Experimental Materials

The M. aeruginosa (FACHB-912 unicellular line) isolated from the cyanobacterial population of Lake Tai (the third-largest freshwater lake in China) was used in this experiment. The M. aeruginosa were inoculated with BG11 medium in a sterile environment and then cultured in a light incubator (LRH-150). The temperature in the incubator was 25ºC, the light-dark ratio was 1:1, and the light intensity was 2400 Lux. The glassware was rinsed with water, soaked in 0.1 mol/L dilute hydrochloric acid, rinsed with sterile water, dried, and autoclaved. The overall experiment was carried out on a sterile bench.

The experimental equipment included a sterile glass incubator, white light color temperature LED lamp, aeration pump, YSI-DSSpro water quality analyzer, and based on this equipment formed a high-frequency culture monitoring device (Fig. 1). YSI-DSSpro is an improved high-frequency water quality analyzer, which can monitor pH, DO, chla, Cond, and other growth indicators of M. aeruginosa at a high-frequency condition. The suitable light conditions and Oxygen conditions for the proliferation of M. aeruginosa. can be provided by the LED lamp and aeration pump. The high frequency culture monitoring device enables both automatic and high survival rate culture work of M. aeruginosa and high frequency monitoring of pH, DO, chla, Cond. In addition, the monitoring data were regularly transferred to a server for storage.

Experimental Method

The M. aeruginosa used in this study were incubated in a light incubator for a short cycle and then used for high-frequency experiments. The M. aeruginosa were mixed and transferred to a sterile glass incubator with the initial chla set at 40 μg/L.

The aeration pump, high-frequency water quality analyzer, LED lamp, and other equipment were turned on after the start of the high-frequency culture experiment and a fixed opening time was set. The monitoring frequency was set to 15 min/time and the aeration times were set at 3:00 a.m., 10:00 a.m., 15:00 p.m., and 22:00 p.m., with each aeration lasting 20 minutes. The light source was turned on at 8:00 a.m. and turned off at 20:00 p.m. The time of aeration and light on/off is fixed every day. The data of the high-frequency culture experiment were initially stored in the high-frequency water quality analyzer and then regularly transmitted to a computer to monitor the growth condition of M. aeruginosa and for further analysis.

Data Analysis

To improve the accuracy and reliability of the data of high -frequency monitoring results, this study used a wavelet transform with adaptive characteristics to remove the noise from the data. A wavelet transform can achieve noise reduction while retaining valid information and ensure the reliability and continuity of pH, DO, Cond, and chla data as much as possible. A wavelet transform can also decompose the time series by different resolutions and decompose the original signal into sub-signals of different frequencies, so that the timing profile and detailed parts of the original signal can be easily identified [20, 21]. Therefore, this type of analysis is appropriate for the continuous processing and cleaning of the data from these culture experiments. The wavelet transform used in this study consisted of the following three processing steps.

Fig. 1.

Schematic of the laboratory high-frequency monitoring device.

(1) Wavelet decomposition: The best wavelet function was selected to decompose the noise signal according to the data characteristics of different variables, the orthogonal wavelet was used for the data x(t), and then the wavelet decomposition coefficients of the l-layer k points were calculated as follows.

where l is the number of wavelet decomposition layers; r = 0, 1, 2, ..., 2l -1 is the number of nodes under the corresponding layers; 0, 1, 2, ..., 2^l is the kth node of the jth layer.

(2) Threshold quantization: The appropriate threshold for the high-frequency coefficients of each layer after the decomposition was selected and the global threshold was used in this study.

(3) Wavelet reconstruction: Wavelet reconstruction was performed based on the high-frequency coefficients in layers 1 to N and the low-frequency coefficients in layer N to finally achieve noise reduction. The first equation that performs single-branch reconstruction, S_l+1^r(k) is the single-branch reconstruction signal at k points in layer l+1 and the second equation is the total reconstructed signal.

Differential Growth of Microcystis aeruginosa in Light and Dark Environment

As shown in Fig. 2a), under indoor culture conditions (populations with stable age structure, unrestricted food and space, and only intraspecific competition), a complete culture cycle of copper green microcysts is about 80 days. From the whole life cycle of M. aeruginosa, the growth of M. aeruginosa was characterized by chla concentration with a growth limit of about 110-120 µg/L (about 2.4x10⁶ million units/L), and the maximum growth rate point appeared on the 23^rd day of culture with a maximum value-added rate of about 1.04x10⁵ units/day. The culture period, growth limit and maximum value-added rate points of M. aeruginosa obtained from this experiment are of great significance and can provide theoretical support for the precise control of cyanobacterial population and actual cyanobacterial salvage.

Fig. 2.

Diurnal differences in growth of M. aeruginosa and logistic model fitting.

From the light and dark environment, there was significant variability in the growth of M. aeruginosa (Fig. 2a). From day 0 to 20 of the experiment, the chla concentrations of the cyanobacteria in the dark environment was higher than in the light environment. From day 20 to 45 of the experiment, the chla concentration in the light environment was gradually higher than that in the dark condition. From day 45 to 80, the chla concentrations in the light environment was similar to the dark environment. A logistic model to the different growth trends of M. aeruginosa during the light and dark environment was fitted (Fig. 2b). The results demonstrated that M. aeruginosa growth in both dark and light environments was consistent with logistic growth. The R² reached above 0.93 and the fitting results were good. This shows that the overall growth is consistent with Logistic growth.

Discovery of Three-Stage Growth Pattern

Depending on the proliferation rate of M. aeruginosa under light and dark conditions (Fig. 3), the growth phase under indoor culture conditions can be divided into three stages. The three different stages are Light adaptation period (LAP), Logarithmic growth period (LGP) and Stabilization period (SP). The LAP represents the stage in which the chla concentration was higher in the dark than the light condition in the early stage of culture. The LGP represents the stage in which the chla concentrations in the light was higher than that in the dark environment in the middle stage of culture, and the growth of M. aeruginosa was logarithmic during this period. For the SP stage, the mean chla concentrations were similar during day and night, and the proliferation rate decreased and gradually tended to 0. The proliferation rate of M. aeruginosa at these three growth stages was different. The M. aeruginosa proliferation rate fluctuated from high to low in the culture cycle, with the slow proliferation of the cyanobacteria at the beginning, an accelerated proliferation rate in the middle, and dynamic equilibrium in the growth at the end of the culture period.

The different growth periods of M. aeruginosa were classified according to the fitting results (Table 1). The results showed that the three growth stages overlapped to some extent with the four logistic stages (start-up, accelerated growth, decelerated, and stabilization periods). The chla concentration was higher under the dark compared to the light conditions during the start-up and accelerated growth period. The chla concentration was lower under the dark compared to the light conditions during the pre-mid decelerated growth period. Finally, the chla concentration was similar under dark and light conditions during the end of the decelerated growth and stabilization periods. The shorter duration of the onset and accelerated growth periods of M. aeruginosa were related to the initial density of M. aeruginosa selected in this study. The initial value of the chla concentrations of M. aeruginosa was about 40 μg/L. Since the initial value of the algal density was high, the M. aeruginosa density could reach the k/2 value quickly.

The Driving Mechanism of the Three-Stage Pattern

Trends in chla Concentrations under Alternating Light and Dark Environments

The chla concentrations trends of M. aeruginosa in the light-dark alternating environment in the three stages of LAP, LGP, and SP were highly variable (Fig. 4). As expected, the LAP showed an opposite trend to the LGP. The chla concentrations in the LAP showed an initially decreasing and then increasing trend after the light was turned on. However, the chla concentrations first increased and finally stabilized after the light was turned off. In contrast, the chla concentrations in the LGP showed a trend of rapid increase after the light was turned on and a significant decrease followed by a small increase after the light was turned off. This was the superficial reason why the chla concentrations in the dark environment were lower than in the light during the LAP. The trend of chla concentrations was not obvious after turning the light on and off during the SP. The diurnal chla concentrations of M. aeruginosa during the SP were the same, although displaying fluctuations related to the density of the initial cyanobacterial culture. The small degree of disturbance that can lead to substantial changes after the algal concentration became high was difficult to avoid.

Fig. 3.

Three-stage pattern schematic.

Table 1.

Relationship between fit and the logistic period.

Three-stage pattern	Growth status	Cultivation time	Logistic period
Light adaptation period (LAP)	Light-free>Light	0-20 day	Start-up and Accelerated growth period
Logarithmic growth period (LGP)	Light>Light-free	20-45 day	Decelerating growth period
Stabilization period (SP)	Consistent	45-80 day	Stabilization period

The significant variability of chla concentrations trends in alternating light and dark environments during the LAP and LGP may be related to the existence of an adaptation process (light or competition) during the pre-culture period. When M. aeruginosa was transferred from the light incubator to the sterile glass incubator, a series of adaptation processes to the new environment were required to ensure normal growth and reproduction. In indoor culture conditions, this adaptation process primarily includes the adaptation to light, the degree of intraspecific competition, and other processes. In this study, photoinhibition and photoadaptation processes in M. aeruginosa were observed. The change of light environment produced a kind of photoinhibition of photosynthesis and the growth and reproduction of M. aeruginosa, reducing the photosynthetic activity and resulting in photodamage. Therefore, this may be the intrinsic reason for the higher chla concentrations in the dark compared to the light environment during the LAP. For the intraspecific competition, the population density was low during the LAP and the LGP, and the degree of intraspecific competition was lower than during the SP. The low intensity of the intraspecific competition may facilitate the process of light adaptation, which may have been one of the reasons for the short duration of the LAP. For the SP, the M. aeruginosa population reached the environmental holding capacity (K), the proliferation and decay of M. aeruginosa reached a dynamic equilibrium at this stage, and the net proliferation rate was essentially 0. After the adaptation process of LAP and the LGP, the proliferation rate of the light and dark environments in the SP reached stability. In summary, we suggest that the adaptation mechanism of M. aeruginosa to light and the intensity of intraspecific competition is one of the intrinsic driving mechanisms of the Three-stage growth pattern in indoor culture.

Fig. 4.

Diurnal growth differences during the three stages.

Variability Analysis of Environmental Drivers

The correlations of different environmental indicators (pH, DO, Cond, and chla) were calculated for three different stages (Fig. 5). The results demonstrated that chla concentrations in the LGP and LAP had a good correlation with Cond and pH. The correlation coefficients of chla with Cond were -0.742 and 0.57 for the LAP and LGP, respectively. The correlation coefficients of chla with pH were -0.508 and -0.46. The correlation of Cond with pH and DO was poor. For the LGP, chla concentrations displayed a negative correlation with Cond, while chla was positively correlated with Cond. Alternatively, the correlation relationship between pH and DO for the three stages was significant, which was related to photosynthesis and respiration of M. aeruginosa.

The differences in the correlation between the environmental factors and chla content at the three stages can partially characterize the differences of the M. aeruginosa proliferation mechanisms. The Cond was closely correlated with the growth of M. aeruginosa during the LGP and LAP. During these two periods, the Cond well-characterized the growth and reproduction of M. aeruginosa. The chla concentrations in the LGP and LAP was associated with faster proliferation, faster increase in algal density, faster uptake of nutrients from the culture medium for life activities, and faster release of metabolic substances to external survival environments. The correlation between chla concentrations and pH, DO, Cond in the SP was poor, which was related to the proliferation state of M. aeruginosa. The proliferation and decay of M. aeruginosa in the Stabilization period reached a dynamic balance, the consumption of nutrients in the medium was accelerated, and the metabolic products were aggregated. The Stabilization period may have been longer because the BG11 medium provided sufficient nutrients for the algae.

Fig. 5.

Environmental factor correlation of the Three-stage pattern.

Fig. 6.

Changes in pH, DO of the Three-stage pattern.

The changes to pH, DO, and Cond were the manifestation of the intrinsic mechanisms of diurnal growth of M. aeruginosa. To further reveal the environmental factors driving the mechanisms of the three growth stages, the diurnal transformation patterns of pH, DO, and Cond at different growth stages were further investigated (Fig. 6). The results demonstrated that the diurnal trends of pH during LAP, LGP, and SP were similar, all showing an increase after the light was turned on and a gradual decrease after it was turned off. However, the pH during the three stages was different, varying following the change of the light conditions. The pH magnitudes were in the following order: LAP>LGP>SP, and the values and magnitudes of the pH changes gradually decreased. The diurnal trend of DO at the different growth stages was similar to the pH diurnal trend, which also increased after the light was turned on and gradually decreased after the light was turned off. The diurnal pattern of Cond varied significantly among the three stages. Cond showed a gradual increase after the light was turned off during the LAP and a substantial decrease after the light was turned off. However, the Cond did not change in the diurnal environment during the SP.

The pH and DO trends can characterize the photosynthesis and respiration processes during M. aeruginosa growth. Under light conditions, M. aeruginosa can photosynthesize and release oxygen by absorbing and using carbon sources in the medium, leading to a decrease in water acidity, a rise in pH and DO. In dark conditions, M. aeruginosa stops photosynthesis and begins respiration, where pH and DO decrease. The high algal activity and strong photosynthesis during the pre-mid culture period were the reasons for the higher pH and DO in the LAP and LGP compared to the SP. The change of Cond can characterize the process of nutrient uptake and metabolite release from the medium by M. aeruginosa. A process of adaptation to the environment during the LAP occurs where sensitivity to environmental changes is enhanced and vital activities are affected, which was the reason for the large changes in Cond during the LAP.

Based on the diurnal patterns of pH, DO, Cond, and chla concentrations, the light environment, photosynthesis, respiration, and intraspecific competition of M. aeruginosa were the driving factors of the Three-stage patterns. Light, nutrient salts, and their interactions were the key factors limiting M. aeruginosa diurnal growth. The LAP contained the adaptation of light as well as competitive conditions to the new environmental adaptation process, during which the growth of M. aeruginosa was primarily affected by light inhibiting the intensity of photosynthesis and initiating photoinhibition. The M. aeruginosa reduced the damage produced via photoinhibition through its light adaptation mechanism, which is an intrinsic mechanism for the above changes in pH and DO. In addition, according to the light-nutrient hypothesis, the growth of M. aeruginosa in the laboratory was influenced by the synergistic effects of light and nutrient salts, rather than by a single factor. Considering that the nutrients such as nitrogen, phosphorus, and carbon in the medium were sufficient to support the vital activities of M. aeruginosa, once the proliferation and decay of M. aeruginosa reached a dynamic equilibrium, the Cond value in the algal solution also reached dynamic equilibrium.

Comparative Analysis of the Three-Stage Pattern and Logistic Pattern

The Three-stage pattern demonstrated some similarities with the four-stage population logistic theory. Roughly, the cut-off points between the LAP and LGP in the Three-stage pattern and between the accelerated and decelerated growth periods in the logistic four-stage were both the points of maximum population proliferation rate (k/2). In this study, the chla concentrations of M. aeruginosa in the dark was higher than in the light condition until the k/2 point was reached. The chla concentrations began to be higher in the light environment compared to the dark when its content reached the k/2 point. In addition, the end of decelerated growth and the Stabilization periods in the logistic growth curve largely overlapped with the Stabilization period of the three-stage pattern. The Stabilization period in the three-stage hypothesis was similar to the growth characteristics of the Stabilization period in the logistic four-stage theory. During these two Stabilization periods, the rates of proliferation and decay of M. aeruginosa reached dynamic equilibrium, and the chla under light and dark conditions gradually became equal. This indicates that the three-stage pattern was compatible with the logistic four-stage theory and that the logistic four-stage theory can somewhat corroborate the Three-stage pattern.

The Three-stage pattern and the logistic four-stage theory did demonstrate some differences. Firstly, the basis for the division between the Three-stage pattern and the logistic four-stage theory is different. The Logistic pattern is based on different growth rates and environmental capacity (k) throughout the life cycle of M. aeruginosa. In contrast, the Three-stage pattern of growth discovered is based on the high-frequency monitoring of diurnal growth variability, proliferation rate, and environmental accommodation for delineation. Secondly, the data context of the Three-stage pattern differs from the four-stage theory of logistic growth curves. Briefly, the logistic four-stage theory focuses on the proliferation rate of a population throughout its life cycle and the data are analyzed in terms of days only. However, the Three-stage pattern is discovered in the context of high-frequency monitoring data, which can be obtained as often as every 15 min.

The Three-stage pattern can clearly and accurately characterize the growth stages of M. aeruginosa under indoor culture conditions. Firstly, the Three-stage pattern has more high-frequency data to support the division of the growth stages of indoor M. aeruginosa at a more microscopic level of diurnal growth differences. Secondly, although only the Three-stage pattern of M. aeruginosa was classified, these covered the trends of multiple processes including adaptation to a new environment (light, competition, etc.), changes in proliferation rate, and the process of diurnal growth variability. It is therefore advantageous to investigate the indoor culture of M. aeruginosa.

The Three-stage pattern has some compatibility with the logistic four-stage theory. The Three-stage pattern encompasses the ecological features of k and k/2 in the logistic four-stage theory and covers the value-added rate changes of M. aeruginosa throughout its life cycle. The extreme value of the stable stages in the Three-stage pattern is represented by k. The LGP and LAP stage were divided by the maximum accretion rate k/2 point. This division model was compatible with the logistic four-stage theory. Alternatively, the three stages of the Three-stage pattern can accurately describe the changes in the algal proliferation rate throughout their life cycle, including the characteristics of diurnal variability of algae in the high-frequency monitoring background.

Logistic models were widely used in growth simulations and predictions of microbial and algal populations. However, microorganisms and algae that fully conform to the logistic four-stage growth characteristics are easily influenced by environmental factors [31]. The three-stage pattern covers both the ecological characteristics of the logistic four stages and incorporates the characteristics of diurnal variability in the context of high-frequency monitoring including the characteristics of the effects of changes in the culture environment. Only for guiding indoor culture and research of M. aeruginosa, the Three-stage pattern has certain applicability. However, the high-frequency monitoring test in this study also had certain drawbacks. Nutrients such as N, P, and C are key nutrient factors for algal growth. High-frequency monitoring of the nutrients was not included in this study because the nutrients in the BG11 medium used for the culture experiments were sufficient for the growth and proliferation of M. aeruginosa and due to the limitations of current high-frequency monitoring equipment. Therefore, to explain M. aeruginosa growth patterns more accurately and thoroughly, this study utilized the high-frequency monitoring system of added nutrients such as N, P, and C. At the same time, the interpretation of cyanobacterial growth patterns only in laboratory culture experiments has certain limitations for the prevention and prediction of cyanobacterial blooms. Therefore, this study aimed to reveal the growth patterns of cyanobacteria in natural water at a more detailed level, to provide scientific suggestions for the development and improvement of current cyanobacterial water bloom prevention, early warning, and control technologies.