Massive carbonate precipitation episodes in the water column, also referred to as “whiting events”, are a well-known phenomenon of modern freshwater (Schultze-Lam et al., 1997; Hodell et al., 1998; Stanton et al., 2023) and marine environment (Shinn et al., 1989; Robbins and Blackwelder, 1992; Larson and Mylroie, 2014). Whitings are caused by large-scale precipitation of micrometre-sized calcium carbonate particles (visible from space) and represent a major sink in the carbon cycle. The particles associated with whitings can make up a major sedimentary constituent of the modern-day and ancient carbonate rock records (Pomar and Hallock, 2008). Whiting events can be triggered by a combination of biological and physicochemical processes. Among the biological mechanisms that have been studied in this context, picocyanobacterial proliferations have often been invoked in the initiation of whitings (Hodell et al., 1998; Thompson, 2000; Obst et al., 2009). Photosynthesis increases pH levels and alkalinity during cyanobacterial blooms, ultimately causing the saturation state of calcium carbonate to rise, thereby leading to its potential precipitation. The role of Synechococcus spp. bloom-forming cyanobacteria in CaCO3 precipitation has been demonstrated in laboratory experiments (Yates and Robbins, 1998; Dittrich et al., 2003; Obst et al., 2009; Bundeleva et al., 2014; Martinho de Brito et al., 2022) and observed in field investigations (Wells and Iling, 1964; Thompson et al., 1990; Dittrich and Obst., 2004). Change in temperature, salinity and CO2 pressure as well as turbulence are some of the physicochemical factors that can lead to the formation of supersaturated solutions and subsequent precipitation of CaCO3, thus initiating the whiting. Even though several possible biogenic and abiotic mechanisms have been identified, the formation of whitings is still poorly understood.

Phytoplankton blooms, including those of picoplankton, are dense accumulations of cells resulting in a visible discolouration of the surface water layers (Reynolds and Walsby, 1975; Huisman et al., 2018). Their occurrence has been recorded worldwide in marine and freshwater bodies (Paerl et al., 2001; Paerl and Huisman, 2008; Ploug, 2008). Light intensity, water temperature, nutrient availability, weather conditions and hydrodynamics are key factors that determine the onset and persistence of a bloom. Blooms are typically seasonal, frequently observed during late spring or summer, and can be dominated by picoplankton (Huisman et al., 2018). Some phytoplankton organisms, notably cyanobacteria, may produce toxins and form large-scale harmful algal blooms (Paerl et al., 2001). The intensity and frequency of cyanobacterial blooms have been increasing due to anthropogenic eutrophication (Heisler et al., 2008; O'Neil et al., 2012), a trend expected to exacerbate due to climate change (Lürling et al., 2018). Cyanobacteria comprise a diverse group of photoautotrophic organisms that play a pivotal role in global primary production and are key players in the biogeochemical cycles of carbon, nitrogen and oxygen (Callieri and Stockner, 2000; Raven et al., 2017). The unicellular cyanobacterium Synechococcus is one of the most abundant photosynthetic microorganisms on Earth (Whitton and Potts, 2012), which contributes substantially to the picoplankton community in marine (Murphy and Haugen, 1985; Coello-Camba and Agustí, 2021) and freshwater environments (Weisse, 1993), and can form dense blooms (Schultze-Lam et al., 1992; Phlips et al., 1999; Dittrich and Obst, 2004).

During the occurrence of dense phytoplankton blooms, high rates of photosynthetic activity lead to a rapid depletion in CO2 in the surface waters, increasing alkalinity. Depending on the buffering capacity of the water, this could result in pH values ≥ 9 to as high as 11 (Ibelings and Maberly, 1998; Zepernick et al., 2021). Consequently, the inorganic carbonate equilibrium shifts towards carbonate (CO32-). Some cyanobacteria possess a carbon concentrating mechanism (CCM) that converts HCO3- to CO2 through the action of carbonic anhydrase enzymes (Price et al., 1998; Badger et al., 2002) and produce hydroxide ions (Kupriyanova and Pronina, 2011). The activity of extracellular carbonic anhydrase (eCA) may contribute to the creation of an alkaline microenvironment in the extracellular polymeric substances (EPSs) surrounding the cyanobacterial cells (Price et al., 2002; Dupraz et al., 2009). When OH- ions are released during photosynthesis it causes the pH to rise, which favours carbonate mineral precipitation, assuming there are enough calcium ions available (Kamennaya et al., 2012). Consequently, during blooms, carbonate minerals can form on EPS or be precipitated in the microenvironment surrounding cyanobacterial cells.

Cyanobacteria are known producers of EPS, especially during blooms (Pannard et al., 2016; Liu et al., 2018). EPSs serve as a boundary between cells and their immediate environment (Whitton and Potts, 2012) and may act as a template for CaCO3 nucleation (Dupraz and Visscher, 2005; Dupraz et al., 2009; Kamennaya et al., 2012). EPSs are high-molecular-weight organic molecules composed of polysaccharides, proteins, nucleic acids and lipids (Pereira et al., 2009; Marvasi et al., 2010; Decho and Gutierrez, 2017). This complex mixture of molecules may contain specific monomer components, such as uronic or sialic acids (monosaccharides), aspartic or glutamic acids (amino acids), or functions (sulfate, phosphate), which carry negative charges in physiological conditions and can therefore bind cations, such as Ca2+, and promote the nucleation of CaCO3 crystals (Trichet and Defarge, 1995; Dupraz et al., 2009; Walker et al., 2019). Conversely, polyanionic EPS in solution can inhibit crystal growth by poisoning the faces of growing nuclei by an adsorption mechanism, according to a classical and accepted view prevailing for other macromolecules of similar charge properties: synthetic peptides (Wheeler et al., 1981), skeletal proteins (Wheeler et al., 1981; Addadi and Weiner, 1985), coccolith-associated polysaccharides (Borman et al., 1982) or natural organic matter dissolved in seawater (Mitterer and Cunningham, 1985). The production and composition of EPS differ among different species of microorganisms and their type of metabolism and depend on the environment in which they live, stressors (e.g. nutrient availability, pH, temperature, light, salinity) and the stage of their growth (Pereira et al., 2009; Pannard et al., 2016; Martinho de Brito et al., 2022). The deprotonation of functional groups at elevated pH enhances the binding capacity of cations such as Ca2+ and controls crystal nucleation and growth by reducing the interfacial energy barrier between the crystal and the EPS substrate (Dupraz et al., 2009; Dittrich and Sibler, 2010). EPSs play a twofold role in carbonate formation by initially inhibiting (through Ca2+ binding) and subsequently promoting carbonate precipitation by releasing calcium ions during EPS alteration and degradation (Dupraz and Visscher, 2005). Furthermore, through specific functional group composition and structural architecture, EPS may also exert control over the mineralogy, morphology and/or abundance of the minerals that are formed (Trichet and Defarge, 1995; Dupraz et al., 2009).

We have previously reported that the pH of Synechococcus cultures increased when grown in a non-buffered medium (Martinho de Brito et al., 2022). In these growth conditions, the production of EPS was enhanced compared to growth in a buffered medium. Furthermore, the EPS from cells grown in non-buffered conditions contained more negatively charged functional groups that impacted the properties of the carbonate minerals that precipitated. The current study further investigates the properties of EPS produced during different growth phases of Synechococcus spp. over an extended incubation time (mimicking a prolonged natural bloom). We aim to better understand the role of cyanobacterial blooms in carbonate precipitation through EPS production and develop a conceptual model of picoplankton-mediated organomineralization to explain the biological origin of whiting events.

Synechococcus PCC7942 was obtained from the Centre de Ressources Biologiques de l'Institut Pasteur (Paris). Cultures were grown in a one-third-strength non-buffered liquid BG-11 medium (Allen, 1968; Rippka et al., 1979). The medium consists of (per litre) 1.5 g NaNO3, 0.04 g K2HPO42H2O, 0.075 g MgSO47H2O, 0.036 g CaCl22H2O, 6 mg citric acid combined with 6 mg ferric citrate, 0.001 g Na2EDTA2H2O and 0.02 g Na2CO3. Trace metal solutions contained (per litre) 2.86 mg H3BO3, 1.81 mg MnCl24H2O, 0.222 mg ZnSO47H2O, 0.39 mg Na2MoO42H2O, 0.079 mg CuSO45H2O and 0.0494 mg Co(NO3)26H2O. Cultures were incubated at room temperature (21 ∘C ± 2), with a cycle of 12 h of light and 12 h of darkness under 36.8 µEm-2s-1 photon irradiance while shaken at 200 rpm with the Cimarec i multipoint stirrer (6 position, 2000 rpm, 3 L per multipoint, 100–240 VAC rotary shaker).

Two independent growth experiments were performed in 1 L glass serum bottles containing 800 mL 1/3 BG-11 medium adjusted to pH 7.5, sealed with silicone caps to allow gas exchange. Cells used for the inoculum (pH = 9.2) were pre-cultured in a full-strength BG-11. Immediately after inoculation (30 mLperbottle), the pH increased to approximately 8.2.

EPSs were extracted from the Synechococcus cultures as previously described by Martinho de Brito et al. (2022). EPSs were harvested after 14, 28 and 56 d of cultivation. Cyanobacterial cells were inspected by microscopy to ensure that no cell lysis had occurred during the extraction process. The pure EPS fractions were obtained by ultrafiltration (retentate > 10 kDa) to reduce volume. The weight of the material was determined after dialysis (using a 1 kDa membrane) and lyophilization; the measurement was carried out on a high-precision analytical balance (Quintix 35-1S, Sartorius, Göttingen, Germany).

The potential of the EPS matrix to interact with the precipitation of calcium carbonate was tested via the diffusion method in the presence of a closed ammonia–CO2-saturated atmosphere (Albeck et al., 1993). A total of 200 µL of the mixture containing pre-filtered (0.22 µm) CaCl2 solution (10 mM) and EPS at increasing concentrations (3, 18, and 36 µgmL-1) were incubated in duplicate in 16-well plates (Lab-Tek, Nunc/Thermo Scientific, Rochester, NY, USA). The EPS concentrations were selected to match the EPS yields at the extraction times (14, 28 and 56 d of cultivation). The plastic covers of the well plates were perforated to allow the reaction between CaCl2 solutions containing EPS and ammonium bicarbonate. The well plates were placed in a desiccator that was incubated at 4 ∘C in the dark for 72 h. At the completion of the incubation period, the pH value was measured in each well, the overlying solutions were carefully removed to dryness and CaCO3 crystals were analysed. Blank experiments were performed without any EPS. The experiment was carried out in duplicates.

All the data concerning Synechococcus growth and EPS production are representative of two independent experiments with two technical replicates (four replicates for EPS extracted at 56 d of culture). The results are reported as the mean ± standard error of the mean. Statistical significance was assessed by performing single-factor ANOVA tests; p values < 0.05 were statistically different.

Cell density and pH values increased over the Synechococcus cultivation period (Fig. 1a and b). The growth of Synechococcus cells showed a typical pattern including a brief lag phase (∼ 6–7 d) followed by a 7 d (Experiment I) and 14 d (Experiment II) exponential phase and finally a stationary phase. The stationary phase (early stationary phase) was reached after 14 and 21 d of growth in Experiment I and II, respectively, and lasted until day 56 of cultivation in both experiments (late stationary phase) (Fig. 1a). Growth experiments I and II started with a similar cell density of approximately 1010 cellsL-1 and demonstrated reproducible growth patterns (p value = 0.91). At the time of inoculation, cell density was 9.5 × 1010 in Experiment I and 7.11010 cellsL-1 in Experiment II (Fig. 1a). Synechococcus grew exponentially until reaching a maximum of 1.7 × 1012 in Experiment I at 14 d of growth and 1.5 × 1012 cellsL-1 after 21 d of growth in Experiment II. At the end of the exponential growth phase, the cell numbers levelled off and achieved a stable growth stage (stationary phase). Typical evolutions of pH values in culture media during the Synechococcus growth experiments are presented in Fig. 1b. As a general trend, pH is linked to the photosynthetic activity of cyanobacteria. The pH levels rose rapidly during the exponential phase in both experiments, reaching around 10, and stayed steady during the stationary phase. While Experiment I experienced significant pH fluctuations during the latter part of the stationary phase, overall, the pH evolution trends for both experiments are comparable (p value = 0.91; Fig. 1b). The p values for pH and cell numbers showed that the two independent growth experiments are not significantly different.

The activity of extracellular carbonic anhydrase (eCA) in solution changed slightly over the growth experiment (Fig. S1 in the Supplement). The highest eCA activity (∼ 1600) was detected after 14 d of culture during the exponential phase. The lowest activity was measured after 56 d of growth in the late stationary phase.

High nitrate concentrations supported exponential growth and high cell density (Table 1). The results show that a major decrease in nitrate and phosphate concentrations occurred during the exponential growth phase and remained slowed down progressively over the stationary phase. At the end of the stationary phase, the phosphate concentration had decreased to approximately 30 % of its initial level. On the other hand, the nitrate concentration was still high, with approximately 67 % of its initial concentration remaining. Ammonium concentration was below the limit of detection (2.22 µM). Calcium concentrations decreased gradually and accounted for a total calcium concentration of 81 % in the late stationary phase. Other medium constituents should be present in excess and were thus not measured.

The recovery yields of the EPS produced (mean ± SD) resulting from the applied extraction method are listed in Table 2. The EPS yields varied from 2.9 ± 0.5 to 18.6 ± 2.1 mgL-1 during exponential and early stationary phases and reached the highest yield of 35.4 ± 4.2 mgL-1 at 56 d of culture in the late stationary phase (Table 2). When the values were normalized per cell yield, results showed that the EPS concentration increased significantly between the exponential and late stationary phases (p value < 0.05) (Fig. 2). Synechococcus continuously produced EPS during the 56 d experiment. In the first 14 d of growth, cells grew exponentially and EPS production was deficient. Between exponential and early stationary phases, EPS production increased by a factor of 5 to 7, reaching a maximum in the late stationary phase, after the 56 d growth experiment.

Forced CaCO3 experiments were performed using a control solution (without EPS) and EPS solutions, at the same pH, with concentrations of 3, 18 and 36 µgmL-1. Each concentration corresponds to the EPS yield at different growth stages: exponential phase (= 3 µgEPSmL-1) as well as early (18 µgEPSmL-1) and late (36 µgEPSmL-1) stationary phases. The crystals formed in the various EPS solutions showed different morphological (Fig. 6) and mineralogical (Fig. S3 in the Supplement) features as well as distinct crystal sizes and distributions compared to those formed in the control solution (Fig. 7).

Macronutrients, such as nitrogen (N) and phosphorus (P), promote the initiation of cyanobacterial blooms (Reynolds and Walsby, 1975; Paerl, 2008; Xu et al., 2015). In our growth experiment, the beginning of the exponential phase (and the persistence of bloom) (Fig. 1a) was positively correlated with the high initial nutrient concentration in the medium (Table 1). Environmental factors such as water temperature, light intensity. hydrodynamics and availability of dissolved inorganic carbon (DIC) are also important determinants of cyanobacteria bloom development (Clark and Flynn, 2000; Dokulil and Teubner, 2000; Havens, 2008). Blooms can dramatically alter the supply of inorganic carbon for photosynthesis, which causes the pH to increase (Ibelings and Maberly, 1998). In the early exponential phase of our batch cultures, the high photosynthetic activity of cyanobacteria cultures resulted in fast pH increase, thereby reducing the total inorganic carbon of the grown medium. Light and CO2 are the sources of energy and carbon for cyanobacteria and are of critical importance for their growth (Takahashi et al., 2004). At pH 9 (Fig. 1b), the concentration of CO2 predicted is close to zero (< 1 µM), and the HCO3- concentration is 475 µM (PHREEQC data). A similar scenario was observed in natural bloom occurrence: the population of cyanobacteria draws down the partial pressure of CO2 (pCO2) in the photic zone, increasing the surface water pH by up to 9–10 (Ibelings and Maberly, 1998; Verspagen et al., 2014), and CO2 concentration can become completely depleted or reach values close to zero (Maberly, 1996). Under extreme conditions, the concentration of HCO3- can also become markedly reduced (Talling, 1976; Maberly, 1996). When the rate of photosynthesis is greater than the combined rate of resupply of CO2 from the atmosphere and DIC in the hypolimnion, deviation from the air equilibrium occurs, favouring CaCO3 precipitation. The pH of most aquatic systems ranges from 7.5–8.1 and keeps inorganic carbon primarily in the form of bicarbonate (O'Neil et al., 2012). In poorly buffered systems, such as highly productive lakes, the pH and speciation of DIC experience large fluctuations which vary widely on a scale from daily (diel) to episodic to seasonal (Maberly, 1996), with diel variations as high as two pH units and 60 µmolDICL-1 (Maberly, 1996). Because CO2 favours the C3 photosynthesis (C3 cycle operation of Calvin–Beson cycle), the high pH of ∼ 10 in our growth medium could be associated with carbon limitation (Ibelings and Maberly, 1998; Verspagen et al., 2014).

To alleviate CO2 limitation, cyanobacteria have developed an efficient CO2-concentrating mechanism (CCM) (Aizawa and Miyachi, 1986; Badger and Price, 1992; Badger et al., 2002; Burnap et al., 2015) and can use bicarbonate as an inorganic carbon source for photosynthesis (Price et al., 1998; Giordano et al., 2005; Sandrini et al., 2016). By activating CCM, cyanobacteria concentrate CO2 by a factor of up to 1000 (Badger and Andrews, 1982; Badger et al., 2002; Price, 2011). CO2-deficient conditions experienced during the exponential phase of our growth experiment, coupled with the continuous cellular demand for inorganic carbon to support photosynthetic carbon fixation, likely led the cells to activate CCM. The predicted concentrations of CO2 and HCO3- in the growth medium (PHREEQC data) in the early and late exponential phase infer that Synechococcus cells were actively transported across the membrane and accumulated DIC into the cell, where the HCO3- pool was utilized to generate elevated CO2 levels around RuBisCO (ribulose-1,5-bisphosphate carboxylase/oxygenase; Badger et al., 2002; Price et al., 2008). The CCM of cyanobacteria accomplishes very high carbon-concentrating factors (Cexternal : Cinternal) at deficient specificity factors of RuBisCO (Tortell, 2000; Tortell et al., 2000). CCM involves bicarbonate transporters in the cell membrane, intracellular (iCA) and extracellular (eCA) carbonic anhydrase enzymes, and concentrated RuBisCO activity located in carboxysomes (Badger et al., 2006; Price et al., 2008; Rae et al., 2013). CA converts HCO3- to CO2 (Badger and Price, 1994), which increases the external pH in close proximity to the cells. In our study, eCA activity was ∼ 1.6–2.0 times higher during the exponential growth phase and reduced gradually through the stationary phase (Fig. S1). The strongly stained band only present in the exponential-phase EPS at around 45–47 kDa (Fig. 4, lanes 2, 3) may be indicative of eCA, as reported by Kupriyanova et al. (2018), but this requires further investigation. Another plausible explanation for the 45–47 kDa band could be the presence of chlorophyll f synthase, which typically migrates at around 46 kDa (Shen et al., 2019). Similarly, Yang et al. (2023) measured the CA anhydrase in solution over a 30 d growth experiment with Synechococcus PCC7942 and reported an increase over the lag phase and large fluctuations over the exponential phase. During the stationary phase, CA did not vary greatly, but a minor decrease was recorded in the late stationary phase (Yang et al., 2023). In our study, the higher eCA activity recorded could explain the strongly stained ∼ 45–47-kDa band that was only identified in our SDS–PAGE gels of EPS produced during the exponential phase (Fig. 4, lanes 2–3). The molecular weight (MW) of this band is similar to a 42–43 kDa eCA previously identified by Kupriyanova et al. (2018) and discussed by Martinho de Brito et al. (2022). As explained in the Results, Sect. 3.4.3, we cannot exclude that the band is chlorophyll f synthase, which seems to show up around 46 kDa. A more substantiated demonstration of the identity of the SDS–PAGE band will require other approaches (beyond the scope of the present study), such as micro-sequencing of the purified 43 kDa band or the use of a CA-specific antibody.

Active uptake of HCO3- and accumulation of Ci species require the input of metabolic energy, e.g. ATP (BCT1 HCO3- transporter), NADPH or reduced ferredoxin (CO2 uptake), or coupling to an electrochemical Na+ gradient (SbtA or BicA HCO3- transport) (Badger et al., 2002; Price et al., 2008). This energetic cost may therefore reflect the growth rates achieved. Synechococcus PCC7942 grows at > 80 % of its maximum growth rate when provided with HCO3- as its main inorganic carbon source (Miller et al., 1984). During the exponential phase, the carbon production from photosynthesis is mainly allocated to biomass production, not to EPS synthesis. During this phase (Fig. 2 and Table 2), the small amount of EPS produced comprises a higher proportion of sugars and lower amounts of protein and GAG compared to EPS produced during the stationary phase (Fig. 3 and Table 3). Our study indicates that rather than proteins, sugars are the major component in all EPS extracts. This finding is supported by the data obtained from FT-IR analysis (Fig. 3). The smaller amount of negatively charged groups of the EPS during the exponential phase (Fig. 4, lanes 2–3) compared to those of EPS from the early and late stationary phases (Fig. 4, lanes 4–7) resulted in weak to moderate inhibitory capacity (Fig. 5a). The main phenomenon observed in the pH-drift assay (Fig. 5) is the initial Ca binding to negatively charged groups in EPS prior to carbonate addition, which initiates CaCO3 precipitation. This results in a decrease in pH. The pH-drift assay showed that EPS from the exponential phase (Fig. 5c) has a larger plateau and thus a lower calcium-binding capacity than the EPS from the stationary phase. This observation was further corroborated by the forced precipitation experiments, which showed that EPS from the exponential phase induced a small amount of mostly large-sized carbonate crystals (> 50 µm), very similar to the negative control experiment (Fig. 7) (Martinho de Brito et al., 2022). The high concentration of Ca2+ in the medium (83 µM) compared to the initial (Ca2+) at the beginning of the experiment (103 µM) indicates that a small amount of calcium ions was bound to negatively charged functional groups of EPS (Table 1; see (Ca2+)).

In our batch experiment, cells continue to grow exponentially for ∼ 20 d of cultivation. At this point, cultures reached the maximum cell density (Fig. 1a), and pH values ranged between 10–11 (Fig. 1b). Based on our calculations, under these alkaline conditions, CO2 was completely depleted (1.7 × 10-3 µM) in the growth medium, whereas HCO3- was extremely low (∼ 79 µM). Thus, the dominant inorganic carbon speciation was CO32- (421 µM). Because cells cannot take up CO32-, and HCO3- concentration seems to be insufficient to cover the carbon demands of cyanobacterial growth, we assume that this may have been the cause of cell numbers starting to level off (Fig. 1a, early stationary phase). Consequently, cultures entered a stationary state due to a lack of inorganic carbon availability required to increase cell population (Miller et al., 1984; Mayo et al., 1989; Verspagen et al., 2014). The excess of nutrients measured in the medium in the late exponential phase (Table 1) suggested that the specific growth rate was not limited by nutrient availability but by a rather low level of CO2 carbon content.

Insufficient CO2 availability is considered to be the external stress factor constraining the growth rate of cyanobacteria (Maberly, 1996; Hein, 1997; Ibelings and Maberly, 1998), and low (HCO3-) could sustain a constant population density for at least ∼ 40 d (see Fig. 1a, stationary phase). Our results suggest that at this point carbon fixation was allocated to EPS synthesis, not to biomass production (Miller et al., 1984). Increased EPS production is usually associated with external stress factors (Rossi and De Philippis, 2015), including high-pH conditions (Martinho de Brito et al., 2022). Moreover, metabolic stress may also alter the composition of EPS (Babele et al., 2019; Martinho de Brito et al., 2022). In the present study, the negative functional group abundance increased, resulting in a higher acidity of EPS (Fig. 4, lanes 4–5) due to an increase in protein and sulfated glycan (GAG) (Table 3). In the pH conditions of the early stationary phase, all the functional groups of the EPS matrix are deprotonated and are able to bind calcium ions present in the solution (Fig. 5b) (Dupraz and Visscher, 2005; Braissant et al., 2007; Dittrich and Sibler, 2010). This EPS binds calcium more efficiently than nanometric nuclei in formation (if their formation is thermodynamically favoured). We suggest that the increased calcium-binding capacity of the EPS probably accounts for the lower Ca2+ concentration measured in the medium (Table 2; see (Ca2+)). In our in vitro-forced precipitation assay, we measure the second effect, the inhibitory one (mineral-binding effect), which results in the production of small-sized calcium carbonate crystals (< 50 µm), in comparison to what happens in the exponential phase (Fig. 7).

As mentioned above, we assume that the continuous increase in EPS production over the late stationary phase, including an overall augmentation of negatively charged functional groups (Fig. 4, lanes 6–7), including GAG content (Table 2), might be a specific response to a stress scenario. As expected, the present study shows that the greater amount of negatively charged functional groups of EPS from the late stationary phase (Fig. 4, lanes 6–7) resulted in a higher Ca-binding capacity than exponential- and early-stationary-phase EPS (Fig. 5c). Our forced precipitation experiments showed that minerals produced in the late-stationary-EPS solutions are smaller and more abundant than those formed in EPS solutions from the early stationary phase (Fig. 7). Under natural conditions, when the Ca2+ supply is continuous, the crystals may or may not continue to grow, depending on the physical space within the EPS matrix (Dupraz et al., 2009). Based on the high concentration of nitrate (4720 µM) measured in the late stationary phase (Table 1), we assume that the abundance of this nutrient supported the persistence of the stationary phase, i.e. similar to a prolonged bloom in natural conditions. The death phase was not observed in our 56 d long experiment. Given that our cultures were continuously stirred, we can assume that light was not limiting cyanobacterial growth. Furthermore, in natural blooms, the increase in population density may affect cells at greater depth through self-shading by decreasing the light available for photosynthesis (Townsend et al., 1994). Yet, cyanobacteria (including Synechococcus) are known to be well-adapted to low-light conditions (Campbell and Carpenter, 1986; Palenik, 2001; Callieri et al., 2011). Additionally, the presence of sulfated constituents on late-stationary-phase EPS contributes to a higher negative charge of the matrix and higher Ca-binding potential (Decho and Kawaguchi, 2003; Skoog et al., 2022) compared to EPS extracted in the exponential phase, which contained significantly lower GAG (Table 2). The present study shows that the greater amount of negatively charged functional groups of EPS from the late stationary phase (Fig. 4, lanes 6–7) resulted in a higher Ca-binding capacity than exponential- and early-stationary-phase EPS (Fig. 5c). Our forced precipitation experiments showed that minerals produced in the late stationary EPS solutions are smaller and more abundant than those formed in EPS solutions from the early stationary phase (Fig. 7), suggesting an increased inhibitory ability of the late stationary EPS.

Our observations made during exponential and stationary phases can be applied to generate a conceptual model of EPS properties during a Synechococcus bloom event (Fig. 8a–c). The onset of a bloom starts with an increase in cell numbers, with high values in spring–summer (exceeding 105–107 cellsmL-1) and lower values in winter months (< 105 cellsmL-1) in both marine (Agawin et al., 1998; Phlips et al., 1999) and freshwater (Maeda et al., 1992; Tai and Palenik, 2009) environments. This resembles the exponential growth phase in our study (Fig. 1, exponential phase). We predict that during the initial phase of a natural bloom, there is little EPS production: cells grow relatively quickly, and the carbon fixed during photosynthesis is predominantly allocated to biomass production (Fig. 8a). The fast growth is followed by a phase during which cell numbers level off, typically due to stress conditions, which is represented by the early stationary phase in our study. Under certain conditions, blooms can be sustained for weeks and possibly longer (Anderson et al., 2002; Havens, 2008; Zhao et al., 2013), similar to what we observed in our growth experiments (Fig. 1a, early stationary phase). The maintenance of a bloom requires continuous input of nutrients, which is also the case in our experiment (Table 1), or in the case of natural systems, a turnover from lysing cells recycled by other microbes. During this phase, we did not observe a significant increase in cell density, but the production of EPS continued at a disproportionately high rate (Fig. 8b and c). Our findings are in agreement with the lab studies using diatom cultures, which show that EPS production is low during exponential growth and increases in the stationary phase (Myklestad and Haug, 1972; Myklestad et al., 1989; Bhosle et al., 1995). These authors reported that nutrient-deficient conditions enhanced the production of EPS over the growth phases. If carbon fixation continues and some critically required nutrient is lacking from the growth medium, most likely the phototrophic organisms produce carbohydrate reservoirs (Ciebiada et al., 2020). These include storage polymers like glycogen and the production of other carbohydrate-rich compounds, including EPS (De Philippis et al., 1996, 2001; Decho and Gutierrez, 2017). The decline in blooms in natural environments is typically associated with nutrients, low or high light intensity, grazing, or viral infection. Under these stressful conditions, an increase in EPS production by the phyto-/picoplankton community may be expected.

Synechococcus spp. blooms can cause whiting events (Thompson, 2000), characterized by the presence of large amounts of CaCO3 minerals in surface water. Various mechanisms have been proposed for this phenomenon, including chemical and physical processes (Shinn et al., 1989; Larson and Mylroie, 2014), as well as biologically mediated precipitation (Thompson and Ferris, 1990; Robbins and Blackwelder, 1992; Stanton et al., 2023). However, no consensus has been reached on the precise cause of these events. Carefully transporting the results from forced precipitation experiments to a whiting event, we suggest that early in the bloom (Fig. 8a), relatively large CaCO3 crystals form, provided sufficient Ca2+ is available (Fig. 8d). As the bloom continues to grow, progressively the larger quantity of negatively charged functional groups in the EPS provides more cation-binding sites and thus inhibits calcium carbonate precipitation largely. Depending on the three-dimensional structure of the EPS and surface properties (Wang et al., 2012), nucleation may yield smaller CaCO3 crystals (Fig. 8). If this occurs, then the production of a more negatively charged matrix (largely contributed by the enrichment in sulfated polysaccharides) may offer some selective advantage to the cyanobacteria population by inhibiting and/or delaying mineral precipitation and by reducing crystal size formed around the cells. This might result in slow sinking rates, extending the residence time of the cyanobacterial community in the photic zone. If the bloom occurrence is short (e.g. similar to 14–28 d in our growth experiment), minerals making up the whiting will be relatively larger. Consequently, the aggregates of cyanobacteria, EPS and CaCO3 minerals may sink faster because mineral precipitation in EPS increases the cyanobacterial-specific density severalfold. The Synechococcus-specific density (ρ) is 1.040 gcm-3 (Reynolds, 1987), near-neutrally buoyant, whereas ρcalcite is 2.710 gcm-3 (Dean, 1999). The production of larger amounts of more negatively charged EPS may act as a protection mechanism against carbonate formation in the vicinity of the cell wall (Martinez et al., 2010; Bundeleva et al., 2012), thus allowing the organisms to reside longer in the photic zone. Interestingly, the production of EPS that contained sulfated groups among bacteria seems to be exclusive to cyanobacteria (Pereira et al., 2009; Maeda et al., 2021). Maeda et al. (2021) reported that the cyanobacterium Synechocystis 6803 produced large amounts of GAG compounds during an experimental bloom formation. The authors suggested that these constituents can be advantageous for the development of surface bloom, as it may increase the buoyancy, permitting cells to migrate upward rapidly when the water column is stable (Walsby et al., 1995). Thus, GAG production may be considered to be an alternative for organisms that lack gas vesicles to remain longer in the photic zone (Maeda et al., 2021). The negative charge of EPS produced containing high sulfated content also protects the community against viral infection (Baba et al., 1988; Ghosh et al., 2009). Therefore, the production of GAG by pelagic cyanobacteria contributes to stress tolerance and viral infectivity, helping in the persistence of bloom. In our growth experiments, a decline in cell numbers was not observed, which would represent the end of the bloom. In the natural environment, nutrient depletion, grazing or viral lysis/infection are the most likely causes of terminating a bloom (Gons et al., 2002). The cell lysis releases organic matter, which supports the growth of heterotrophic bacteria (Kjelleberg et al., 1987; Hagström et al., 1988; Kieft et al., 2021). Photosynthetically derived organic carbon is one of the major carbon and energy sources for heterotrophic bacteria (Allgaier et al., 2008). These heterotrophs can degrade EPS and liberate bound Ca2+ (Visscher et al. 1998; Ionescu et al., 2015; Diaz et al., 2017). In addition, microbial respiration will produce HCO3-/CO2, increasing the saturation index of CaCO3, and may enhance the whiting (Fig. 8). Although our model is (somewhat) largely theoretical at this stage and explores the role of one picoplankton species, its merit is to focus on an overlooked actor of whiting events, the EPS. Furthermore, it provides a conceptual framework to work with, for designing novel experiments and measurements both in natural systems and at the lab bench, to validate the molecular mechanisms involved in microbial-bloom-associated CaCO3 formation in marine and lacustrine models.