Effects of the nutrient inhibition on the yield of DBPFPs by Microcystis aeruginosa

Eutrophication has become a worldwide problem, and the water bloom caused by the overgrown algae and cyanobacteria in eutrophic lakes and reservoirs not only results in the deterioration of aquatic ecosystems but also causes a serious threat to the safety of water supply when these waters serve as an urban water source (Zhang et al. 2021). The phytoplankton in eutrophic waters can secrete some odorants, e.g., geosmin and 2-methylisoborneol, which emit an undesirable smell, are often a cause of consumer complaints (Liu et al. 2019), and produce some algal toxins, e.g., microcystin, nodularin and anatoxin, which may promote tumors, impair the function of multiple organs and result in neurodegenerative disorders after long-term exposure (Bilibana et al. 2022; Lad et al. 2022). Algae and cyanobacteria as well as their metabolites are also the precursors of disinfection byproducts (DBPs), which may react with chlorine to form various toxic halogenated organic compounds, such as trihalomethane (THM), haloacetic acid (HAA) and some nitrogenous DBPs (Tomlinson et al. 2016). These DBPs may increase the risk of bladder cancer, early-term miscarriage, and birth defects (Richardson & Kimura 2017). While the algal toxins can be degraded by chlorine during the disinfection process, the precursors of DBPs will be transferred into DBPs and pose a persistent threat to the safety of drinking water (Brookes & Tomlinson 2019). Since disinfection is critical to eliminate water-borne diseases, the formation of DBPs is inevitable and has become the most serious security issue in the treatment of algal-laden water in drinking water treatment plants (DWTPs).

Previous studies usually use DBP formation potential (DBPFP), the measured concentration of DBPs in water samples after a long-term reaction with excessive chlorine, to evaluate the disinfection risks. It is found that algae and cyanobacteria are typical producers of DBPFPs, and both the extracellular and intracellular algal organic matter (EOM and IOM) could react with chlorine to form DBPs (Plummer & Edzwald 2001). The IOM usually contains much more DPBFPs than EOM. However, the DBPFPs in EOM may pose a more serious threat to the safety of the water supply than those in IOM, because the IOM can be well treated with the removal of intact algal cells by the conventional water treatment process (Liu et al. 2011; Liu et al. 2014). It is also reported that hydrophilic algal organic matter (AOM) contributes more DBPFPs than hydrophobic ones, but the latter may have a higher reactivity with chlorine, i.e., hydrophobic algal DBPFPs are more easily converted to DBPs during the disinfection process of DWTPs (Huang et al. 2019). Nevertheless, since the hydrophilic compounds are more recalcitrant to the conventional water treatment process, some researchers indicate these hydrophilic compounds derived from algae could increase the disinfection risk even after the treatment by DWTPs (Tomlinson et al. 2016). Overall, the actual disinfection risk is highly relative to the characteristics and distribution of DBPFPs in algal metabolites.

It is essential to reveal the relationship between the category of AOM and DBPFPs for the management of disinfection risks during the treatment of eutrophic source water. Previous studies have shown that proteins, carbohydrates and fatty acid could be the precursors of DBPs (Navalon et al. 2008; Hong et al. 2009; Philippe et al. 2010; Tomlinson et al. 2016). Researchers also use fluorescence spectroscopy to analyze the optical properties of DBPFPs and indicate protein-like and humic-like organics are the most important precursors of DBPs (Hua et al. 2018; Hua et al. 2020; Yao et al. 2022). Some organic indicators, such as dissolved organic carbon (DOC), the ultraviolet absorbance at 254 nm (UV₂₅₄) and specific UV absorbance (SUVA), can also be used as a quick surrogate for DBPFPs (Lin et al. 2022; Yao et al. 2022).

The productivity of DBPFPs by phytoplankton, however, is still ambiguous. Generally, the concentrations of DPBFPs are proportional to the densities of algae and cyanobacteria (Liu et al. 2011; Tomlinson et al. 2016). The productivity of DBPFPs is also sensitive to algal species. It is reported that cyanobacteria could produce more DPBFPs than green algae and diatom (Liao et al. 2015; Son & Hwang 2015). Even among the cyanobacteria, the productivity of different species could still vary significantly. It is found that the specific THM formation potential (THMFP), which is expressed as yield/unit DOC, by Microcystis aeruginosa is slightly higher than that by Anabaena flos-aquae, while the specific HAA formation potential (HAAFP) by the former has approximately doubled as compared with that by the latter (Huang et al. 2009). Furthermore, the same species in different growth phases can have various performances on the DBPFP productivity. The algae and cyanobacteria at the decline phase usually produce more DBPFPs with a lower cell density than those at the stationary or exponential phases (Chen et al. 2017; Tang et al. 2020). The dead cells may release IOM into the water column and result in a soaring concentration of DBPFPs in EOM, which significantly increases the disinfection risk as discussed above (Park et al. 2021). It is believed that these variations should be strongly linked to the changed characteristics of metabolites at different growth phases (Chen et al. 2017; Park et al. 2021). The growth of algae and cyanobacteria may also be easily affected by environmental conditions, e.g., nutrient level, but the impacts of nutrient level on the production of DBPFPs by algae and cyanobacteria are rarely reported. The previous study has shown that nutrients could inhibit the growth of algae and indirectly reduce the DBPFP concentration (Tomlinson et al. 2016), but the direct impact, i.e., the productivity of DBPFPs, has not been revealed. In fact, the characteristics of AOM can vary significantly with the changed nutrient concentration (Han et al. 2016; Hua et al. 2018), which should be highly relative to the formation of DBPs. The algae from lakes with diverse trophic levels may perform different disinfection characteristics, which could affect the strategies of DWTPs for dealing with the disinfection risk. Thus, it is important to understand the response of algae and cyanobacteria to nutrient stress from the perspective of the management of DBPFP.

In the present study, the impacts of nitrate and phosphate on the growth of M. aeruginosa, a typical dominant cyanobacterium in eutrophic waters, as well as the productivity of THMFPs and HAAFPs were investigated. The algal metabolites were fractionated into five groups, i.e., hydrophilic EOM (HPI-EOM), hydrophobic EOM (HPO-EOM), hydrophilic IOM (HPI-IOM), hydrophobic IOM (HPO-IOM) and cell debris, and then the contributions of these five groups to the total DBPFPs and the structure of DBPFPs under different nutrient levels were compared. The results may provide a deep insight into the production of DBPFPs by phytoplankton under complex environmental conditions.

M. aeruginosa (FACHB-836), purchased from the Freshwater Algae Collection in the Institute of Hydrobiology, Chinese Academy of Sciences, was incubated in a BG-11 medium according to the method from Stanier et al. (1971). The culture was incubated in an illumination incubator (GZX250E, Taisite Instrument, Tianjin, China) with a constant temperature of 25 ± 1 °C and an illumination intensity of 3,300 lx in a 12-h light/12-h dark cycle. The stock culture was maintained in the late exponential or the early stationary phases by refreshing the BG-11 medium every 15–20 days. The experimental culture was prepared and incubated by diluting the stock culture to a cell density of approximately 0.25 × 10⁶ cells/mL with a modified BG-11 medium. In this medium, citric acid was not added and ammonium ferric citrate was replaced by ferric trichloride with the same molar concentration of Fe. Our pre-experimental results showed that both citric acid and ferric trichloride could react with excessive chlorine to form THMs and produce a false positive result. The concentrations of nitrate and phosphate were adjusted around the standard concentrations of these two chemicals in the BG-11 medium according to the experimental design.

The experimental culture was harvested after a 7-day incubation. The EOM was separated from algal cells by centrifugation at a speed of 8,000 rpm for 10 min (TGL-15B, Shanghai Anting Scientific Instruments Factory, Shanghai, China), and then it was filtered with a 0.45 μm polyvinylidene fluoride (PVDF) membrane to remove residual suspended solids or unprecipitated cells. The pellets were washed with deionized water (DI water) (Milli-Q, Millipore Ltd, Bedford, MA, USA) and subjected to sonication (VCX130, Sonics & Materials Inc., Newtown, CT, USA) in an ice bath with the amplitude of 99% for 1 h at a 12-s on/2-s off interval. The IOM was separated from cell debris by a 0.45 μm PVDF membrane, and the IOM in the filtrate was diluted to the same volume of initial algal suspension (usually 100 mL). The debris on the membrane was resuspended with DI water in an ultrasonic cleaner (ST36A, GT Sonic Ltd, Guangdong, China) for 20 min and was chlorinated for the measurement of DBPFPs.

The EOM or IOM was further fractionated into HPI and HPO fractions by XAD-8 resins according to the methods of Her et al. (2008) and Labanowski & Feuillade (2011). The named HPI fraction, in fact, contained some transphilic (TPI) organics. Because the TPI fractions only account for approximately 10% of the total DOC (Wei et al. 2021; Yao et al. 2022), the TPI fractions may not have a significant impact on the production of DBPFPs and were not further separated from the HPI fractions in the present study.

The samples of DBPFPs, including the HPI and HPO fractions or the cell debris, were chlorinated for a 7-day reaction with excessive chlorine at 25 ± 1 °C in a dark place according to our previous method (Huang et al. 2019). After the chlorination, the concentrations of THMs were measured by liquid–liquid extraction and gas chromatography with an electron capture detector (GC/ECD, Agilent 7890A, Agilent, Santa Clara, CA, USA) according to the method of EPA 551.1, while the concentrations of HAAs were measured by liquid–liquid extraction, derivatization with methanol and GC/ECD according to the method of EPA 552.2. Because of the absence of bromine and iodine, only chlorinated byproducts, including trichloromethane (TCM), monochloroacetic acid (MCAA), dichloroacetic acid (DCAA) and trichloroacetic acid (TCAA), were detectable. Furthermore, the standard substance of MCAA was unavailable to us because of the national regulation, and the concentration of MCAA was not quantitative.

DOC was measured with a TOC analyzer (TOC-VCPN, Shimadzu (China) Co. Ltd, Shanghai, China). UV₂₅₄ was analyzed with a UV spectrophotometer (T6, Persee General Instrument Co. Ltd, Beijing, China), and SUVA was calculated by dividing UV₂₅₄ by DOC and multiplying by 100. Proteins were measured by the Bradford method, and carbohydrates were detected by the phenol–sulfuric acid method (Dubois et al. 1956).

The experiments were carried out in duplicate, and the results were expressed as mean ± standard deviation.

Furthermore, the contributions of these five algal fractions to the total DBPFPs were highly dependent on the species of DBPFPs. TCMFP derived from EOM, IOM and cell debris accounted for approximately 19.8, 53.5 and 26.7% of the total TCMFP, respectively, while for DCAAFP, the proportions were 20.1, 57.5 and 20.4%, respectively. The proportion of TCAAFP from cell debris further reduced to less than 7%, and it increased to approximately 36% from EOM, rather than IOM, which is different from the case of DCAAFP. Because the conventional water treatment process in DWTPs cannot remove dissolved hydrophilic organics as efficiently as the hydrophobic ones and the solids (Liu et al. 2011,  2014; Tomlinson et al. 2016), the TCAAFP derived from EOM, especially from HPI-EOM, should be more difficult to remove and may threaten the safety of water supply.

The variation of nitrate concentration could affect the yield of DBPFPs. M. aeruginosa grown at the optimized nitrate concentration (247 mg N/L) usually produced the most DBPFPs, and excessive nitrate could reduce the concentrations of the three DBPFPs slightly (approximately 5–8%). However, the high nitrate concentration significantly inhibited the growth of M. aeruginosa during a 7-day incubation, and the final cell density of the culture with 494 mg N/L was 1.0 × 10⁶ cells/mL, approximately two-thirds of those with 124 or 247 mg N/L (approximately 1.5 × 10⁶ cells/mL). Thus, the productivities of these three DBPFPs by M. aeruginosa with 494 mg N/L, which were expressed as pg DBPFPs/cell, increased by approximately 40% as compared to those with 247 mg N/L. Obviously, the increased productivity of DBPFPs with high nitrate conditions would raise the disinfection risk during the treatment of algae-laden water.

Figure 1 also indicated that the concentration of nitrate could change the distribution of DBPFPs in algal fractions. As compared with the case of 247 mg N/L, reducing the concentration of nitrate did not change the TCMFP and TCAAFP in the five fractions, but it halved the concentration of DCAAFP in HPI-IOM (from 155.9 ± 17.1 to 79.5 ± 8.7 μg/L). This effect reduced the proportion of DCAAFP in HPI-IOM from 37.4 to 31.5%. Since the change primarily occurred in IOM and algal cells can be integrally removed by the coagulation and sedimentation process in DWTPs (Liu et al. 2014), the varied proportion would not affect the disinfection process significantly. Increasing the concentration of nitrate could promote the productivity of TCMFP by M. aeruginosa in EOM significantly (from 643.0 ± 43.3 pg/cells to 1,018.6 ± 67.8 pg/cells) but restrain that in IOM (from 1,737.2 ± 136.6 to 1,277.7 ± 125.6 pg/cells). Thus, the proportion of TCMFP in EOM increased from 19.8 to 34.1% accordingly. There were similar situations for DCAAFP and TCAAFP, and the proportion of these two DBPFPs in EOM increased by approximately 5%. As discussed above, the increased proportion of DBPFPs in EOM may be difficult to remove and raise the disinfection risk.

The phosphate inhibition showed a similar pattern to the nitrate inhibition. Excessive phosphate also made the growth of M. aeruginosa slow down, and the density after a 7-day incubation was only 0.75 × 10⁶ cells/mL, approximately half of that with P concentration of 7.1 mg/L in a standard BG-11 medium. Because of the growth inhibition, the concentrations of these three DBPFPs were reduced by approximately 35% as shown in Figure 2(a), 2(c) and 2(e). The productivities of TCMFP, DCAAFP and TCAAFP, however, increased by 32.4, 30.0 and 27.0%, respectively, which were also by the case of nitrate inhibition. Nevertheless, the variations in the distributions of DBPFPs in the five fractions were different from the case with nitrate inhibition. The proportion of TCMFP in EOM remained virtually unchanged and the proportion of TCAAFP in EOM even decreased, which implied that the inhibition mechanism of nitrate and phosphate may be different.

It was interesting that the mechanism of the nutrient inhibition may be quite different, though the proportion of DBPFP did not change under the nitrate and phosphate inhibition conditions. The algae may change their metabolic activity to adapt to the severe environment, including producing more AOM and changing the structure of AOM (Staehr & Birkeland 2006; Zhao et al. 2015; Desmond et al. 2018), both of which may affect the productivity of DBPFPs.

The nutrient inhibition promoted M. aeruginosa to produce more metabolites, which expanded the pool of DBPFPs. As compared with the cyanobacteria grown in a standard BG-11 medium, the nitrate inhibition increased the productivity of DOC and UV₂₅₄ by M. aeruginosa, which was expressed as the concentration divided by cell density, by 18.3 and 52.5%, respectively, while the phosphate inhibition increased the productivity by 88.2 and 74.9%, respectively (Supplementary Fig. S1). This variation may be attributed to the physiological response of algae and cyanobacteria to severe environmental conditions. It is found that Chlorella pyrenoidosa, a typical green alga, can secrete more EOM under a low-temperature condition (Zhao et al. 2015). Han et al. (2016) also indicate that the increase of nitrate or phosphate can enhance the yield of DOC by Nitzschia palea, a species of common diatom. The increased total amounts of AOM could provide more DBPFPs and pose a more severe disinfection risk to the consumers of drinking water.

It could be concluded that both the nitrate and phosphate inhibitions could enhance the productivity of DBPFPs by M. aeruginosa, but the mechanisms were different. The nutrient inhibition could promote M. aeruginosa to secrete more metabolites, but only a part of these metabolites could react with chlorine to form DBPs, i.e., DBPFPs. As compared with the case grown in a standard BG-11 medium, the cyanobacteria under the nitrate inhibition condition tended to secrete more DBPFPs than non-DBPFPs, which caused the increased specific DBPFPs. However, M. aeruginosa under the phosphate inhibition condition tended to secrete more non-DBPFPs, though the production of DBPFPs was also enhanced, which resulted in the decreased specific DBPFPs but a raised productivity of DBPFPs. Although the variations of carbohydrates in EOM and proteins in IOM (Supplementary Fig. S2) supported the inference, the reason that M. aeruginosa showed quite different responses to the nitrate and phosphate inhibitions and the algal metabolites dominated the variation was still unclear, which may be an interesting topic for the researchers in phycology.

In the present study, the effects of the nitrate and phosphate inhibitions on the yield of DBPFPs by M. aeruginosa were investigated, and the key findings were listed below:

• The nitrate inhibition with the nitrate concentration of 494 mg/L could enhance the productivity of DBPFPs by approximately 40% as compared with that grown in a standard BG-11 medium, and the increased DBPFPs primarily came from EOM.

• The phosphate inhibition also could enhance the productivity of DBPFPs by M. aeruginosa with a lesser range, but the proportion of DBPFPs in the five fractions (HPI-EOM, HPO-EOM, HPI-IOM, HPO-IOM and cell debris) did not change significantly.

• The nutrient inhibition could promote M. aeruginosa to secrete more metabolites. The cyanobacteria under the nitrate inhibition condition tended to produce more DPBFPs, while those under the phosphate inhibition condition tended to produce more non-DPBFPs. This difference resulted in an increased specific DBPFP under the nitrate inhibition condition and a decreased trend under the phosphate inhibition condition.

This work has been financially supported by the National Natural Science Foundation of China (No. 52270002), Guangdong Basic and Applied Basic Research Foundation (2021A1515010508), Science and Technology Project of Guangzhou (202102010448), and Hundred Talent Program of Guangzhou University (RQ2020050).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.