On 20 March 2023, a phytoplankton net with a mesh size of 63 µm was used to collect a thin layer of surface scum composed of cyanobacteria that was collected during the early stages of a spring bloom in Guanqiao fish pond located in Wuhan, China. The collected cyanobacterial colonies were quickly transported to the laboratory. They were immediately filtered through a 300-µm mesh filter followed by a 45-µm mesh filter to eliminate coarse impurities. The filtered and concentrated cyanobacterial colonies were subjected to a pre-culturing process in 10% BG11 medium, at a temperature of 25°C, under a light/dark cycle of 12 h/12 h with light exposure of 25 µmol photons m⁻² s⁻², for a duration of 3 days (Zhao et al., 2019; Gao et al., 2020; Xu et al., 2023). Morphological identification was conducted, which revealed the predominance of Microcystis species ( Supplementary Figure S1C ).

To isolate cells from Microcystis colonies, the Microcystis colonies were subjected to pretreatment using an Ultrasonic Processor (VCX150, Sonics & Materials Inc., USA). The colonies were dispersed into individual cells ( Supplementary Figures S2B and S3B ) through the application of ultrasonic waves (Wu et al., 2012; Zhang et al., 2021). The ultrasonic amplitude was set at 50%, and a total energy of 108–112 J was released over a 4-min duration. The single cells obtained from ultrasonic dispersion were stained with fluorescein diacetate (FDA) at 100 µg mL⁻¹ and subsequently kept in darkness for 5 min (Chen et al., 2005). Upon exposure to an excitation light source, the cells exhibited a vibrant green fluorescence ( Supplementary Figure S3B ) confirming that the ultrasonic treatment employed did not induce cellular mortality (Yang et al., 2021). Both the cells and EPS were isolated by centrifugation at 9,000 rpm for 10 min. The experimental cells were obtained by subjecting them to three rounds of resuspension and centrifugation in 10% sterile BG-11 medium.

The resulting supernatant from centrifugation at 9,000 rpm was collected and subsequently filtered using GF/F membranes (0.22 µm; Waterman, UK) to obtain Microcystis EPS that was devoid of cells. This collected EPS is called homologous EPS. To prevent the EPS from rapidly deteriorating at ambient temperature, the separated homologous EPS was promptly refrigerated and stored for later use.

The homologous EPS collected during pre-treatment was added to cells at different concentrations, determined by the dissolved organic carbon (DOC) concentration. This resulted in EPS concentrations of 0 mg L⁻¹ (CK), 0.66 mg L⁻¹ (Low), and 12 mg L⁻¹ (High) within the culture system ( Supplementary Table S1 ). During the collection of Microcystis in the field, the pond’s DOC content was measured to be 12 mg L⁻¹ at a depth of 0.1 m below the water surface.

All groups were cultured in Erlenmeyer flasks containing 10% sterile BG11 medium. In all groups, the starting cell density was consistent at 6.5×10⁸ cell L⁻¹. The experiments were conducted in an Illuminated Incubator (PGX-100A-LED, Jiangsu Tianling Instrument Co., Ltd., Yancheng, China). Culture conditions mirrored those of wild Microcystis colonies, with gentle agitation provided by manual shaking two to three times per day (Gao et al., 2020; Xu et al., 2023). Other culturing conditions were the same as those in Section 2.1. Samples were collected every 3 days during the initial phase when cells reaggregated into microcolonies and every 5 days when the size of reaggregated colonies stabilized. Nutrient concentrations within the culture system were monitored throughout the sampling period to ensure adequate nutrient availability. All experimental groups were conducted in quadruplicate, and the results were reported as averages.

The Microcystis population is enumerated after the alkaline hydrolysis using an optical microscope (BX43, Olympus Corporation). The relative growth rate (ν) of Microcystis was calculated as follows (Xu et al., 2023):

where C_ti and C_tj (cells L⁻¹) are the cell densities of Microcystis at two consecutive sampling times.

Chlorophyll a (Chla) was measured using a 4.8-mL sample. The Phytoplankton Analyzer (Phyto-PAM, Walz Co., Erlangen, Germany) was utilized for the assessment of Microcystis photosynthetic activity (Ye et al., 2023). The Fv/Fm parameter represents the potential maximum conversion efficiency of Microcystis photosystem II, as it is overwhelmingly dominant (Wang et al., 2018). Photos were taken using software (HaoKangBioImaging) on a microscope.

The zeta potential of Microcystis was measured with a Malvern Zetasizer (Malvern-Nano-ZSMalvern, UK) (Tattibayeva et al., 2022). The pH was determined with a pH meter (pH100A, EcoSence, China). Additionally, the dissolved oxygen (DO) in the Microcystis fluid was quantified using a YSI instrument (YSI Pro ODO, USA).

The study utilized a Laser In-Situ Scattering and Transmissometer instrument (LISST-200X, Sequoia Scientific Inc., Bellevue, WA, USA) to measure colony size, cell size, and biovolume concentration. The median colony size (D₅₀ ), representing the median biovolume concentration location, was used to characterize the diameter of Microcystis cells and colonies. Before measuring, the Microcystis sample was appropriately diluted with distilled water. Then, 10 mL of the diluted sample was used to determine the size distribution and biovolume concentration (V_colony in µL L⁻¹) of the original Microcystis colony. After the measurements, the Microcystis fluid was retrieved, and the colonies were disaggregated into single cells through alkaline hydrolysis at 85 °C for 6–8 min (Wang et al., 2015). The size distribution and biovolume concentration (V_cell in µL L⁻¹) of Microcystis cells were subsequently assessed using the LISST-200X. The volume ratio (VR) of the cells to colonies was calculated according to established laboratory protocols from prior studies (Wu et al., 2020; Xu et al., 2023), and it indicates the relative compactness of colonies:

where V_cell (µL L⁻¹) and V_colony (µL L⁻¹) are the mean biovolume concentrations of single cells and of colonies in a sample.

Coomassie Brilliant Blue G-250 and phenol-sulfuric acid were used to measure polysaccharides and proteins, respectively (Wang and Xing, 2009; Ma et al., 2021; Duan et al., 2022) in this study. The sample was first centrifuged at 5,000 rpm for 15 min. After removing the supernatant, the remaining sediment was resuspended using a 0.05% NaCl solution. The suspended algae solution was then centrifuged at 5,000 rpm for LB-EPS measurement (Xiao et al., 2018). The residual sediment was re-suspended with a 0.05% NaCl solution at an adjusted pH of 10 and heated at 45°C for 4 h. Subsequently, centrifugation was performed at 11,000 rpm for 15 min, and the resulting supernatant was collected for the measurement of polysaccharides and proteins (Yang et al., 2008; Xiao et al., 2012, 2019). LB-EPS encompasses loosely bound proteins and polysaccharides. TB-EPS comprises tightly bound proteins and polysaccharides.

The content of dissolved polymeric substances (S-EPS) was represented by DOC in this study. The DOC in liquid samples filtered through GF/F membranes was quantified using a Total Organic Carbon Analyzer (TOC-VCPH, Shimadzu Co., Kyoto, Japan).

SPSS statistics software (version 27, Chicago, IL, USA) was used to determine the differences in physiological regulation of Microcystis at different periods of colony formation and the differences in colony formation caused by homologous EPS. Values with p< 0.05 were considered significant, while those with p< 0.01 were considered very significant. All data images were produced using Origin 2021 software (OriginLab, Northampton, MA, USA).

As shown in Figure 1A , there was no significant difference in Microcystis cell density during the first 0–6 days. However, the cell density in group High remained consistently higher than the other two groups for the following 42 days. In Figure 1B , the growth rate of group CK had a small peak of 0.46 day⁻¹ during 3–6 days followed by a small decrease and a rise to the second peak of 1.05 day⁻¹ during 27–32 days. This pattern repeated, reaching the third peak of 1.05 day⁻¹ during 42–47 days. During the final 20 days of the experiment, the rapid increase in growth rate of group CK ceased and then decreased to a stagnant growth state. It showed a relatively stable and low cell density curve in Figure 1A . The growth rate of group Low rose and fell during the experiment in general. But it decreased to 0.25 day⁻¹ between days 27 and 32 before peaking at 0.84 day⁻¹ in the following 5 days. Group High had higher growth rate at the beginning and end of experiment. Specifically, the peak values observed were 0.69 and 1.17 day⁻¹ on days 6–9 and 42–47. At the same time, it was observed that the dissolved total phosphorus (DTP) in this set of systems was almost depleted by the end of day 12, and the medium was replenished in a timely manner ( Supplementary Figure S4 ). The growth rate of group High reached a second small peak of 1.7 on days 22–27. A third peak in growth rate was observed after 47 days. The growth rate of group Low reached a small peak on day 22 and increased rapidly to 7.4 on days 32–52. In the long term, the addition of homologous EPS is beneficial to Microcystis population.

The Chla per cell in group CK was significantly lower than that in group Low (p< 0.05). In comparison to group Low, the difference in Chla content in group High was much more significant (p< 0.01). The Chla content per cell increased during the initial 6 days, followed by a decrease as the days progressed, and eventually stabilizing at a relatively constant value. The Chla content was similar among the different treatment groups ( Figure 1D ).

During the first 6 days, the photochemical efficiency of PSII (Fv/Fm) significantly decreased in all treatment groups. The magnitude of the decrease was directly proportional to the amount of added homologous EPS. The group with high homologous EPS addition showed a reduction of 0.34, which was two to five times lower than the low addition group and group CK ( Figure 1C ). Overall, group CK had the highest Fv/Fm value, while the group with high homologous EPS addition had the lowest value.

It was observed that the absolute value of the Zeta potential in group CK increased in the first 0–3 days before decreasing. Groups Low and High experienced a decrease in the absolute value of Zeta potential until 12 days ( Figure 2A ). From days 10 to 40, the absolute value of the Zeta potential of group CK remained at 20 to 24, which was lower than that of the other two groups. There is no significant difference in the Zeta potential between group Low and group High. The concentration of DO and pH of group High were higher than those in group Low. Similarly, group Low had higher DO and pH than group CK in the first 20 days ( Figures 2B, C ).

In this experiment, the EPS of Microcystis were analyzed by their content of polysaccharides and proteins as representative structural EPS. The ultrasonic treatment used to extract the EPS was not entirely thorough, as evidenced by the incomplete extraction of TB-EPS ( Figure 5B ). The addition of homologous EPS resulted in a rapid response from Microcystis cells, as indicated by the production of LB-EPS. The strength of the positive response was directly proportional to the amount of homologous EPS added ( Figure 5A ). Initially, the added homologous EPS mainly consisted of soluble EPS, with a smaller proportion of LB-EPS. Among them, proteinaceous substances had a higher proportion than polysaccharides ( Figures 5A, B, D, F ). During the experiment, the LB-EPS on individual cells in both groups Low and High first decreased from 0 to 6 days and then increased. In groups CK, Low, and High, the LB-EPS content on individual cells reached its highest levels at approximately 32 days, with values of 0.045, 0.048, and 0.065 pg cell⁻¹, respectively ( Figure 5A ). In contrast to the immediate decrease in LB-EPS content in groups CK and Low, group High maintained its peak levels for 10 days. Over the course of 0–20 days, the TB-EPS on individual cells in group Low was significantly higher than in groups CK and High. In Figure 5B , the TB-EPS peak values were 0.092, 0.111, and 0.139 pg cell⁻¹ for groups Low, CK, and High, respectively. The peak order of B-polysaccharides on cells was similar to that of TB-EPS, with only group High showing a peak of up to 0.114 pg cell⁻¹ at 32 days. According to the trend of extracellular substances on cells from 0 to 6 days, polysaccharides are suggested to be mainly TB-EPS, while proteins are suggested to be mainly LB-EPS ( Figures 5A–D ). The trends in B-EPS on individual cells in each group were generally similar, with group CK maintaining lower B-EPS levels than the other two groups ( Figure 5A ), resulting in the formation of denser Microcystis colonies in group CK. The DOC level in group CK consistently remained low, with a slight decrease after a gradual increase ( Figure 5F ). In group Low, the DOC level was above 0.003 pg cell⁻¹ after the ninth day and continued to decline after approximately 20 days. The DOC level in group High rapidly decreased from 0.018 to 0.003 pg cell⁻¹ within the first 3 days and has since fluctuated between 0 and 0.003 pg cell⁻¹.