Picoplankton seasonal variation and community structure in the northeast Adriatic coastal zone

Abstract

The bacterial community in coastal waters of northeastern Adriatic Sea was dominated by SAR11 and Sulfitobacter taxa throughout the year. The seasonal distribution of bacterioplankton taxa showed continual differences between surface (0 m) and bottom (27 m) layers. The surface assemblage was represented by Actinobacteria,Cyanobacteria, Alphaproteobacteria, and Gammaproteobacteria, while the bottom assemblage was made up of Bacteroidetes, Cyanobacteria and Alphaproteobacteria. As SAR11 was more dominant in the bottom layer, its appearance may be linked to northward transport of oligotrophic waters of higher salinity from the south. Gammaproteobacteria appeared only in the surface layer during summer, influenced by higher amounts of nutrients, brought in by the Po River. Synechococcus was the most abundant taxon at the genus level. Dominance of Synechococcus during the whole season agrees with its dominance in terms of abundance determined by flow cytometry, and confirms its utmost importance in the picoplankton community of this area. We found two different types of Synechococcus: one type with high similarity to SynechococcusCC9902, present in the surface and bottom layers, and another one similar to SynechococcusWH7803, present only in the surface layer. Oligotrophic conditions together with complex hydrological features of this area were reflected in diversification and dynamic shifts of surface and bottom assemblages.

Northern Adriatic, bacterioplankton, DGGE, Synechococcus, SAR11

Introduction

Great effort has been made in the last decades to assess the diversity of marine microbial communities using different molecular techniques. The application of such methods has expanded our understanding of marine microbial evolution, metabolism, and ecology (De Long & Karl, 2005). Yet, much remains to be learn about their diversity, ecology and distribution. In addition, identifying patterns in microbial communities living in different habitats will establish a more comprehensive picture of microbial and biogeochemical processes in marine systems. Some of the studies have addressed spatial variability in microbial structure (Murray et al., 1998; Schauer et al., 2000; Yokokawa et al., 2010; Celussi et al., 2011) pointing out to their large scale stability. Temporal studies have been mainly carried out in coastal waters (Murray et al., 1996, 1998; Pinhassi & Hagström, 2000; Schauer et al., 2000, 2003; Kan et al., 2006; Alonso-Sáez et al., 2007; Celussi & Cataletto, 2007; Celussi et al., 2011) because of the difficulty of periodic sampling in open sea. Coastal zones are usually very dynamic and subject to environmental perturbations; consequently, different parts of the same area can have different factors shaping the ecosystem and its microbial community structure. The northern Adriatic is a dynamic area with highly variable circulation (Mihanović et al., 2006 and references therein), influenced by strong winds, freshwater inputs, and oligotrophic water brought in from the south Adriatic Sea. The Po River outflow and meteorological factors trigger stratification or vertical mixing of the water column (Socal et al., 2008), consequently shaping microbial community structure through the whole water column. The spatial and temporal extension of seasonal stratification shows interannual variability (Solidoro et al., 2009). The main contribution to the circulation and to the trophic state of the basin is the Po River and its strong seasonal and interannual variability in water loads (Socal et al., 2008). Usually, in the presence of vertical stability and strong Bora winds, a branch of the Po River plume can travel eastward reaching Istria, following a semi-closed circulation pattern (Artegiani et al., 1997). This Po River branch reaching coastal northeastern part of Adriatic Sea has an important influence on plankton community because it brings nutrients and triggers vertical mixing. This complex hydrodynamism of the studied area makes it quite different from other coastal areas where the water column is homogeneous, and vertical mixing is well established during most of the season. Thus, the main aim of this study was to (1) obtain detailed information concerning microbial assemblages in terms of their composition and temporal dynamics in coastal waters of northeastern Adriatic Se a and (2) to correlate microbial community structure to physical and environmental variables so as to identify the principal factors influencing bacterial dynamics.

Methods

Sample collection and preparation

Seawater samples were collected from five depths (0, 5, 10, 20, and 27 m) at coastal station RV001, one nautical mile from the shore in northern Adriatic Sea (13°61′E, 45°08′N). Samples were collected every 2 weeks between September 2008 and October 2009. Temperature and salinity were measured with a SBE 25 Sealogger CTD probe (Sea–Bird Electronics, Inc., Bellevue, WA) in situ, while samples for nutrients and chlorophyll a were collected in polycarbonate bottles and processed on board. Subsamples for the determination of dissolved nutrients such as nitrate, nitrite, ammonium, and phosphate (PO₄) were analyzed immediately after collection (Parsons et al., 1984; Ivančić & Degobbis, 1984). Subsamples of 500 mL for the determination of chlorophyll a (pico-, nano-, and microfractions) were filtered onto Whatman GF/F filters. Subsamples (500 mL) were filtered directly onto GF/F filters (for total chl a), through 20-μm net to GF/F filters (for nano-chl a), and through 3 μm polycarbonate Nucleopore filters (47 mm diameter) to GF/F filters (for pico-chl a). A filtration vacuum of < 2 cm Hg was used for all filtration steps. Filters were frozen (−18 °C) and analyzed following the fluorometric procedure after Parsons et al. (1984). The micro-chl a fraction was obtained by subtracting total chl a from < 20 μm fraction. The nano-chl a fraction was obtained by subtracting nano-chl a from pico-chl a fraction.

Subsamples (2 mL) for heterotrophic bacteria counts were fixed with formaldehyde (2% final concentration) and stored at 4 °C until returning to the laboratory. Subsamples (4 mL) for picophytoplankton counts were fixed with 0.5% glutaraldehyde for 10 min, frozen in liquid nitrogen, stored at −80 °C. For denaturing gradient gel electrophoresis (DGGE) analysis, 5 L of seawater was collected once per month at surface (0 m) and bottom depth (27 m) and filtered through 0.2 μm pore diameter filters (Nucleopore PC) with a vacuum pump at ≤ 150 mmHg. Filters were placed in cryo-vials, filled with 1.8 mL of lysis buffer (40 mM EDTA, 50 mM Tris–HCl, 0.75 M sucrose) and stored at −80 °C.

Flow cytometry

Samples were analyzed as described earlier (Šilović et al., 2011) using a Partec PAS III flow cytometer, equipped with an Argon laser (488 nm). Instrumental settings were standardized for all parameters by using 1 and 3 μm fluorescence polystyrene calibration beads. Data were collected in listmode files using FL3 as a trigger parameter and processed with software FloMax (Partec). Synechococcus and picoeukaryotic cells were distinguished by their autofluorescing chlorophyll (FL3) and phycoerythrin (FL2) content of the cells as well as by the cells' side-angle light scatter (SSC) as a proxy of their size.

Heterotrophic bacteria abundance

Samples were counted by epifluorescence microscopy (Leitz Laborlux D and Zeiss: imager. Z1 at a magnification of 1000×) after staining with 4,6-diamidino-2-phenylindol (DAPI; 1 μg mL⁻¹, final concentration).

Statistical analysis

A logarithmic transformation [log₁₀ (x + 1)] was used on the band intensity data prior to statistical analyses to obtain normal distributions. A standard Pearson correlation using the program systat 10.2 was used to quantify direct correlations between picoplankton abundances and environmental parameters. The statistical package primer 6 (Clarke & Gorley, 2006) was used for principal component analysis (PCA) of physical and chemical variables, picoplankton abundances and band intensity data, as well as for multidimensional scaling (MDS) and hierarchical agglomerative clustering with group-average linking (CLUSTER) and an associated Similarity Profiles (SIMPROF) test of band intensity data.

DNA extraction

DNA was extracted as previously described (Boström et al., 2004). Briefly, cells were treated with lysozyme, proteinase K, and sodium dodecyl sulfate followed by phenol-chloroform-isoamyl alcohol extraction. Extracted DNA was desalted and treated with concentrated ethanol and 5 M sodium acetate. DNA was dilluted in 50 μL of MQ water.

PCR and DGGE

A 16S rRNA gene fragment (approximately 550 bp long) was amplified by PCR, using the bacterium-specific primer 358f (5′CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGCCTACGGGAGGCAGCAG) that is complementary to positions 341–358 (Escherichia coli numbering) and has a GC clamp (underlined) and the universal primer 907rm [5′-CCGTCAATTC(A/C)TTTGAGTTT] that is complementary to positions 927–907. The reaction mixture volumes were 50 μL, containing 20 ng of template, 200 μM concentration of each deoxynucleoside triphosphate, standard 1× PCR buffer, 2 mM MgCl₂, 0.25 μM concentration of each primer, bovine serum albumin (0.15–0.30 μg mL⁻¹), and 1.25 U of Taq polymerase (Invitrogen). After an initial denaturation step (5 min at 94 °C), samples were amplified with 10 touchdown cycles, with one cycle consisting of denaturation (1 min at 94 °C), annealing [1 min at 65–55 °C (temperature of 65 °C in the first cycle, with the temperature decreasing 1 °C each cycle and ending at 55 °C in the last cycle)], and extension (3 min at 72 °C). This was followed by 20 standard cycles, with one cycle consisting of denaturation (1 min at 94 °C), annealing (1 min at 55 °C), and extension (3 min at 72 °C), and a final extension step (7 min at 72 °C). PCR products were quantified by agarose gel electrophoresis with a molecular size standard in the gel (Low DNA Mass Ladder, Invitrogen).

A total of 800 ng of PCR product for each sample was loaded on a 6% (w/v) polyacrylamide gel (acrylamide and N,N′-methylene bisacrylamide at a ratio of 37 : 1) with a denaturing gradient that ranged from 40% to 80% (where 100% is defined as 7 M urea and 40% deionized formamide). Electrophoresis was carried out with a DGGE-2000 system (CBS Scientific Company). Gels were run at 100 V for 16 h at 60 °C in 1× TAE running buffer [40 mM Tris (pH 7.4), 20 mM sodium acetate, 1 mM EDTA]. Gels were stained with the nucleic acid stain SYBR Gold Safe (Molecular Probes) for 20 min, rinsed with 1× TAE running buffer. Images were processed in Quantity One Software by ChemiDoc System. The ChemiDoc software detects the bands and calculates their relative contribution to the total lane intensity. Each lane represented one sample, and gained values were used for building dendrogram while the distance matrix of gained values was used for MDS diagram with the statistical package primer 6.

Sequencing of DGGE bands

DGGE bands were excised using a sterile razor blade and eluted in 20 μL of MilliQ water overnight at 4 °C, followed by a freeze-thaw cycle. A total of 3 μL of the eluate was used for reamplification with the original primer set. Ten to 20 ng of the PCR product was sent for commercial sequencing (Macrogen, the Netherlands). Obtained sequences were compared with public database using blast to determine their taxonomic affiliation.

Results

Environmental parameters

Seawater temperature at station RV001 ranged from 9.3 °C (February, 5 m) to 27.5 °C (August, surface), whereas salinity varied between 34.22 (June, surface) and 38.18 (September 2009, bottom). Water temperature and salinity values showed a regular seasonal trend with a sharp stratification period starting at the end of May, and lasting until the end of September (Fig. 1). Salinity was rather constant during the year with a marked decrease in the subsurface layer in spring–summer period (Fig. 1b), typical for this area (Ivančić et al., 2010).

Temperature (a) and salinity (b) at station RV001 from September 2008 to October 2009.

Nutrient concentrations showed the lowest values in the surface layer (0.03 μM for in December, 0.01 μM for in August, at 5 m and 0.12 μM in October at 0 m) (Fig. 2a–c). Phosphate concentrations were generally very low, particularly in February and March when their concentrations were below detection limit (< 0.01 μM) through the whole water column (Fig. 2d). The highest value of phosphates (0.26 μM) was recorded in April at 20 m. During stratification, nutrient values gradually increased reaching the maximal values in the bottom layer (except for with maximal values of 10.34 μM at 5 m in June), 0.26 μM for in April, 5.10 μM for in May at 20 m, 2.07 μM for in October at 27 m (Fig. 2a–d).

Inorganic nutrient concentrations at station RV001 from September 2008 to October 2009 for: (a) ; (b) ; (c) ; (d) .

Total chl a concentration was generally very low and varied between 140 and 850 ng L⁻¹ showing the maximum in September 2008 (Fig. 3a). In September 2008, the mean value of the whole water column was 720 μg L⁻¹, while in September 2009 that value was only 450 μg L⁻¹. Pico-chl a (< 3 μm) ranged from 10 to 570 ng L⁻¹ with the highest value in the bottom layer in August 2009 (Fig. 3b). Comparing pico-, nano-, and micro-chl a fractions, pico-chl a dominated through the whole sampling period, with highest contribution in March and in August 2009 (Fig. 4). The only exception was the period of phytoplankton bloom in fall (September 2008) when micro-chl a accounted for more than 50% of total chl a (Fig. 4). No significant correlation was found for total or pico-chl a distribution with Synechococcus or picoeukaryote abundances, while heterotrophic bacteria significantly correlated with pico-chl a (P < 0.05, r = 0.286, n = 107).

Chlorophyll a concentration at station RV001 from September 2008 to October 2009 for (a) total chl a and (b) pico-chl a.

Percentage contribution of pico-, nano-, and micro-chl a to total chl a from September 2008 to October 2009.

Picoplankton abundance

Heterotrophic bacteria numerically dominated the picoplankton community. Their abundance varied between 1.5 × 10⁵ cells mL⁻¹ (in May at 0 m) and 27 × 10⁵ cells mL⁻¹ (in July at 0 m) (Fig. 5a). Synechococcus dominated the autotrophic community (generally > 80%) with concentrations ranging from 2.4 × 10⁴ cells mL⁻¹ (in December at 10 m) to 21 × 10⁴ cell mL⁻¹ (in September at 0 m) (Fig. 5b).

Abundance of different picoplankton groups at station RV001 from September 2008 to October 2009 for: (a) heterotrophic bacteria, (b) Synechococcus, and (c) picoeukaryotes.

Heterotrophic bacteria and Synechococcus abundances showed a significant positive correlation with temperature (P> 0.001, r = 0.70, n = 129; P< 0.001, r = 0.60, n = 130, respectively) and both correlated negatively with salinity (P< 0.001, r = −0.59, n = 130; P< 0.001, r = −0.49, n = 130, respectively). A significant negative correlation was also found for Synechococcus and nitrites (P< 0.001, r = −0.45, n = 131).

Picoeukaryotes were generally found in lower numbers, with highest values in winter, up to 10⁴ cells mL⁻¹ at 5 m in January (Fig. 5c). They showed significant positive correlation with salinity (P < 0.05, r = 0.308).

Bacterial community structure

The bacterial community structure was analyzed according to the intensity matrix (presence/absence of bands in the DGGE gels, combined with their intensity) by MDS and cluster analysis. According to MDS analysis, the composition of bacterial assemblage followed a seasonal cycle (Fig. 6a and b) in both surface and bottom samples. The only exception was October 24 (2008) at the surface, when the assemblage was relatively similar to the November sample, but remarkably different from the anterior and posterior samples.

MDS diagram representing changes in bacterial community from October 2008 to October 2009 for: (a) surface assemblage and from October 2008 to September 2009 for: (b) bottom assemblage.

The number of different bands defined as OTUs was larger at the surface samples (61 band types), than at the bottom (51 band types). Detected OTUs were different in spring–summer vs. fall–winter samples, indicating differences in bacterial assemblages in periods of stratification and vertical mixing (Fig. 7a and b).

Cluster diagram showing differences in bacterial community from October 2008 to October 2009 for: (a) surface assemblage and from October 2008 to September 2009 for: (b) bottom assemblage. (Branches marked with lighter gray lines, indicated that SIMPROF can find no statistical evidence for any sub-structure within these.)

Cluster analysis was carried out with the presence/absence matrix on surface and bottom samples. Surface samples were grouped into two main groups indicating two different periods of water column stability: stratification and mixing. ‘Stratification group’ was further divided into three subgroups: late spring cluster and summer–fall cluster. ‘Mixing group’ contained two subgroups: the first one was late autumn cluster and second one was divided into early spring and winter clusters. Bottom samples were divided into three main groups that were different from surface groups. The two main groups in bottom samples indicated periods of stratification and mixing as well. Mixing group contained winter and early spring cluster, while stratification group contained fall–summer in one group and late spring and winter in other group (Fig. 7a and b).

We excised and sequenced 19 bands from the surface and 14 bands from the bottom samples. A blast search was used to determine their closest sequences. Most of the bands showed relatively high similarities (91–100%) with cultured strains from GenBank (Table 1). Of 19 surface bands, 12 belonged to the Alphaproteobacteria, including nine Rhodobacterales, two SAR11, and one SAR116. Of 14 bottom bands, eight corresponded to Alphaproteobacteria (six Rhodobacterales, one SAR11, and one SAR116). The surface community was made up from four different (taxonomic) groups: Actinobacteria (Actinomyceta), Cyanobacteria (Synechococcus sp.), Alphaproteobacteria (Rhodobacterales, SAR11, SAR116, and Sphingomonadales), and Gammaproteobacteria (Pseudomonadales). Class Alphaproteobacteria dominated the bacterial community, but considering lower taxonomic groups, Synechococcus was the dominant genus level taxon with the highest band intensity in total and during the whole sampling period (except in April when it was outnumbered by Rhodobacterales and February, July, August and October by SAR11) (Fig. 8a). A few bands corresponding to eukaryotes were observed as well and they were related to Micromonas pusilla plastids (Prasinophyceae) with 9.5% of total observed band intensity. The bottom assemblage was made up by these groups: Bacteroidetes (Flavobacteria), Cyanobacteria (Synechococcus sp.), and Alphaproteobacteria (Rhodobacterales, SAR11, and SAR116). According to total band intensity, Synechococcus had the highest contribution to bottom assemblages (25.1%). It was the dominant population in January, March, April, 21st of May and July, while it was not detected in October and November (Fig. 8b). The part of the season in the bottom assemblage not dominated by Synechococcus was dominated by Alphaproteobacteria, mainly SAR11 (except 6th of November and 7th of May with Rhodobacterales and February with SAR116).

Relative abundance of different phylogenetic groups (based on DGGE band intensity) from October 2008 to October 2009 at station RV001 for: (a) surface assemblage and from October 2008 to September 2009 for: (b) bottom assemblage. 1 – Actinobacteridae; Acinomyceta; 2 – Bacteroidetes; Flavobacteria; 3 – Cyanobacteria; Chroococcales; Synechococcus; 4 – Alphaproteobacteria; 5 – SAR11; 6 – SAR116; 7 – Alphaproteobacteria; Rhodobacterales; 8 – Alphaproteobacteria; Sphingomonodales; Erythrobacteraceae; 9 – Gammaproteobacteria; Pseudomonadales; Pseudomonadaceae; 10 – Eukaryota; Chlorophyta; Prasinophyceae.

Phylogenetic affiliation of sequences in DGGE bands and bands relative intensity within different samples

Band number	Closest match (environmental or culture)	GeneBank Accession number	Number of bases	Sequence similarity with closest culture match (%)	Taxonomic group	Closest cultured match
RV-24	Yonghaparkia sp. MOLA 360	AM945590.1	377	96.00	Actinobacteria	Yonghaparkia sp. MOLA 360
RV-57	Microbacteriaceae bacterium CL-Dokdo102	FJ214966.1	464	94.93	Actinobacteria	Microbacteriaceae bacterium
RV-43	Synechococcus sp. CC9902	CP000097.1	534	99.53	Cyanobacteria	Synechococcus sp.
RV-45	Synechococcus WH7803	CT971583.1	530	95.83	Cyanobacteria	Synechococcus sp.
RV-5	Uncultured marine bacterium clone MOLA	GU204718.1	513	100.00	Alphaproteobacteria	Uncultured
RV-30	Rhodobacteraceae bacterium RCA23	GQ468661.1	629	98.49	Alphaproteobacteria; Rhodobacterales	Rhodobacteraceae bacterium
RV-48	Rhodovulum strictum strain MB-G2	NR_025845.1	593	86.02	Alphaproteobacteria; Rhodobacterales	Rhodovulum strictum
RV-34	Thalassobius aestuarii strain F84013	HQ908719.1	627	92.12	Alphaproteobacteria; Rhodobacterales	Thalassobius aestuarii
RV-37	Sulfitobacter dubius strain F71056	HQ908665.1	518	99.53	Alphaproteobacteria; Rhodobacterales	Sulfitobacter dubius
RV-11	Candidatus Pelagibacter ubique clone fosmid	EU410957.1	643	97.13	SAR 11	Canddatus Pelagibacter ubique
RV-13	Uncultured SAR11 cluster alpha proteobacterium	AM748227.1	526	97.00	SAR11	Uncultured
RV-28	Candidatus Puniceispirillum marinum IMCC1322	CP001751.1	390	88.01	SAR116	Candidatus Puniceispirillum marinum
RV-39	Erythrobacter aquimaris strain D4017	FJ161254.1	523	93.87	Proteobacteria, Alphaproteobacteria	Erythrobacter aquimaris
RV-43	Uncultured Pseudomonas	AF195482.1	406	93.58	Gammaproteobacteria	Pseudomonas mendocina (EU043329.1)
RV-23	Micromonas pusilla strain P7/1, plastid.	EF051747.1	262	98.83	Eukaryota; Prasinophyceae	Micromonas pusilla
RV-12	Uncultured Flavobacteria bacterium, CLONE	FN433329.1	461	98.68	Bacteroidetes	Uncultured
RV-10	Uncultured alpha proteobacterium clone, PARTIAL SEQ	AY663892.1	420	98.22	Proteobacteria; Alphaproteobacteria	Uncultured
RV-8	Candidatus Pelagibacter ubique clone fosmid	EU410957.1	423	98.00	SAR 11	Canddatus Pelagibacter ubique
RV-33	Rhodobacteraceae bacterium	FN811289.1	544	87.65	Alphaproteobacteria; Rhodobacterales	Rhodobacteraceae bacterium
RV-16	Micromonas pusilla strain, plastid	EF051747.1	542	86.74	Eukaryota; Prasinophyceae	Micromonas pusilla

We found two different types of Synechococcus: one type with high similarity to Synechococcus CC9902, present in the surface and bottom layers, and another one similar to Synechococcus WH7803, present only in the surface layer during the whole sampling period, except in January, February, and March.

PCA carried out on surface and bottom environmental data (physical, chemical parameters, and band intensities of sequenced DGGE bands) (Fig. 9a and b) revealed five principal components (PCs) with eigenvalues < 1 that accounted for 83.1% and 87.9% of total variance for surface and bottom samples, respectively. In surface samples, the first principal component (PC1) accounted for 34% of the total variance and was mainly explained with physical parameters: negative temperature and positive salinity. Actinobacteria, Alphaproteobacteria, WH7803, and SAR116 showed high correlation with temperature, while CC9902 and Prasinophyceae had positive coefficients like salinity. The second principal component (PC2) accounted for 22.8% of the total variance and was mainly related to nutrients: nitrates and nitrites with high coefficients found in Prasinophyceae, Erythrobacteraceae, SAR11, and SAR116. In bottom samples, PC1 accounted for 32.4% of the total variance and was explained with salinity and nitrites. Alphaproteobacteria and SAR11 showed the highest positive coefficients, while CC9902 and Rhodobacterales had negative coefficient values. PC2 in bottom samples accounted for 21% and was explained by positive phosphates and chl a. Alphaproteobacteria and Rhodobacterales showed the highest positive coefficients. Bacteroidetes and Prasinophyceae had negative coefficient values correlating with nitrites and salinity.

PCA showing relationship between physical and chemical parameters and bacterial community from October 2008 to October 2009 (after excluding outliers (K – July 2009; L – August 2009) for: (a) surface assemblage and from October 2008 to September 2009 for: (b) bottom assemblage. A – October 2008; B – 6 November 2008; C – 17 November 2008; D – December 2008; E – January 2009; F – February 2009; G – March 2009; H – April 2009; I – 7 May 2009; J – 21 May 2009; K – July 2009; L – August 2009; M – September 2009; N – October 2009; T – temperature; S – salinity; – phosphates; – nitrates; – nitrites; – ammonium; Chl a – chlorophyll; Act – Actinobacteridae; Bacter – Bacteroidetes; WH7803 – SynechococcusWH7803; CC9902 – SynechococcusCC9902; Alpha – Alphaproteobacteria; SAR11; SAR116; Rhod – Rhodobacterales; Eryth – Erythrobacteraceae; Gamma – Gammaproetobacteria; Pras – Prasinophyceae.

Discussion

The northern Adriatic Sea was for many years characterized as a highly productive area mainly due to some occasional episodes of eutrophication followed by diatom blooms and “mucilage” appearance. Reduced Po River discharges in the mid-1980s resulted in significant decrease in nutrient concentrations, particularly phosphorus and ammonia (Mozetič et al., 2009). Nutrient limitation, together with a decrease in chl a levels led to the general oligotrophication in the northern Adriatic Sea (Mozetič et al., 2009; Ivančić et al., 2010). Hydrological features and atmosphere–sea interactions tend to be the main drivers of northern Adriatic Sea dynamics (Celussi et al., 2011). According to chl a values, picoplankton was the most important size class during 2009, showing that the system had shifted toward a smaller size class-microorganism. Among the observed environmental parameters, only temperature and salinity showed significantly important influence on picoplankton populations. Surface salinity decreased in summer months (Fig. 1b), together with higher temperatures, resulting in heterotrophic bacteria and Synechococcus peaks, while picoeukaryotes decreased, showing their preference for saltier waters. The Synechococcus dominated over picoeukaryotes in terms of abundance, which is in line with previous observations in the Adriatic Sea (Ninčević-Gladan et al., 2006; Bernardy-Aubry et al., 2006; Radić et al., 2009; Šilović et al., 2011). The only period when picoeukaryotes peaked was December (Fig. 5c), likely as a consequence of vertical mixing that provided nutrients throughout the water column. During stratification, picoeukaryotes accumulated in deeper layers, particularly in summer because of higher nutrient levels. Picoeukaryotes showed a positive response to increased nitrites availability, similar to Prasinophyceae (recovered from DGGE gel) (PCA; Fig. 9a and b). The peak of picoeukaryotic abundance appeared in January 2009 and was followed by high Prasinophyceae abundances (Fig. 8a).

The seasonal pattern in bacterioplankton community found in this study was similar to those in other coastal areas, like Blanes Bay (Schauer et al., 2003; Alonso-Sáez et al., 2007), Ría de Vigo (Alonso-Gutiérrez et al., 2009), Chesapeake Bay (Kan et al., 2006), or Gulf of Trieste (Celussi & Cataletto, 2007), but we observed certain differences between surface and bottom assemblages. The vertical distribution of temperature and salinity showed a clear change in the vertical structure of the water column from stratification in summer to the vertical mixing in fall (Fig. 2). Consequently, we expected that the most important difference in bacterial community structure would be between those two periods. On the contrary, the main difference observed was not between ‘stratification’ and ‘mixing’ communities but between surface and bottom assemblages (independently of vertical mixing). We assume that the reason for such differences lies in (1) variable influence of freshwater inputs in surface waters and (2) influence of oligotrophic water from the south in the bottom layer (Degobbis et al., 2000). The Po River influence is most evident in summer, when due to high flow and changes in circulation pattern its plume can reach the Istrian coastal zone (Supić et al., 2003). This happened in 2009. The Po River had the strongest outflow in May 2009 (average value of 4033.63 m³ s⁻¹; source: ARPA-Romagna Report, daily data for September 2008–2009), when it reached the Istrian coast in June and brought nutrients with it (Figs 1b and 2a). This event was exceptional and can explain differences in environmental parameters during fall periods in 2008 and 2009. We observed a certain decrease in Synechococcus and an increase in SAR11 abundance from fall 2008 to fall 2009, while in bottom samples, we noticed the opposite trend. SAR116 had the increasing trend from 2008 to 2009 in both layers.

These complex interactions made this northeastern microbial community quite different from that found in previous studies in the Adriatic Sea (Celussi & Cataletto, 2007; Celussi et al., 2011). The highest number of OTUs found by Celussi & Cataletto (2007) in Gulf of Trieste was in summer and lowest between winter and spring. In our study, highest number of OTUs in surface samples appeared in winter, while the lowest were found in the fall. In bottom samples, highest number of OTUs appeared in fall and the lowest in spring. So, although both layers followed a seasonal cycle, there was a clear difference between these two areas of the Adriatic Sea, not only in number of OTUs but also in their composition, even during periods of vertical mixing.

The main groups of bacteria appearing in our study, Cyanobacteria (Synechococcus sp.) and Alphaproteobacteria, are typical for coastal areas, as reported before from DGGE analyses (Celussi & Cataletto, 2007; Schauer et al., 2003; Kan et al., 2006) or from clone libraries (Alonso-Sáez et al., 2007 and references therein; Alonso-Gutiérrez et al., 2009). Alphaproteobacteria class was dominated by Rhodobacterales and SAR11 taxa. SAR11 is a ubiquitous clade that dominates surface waters worldwide (Giovanonni et al., 1990; Giovanonni & Rappé, 2000; Morris et al., 2002; Zubkov et al., 2002; Mary et al., 2006). In our study, SAR11 appeared more dominant in the bottom layer and we assume that this is linked to intrusions of oligotrophic water (Fuchs et al., 2005) from the southern Adriatic Sea (confirmed by PCA with high correlation with salinity). Other studies have not detected SAR11 in the Adriatic Sea, but perhaps the primers used were missing this normally abundant phylotype (Sánchez et al. 2009). The oligotrophic characteristic of this area and low freshwater influence during the period of study are correlated with the low abundance of Bacteroidetes, Betaproteobacteria, and Gammaproeteobacteria. The presence of Bacteroidetes only in the bottom layer during stratification points out to their demand for higher nutrient levels than were present in the area (PC2, Fig. 9b), as higher nutrient levels favor their growth (Alonso-Gutiérrez et al., 2009 and references therein).

Dominance of Synechococcus during the whole season in the DGGE agrees with their dominance in terms of abundance (by flow cytometry) and confirms their importance in the picophytoplankton community of this area. The presence of two different Synechococcus subgroups in northern Adriatic Sea had been observed in culture experiments (Paoli et al., 2008), but their taxonomy or physiology had not been explored. We found two different Synechococcus types, CC9902-like present in both layers, and WH7803-like present only in the surface layer. Type CC9902 (clade IV) is predominantly found in coastal boundary zone in a broad range of nitrate and phosphate concentrations (Zwirglmaier et al., 2008). Interestingly, according to PCA, CC9902 correlated positively with salinity and nitrites in the surface layer but negatively in the bottom layer. On the other hand, the WH7803 type (clade V) is observed during transition periods between mixing and stratification (Post et al., 2011). Genotypes belonging to clade V tend to dominate in upwelling zones (Fuller et al., 2006), suggesting their preference for high nutrient availability. The absence of clade V in winter and their complete absence from bottom waters can be connected to their preference for higher temperatures (Fig. 9a), as was previously observed in coastal California (Tai & Palenik, 2009). Celussi and colleagues carried out seasonal studies similar to the present one in the Gulf of Trieste, northeast Adriatic Sea (Celussi & Cataletto, 2007; Celussi et al., 2011). The specific hydrodynamics of the Gulf of Trieste shapes its bacterial community structure, making it rather different from the area studied here, despite its proximity. In effect, the physical processes occurring along the eastern and western Adriatic coasts differ greatly in their characteristics. Water exchange between the semi-enclosed basin of the eastern coast and the open sea is mainly forced by the local wind. Conversely, the shelf area along the western coast is dominated by the Po River outflow, which in winter remains mostly confined to a coastal boundary layer, whereas in summer, spreads to the open sea as well (Orlić et al., 1992). Considerable freshwater inputs, together with particular circulation patterns, make the Italian coast quite eutrophic. On the other hand, the northeast Croatian coast is quite oligotrophic (Mozetič et al., 2009; Ivančić et al., 2010) and much more dynamic (depending on prevailing freshwater/oligotrophic water influence). These differences in hydrodynamics of the two areas are the most likely cause of the differences in the bacterial assemblages found in both studies.

Acknowledgements

We thank the captain and crew of ‘Vila Velebita’ and ‘Burin,’ particularly Paolo Krelja and Margareta Buterer for invaluable help during the field work. We are grateful to Paolo Paliaga for his help in the field and in the laboratory. This work was supported by Croatian project 098-0982705-2729 and UKF Grant Agreement No. 70/10. S. Orlić is grateful to the Croatian Science Foundation (BABAS project). The work in Barcelona was supported by grant GEMMA (CTM2007-63753-C02-01/MAR) from the Spanish Ministry of Science and Innovation.

Author notes

Editor: Riks Laanbroek