Dionex, Germany) system. The Acquity UPLC BEH C18 (2.1 × 100 mm with 1.7 µm particles) (Waters, Milford, Massachusetts, USA) column was maintained at 50 ºC while a gradient elution was employed. The solvent system comprised solution A: LC-MS-grade methanol:ultrapure water (1:1, v:v; 10 mM ammonium-acetate/0.1 % formic acid) and solution B: LC–MS-grade isopropanol (10 mM ammonium-acetate, 0.1 % formic acid). The gradient started with 55% A/45% B, reached 90% B in 40 min, 99% B in 2 min and remained there for 10 min, then 45% B in 1 min, followed by equilibration for 22 min. The flow rate was 0.15 mL/min and sample mixture injected volume was 10 µL. Immediately before the analysis, dichloromethane was evaporated and the sample was redissolved in a solution of methanol:chloroform (1:2, v:v). The HPLC system was connected online to the amaZon ETD ion trap mass spectrometer (Bruker Daltonik, Bremen, Germany) for analyzing the fatty acid composition. The mass spectrometer was equipped with a standard electrospray ionization ion source (nebulizer pressure 8 psi; drying gas flow rate 5 L/min; drying gas temperature 250 ºC; the potential on the capillary –/+ 4500 V). Lipid profiling was performed in both positive and negative ion mode. Data were collected in the mass range of m/z = 100–1200. ESI MS/MS was performed using collision energy of 1 eV. The MGDG species were identified as [M+NH4]+ ions in the positive mode. The derived elemental compositions were matched with an internally compiled lipid library from LIPID MAPS (http://www.lipidmaps.org/).
2.6Data analysis
Principal-component analysis (PCA) was performed to determine how the environmental variables (salinity (S), temperature (T), DIN and PO43¯) related to the accumulation unsaturated MGDG. It was performed using Origin 7 computer software. Linear fits (Origin 7 computer software, Origin Lab) were performed to analyze the correlations of interest.
3RESULTS
3. 1Environmental conditions
The environmental conditions differed significantly between the estuaries studied (Figure 2 and Table S1). Much higher T were measured in the Wenchang River Estuary (27.9-31.5 °C) compared to the Krka River Estuary (21.5-26.2 °C). The Wenchang River Estuary was enriched in DIN (4.0-154.9 µmol L-1), while PO43- concentrations were much lower (0.08-2.99 µmol L-1), as previously observed (Liu et al., 2011). Both DIN (1.0-5.8 µmol L-1) and PO43¯ (0.21-0.70 µmol L-1) were notably lower in the Krka River Estuary than in the Wenchang River Estuary.