Strategies adopted by Aphanizomenon flos-aquae in response to phosphorus deficiency and their role on growth

The N2-fixing cyanobacterium, Aphanizomenon flos-aquae is a globally distributed bloom causing species that degrades water quality of fresh and marine water bodies. Overcoming phosphorus (P) deficiency is one of the ecological advantages for bloom-forming cyanobacteria. It remains unclear to what extent can A. flos-aquae alleviate P deficiency by regulating P using strategies. Based on in situ observations of extracellular alkaline phosphatase (APase) in A. flos-aquae via enzyme-labeled fluorescence in freshwater bodies in China, Poland and Czechia, we further investigated responses of isolated A. flos-aquae to different P supplies (dissolved inorganic P (Pi) as +DIP, dissolved organic α-glycerophosphate and β-glycerophosphate as +DOPα and +DOPβ, P-free condition as P-depleted). The significantly negative relationships between percentage of APase producing cells and soluble reactive P concentration in both fields and cultures suggested that the excretion of APase in cyanobacterium was regulated by ambient Pi supply. Suffering from P deficiency in the P-depleted treatments, A. flos-aquae showed the highest APase activity but a vigorous growth at the early culture stage, which might also benefit from the formation of polyphosphate body (PPB) and the decrease of cell P quota. In the +DOP treatments, the coordination of dissolved DOP mineralization and continuous prompt utilization of PPB might contribute to a maintenance but not reproduction of A. flos-aquae when relying on DOP, since the specific growth rate kept around 0 cells L−1 day−1 at the second half culture period and the highest cell density reached only 13.38% of that in +DIP treatments while photochemical efficiency was comparable during the whole experiment. Luxury uptake of phosphate as PPB in the +DIP treatments was consistent to the canonical view of polyphosphate as P storage. Aphanizomenon flos-aquae could achieve an instantaneous growth in response to P deficiency with the coordination of P utilization strategies, while it maintained a long-term sustainable growth but not reproduction under sole DOP supply. Persistent and active reproduction could only be achieved in high Pi supply, which implying that an effective consequence can be expected for combating the bloom of A. flos-aquae when controlling P supply.


Background
Filamentous cyanobacterial blooms of the diazotrophic genus Aphanizomenon occur worldwide and pose deleterious effects on aquatic ecosystem [1][2][3][4]. Theoretically, the nitrogen demand of the developing population of Aphanizomenon can be provided by nitrogen fixation in heterocysts [1,5,6]. Thus, the phosphorus (P) scarcity can most often lead to its growth limitation. For example, the first appearance of Aphanizomenon in Lake Kinneret in August 1994 was apparently elevated by relatively high total dissolved P concentrations [1]. Aphanizomenon behaved sensitively when exposed to changes in ambient dissolved inorganic P (Pi) so that it might be adapted to environments with higher P concentrations [7]. Orthophosphate in low concentrations led to a strong decrease in the growth of A. issatschenkoi, A. aphanizomenoides and A. gracile [8]. Correspondingly, bloomforming cyanobacteria developed various strategies to overcome P deficiency. A strategy for cyanobacteria to survive in P deficiency status was that they store P in the form of polyphosphate bodies (PPB) in cells when P is abundant, and breakdown it upon P stress [9]. Polyphosphate granules observed in cells of Aphanizomenon sp. from blooms in Baltic Sea suggested that the population was not subjected to P starvation [10]. Furthermore, Aphanizomenon had the ability to decrease cell P quota when P in their environment is scarce [7]. In addition, Aphanizomenon could meet its P demand by expressing extracellular alkaline phosphatase (APase), enzymes hydrolyzing bio-available phosphate from dissolved organic P (DOP) molecules [7, [11][12][13]. Like the Aphanizomenon in Lake Kinneret, whose biomass increased accompanying high enzymatic alkaline phosphatase activity (APA) in phosphate limited summer and autumn [1]. Aphanizomenon sp. which is one main bloom-forming filamentous diazotrophic species in Baltic Sea showed cell-surface-bound phosphatase activity as indicated by enzyme-labeled fluorescence (ELF) assays [13]. On the contrary, a combination of field and lab-cultures study showed that A. ovalisporum excreted APase only under severe P deficiency, it firstly enslaved another phytoplankton other than itself to produce APase [14]. In summary, APase production in the development of Aphanizomenon bloom at the single cell level in freshwater is rarely studied. Furthermore, it is still unclear to what extent can A. flos-aquae alleviate P deficiency by regulating P using strategies.
In this study, single-cell assays of APase were conducted at species level using the ELF technique in 2 Chinese lakes, 1 Czechia reservoir and 1 Polish reservoir. Furthermore, laboratory studies were performed with batch cultures, and three strains of Aphanizomenon flosaquae (A. flos-aquae FACHB 1249, A. flos-aquae FACHB 1171 and A. flos-aquae FACHB 1260) isolated from three Chinese shallow lakes such as Lake Chaohu, Lake Taihu and Lake Dianchi were inoculated in medium with four P status with additions of Pi and two DOP species, respectively, as well as without P addition. Multiple variables related to algal growth such as cell density and photosystem II quantum efficiency (Fv/Fm), P cycling such as bulk and size-fractionation APA, concentrations of soluble reactive P (SRP) and DOP in culture medium, as well as the ELF labeling and PPB containing cells of the strains were determined at different intervals. Objectives of this study were (1) to investigate whether A. flos-aquae could employ its own APase in acquiring P from DOP for cell growth and reproduction, and (2) to evaluate to what extent can A. flos-aquae abate the P limitation with various P using strategies.

Field sample collection and analysis
Field samples of surface water (0-50 cm) from Lake Chaohu, Lake Zhiyin (China), Reservoir Orlík (Czechia) and Reservoir Stefanski (Poland) were collected for chemical, biological and biochemical analysis on 05th Jun. 2018, 18th Jan. 2018, 22nd Oct. 2014 and 26th Aug. 2010, respectively (Additional file 1: Fig. S1, Table S2). 2 L of samples were taken and transported to the laboratory immediately. Fresh phytoplankton samples were inspected to minimize difficulties in determining species on the membrane filters. SRP concentrations were determined according to Murphy and Riley [15] after filtering with cellulose acetate membrane (0.45 μm pore diameter size). ELF method was used to quality detect APase on single-cell level. Samples obtained from field were treated using ELF ® 97 phosphate (ELFP, Invitrogen ™ ) on the basis of the protocol described in Strojsova et al. [11]. 0.5 mL incubations were started by adding the ELFP solution (final concentration 27 μM) and samples were incubated at 25 °C for 2.5 h. Each incubation was terminated by transferring the sample to a filter holder (diameter 7 mm) with a membrane filter (Millipore; 0.22 μm pore size). The filter with retained algae were placed on a microscope slide, embedded it with the anti-fading reagent Citifluor AF1 (Citifluor, London, UK), and covered with a cover slide for microscopic inspection. ELF-labeling cells were inspected by epifluorescence microscopy (Olympus BX51FL). 20-50 (depending on the cell density) images were taken randomly on the filters. The active and inactive cells with fluorescence were recorded in each image to quantify the percentages of cells with ELF labeling. were harvested in the exponential growth period by centrifugation (3000 rpm, 10 min), rinsed 3 times with P-free BG11 medium and then cultivated in P-free BG11 medium for 48 h to consume P stored in cells to a low level and equal basis [16]. After that, the culture was inoculated in BG11 [17] medium with modified P concentrations and the modifying sets were as follows:

Experiment design in laboratory culture
A culture using K 2 HPO 4 as the sole P source was defined as +DIP treatment. In P-depleted treatment, K 2 HPO 4 was replaced by KCl. α-glycerophosphate and β-glycerophosphate alternated K 2 HPO 4 with the same amount of P in the treatments of +DOP α and +DOP β . We conducted the experiments in sterile condition to avoid exotic bacterial contamination.
Initial cell density of A. flos-aquae in each culture was roughly 5 × 10 8 cells L −1 . All cultures were incubated in conditions suitable for cyanobacterial growth: constant white light intensity of 30 μmol photons m −2 s −1 irradiance, 12 h of light: 12 h of dark, and 25 °C.
Subsamples were collected at different intervals over 59 days for chemical and biochemical analysis, e.g., SRP and DOP concentrations, cell density, photochemical efficiency, cell P quota, quality and quantity determination of APase, P storage detection, etc.
Cell density, specific growth rate and photochemical efficiency of photosystem II Samples for cell density estimation were preserved with Lugol's solution and counted with an Olympus BX 41 microscope (Olympus Corporation, Japan) [18]. The specific growth rate was calculated according to the following equation: μ = (ln Ct2 − ln Ct1)/(t1 − t0), where Ct2 and Ct1 are cell densities at times t2 and t1, respectively [19].
Fv/Fm is one of the chlorophyll fluorescence parameters which could reflect the maximum quantum yield of the photochemistry. The value of Fv/Fm will decrease when plant was stressed by environmental conditions so that it is often used to evaluate whether the plant is subjected to conditional inhibition [20]. It was examined by Water PAM2100 (Walz, Germany) and determined according to Ting and Owens [21].

P concentrations and cell P quota
SRP concentration was determined using the same method as field investigation. Total phosphorus (TP) and total dissolved phosphorus (DTP) concentrations were measured following digestion according to the measurement reported by Beattie et al. [22]. DOP concentration was calculated as the difference between DTP and SRP concentrations.
Aliquots of 5 mL from individual cultures were harvested onto 25 mm GF/F glass fiber filters (Waterman, UK) for the measurement of TP in samples to determine the specific total P content of cells according to Vahtera et al. [12].

Size-fractionation of APA
APA was quantified by fluorometric assay [23]. It was determined in unfiltered sample (APA T ) and the filtrates through 3.0 μm and 0.45 μm membrane filters (APA <3.0μm and APA <0.45μm, respectively). 150 μL of Tris-HCl (pH = 8.5) was added to 2.7 mL subsample, after pipetting aliquots of fluorogenic substrate (MUFP-phosphate, final concentration of 100 μM), fluorescence was measured immediately, after incubation at room temperature for 2 h, fluorescence was read again. The contribution of APA to the algal (APA >3.0μm ) and bacterial (APA 0.45-3.0μm ) fractions were calculated as follows: in which APA <0.45μm represented the dissolved phosphatase activity [24].

Fluorescence measurements and microscopic inspections
In addition to determining the quality of algal APase by ELF method, we detected PPB in cultures of A. flos-aquae by DAPI (4′,6-diamidino-2-phenylindole) according to the procedure of Bar-Yosef et al. [14]. 0.5 mL incubations were started by adding the DAPI solution (final concentration 15 μM) and samples were incubated at room temperature in dark for 5 min to detect PPB accumulation. Incubation termination, slide embedded protocol and inspection were the same as the stain of ELF. DAPI could bind to polyphosphate and show a yellow fluorescence at 526 nm under UV excitation [25,26].

Statistical analysis
Pearson's correlation analysis and linear regression were carried out to analyze the relationships between SRP concentration and ELFp, APA and ELFp. The significant differences among four treatments were determined by ANOVA analysis and the differences were considered significant at P < 0.05 or P < 0.01. SPSS statistical software (version 18.0, Chicago, IL, USA) was used for those analyses. Data were log transformed to fit normal distribution. The plots of regression were drawn by SigmaPlot 10.0.

Differences in growth of A. flos-aquae in various P supply
The three A. flos-aquae strains of the four treatments yielded similar cell density increasing from around 5.0 × 10 8 cells L −1 to 1.5 × 10 9 cells L −1 at first 8 days (Fig. 1). The cells in the P-depleted and +DIP treatments were then rapidly growing until the day 20 as indicated by the gradually increasing of cell densities and positive specific growth rates (Figs. 1a, d, 2a, d). Moreover, the specific growth rates in the P-depleted treatments had the highest values among all the treatments in the first 20 days (Fig. 2a). After that, specific growth rates fluctuated negatively and positively leading the peak cell densities at day 31 (4.6, 6.6 and 6.5 × 10 9 cells L −1 in A. flos-aquae 1249, A. flos-aquae 1171 and A. flos-aquae 1260, respectively), followed by specific negative growth rates (Fig. 2a) together with remarkable drop in cell densities (Fig. 1a) till the end of the experiment. On the contrary, the growth rates kept constantly positive in the +DIP treatments (Fig. 2d). As a result, the highest cell densities were observed at the end of the experiment (1.75, 2.55 and 2.86 × 10 10 cells L −1 in A. flos-aquae 1249, A. flos-aquae 1171 and A. flos-aquae 1260, respectively) (Fig. 1).
The specific growth rates in the +DOP α and +DOP β treatments were the lowest compared with other two treatments in the first 20 days, after then they fluctuated around zero till the end of the experiment (Fig. 2b,  c). Furthermore, the specific growth rates in the +DOP treatments were significantly lower than that of +DIP  treatment among three culture strains (P < 0.05). In terms of cell density, DOP treatments gave comparable value firstly. After day 8, the lowest cell density was observed in the DOP treatments spatially. In details, the maximum cell density in the +DOP α and +DOP β treatments was 2.3 × 10 9 cells L −1 for the three strains, reaching only 47.16% and 50.82% for A. flos-aquae 1249, 28.94% and 30.43% for A. flos-aquae 1171, 26.84% and 29.02% for A. flos-aquae 1260 of that in P-depleted treatments, respectively. When compared with +DIP treatments, the maximum cell densities in the +DOP α and +DOP β treatments were 12.42% and 13.38% for A. flos-aquae 1249, 7.64% and 8.04% for A. flos-aquae 1171, 6.1% and 6.66% for A. flos-aquae 1260 of that, respectively. Noticeably, the cell densities in the +DOP α and +DOP β kept increased before day 11, then they stabilized ranging from 1.5 × 10 9 cells L −1 to 2.0 × 10 9 cells mL −1 (Fig. 1b, c).

Photochemical efficiency of photosystem II in different treatments
The photochemical efficiency of photosystem II (Fv/ Fm) values remained relatively consistent at 0.400 to 0.538 during the whole incubation period in the +DIP treatments (Fig. 3d). So did the Fv/Fm values in the +DOP α and +DOP β treatments fluctuating from 0.251 to 0.590 (Fig. 3b, c). In the P-depleted treatments, Fv/ Fm values were comparable with that in the +DIP treatments before day 31, while it dropped sharply at the end of the experiment (0.014, 0.077 and 0.037 for A. flos-aquae 1249, A. flos-aquae 1171 and A. flos-aquae 1260, respectively; Fig. 3a).

SRP and DOP concentrations in different treatments
SRP concentrations were determined in all treatments while DOP concentrations were determined only in +DOP α and +DOP β treatments. The P-depleted treatments always showed the lowest SRP concentrations among the treatments (Fig. 4a). In the +DOP α and +DOP β treatments, the SRP concentrations rapidly peaked the value of around 3500 μg L −1 at the first 15 days with the DOP concentrations decreasing to almost 0 μg L −1 . The SRP concentrations stabilized around 2700 to 3000 μg L −1 after then (Fig. 4b, c). In the +DIP treatments, the SRP concentrations declined gradually until day 20 and more rapidly thereafter. Moreover, they decreased under the levels of detection limit at day 46 in the strains of A. flos-aquae 1171 and A. flos-aquae 1260 (Fig. 4d).

Expression of APase and its relationship with ambient P
The active ELF signals, indicating extracellular APase expression, in cells of A. flos-aquae were observed in the samples from different subtropical and temperate water bodies in Czechia, Poland and China, as well as in the studied strains (Fig. 5). Noticeably, the ambient SRP concentration was 681 μg L −1 in Reservoir Stefanski in Poland, which was 13-49 times higher than that in the other three field samples (Additional file 1: Table S1). In addition, the ELFp of A. flos-aquae was negatively related to the SRP concentration in Lake Chaohu (r = − 0.1159, P < 0.05, Fig. 6a). The two parameters also correlated negatively in the cultures among four treatments in the lab-culture (r = − 0.2590, − 0.2391 and − 0.1694 for A. flos-aquae 1249, 1171 and 1260, respectively; P < 0.01, Fig. 6b).
In the cultures, the ELFp exhibited significantly positive relationship with APA >3.0μm for all strains (r = 0.8145, 0.7118, 0.834 for P-depleted, +DOP α and +DOP β treatments, respectively; P < 0.01) (Additional file 1: Fig. S2), and the APA >3.0μm accounted for the largest part of APA T (round to 50-70%) for each treatment (Additional file 1: Fig. S3). In the P-depleted treatments, APA >3.0μm together with the percentages of cells with ELF labeling (ELFp) were significantly higher than those in the other treatments for all strains (P < 0.01) (Additional file 1: Figs. S3, S4). In addition, the ELFp in +DOP α and +DOP β treatments were significantly lower than that in the P-depleted treatments (6.6 to 9.5% of that in the P-depleted treatments; P < 0.01) while they were significantly higher than that in +DIP treatments (17.8-to 25.6-fold of that in the +DIP treatments; P < 0.01) (Additional file 1: Fig. S4).

Cell P quota and PPB containing cells in different culture treatments
The cell P quota in the P-depleted treatments showed significant lower values compared to that in the other ones (P < 0.01) (Fig. 7). The PPB were observed in each culture, with the initial percentages of PPB containing cells (PPBp) being around 0.05%. PPBp in the P-depleted treatments was significantly lower than that in the other treatments (P < 0.01) (Fig. 8). Noticeably, PPB containing cells could be observed at first day 15 (Fig. 8a). PPBp began to increase at first (for A. flosaquae 1171, 1260) or second (for A. flos-aquae 1249) day and peaked at day 5. It rapidly declined to 0 at day Though demonstrating the same pattern, PPBp in A. flos-aquae 1249 was obviously lower than that in the other two strains. In the +DIP treatments, PPBp increased from day 1 and reached maximum at day 2, and remaining constant (85-93%) until day 42. In terms of strains, the PPBp decreased in A. flos-aquae 1171 and A. flos-aquae 1260 at day 46 to 74% and 78%, respectively (Fig. 8d). In the +DOP α and +DOP β treatments, the PPBp fluctuated regularly with incubation time. In details, the PPBp reached a level (over 80%) comparable to that in the +DIP treatments at first days and decreased to 10-40% several days later and went up and down later on (Fig. 8b, c). Furthermore, PPB numbers per cell varied greatly among the different cultures (Fig. 9). It was rich and constant in the +DIP treatments ( Fig. 9e and f ) while fluctuated in +DOP α and +DOP β treatments ( Fig. 9a-d).

Discussion
Whether cyanobacteria alleviated P deficiency by producing APase is rarely studied, despite numerous studies having showed that it did so in phytoplankton in freshwater and ocean by evaluating the APA in situ and cultures [27][28][29][30][31][32]. The current study illustrated the P utilization adoption of A. flos-aquae in response to various P supply and its role in regulating the growth and reproduction. A vigorous growth at the early stage in the P-depleted treatment might be owing to the cooperation of P utilization strategies in A. flos-aquae, e.g., producing APase, increase PPB formation and decrease cell P quota. Unexpectedly, A. flos-aquae in the treatment of DOP gave the cell density reaching around 13% of that in replete Pi control, even though the quantity and quality of APase were observed and the substrate monoester was sufficient. No growth rate could be determined, while the photochemical efficiency was comparable to the treatment of +DIP, indicating that A. flos-aquae maintained its metabolic activity by enzymatic hydrolyzing P but not reproduced when DOP was supplied as a sole source of P. The ELF signals in cells of A. flos-aquae in the field and lab-culture and the inverse relationship between the percentage of active cells (or the APA contributed by algal cells) and the SRP concentration indicated that the APase excretion by A. flos-aquae was regulated by the ambient P concentration, which was commonly observed so that high APA was recommended as an indicator of P stress or limitation [28,33,34]. It is reasonable that A. flosaquae produced APase in Lake Chaohu, Zhiyin and Reservoir Orlík where the SRP concentration varied between 14 and 51 μg L −1 . Unexpectedly, it still produced APase where SRP concentration was 13-49 times higher in Reservoir Stefanski in Poland (Additional file 1: Table S1). A possible explanation for these observations might be that the threshold for APase initiation may be strain-specific so as to response to local environmental conditions. Phytoplankton species would respond to P deficiency at various concentrations of SRP. For example, Prorocentrum minimum began to express APA at a higher SRP concentration than Chaetoceros neogracile, Chlorella autotrophica, Isochrysis sp. and Tetraselmis chui did using the ELF method [35]. In addition, a repression of the synthesis of APase was observed over an orthophosphate concentration threshold between 0.4 and 1 μM in Alexandrium catenella in the France lagoon [36]. In time-course experiment, the APA was induced at orthophosphate concentration at 0.43 mM for Alexandrium tamarense isolated from Hiroshima Bay [37]. These results indicated that the threshold for producing APase is species specific, while our field and culture observations indicated that it might be strain-specific among A. flos-aquae. On the other hand, scientists also referred to APA's contribution to phytoplankton P absorption and growth in mineralizing DOP [31,38]. It was quantified that up to 89% of the total phytoplankton P uptake was represented by APA in Lake Wivenhoe of Australia [39]. In addition to algae, free living and attached bacteria were the main contributors to APA [40][41][42]. In our experiments, APA >3.0μm accounted for the most part of APA T (see Additional file 1: Fig. S3). We could not exclude the contribution of the attached bacteria to APA >3.0μm . But the positive relationship between ELFp and APA >3.0μm (see Additional file 1: Fig. S2) indicated that an anomaly thriving growth of A. flos-aquae in the P-depleted treatments at the first . Therefore, the new formation of PPB in P-depleted treatments was probably from the hydrolysis of DOP released from A. flos-aquae as evidenced by the fluctuation of PPBp (Fig. 8a). This pattern of PPB forming was contradicted to the canonical point of polyphosphate as a luxury P storage molecule. In accordance with our study, polyphosphate accumulation was observed in cells of A. flos-aquae obtained from a summer bloom in the central Baltic Sea, where only small amounts of phosphate were available in the ambient environment, which could help to compete with other microorganisms for phosphate during bloom [44]. In addition, the phytoplankton in the western North Atlantic was enriched in polyphosphate in the P-depleted subtropical Sargasso Sea compared with nutrient-replete temperate waters, proposing that the polyphosphate cycle might produce a feedback loop, which reduces the P export when it is lacking and makes contribution to bio-available P for primary production [45]. Another explanation for this transitory growth is that A. flosaquae decreased the requirement of cellular P and alleviated P deficiency to some extent. The lowest cell P quota was found in this treatment (Fig. 7). Consistently, the N:P ratios and intracellular P demand in A. flos-aquae TR183, Nodularia spumigena AV1 and Trichodesmium erythraeum IMS101 can be remarkably plastic in response to P deficiency [12,46]. It was, therefore, inferred that the A. flos-aquae utilized some strategies such as the promotion of excreting APase, elevation of PPB formation and reduction of P demand to gain maximum growth at the circumstance of P deprivation. This hypothesis was verified by the field observation in Lake Chaohu that high percentage of ELF labeling was accompanied by high biomass. Our previous work about another N 2 -fixing genus Dolichospermum flos-aquae also suggested that it suffered from severe P limitation at the end of the bloom [24]. An obviously lower cell P quota was observed in cyanobacteria during their bloom (unpublished data) and its underlying mechanism needs further investigation.
Unlike the pattern of PPB forming in the P-depleted treatments, it seemed following the classical view as a luxury P storage molecule in the +DIP treatments. The types of PPB may explain this variation. Polyphosphate can be classified as acid-soluble polyphosphate (ASP) and acid-insoluble polyphosphate (AISP). ASP is considered to be used for metabolism and production of DNA and protein, it can also act as a function of shortterm P pool. In contrast, AISP appears to be a form of P storage that could be utilized by the cell for growth when the P concentration is limiting in external environment [47,48], which enables cell divisions in the absence of any exogenous phosphate [49,50]. Additionally, this polyphosphate was only utilized by microalgae when phosphate concentration of the waste-water reached a growth limiting level [48]. The three strains of A. flos-aquae in the +DIP treatments remained active growth after the lag phase and gave the highest cell density with the PPBp as well as number of PPB per cell in constant levels (Figs. 8d, 9e, f ). A. flos-aquae appeared to use the storage P to achieve active growth in relation to the positive growth rate after day 46 when the SRP concentrations were decreased into the levels of that in the P-depleted treatments in the strains A. flos-aquae 1171 and A. flos-aquae 1260, where the PPBp declined then (Fig. 8d). The main component of the PPB in the +DIP treatments might be AISP which was used upon P-deficient environment.
It is expected that APase might overcome P deficiency together with amply supply of the substrate. The use of glycerophosphate and other DOP gave quite similar growth rates to that obtained with inorganic ones [51][52][53]. In contrast to the above results, the lowest cell densities in the glycerophosphate treatments were detected, with their growth even not as good as that in the P-depleted treatments with P-free condition. The APA in +DOP α and +DOP β treatments were significantly lower than that in the P-depleted treatment. Obviously, APA was not regulated by substrate DOP, which was also observed in the culture of Microcystis aeruginosa [32]. Low activity in the +DOP treatments could be resulted from feedback inhibition of APA by the phosphate produced in the glycerophosphate hydrolysis [53]. For example, low levels of external phosphate were sufficient to induce repression of APA in Gelidium latifolium [54]. Despite the repression of APA in +DOP α and +DOP β treatments, as well as the significantly lower exhibition of ELFp compared with the values in the P-depleted treatments, ELFp were much higher than that in +DIP treatments (round 18-to 26-fold), irrespective of the strains studied (Additional file 1: Fig. S4). Consistently, the Coccomyxa isolated from Lakes Plesne and Hromnice in Czechia increased in cell-specific phosphatase activity measured combining the ELF assay with image analysis at low P concentration and the cultures treated with DOP (β-glycerol phosphate), which generated remarkably higher activity than those supplied with the same Pi concentrations [55]. Therefore, APase in A. flosaquae could perform its function of hydrolyzing organic P, as illustrated by significant and positive relationship between the ELFp and APA >3.0μm (Additional file 1: Fig.  S2). Furthermore, the SRP concentrations in the +DOP α and +DOP β treatments appeared increasing before day 15 with the expense of the DOP concentrations (Fig. 4b,  c), which was an additional evidence that the DOP had been decomposed by the enzyme. On the other hand, it was noted that three strains of A. flos-aquae in +DOP α and +DOP β treatments showed fluctuations in both PPBp (Fig. 8b, c) and number of PPB per cell (Fig. 9a-d) along the culture period. The PPB in the +DOP treatments might mainly be the ASP, which was rapidly utilized after approximate three days by microorganisms in waste-water [48]. Noticeably, no growth rate was determined and the photochemical efficiency in terms of Fv/ Fm values was comparable during the whole experiment in the treatments of +DIP and +DOP. Shortly, the effective DOP mineralization and continuous processes of production and consumption of PPB in the +DOP treatments supported an active and sustainable growth but not reproduction for A. flos-aquae in +DOP treatments. Consistent to our observation, the growth rate of A. flos-aquae TR183 isolated from Baltic Sea in P-deplete treatment was a little higher than that in glycerol phosphate culture and it yielded lower biomass in glycerol phosphate culture when the Pi in the medium was not depleted than that in Pi-replete condition [12]. This suggested that A. flos-aquae is a species that grows better in high Pi pulse supply. The coordination of DOP mineralization and prompt utilization of PPB in A. flos-aquae might contribute to this maintaining growth, which was in accordance with the statement raised by Strojsova et al. [56] that the ability to decompose organic polymers and particles with extracellular phosphatases was beneficial for longer persistence of given population in a Piscarce circumstance.

Conclusion
Our findings provided evidence that P utilization strategies in A. flos-aquae, including producing APase, increase PPB formation and decrease cell P quota, played an important role in response to P stress. DOP was an important alternative P source for longer maintenance in filamentous cyanobacteria under Pi deficiency. Persistent and active reproduction could only be achieved in high pulse Pi supply, implying that an effective consequence can be expected for combating the bloom of A. flosaquae when controlling P supply.