Introduction

Non-pigmented bacteria and photosynthetic picoplankton are ubiquitous and represent numerous and active components of the marine ecosystem. Bacteria play an important role through their assimilation of dissolved organic matter (Cole et al. 1988) and by the decomposition of organic matter and through the transformation of inorganic compounds into forms suitable for primary producers (Ducklow et al. 1986). Photosynthetic picoplankton appears to be an essential component of marine ecosystems and one of their representatives, cyanobacteria Synechococcus (Waterbury et al. 1979) and Prochlorococcus (Chisholm et al. 1988) significantly contribute to phytoplankton biomass, especially in oligotrophic areas like the Mediterranean Sea (Magazzù and Decembri 1995; Agawin and Agustí 1997). In addition, the application of flow cytometry has expanded the knowledge of the microbial community members and has permitted discrimination within cyanobacterial cells (Synechococcus and Prochlorococcus) and non-pigmented bacterial cells based on different deoxyribonucleic acid content (Olson et al. 1993; Gasol and Del Giorgio 2000). There is much literature concerning the distribution and dynamics of non-pigmented cells within the investigated area of the Adriatic Sea. A few reports of Synechococcus and Prochlorococcus exist (Fuks et al. 2005; Ninčević et al. 2006; Radić et al. 2009; Šilović et al. 2011; Šantić et al. 2011), but simultaneous observations of Synechococcus, Prochlorococcus and non-pigmented bacteria are absent. The objectives of this study were to describe the seasonal cycles of non-pigmented prokaryotes, indicate the possible role of HNA and LNA bacteria in bacterial activity in the investigated areas, show the relationship between biomasses of autotrophic and non-pigmented prokaryotes and identify the role of heterotrophic nanoflagellates in controlling the bacterial population.

Materials and methods

Study area

The Adriatic Sea is the northernmost basin in the Mediterranean, 800 km long and 200–250 km wide. Bathymetry divides the basin into three parts: a broad northern Adriatic shelf with an average depth of 40 m, the central Adriatic with depressions as deep as 280 m, and connected to the southern Adriatic circular basin over the Palagruža Sill. The coastal area investigated is located in the central and southern Adriatic basin, covering the coastal zone from Zadar to Dubrovnik and is partly under the influence of the karstic rivers: Krka (Šibenik area), Jadro (Kaštela Bay), Žrnovnica, Cetina (Split area), Neretva (Ploče area) and Trebišnjica (Dubrovnik area). In the open sea, samples were collected from station CA001 located in the Jabuka Pit, CA003 located in the open sea area of Šibenik and at station CA009 located near the island of Vis (Fig. 1).

Fig. 1
figure 1

Study area with sampling stations. a Coastal areas: Zadar (Z1–Z3), Šibenik (Š1–Š5) and open sea area (CA001, CA003) b coastal areas: Kaštela Bay (ST101–ST103), Split (S1–S5, CA007) and open sea area (CA009) c coastal areas: Ploče (P1–P3), Dubrovnik (D1–D3)

Full size image

Sampling

Sampling was mostly at monthly intervals from January 2005 to December 2005 on RV Bios using Niskin bottles (5 l). Samples were collected at several depths between the surface and the bottom (at 5- to 10-m intervals for the upper 50 m and at 75, 100, 168, 200 and 260 m at deepest station). A SeaBird 25 CTD profiler recorded temperature and salinity data. Nutrient concentrations (NO3 −, NO2 −, NH4 +, total dissolved inorganic nitrate/DIN/and soluble reactive phosphate/SRP) were determined using the autoanalyser modified method by Grasshof (1976).

Flow cytometry

Abundances of Synechococcus, Prochlorococcus and non-pigmented bacteria were determined using flow cytometry (Marie et al. 1997). For flow cytometry counts of autotrophic cells, 2 ml of preserved samples in 0.5 % gluteraldehyde was frozen at −80 °C and stored until analysis (5–10 days), while samples for bacteria were preserved in 2 % formaldehyde and stored at 4 °C until analysis (5–10 days).

Abundances of SybrGreen I-stained non-pigmented bacteria were determined using flow cytometry (Marie et al. 1997). Samples of 1 ml without replicas were analysed on a Beckman Coulter EPICS XL-MCL with a high flow rate from 1 to 1.2 μL s−1, and the analysed volume was calculated by acquisition time. To standardise the fluorescence intensity of the cells, 1.0-μm yellow-green beads were added (Level-III Epics Division of Coulter Corporation Hialeah, Florida). Two groups of bacteria were distinguished according to their relative green fluorescence as a proxy for the nucleic acid content (Jochem 2001), referred to as high nucleic acid (HNA) and low nucleic acid bacteria (LNA) and light scattering.

Autotrophic cells were separated into two groups of cyanobacteria (Synechococcus and Prochlorococcus) and were distinguished according to light scattering, red emission of cellular chlorophyll content and orange emission of phycoerythrin-rich cells. Biomasses of Synechococcus, Prochlorococcus and heterotrophic bacteria were calculated by using the following volume-to-carbon conversion factors: 250 fg C cell−1 for Synechococcus, 53 fg C cell−1 for Prochlorococcus and 20 fg C cell−1 for non-pigmented bacteria (Zhang et al. 2008).

Bacterial cell production was measured from DNA synthesis based on incorporation rates of 3H-thymidine (Fuhrman and Azam 1982). Conversion factors (CF) for bacterial production were calculated from bacterial cell numbers and 3H-thymidine incorporation during bacterial growth in 1-μm pre-filtered seawater (Riemann et al. 1987): CF = (N 2−N 1)/3H where N 1 and N 2 are the numbers of bacteria at the beginning and the end of the experiment, and 3H is the integrated 3H-thymidine incorporation rate during the experiment. The number of heterotrophic nanoflagellates (HNF) was estimated using epifluorescence microscopy. Samples were stained with 4′-6-diamidino-2-phenylindole (DAPI) for 10 min and filtered through polycarbonate filters with 0.8 μm pore diameters (Milipore, Ireland). Microscope slides were examined with an Olympus microscope under UV light at a magnification of 1,000X (Porter and Feig 1980).

Methods for determination of regulation mechanisms in the prokaryotic community

Relationships between abiotic and biotic factors and the abundance of autotrophic and heterotrophic prokaryotes were determined using the Pearson’s rank correlation index.

To examine the regulation of bacteria by predation, data were analysed using an empirical model (Gasol 1994). The simultaneous observations of the abundance of bacteria and heterotrophic nanoflagellates (HNF) are plotted on a log–log graph, which provides information about coupling between bacterial and HNF abundance. The graph includes an empirically determined maximum attainable abundance (MAA) line (based on a large database from a variety of systems) depicting the HNF abundance that could be attained at a given bacterial abundance (max log HNF = −2.47 + 1.07 log bacterial abundance) and a mean realised abundance (MRA) line. The points close to the MAA line indicate strong coupling between the bacteria and HNF, which according to Gasol (1994), could be interpreted as strong predation on the bacteria. The points that lie well below the MRA line indicate conditions when bacterial abundance was not controlled by HNF grazing. Therefore, distance (D) values (differences between the maximum and realised HNF abundances at different bacterial concentrations) represent a good indicator of the importance of HNF predators in controlling bacterial abundance. The distance (D) between the maximum and actual measured HNF abundance, for each bacterial abundance value, shows the degree of uncoupling between bacteria and their predators and is surrogate for the grazing pressure of HNF on bacteria. Low D values mean strong coupling between the abundance of bacteria and HNF, while high D values mean no or low coupling between bacteria and HNF.

Results

Physicochemical conditions of seawater

In the investigated areas, an isothermal period was obtained from January to April and again in December, while thermal stratification of the water column began in May and lasted until September (Table 1). The lowest values and widest ranges of temperature and salinity were recorded in the areas of Šibenik, Kaštela Bay, Split and Ploče, mostly due to the influence of the river Krka and to a lesser extent, the rivers Jadro, Žrnovnica, Cetina and Neretva (Morović et al. 2006). Nitrates and nitrites in the investigated areas showed increased values in January to February and in December. Values of soluble reactive phosphates at the coastal sea stations were uniform during the isothermal and stratified periods. The highest concentration of ammonium in the coastal areas was detected in November, but high values were also recorded in May and August. In the open sea areas, slightly increased values of ammonium were detected in the period from April to May. Increased values of chlorophyll a were determined in the period from September to December with higher concentrations in coastal areas compared with the open sea (Table 1).

Table 1 Mean monthly value with standard deviation (±) of temperature (°C), salinity, nutrient concentrations (μM) and chlorophyll a (mg m−3) at coastal and open sea area
Full size table

Abundance of non-pigmented bacteria

Average monthly abundance of bacteria, obtained as the average value from the surface to the bottom layers of the investigated stations, ranged from 0.24 × 106 to 1.3 × 106 cells mL−1 in the central and southern coastal areas and in the open sea ranged from 0.23 × 106 to 0.63 × 106 cells mL−1 (Fig. 2). Seasonal distribution of non-pigmented bacteria in the coastal areas has shown an increased abundance during the warmer seasons and lower abundance during the colder seasons, with pronounced variations in bacterial densities in relation to open sea area (Fig. 2). Comparing all investigated areas, the highest numbers of non-pigmented bacteria were found in the Kaštela Bay area (Table 2).

Fig. 2
figure 2

Seasonal fluctuations of non-pigmented bacteria. a Coastal areas b open sea station. Average values (column, line with markers) and positive standard error (bars) are presented. The missing sampling points show that there was no sampling

Full size image
Table 2 Abundance of non-pigmented bacteria, % HNA and bacterial production at coastal areas Zadar, Šibenik, Kaštela Bay, Split, Ploče, Dubrovnik and open sea stations (CA001, CA003 and CA009)
Full size table

Percentage of HNA bacteria in the community of non-pigmented bacteria

The average monthly percentage of HNA bacteria in the central and southern coastal areas ranged from 26 to 83 %, while in the open sea HNA ranged from 28 to 65 %. Different temporal patterns were found for HNA and therefore for LNA bacterial groups. Seasonal distribution mostly showed the prevalence of the HNA group during the colder seasons and a prevalence of LNA bacteria, or equal proportions of the LNA and HNA groups, in the bacterial community during spring and summer. However, a prevalence of the HNA group in the bacterial community during all four seasons was found in the coastal areas directly influenced by rivers (Fig. 3), specifically in the Kaštela Bay (Jadro) and Šibenik areas (Krka). Comparing the investigated areas with the annual average, variations in the proportion of HNA bacteria were mostly more pronounced in the coastal sea areas compared with the open sea area. The annual average at the open sea stations CA001 and CA003 showed percentages of HNA bacteria less than 50 % (Table 2).

Fig. 3
figure 3

Seasonal fluctuations of HNA %. a Coastal areas b open sea stations. Average values (column, line with markers) and positive standard error (bars) are presented. The missing sampling points show that there was no sampling

Full size image

Relationship of bacterial production and the contribution of HNA bacterial group

The average monthly values of bacterial production in the central and southern coastal areas ranged from 0.03 × 104 to 2.8 × 104 cells mL−1h−1 and in the open sea ranged from 0.06 × 104 to 2.71 × 104 cells mL−1h−1 (Fig. 4). Seasonal distribution of bacterial production in the coastal areas from Zadar to Split shows increased bacterial productivity during the summer. The second highest values were found in the area of Šibenik during the spring and in the area of Split during the winter. In the area of Ploče, high values of bacterial production were found in January and December. In the southern coastal region, increased bacterial production was determined during the seasons of spring and autumn. Distribution at the open sea stations CA001 and CA003 showed increased bacterial productivity in the spring and at CA009, during winter and autumn (Fig. 4).

Fig. 4
figure 4

Seasonal fluctuations of bacterial production. a Coastal areas b open sea stations. Average values (column, line with markers) and positive standard error (bars) are presented. The missing sampling points show that there was no sampling

Full size image

Comparing the investigated areas with the annual average, variations in bacterial production in the coastal and open sea areas were more pronounced in the area influenced by rivers Krka (Šibenik area) and less pronounced in the southern part of coastal sea (Table 2).

In order to examine bacterial activity, this study observed bacterial productivity in relation to seasonal distribution of the HNA and LNA bacterial groups. The analysis of these two bacterial groups in relation to bacterial productivity in the coastal and open sea regions showed a simultaneous increase in the percentage of the LNA group and bacterial production (Figs. 3, 4). However, exceptions with a prevalence of the HNA group in the bacterial community during the highest values of bacterial productivity were found in those areas influenced by the Krka and Jadro Rivers (Figs. 3, 4). A comparison of all of the investigated areas of the one-year survey showed that the HNA bacterial group was dominant in the bacterial population in coastal areas over values of bacterial production ranging from 0.27 × 104 to 0.89 × 104 cells mL−1h−1. However, in the open sea areas, the LNA bacterial group was mostly prevalent in the bacterial population over values of bacterial production ranging from 0.53 × 104 to 0.89 × 104 cells mL−1h−1 (Table 2).

Biomass of non-pigmented bacteria in relation to biomasses of Synechococcus and Prochlorococcus

The biomass of the prokaryotic community throughout the studied area was mainly dominated by non-pigmented prokaryotes, except during the spring in the Dubrovnik area and during winter at the open sea station CA003, when the biomass of autotrophic prokaryotes predominated over non-pigmented prokaryotes (Fig. 5). Average seasonal values of non-pigmented prokaryotic biomass in the coastal area ranged from 7.87 to 22.27 μgC L−1 and in the open sea from 4.60 to 9.60 μgC L−1. The values of the biomasses as well as their abundances showed seasonality, mostly demonstrating higher values during the warmer seasons (Fig. 5).

Fig. 5
figure 5

Seasonal fluctuations of Synechococcus, Prochlorococcus and bacterial biomasses. Coastal areas: Zadar (Z), Šibenik (Š), Kaštela Bay (KB), Split (S), Ploče (P), Dubrovnik (D) and open sea stations (CA001, CA003, and CA009). Average values (stacked column) are presented

Full size image

Influencing factors on prokaryotic picoplankton

In the investigated area, correlations between bacterial parameters with temperature and inorganic nutrients were weak or not statistically significant (Table 3). The relationship between HNF as a main predator of bacteria and the bacteria was described by distance points (D) obtained by the Gasol model (1994). Thus, lower D values show higher predation pressure and vice versa. Results indicate that the impact of HNF was more pronounced during the warmer season (Fig. 6). Comparing coastal and open sea areas, the strongest associations between HNF and the bacteria were found in the coastal areas of Šibenik, Ploče and Dubrovnik (Fig. 6).

Table 3 Pearson’s correlations between the nutrient concentrations, chlorophyll a, Prochlorococcus abundance and the abundance of non-pigmented bacteria (B) and bacterial production (BP) during isothermal and stratified water column period
Full size table
Fig. 6
figure 6

Relationship between HNF and heterotrophic bacteria abundance according D values

Full size image

In addition to a predator–prey relationship, other relationships within the prokaryotic community in the study area were also identified. In both thermal conditions of the water column, bacterial abundance was positively correlated with Prochlorococcus abundance (Table 3). During colder periods, in the coastal sea areas, a positive correlation between bacterial abundance and concentration of chlorophyll a (chl a) was determined. During warmer periods, in the open sea areas, a negative correlation between bacterial production and concentration of chl a was observed.

Discussion

The abundance of non-pigmented bacteria in most parts of the study area was lower than 106 cells mL−1. According to bacterial density, the investigated area of the Adriatic Sea is mainly oligotrophic (Cotner and Biddanda 2002). The exceptions are those areas under direct influence of the Rivers Krka (Šibenik area) and Jadro (Kaštela Bay), where abundances of non-pigmented bacteria were more than 106 cells ml−1. The obtained values of chlorophyll a were in a range from 0.02 to 4.73 chl a m−3 with the highest values in Kaštela Bay (1.53 chl a m−3) and the Šibenik area (4.73 chl a m−3), which describes the investigated area as mostly oligotrophic with mesotrophic and eutrophic parts (Marasović et al. 2006). During the investigation, seasonal distribution in the bacterial community in the coastal areas was determined with maxima in the spring–summer period and minima during the winter, which confirms previous reports on the central Adriatic (Krstulović 1992; Šolić et al. 2001). Seasonal distribution of non-pigmented bacteria at the open sea stations was not as pronounced as the seasonal distributions of bacteria in the Bay of Biscay (Calvo-Díaz and Morán 2006).

Generally, within the bacterial community, the HNA group predominated in eutrophic and mesotrophic areas, whilst the LNA group predominated in oligotrophic environments (Zubkov et al. 2004; Morán et al. 2007). The predominance of the LNA group in oligotrophic conditions can be explained by the high surface area to volume ratio of the cell (Button 1998; Jochem et al. 2004), which facilitates successful survival in poor conditions. However, the prevalence of an individual bacterial group in different environmental conditions could be the result of different life strategies. Members of the HNA bacterial group with large genomes are able to cope with a wide variety of environmental conditions, whilst the LNA group with a small genome occupy narrow ecological niches (Schattenhofer et al. 2011). In this paper, the domination of the HNA bacterial group in the bacterial community was established in areas that have a high trophic level and which are directly influenced by rivers. Thus, our finding is consistent with other research that found that rivers directly influence the dominance of the HNA over the LNA group in eutrophic areas (Li et al. 1995; Šolić et al. 2009). Furthermore, the annual average of the investigated oligotrophic open sea stations showed percentages of HNA bacteria below 50 %. This is in accordance with the predominance of the LNA group over HNA, which has also been established for other oligotrophic areas, such as the Celtic Sea, Gulf of Mexico and the Atlantic (Zubkov et al. 2001; Jochem et al. 2004; Andrade et al. 2007).

The seasonal distribution in all investigated oligotrophic areas showed the dominance of the LNA group in the bacterial community mostly during the summer season and dominance of the HNA group during the colder seasons, as described previously in the literature (Grégori et al. 2001; Calvo-Díaz and Morán 2006). The domination of LNA cells during the warmer seasons can be attributed to selective grazing (Šimek and Chrzanowski 1992; Jürgens and Matz 2002), or better adaptation to oligotrophic conditions (Kjelleberg et al. 1993).

Historically, it was thought that the HNA bacteria were the active part of the microbial community, and the LNA were the inactive or dead part (Zweifel and Hagström 1995; Gasol and Morán 1999), and recent studies have confirmed the importance of the activity of the HNA group in the NW Mediterranean Sea (Mével et al. 2008). However, a number of papers have shown that the LNA group also contributes to the activity of bacterioplankton in the marine ecosystem (Grégori et al. 2001; Li and Dickie 2001; Zubkov et al. 2001; Jochem et al. 2004). Some authors have also shown that the LNA group successfully consumes nutrients and amino acids (Zubkov et al. 2006; Mary et al. 2006) and demonstrates a high-specific growth rate that can exceed the growth rate of HNA bacteria in oligotrophic conditions (Longnecker et al. 2005; Mary et al. 2006). In this study, bacterial activity was observed through seasonal distribution of bacterial productivity in relation to the percentage of the HNA and LNA bacterial groups. At the same time, it was determined that LNA values were higher than 50 % with increased values of bacterial production, which is consistent with previous results from the Atlantic (Morán et al. 2002, 2004). However, in the areas directly influenced by karstic rivers, the HNA bacterial group was dominant throughout the study as well as in areas directly influenced by inflowing rivers (Casotti et al. 2000; Jochem et al. 2004) during all four seasons, as reported by Longnecker et al. (2006). Our results indicate the relative importance of both bacterial groups in bacterial production and activity.

Data regarding abundance and seasonal distribution of Synechococcus and Prochlorococcus for the same period and same investigated stations have been previously published (Šantić et al. 2011). This paper showed the relationship between biomasses of autotrophic and non-pigmented prokaryotes. Analysis of biomass in the prokaryotic community usually shows an increase of autotrophic picoplankton biomass in investigated sites, as well as in some research areas of the Mediterranean Sea, during the warmer season (Agawin et al. 2000). In relation to coastal areas, an increase of autotrophic prokaryotic biomass was found in the open sea as well as in the Mediterranean Sea (Magazzù and Decembri 1995). Within the autotrophic fraction, it was found that Synechococcus biomass as well as abundance predominated over Prochlorococcus, as reported in the literature for the summer period in the surface layers of the Mediterranean Sea (Llabrés et al. 2010). This can be compared with data from the Coral Sea (Crosbie and Furnas 2001), Pacific and the Atlantic (Zhang et al. 2008). However, our results are contrary to those from different regions of the Mediterranean Sea (Mella-Flores et al. 2011), where abundances of Prochlorococcus mostly predominated over Synechococcus. The difference in domination of Synechococcus over the Prochlorococcus densities in the investigated areas of the Adriatic Sea could be explained by the fact that generally genus Synechococcus is known as a eurythermal organism (Waterbury et al. 1986; Shapiro and Haugen 1988; Neuer 1992). It is adapted on water-column mixing (Lindell and Post 1995) and is able to consume different types of nitrogen and phosphorus sources, which possibly gives it an advantage over genus Prochlorococcus in the coastal study area of the Adriatic Sea.

Our results show that predation has an important role in the observed seasonal and spatial distributions of non-pigmented bacteria. We identified changes in the degree of predation as a type of control on the seasonal scale, in the coastal areas where bacteria were more controlled by HNF during the warmer period, as also reported by Šolić et al. (2009). On the spatial scale, our results show the strongest correlation between bacteria and HNF in the coastal area, especially in the region of Šibenik and Ploče, which is described as the area richest in nutrients (Cetinić et al. 2006; Svensen et al. 2007; Šolić et al. 2008). Our obtained results agree with Hobbie and Cole (1984) who highlighted the impact of predation by HNF in nutrient-rich environments.

In oligotrophic ecosystems, bacterial abundance is mostly controlled by the supply of nutrients (Cole et al. 1988; Ducklow 1992). According to the literature, the correlation between bacteria and phytoplankton excretion, like food source, is well known (Carlson et al. 1996; Kirchman and Rich 1997). Similar correlations with the concentration of chlorophyll a determined in this paper have also been reported in previous papers for the oligotrophic central Adriatic (Krstulović et al. 1997; Šestanović et al. 2004). The very weak relationship between bacteria and inorganic nutrients obtained in this paper could be explained by the fact that, during the investigated period, the warmer and more saline Levantine Intermediate Water influenced the deep waters in the eastern middle Adriatic Sea. This brings higher concentrations of inorganic nutrients (Šolić et al. 2008), and therefore, bacterial expansion was not limited by low concentrations of inorganic nutrients.

Conclusions

The proportions of HNA bacteria in total abundance of non-pigmented bacteria were highest in those areas with the highest bacterial production (in the areas influenced by karstic rivers) and the higher bacterial densities, during all investigated seasons. In oligotrophic areas, the LNA group dominated for a period of high bacterial production, indicating the importance of both bacterial groups in bacterial production and activity. The biomass of non-pigmented prokaryotes was mostly predominant in the prokaryotic community. Within the autotrophic community, Synechococcus biomass mostly predominated. Predation of heterotrophic bacteria by HNF was pronounced during the warmer period in the coastal areas.