Introduction

Summer blooms by filamentous, N2-fixing cyanobacteria are dominated by Nodularia spumigena, Aphanizomenon sp. and Anabeana sp. in the Baltic Sea. These large cyanobacteria form colonies and contain gas vesicles providing buoyancy, whereby they can accumulate at optimal depth to sustain energy and nutrient demands for cellular growth and limit sedimentation losses (Walsby et al., 1997). N. spumigena accumulate close to the sea surface, whereas Aphanizomenon sp. often occurs deeper in the water column (Hajdu et al., 2007). Nodularia sp. is highly tolerant to ultraviolet radiation through the production of mycosporine-like amino acids and carotenoids (Wulff et al., 2007; Mohlin and Wulff, 2009), and mm-thin Nodularia layers accumulating at the sea surface have recently been shown to be highly productive microenvironments with steep O2 and pH gradients during light as well as during darkness (Ploug, 2008).

The large, filamentous cyanobacteria are thought to be the main N2-fixing organisms in the Baltic Sea (Boström et al., 2007; Degerholm et al., 2008; Ploug et al., 2010). Budget calculations of combined nitrogen as well as experimental studies have shown that these cyanobacteria must fix excess N2 relative to their own nitrogen demand and channel the surplus nitrogen through the microbial food web and to picocyanobacteria (Ohlendieck et al., 2000; Larsson et al., 2001; Stal et al., 2003). Using 15N tracers combined with nanoscale secondary ion mass spectrometry (nanoSIMS) and elemental analysis isotope ratio monitoring mass spectrometry, it was recently shown that Aphanizomenon sp. releases a significant fraction (35%) of its newly fixed N2 as NH4+ to the surrounding waters in the Baltic Sea (Ploug et al., 2010). Hence, these large cyanobacteria can supply other algae and (micro)organisms with nitrogen when bulk concentrations of dissolved inorganic nitrogen are low. The significance of N2 fixation and NH4+ release by N. spumigena in the Baltic Sea is still unknown.

In this study, we used a combination of microsensors and isotope tracer experiments combined with nanoSIMS, and fluorometry to measure O2 fluxes, C assimilation, N2 fixation and NH4+ release in a field population of N. spumigena in the Baltic Sea. Distributions of photosynthesis and respiration across the size spectrum of colonies were analyzed in relation to diffusion limitation of photosynthesis and respiration by inorganic carbon and O2, respectively. Functional similarities between N. spumigena and Aphanizomenon sp. and their impact on the Baltic Sea ecosystem are discussed.

Materials and methods

Sampling

Bulk water samples including N. spumigena colonies were directly collected at the sea surface using 1 l Duran bottles or a bucket in the upper meter of the sea at station B1 (N 58°48′28, E 17°37′60) in the Stockholm archipelago in August 2009. The salinity was 6 and the temperature was 19°C. Samples were immediately brought to the laboratory where they were incubated with stable isotopes or analyzed by the use of microsensors and fluorometry. All experiments were carried out at in situ temperature.

N. spumigena cell sizes, abundance and chlorophyll (Chl.a) colony structure

Cell sizes and abundance were studied using an inverted microscope (Zeiss, Bern, Switzerland) equipped with a digital camera (Axio Cam, Zeiss) connected to an image analysis program (AxioVision, Zeiss). Cell abundance and cumulative trichome length (m l−1) were measured in Lugol-fixed samples in a gridded Sedgewick Rafter counting chamber (Wildlife Supply Company, Buffalo, NY, USA). Triplicate analysis of Chl. a in 0.5 l bulk water samples filtered onto GF/F filters (extracted in 15 ml Falcon vials containing 90% methanol) or in single N. spumigena colonies (extracted in 1.5 ml Eppendorf vials containing 90% methanol), and blanks were performed according to Meeks and Castenholz (1971).

Dissolved inorganic carbon

Samples were injected into an acid stream (30 mM HCl) to convert dissolved inorganic carbon (DIC) into CO2. Afterwards, samples were analyzed in a flow injection system coupled to a conductivity cell (Hall and Aller, 1992). Final concentrations of 13C-bicarbonate in isotope incubation experiments (see below) were measured on a gas chromatography-isotope ratio monitoring mass spectrometry according to Assayag et al. (2006).

Incubations with stable isotopes

Bulk water samples were incubated in 1 l bottles. The bottles were closed with rubber stoppers through which 13C-bicarbonate (9% final labeling) and 15N2 (Sigma) (10% final labeling; Mohr et al., 2010) was injected to the sample, except for three control bottles. The syringe was flushed with Argon gas between injections with 15N2. Three replicates were incubated for 0, 3 or 6 h at 270 μmol photons m−2 s−1 as measured using an LiCOR irradiance sensor, or in dark at in situ temperature (19 °C) in a thermostated room. The light sources were neon lamps. The incubations were stopped by filtration of samples onto pre-combusted GF/F filters (Montoya et al., 1996). The filtrates were filled into 12 ml gas-tight exetainers to which 100 μl saturated HgCl2 solution was added. Subsamples for nanoSIMS analysis were fixed with 2% paraformaldehyde and washed after filtration onto gold-palladium-coated GTTP filters (pore size 0.22 μm; diameter 25 mm; Millipore, Schwalbach, Germany).

Elemental analyzer-isotope ratio monitoring mass spectrometry analysis

The GF/F filters were analyzed through mass spectrometry, using N2 and CO2 released by flash combustion in excess oxygen at 1050 °C to determine the amount of 15N gas and 13C-bicarbonate incorporated into biomass. An automated elemental analyzer coupled to a Delta Plus Advantage mass spectrometer was used (Warembourg, 1993). Before analysis, GF/F filters were freeze-dried and dehydrated by HCl smoke in a desiccator overnight. Round punch-outs (Ø1 cm) of the GF/F filters were packed into tin cups, tightly pressed into a pill-shaped form and loaded onto an autosampler flushed with helium. As a standard, caffeine was used for calibration.

NanoSIMS analysis

For nanoSIMS analysis of single N. spumigena cells, the GTTP filters containing chemically fixed water samples were cut with a round stencil (Ø5 mm) and mounted onto a sample holder. The analysis was performed using a nanoSIMS 50 l manufactured by Cameca (Gennevilliers, France). For each individual cell, we recorded simultaneously secondary ion images of 12C, 13C, 12C14N and 12C15N using four electron multipliers. The measurements and the image and data processing were performed as described previously (Musat et al., 2008; Ploug et al., 2010).

Photosynthesis and respiration measurements in colonies using microsensors

Freshly sampled colonies, which appeared relatively homogeneous in density, were transferred to a Petri dish coated by a 3-mm-thick agar layer (1% w:w) at the bottom and covered by water from the sampling site. The diffusion of gases, ions and solutes in 1% agar is close to that in sea water (Libicki et al., 1988; Revsbech, 1989). Artifacts from a solid boundary, which limits solute exchange between colonies and the surrounding water, was hereby minimized (Ploug and Grossart, 1999). The Petri dish was placed in a thermostated container at in situ temperature. Oxygen concentration gradients were measured using a Clark-type oxygen microsensor (Revsbech, 1989) attached to a micromanipulator. The current was measured by a picoamperemeter (Unisense, PA2000) connected to a strip-chart recorder (Kipp en Zonen). The electrode was calibrated at anoxic conditions and at air saturation. Its 90% response time was <1 s and the stirring sensitivity <0.3%. Its tip was 2 μm wide and its position was observed through a dissection microscope. Three replicates of the O2 concentration gradient at the colony–water interface were measured at 50 μm step increments at steady-state in light and in dark, respectively. The light source was a Halogen Schott lamp (KL 1500 LCD) equipped with an infrared cutoff filter and calibrated using an LiCOR irradiance sensor. The O2 fluxes and the exact position of colony surface relative to the center of the colony were calculated from the measured concentration gradients using the analytical solutions for diffusion in aggregates (Ploug et al., 1997). The molecular diffusion coefficient for O2 of 2.02 × 10−5 cm2 s−1 at a salinity of 6 and at 19 °C was used in the calculations (Broecker and Peng, 1974). O2 fluxes were converted into carbon assimilation rates assuming a photosynthetic quotient of 1.2 (Masotti et al., 2007). The colony dimensions were measured under a dissection microscope with a calibrated ocular micrometer. Surface area and volume was calculated as for ellipsoids (Mass, 1994).

The theoretical limits of DIC and O2 fluxes to colonies were calculated as a function of colony size as described in Ploug et al. (1997). For calculations of DIC fluxes, a bulk concentration of 1.35 mM (as measured) and a diffusion coefficient of 9.4 × 10−6 cm2 s−1 (for HCO3−, the dominant species at pH of the bulk water) were used (Li and Gregory, 1974). For calculations of O2 fluxes, a bulk concentration of 300 μM O2 and a diffusion coefficient of 2.02 × 10−5 cm2 s−1 were used. An apparent diffusivity of gases and solutes within colonies being 0.95 times the molecular diffusion coefficient in sea water was assumed (Ploug et al., 2008). Both fluxes were calculated for a Sherwood number of 1, that is, no effective difference between the motion of the colony and that of the surrounding water.

Ammonium analysis

The net production of 15NH4+ within the plankton community of cyanobacteria and other (micro) organisms was quantified in the filtrate after incubations with 15N2 during 0, 3 and 6 h in light or in darkness using the method by Warembourg (1993). Samples were analyzed on a gas chromatography-isotope ratio monitoring mass spectrometry (VG Isogas Limited, Middlewitch, UK).

Colonies of different sizes, in which photosynthesis at 270 μmol photons m−2 s−1 and dark respiration had been measured, were incubated afterwards for measurements of ammonium release during early evening and night. We used 15 ml acid-washed Falcon tubes rinsed with UV-treated MilliQ water. One single colony was added to each vial with 0.2-μm-filtered sea water from the sampling site, and closed without air bubbles. The vials were positioned on a roller table to avoid diffusion limitation during incubations. The colonies were first incubated at 270 μmol photons m−2 s−1 in <9 h at in situ temperature in a thermostated room. Afterwards, the same colonies were harvested and re-incubated in darkness during 11 h in new vials containing 0.2-μm-filtered sea water from the sampling site. Three controls (nine vials) without colonies were incubated during 0, 9 and 20 h. Ammonium concentrations was measured on a fluorometer (Turner Designs, TD-700) using the method by Holmes et al. (1999). Ammonium concentration distributions within and around colonies were modeled from the measured NH4+ release rate and colony size using the analytical solutions for diffusion in aggregates (Ploug et al., 1997). A molecular diffusion coefficient for NH4+ of 1.62 × 10−5 cm2 s−1 at a salinity of 6 and 19 °C was used in the calculations (Li and Gregory, 1974).

Results

DIC, NH4+ and Chl.a concentrations in the bulk as well as total net C- and N2-fixation rates, and net NH4+ release rates by the bulk plankton community are shown in Table 1. DIC decreased significantly within the beginning of the cyanobacterial bloom, which was composed of Aphanizomenon sp., Anabaena sp. and of Nodularia sp. occurring as free-living trichomes as well as in mm-size colonies. Dinoflagellates and picoplankton were also present at high abundance. Chl.a concentrations reached 9.2 μg l−1 during our study. Nitrate concentrations were <100 nM (Swedish monitoring program), and ammonium concentrations were <200 nM. Total net C- and N2-fixation rates were 896 nmol C l−1 h−1 and 41 nmol N l−1 h−1, respectively, in light-exposed samples and 10-fold lower during darkness. The net production of NH4+ of the newly fixed N2 (traced as 15NH4+ in the bulk after 3–6 h incubation with 15N2) within the bulk plankton community was on average 1.0 and 1.3 nmol N l−1 h−1 during light and dark, respectively.

Table 1 Bulk measurements of DIC, NH4+ and Chl. a during the course of time
Full size table

An example of the isotopic compositions of 13C/12C and 15N/14N in a Nodularia trichome after incubations with stable isotopes during 6 h is shown (Figure 1). Vegetative cells were highly labeled by 13C and 15N after 6 h. In contrast, heterocysts were only faintly labeled by 13C relative to that by 15N. Total net C- and N2-fixation rates were highly variable for individual cells (Table 2). The average net C-fixation rate was 81.6±6.7 fmol C per cell per h, whereas the average net N2-fixation rate was 11.4±0.9 fmol N per cell per h. Paired measurements revealed an average net C:N fixation ratio of 8.0±0.8 (s.e.; n=47). Total net C assimilation by free-living Nodularia trichomes was 38±3 nmol C l−1 h−1 and total net N assimilation was 5.4±3.0 nmol N l−1 h−1. Total average net carbon assimilation rate by colonies (>1 mm) was 50±16 nmol C per colony per h and net N2-fixation rate was 7.4±0.6 nmol C per colony per h. Hence, 18% of the total net bulk N2 fixation was on average associated with one single Nodularia colony per l, and total net N2 fixation by N. spumigena (free-living trichomes and colonies) comprised 31% of total net N2 fixation in the surface waters of the Baltic Sea in light. Dark N2 fixation by N. spumigena was twofold lower than that of N2 fixation in light, and it explained 78% of the bulk N2 fixation during darkness.

Figure 1
figure 1

Distribution of 12C14N−, 13C/12C and 15N/14N in single Nodularia cells as measured by nanoSIMS after 6 h incubations with 13C and 15N2.

Full size image
Table 2 C- and N-fixation rates in Nodularia sp. cells, trichomes and colonies measured by nanoSIMS and microsensors
Full size table

An example of a Nodularia colony is shown (Figure 2). Oxygen concentrations within a 1.3 mm large colony during the early stage of the bloom (31 August) varied between 220 and 510 μM in dark and light at 270 μmol photons m−2 s−1, respectively (Figure 3). Dark respiration rate calculated from the O2 distribution was 5.5 nmol O2 per colony per h. Net photosynthesis was 8.5 and 18 nmol O2 per colony per h at 120 and 270 μmol photons m−2 s−1, respectively. Photosynthesis was saturated in the entire colony at 270 μmol photons m−2 s−1 (data not shown). Volume-specific net photosynthesis was 8.5 and 18 nmol O2 mm−3 h−1 at 120 and 270 μmol photons m−2 s−1, respectively, because the colony volume was 1 mm3. The colony appeared compact and its Chl.a content was 0.41 μg per colony. The Chl. a-specific photosynthesis in this colony was 0.44 mg C mg Chl. a per h at 270 μmol photons m−2 s−1.

Figure 2
figure 2

N. spumigena colony collected in the Baltic Sea.

Full size image
Figure 3
figure 3

O2 distributions measured in a N. spumigena colony at different light conditions. All data points represent the mean value of three series of measurements with the standard of the mean value shown as bars (n=3).

Full size image

Dark respiration measured as a function of gross photosynthesis at 270 μmol photons m−2 s−1 in the same colonies is shown (Figure 4). Dark respiration was a significant linear function of gross photosynthesis during the early stage of the bloom. It was described by the equation: R=0.26 × GP−0.6 (R2=1.00), where dark respiration and gross photosynthesis is measured as nmol O2 per colony per h. After few days, this relationship was described by: R=0.26 × GP+3.3 (R2=0.83), which indicates a higher background respiration by attached microbiota.

Figure 4
figure 4

Dark respiration as a function of gross photosynthesis measured in differentially sized colonies during the bloom. All data points represent the average value of three to five series of measurements of O2 gradients during light and dark, respectively (>250 O2 gradients).

Full size image

Net NH4+ release from colonies to the 0.2-μm-filtered bulk water was measured during light and darkness in colonies of which size, photosynthesis and respiration had previously been measured. The average NH4+ release was sixfold higher in light as compared with that measured in darkness. The average net NH4+ release comprised 8–33% of the estimated gross N2-fixation rate in light, when the ratio of dark respiration to gross photosynthesis was highest (Table 3). The NH4+ concentration gradients corresponding to the average value of measured net NH4+ release rates in light (0.77 nmol NH4+ per colony per h equal to the average value measured on 7 August; Table 3) and in dark-exposed 2 mm large colonies are shown (Figure 5). The maximum concentration varied up to 9 μM in the center of colonies during light when the bulk concentration was 150 nM, thus only N. spumigena colonies were on average <60-fold enriched in NH4+ as compared with the bulk water during light. The net ammonium release and NH4+ gradient were substantially smaller in dark-exposed colonies.

Table 3 Changes in volume, photosynthesis, respiration and ammonium release of Nodularia sp. colonies during time
Full size table
Figure 5
figure 5

NH4+ distributions during dark and light at 270 μmol photons m−2 s−1 modeled from measured net fluxes (Table 3).

Full size image

We compared data of gross photosynthesis measured in Aphanizomenon sp. colonies during the same season and at the same temperature and station in the Baltic Sea, but exactly 1 year earlier (Ploug et al., 2010) with those measured in N. spumigena colonies sampled during this study (Figure 6). Concentrations of orthophosphate and inorganic nitrogen were similarly low during the two studies (Swedish monitoring program). Photosynthesis was saturated during all measurements performed at 270–300 μmol photons m−2 s−1 of both Aphanizomenon sp. and N. spumigena Gross photosynthesis varied almost four orders of magnitude in colonies, with volumes varying five orders of magnitude. It ranged between 0.5 ng C per colony per h and 3.5 μg C per colony per h, and was significant functions of colony volume. For small colonies (<1 mm) it could be described by: GP=250 × (vol)0.81 (R2=0.91, P<0.001; n=20), where gross production is measured as ng C per colony per h and volume is measured as mm3. For large colonies (>1 mm), it was described by: GP=201 × (vol)0.51 (R2=0.51, P<0.001; n=31). Hence, the volumetric gross photosynthesis was lower in large colonies relative to that in small colonies.

Figure 6
figure 6

Gross photosynthesis as a function of colony size of Aphanizomenon sp. (squares) and N. spumigena (circles). All data points represent the average value of three series of measurements.

Full size image

Diffusion to a stagnant sphere increases proportionally to the sphere radius, whereas biological rates of photosynthesis and respiration may increase proportionally to r3, that is, colony volume, as shown in Aphanizomenon sp. (Ploug et al., 2010). Larger colonies may, therefore, be subjected to increasing diffusion limitation of biological processes at a certain colony size. The volumetric net photosynthesis rates measured at saturating light intensities are shown as a function of colony volume (Figure 7a). It was high and invariable with size for small colonies (<1 mm3 equivalent to ∼1.25 mm diameter) of both Aphanizomenon sp. and N. spumigena and hence independent of the potential DIC flux. In larger colonies composed by N. spumigena, however, the volumetric net photosynthesis rates decreased likely as a result of DIC limitation. Prognostic modeling showed that C-fixation rates due to photosynthesis were close to the theoretical maximum fluxes of DIC from the surrounding water to colonies (>1 mm) in the Baltic Sea.

Figure 7
figure 7

(a) Volumetric net photosynthesis rate as a function of colony size of Aphanizomenon sp. (squares) and N. spumigena (circles). The maximum diffusive flux of DIC to colonies is indicated (see text). (b) Volumetric respiration rate as a function of colony size of Aphanizomenon sp. (squares) and N. spumigena (circles). The maximum diffusive flux of O2 to colonies is indicated (see text). All data points represent the average value of three series of measurements.

Full size image

The volumetric respiration rates measured in colonies composed by Aphanizomenon sp. and N. spumigena, respectively, are shown as a function of colony volume in Figure 7b. As in the case for photosynthesis, the smallest colonies by Aphanizomenon sp. and N. spumigena show uniform volumetric respiration rates independent of colony size. In Aphanizomenon sp., these respiration rates were up to 100-fold lower than that needed to create anoxia in the center of a sphere (Figure 7b). Measured respiration rates in Nodularia colonies (>1 mm3) decreased in parallel to the theoretical respiration limit set by diffusion limitation by O2 from the surrounding water to the colonies. Community respiration by N. spumigena with their associated (micro)organisms in these large colonies were, therefore, close to the theoretical concentration limit of respiring biomass under aerobic conditions in the Baltic Sea.

Discussion

Combining microsensor techniques with stable isotope enrichments and nanoSIMS, we have recently shown that Aphanizomenon sp. is a key player in biogeochemical C and N fluxes in the Baltic Sea (Ploug et al., 2010). The present study was focused on the contribution by trichomes and colonies formed by N. spumigena to C and N fluxes in the Baltic Sea. NanoSIMS analysis showed a high enrichment of 15N, which had been transferred from heterocysts to vegetative cells within hours. In contrast, heterocysts were only faintly labeled with 13C as compared to vegetative cells within the same time course, presumable due to the fact that no oxygenic photosynthesis occurs within heterocysts. Similar patterns have been observed in Aphanizomenon sp. (Ploug et al., 2010) as well as Anabaena sp. (Popa et al., 2007). Both the cell-specific net C- and N2-fixation rates by N. spumigena were ca. 2.4-fold higher than that of the respective rates reported for Aphanizomenon sp. in light (Ploug et al., 2010). Cell-specific dark N2 fixation by Nodularia sp. was 20-fold higher than that reported for Aphanizomenon sp. Dark C-fixation was not detectable, and respiration of stored carbon presumably supported this N2 fixation. The volume of N. spumigena cells is approximately 2.5-fold larger than that of Aphanizomenon sp., and volume-specific net C- and N2-fixation rates at light saturation of photosynthesis in N. spumigena were thus slightly lower than those reported for Aphanizomenon sp. The abundance of filamentous cyanobacteria is, at times, measured as meter filament per liter. The net C- and N2-fixation rates associated with 1 m cumulative length of N. spumigena trichomes was almost sixfold higher than those reported for Aphanizomenon sp. (Ploug et al., 2010). The abundance of N. spumigena and Aphanizomenon sp. varies between 1 and 50 m l−1 in the Baltic Sea (Hajdu et al., 2007; Rolff et al., 2007). In this study, the abundance of free-living trichomes of N. spumigena was 1.5 m l−1, only. N. spumigena colonies did not disintegrate in Lugol's solution and their cell content could, therefore, not be quantified as in Aphanizomenon sp. colonies (Ploug et al., 2010). In that study, cell-specific net carbon assimilation rate in Aphanizomenon sp. colonies measured with microsensors was similar to that measured by nanoSIMS. Total net C- and N2-fixation by N. spumigena of this study was dominated by colonies. Both photosynthesis and respiration, which provide the energy for N2-fixation, were so high that they reached the physical limits set by diffusion limitation of bicarbonate and oxygen, respectively, in the Baltic Sea. Hence, despite their small volume and low abundance, N. spumigena colonies can contribute substantially to total N2-fixation and NH4+ release during cyanobacterial blooms in surface waters of the Baltic Sea.

Shear and fluid motion as well as floatation to the air–water interface may enhance mass transfer of O2 and CO2 to colonies when photosynthesis and respiration are limited by the diffusive supply of these gases from the surrounding water to colonies (Ibelings and Maberly, 1998; Ploug et al., 1999; Ploug, 2008). Gross photosynthesis in 1–6 mm thin layers of primarily N. spumigena accumulating at the sea surface has previously been shown to vary between 1 and 7 nmol O2 mm−3 h−1 (Ploug, 2008). The average gross photosynthesis in single colonies of this study varied between 4 and 32 nmol O2 mm−3 h−1 depending on colony size and time during the bloom (Table 3). The average gross production rate by much smaller Aphanizomenon sp. colonies in the Baltic Sea was 78 nmol O2 mm−3 h−1 (Figure 7). The measured carbon-specific photosynthesis of Aphanizomenon sp. collected in the Baltic Sea (Ploug et al., 2010) as well as the Chl.a.-specific photosynthesis of N. spumigena colonies of this study were in the same range as those measured in cultures of the respective cyanobacteria isolated from the Baltic Sea (Moisander et al., 2007; Degerholm et al., 2008).

This study suggests that bicarbonate in addition to CO2 must be an important carbon source in N. spumigena colonies, because the net photosynthesis rate can be as high as the potential diffusive fluxes of DIC to colonies. Net photosynthesis in the smaller Aphanizomenon sp. colonies was 10- to 100-fold lower than that of the potential DIC flux, although these colonies belong to the most productive microenvironments reported in aquatic systems (Ploug et al., 2010). Bicarbonate uptake in cyanobacterial cells must be accompanied either by uptake of H+ or by excretion of OH− to keep balance of charges (Wolf-Gladrow et al., 2007). High uptake rates of H+ are problematic at alkaline pH as shown using a kinetic model of the marine carbonate system (Wolf-Gladrow and Riebesell, 1997; Wolf-Gladrow et al., 1999). The uptake of HCO3− in combination with efflux of OH− leads to an alkaline microenvironment in the colonies. The pH within surface accumulations of N. spumigena has previously been shown to vary between 7.4 during darkness and 9.0 in light (Ploug, 2008). The high photosynthesis rates associated with N. spumigena colonies suggests that this cyanobacterium is adapted not only to high pH in the interior of colonies, but may also benefit from CO2 release through respiration by attached heterotrophic bacteria during photosynthesis.

Previous studies have shown that N. spumigena colonies are microbial consortial communities, and it has long been thought that the close associations of bacteria as well as protozoans and metazoans associated with these colonies are important in nutrient, O2 and CO2 exchange (Paerl and Kuparinen, 2002). Dark respiration and gross photosynthesis was tightly coupled in colonies with a ratio of 0.26. Dark respiration comprised 16% of gross photosynthesis in Aphanizomenon sp. colonies during the same season and at the same temperature in 2008 (Ploug et al., 2010). This difference may be explained by higher dark respiration rates by N. spumigena cells during higher dark N2 fixation as well as by higher densities of heterotrophic bacteria on the N. spumigena trichomes as compared with Aphanizomenon sp. N. spumigena may also harbor heterotrophic N2-fixing bacteria, which have been shown to play a significant role in the Baltic Sea (Boström et al., 2007). Cyanobacterial aggregates have earlier been shown to be anoxic in their interior during an ephemeral stage (12 h) of their decay (Ploug, 2008). During such relatively short periods with very high respiration rates, however, a substantial fraction of organic matter must be respired in aggregates (Ploug et al., 1997). During a light–dark cycle new organic matter is synthesized during daytime. Anoxic conditions have long been suspected to occur within cyanobacterial colonies (Pearl and Bebout, 1988; Hietanen et al. 2002). This study shows that anoxic conditions potentially occur within colonies >1 mm in the Baltic Sea (Figure 7b). The colonies composed by Aphanizomenon sp. are too small and poorly colonized by bacteria for which respiration is too low to develop anoxia inside colonies when the bulk O2 concentration is at air saturation. This was not the case for the larger N. spumigena colonies.

The net C:N fixation ratio varies between 1 and 40 in the bulk plankton community during a developing cyanobacterial bloom in the Baltic Sea, presumably partly due to a variable NH4+ release by the large cyanobacteria (Gallon et al., 2002). In our study, the C:N fixation ratio in the bulk was 20, whereas that of N. spumigena was 8 during light. During a 16:8 h light dark cycle, the net C:N fixation ratio would be (16 × 82−8 × 28):(16 × 11.4+8 × 5.4) =4.8 assuming dark respiration of the cyanobacteria to be 26% of gross photosynthesis. If 25% of the fixed N2 is released as dissolved organic nitrogen and NH4+ to the surrounding water, the resulting C:N fixation ratio would be 6.4 integrated over the diel cycle. Release of dissolved organic nitrogen and NH4+ to the surrounding water, during the early and late light phase, has recently been shown in cultures of N. spumigena (Wannicke et al., 2009), and 15N2 fixed by large cyanobacteria (Aphanizomenon sp. and N. spumigena) in the Baltic Sea has been traced in organisms represented by the 2–5 μm size class, for example, picocyanobacteria (Ohlendieck et al., 2000). N. spumigena appears to use NH4+ inefficiently, but N2-fixation in N. spumigena cultures decreases when bulk NH4+ concentrations are very high (250 μM) (Vintila and El-Shehawy, 2007; Vintila et al., 2010). In our study, the NH4+ release from light-exposed N. spumigena colonies to 0.2-μm-filtered bulk water comprised 8–33% of the estimated gross N2-fixation rates. These are minimum estimates, because dissolved organic nitrogen may be an additional nitrogen source released by the cyanobacteria that attached biota presumably assimilate N for growth. The lowest NH4+ release was measured in colonies with the highest respiration relative to gross photosynthesis (Table 3). N. spumigena is usually colonized by a diverse community of bacteria exhibiting high ecto-enzymatic activities (Hoppe, 1981; Stoecker et al., 2005; Tuomainen et al., 2006), and a significant fraction of the primary production, measured by use of H14CO3 in N. spumigena colonies from the Baltic Sea, was incorporated into the associated bacteria (Hoppe, 1981). Many pelagic bacteria are motile, show chemotaxis and their colonization rates of particles are in the order of minutes (Kiørboe et al., 2002). The nitrogen sources leaking from Nodularia colonies may sustain growth of other biota when bulk concentrations of dissolved inorganic nitrogen are low. A similar substantial NH4+ release was shown to be associated with Aphanizomenon sp. in the Baltic Sea (Ploug et al., 2010). The net production of 15NH4+ within the whole plankton community was 2% and 26% of net 15N2 fixation during light and dark, respectively (Table 1). The low 15NH4+ traced in the bulk sample during 3–6 h incubation in light provides additional evidence that this nitrogen source is rapidly used by other organisms in the plankton community and channeled through the (microbial) food web during daytime when C- and N-fluxes within the plankton community are highest.

Aphanizomenon sp. is a key player in biogeochemical C- and N-fluxes owing to its high productivity and summer blooms lasting 2–3 months in the Baltic Sea (Larsson et al., 2001; Rolff et al., 2007; Ploug et al., 2010). This study shows that N. spumigena, due to its large size and productivity, including high dark N2 fixation, can contribute substantially to total C and N fluxes even at low cellular abundance in the Baltic Sea.