Microcystis aeruginosa and microcystin-LR

Microcystis aeruginosa (from the Canadian Phycological Culture Centre [CPCC] at the University of Waterloo, Waterloo, ON, Canada; strain 300; CPCC 300) was cultured at the Ontario Ministry of Environment, Conservation and Parks (MOECP) in Etobicoke, Ontario, Canada in a BG-11 media (Acreman n.d) following the method of Shahmohamadloo et al. (2019). The concentration of microcystin-LR, microcystin-dmLR, nodularin, anatoxin-A, microcystin-dmRR, microcystin-RR, microcystin-LA, microcystin-LF, microcystin-LY, microcystin-Hi1R, microcystin-LW, microcystin-YR, microcystin, HtYR, and microcystin-WR produced by the M. aeruginosa culture was measured using on-line solid phase extraction and liquid chromatography at the MOECP according to the methods of Ortiz et al. (2017). A sufficient volume of M. aeruginosa to perform each experiment was isolated from the M. aeruginosa culture and stored at 4 °C before being used in the experiment. Nominal treatment concentrations for both juvenile and glochidia tests were calculated based on the measured concentrations in the Microcystis culture. Microcystis aeruginosa were transported to the University of Guelph (Guelph, ON, Canada), where they were kept in a stationary phase in the dark at approximately 6 °C to prevent growth. To account for fluctuations in the concentration of microcystins produced by M. aeruginosa CPCC 300, water samples were taken throughout both juvenile mussel and glochidia experiments and transported back to the MOECP, where their concentrations were confirmed.

Chronic toxicity juvenile test

Juvenile fatmucket mussels were selected due to their fairly abundant dispersal across Ontario's freshwater systems. Juvenile mussels were obtained from the laboratory of C. Barnhart at Missouri State University (Springfield, MO, USA). Glochidia were collected from female mussels from Silver Fork of Perche Creek (Boone County, MO, USA). Largemouth bass (Micropterus salmoides), originating from Chesapeake Fish Hatchery (Mt. Vernon, MO, USA), were used as a host species for the glochidia in the Barnhart laboratory. Juvenile mussels were released from the gills of the fish between 7 and 18 July of 2018. Juveniles were then shipped overnight to the University of Guelph on 5 September 2018. On arrival, mussels were approximately 1.5 to 2 mo old. At the University of Guelph Hagen Aqualab, mussels were transferred to Petri dishes filled with approximately 2 mm of sand, which were placed in an aerated communal tank. In the communal tank, the mussels underwent a 4-d acclimation period. During the acclimation period, mussels were fed daily 200 µg/L of a mixture of 0.7% Nanno 3600^TM (Nannochloropsis sp.) and 2.7% Shellfish Diet 1800^TM (Isochrysis, Pavlova, Thalassiosira, and Tetreselmis sp.; Reed Marticulture, San Jose, CA, USA).

At the end of the acclimation period, mussels were exposed to nominal microcystin concentrations of 2.5, 5, 10, 25, 50, and 100 µg/L. The concentrations of the microcystin solutions were made based on the concentration of microcystin-LR measured in the M. aeruginosa culture before a sufficient volume of cyanobacteria was isolated from the M. aeruginosa culture for testing and by diluting with culture water from the Hagen Aqualab (Supplemental Data, Table S1). For each concentration, 3 1-L beakers were filled with approximately 2 mm of sand and 700 mL of microcystins solution. Two control treatments with 3 replicates each were used in the experiment, a negative control, and a BG-11 media control to ensure that any observed effects were not due to the M. aeruginosa culture media. Each beaker contained 10 juvenile mussels that were aerated and given 200 µL of algae food mixture, as described above, per beaker, per day. Beakers were randomized within a growth chamber at the Hagen Aqualab with a 16:8-h light:dark photoperiod and a temperature of 20 ± 1 °C.

Water was changed weekly to refresh the microcystin concentrations. Weekly water changes occurred on days 7, 14, and 21. During water changes, mussels were removed from sediment and counted as alive or dead under a dissecting microscope. Mussels were considered alive if foot movement was observed in a 2-min time period. During water changes, water samples were taken, and water chemistry parameters were measured at the beginning and end of each 7-d period (Supplemental Data, Table S2). Beakers were rerandomized after each water change. Water samples were frozen and analyzed for microcystin concentration at the MOECP.

Acute toxicity glochidia test

The procedures of the glochidia test followed the methods of ASTM International (2013). Gravid female L. siliquoidea mussels (~8 cm in length) were collected from the Maitland River in Ontario, Canada. The mussels were kept in a flow-through system at the Canadian Centre for Inland Waters (Burlington, ON, Canada) at approximately 14 ± 2 °C. Glochidia testing followed the procedures outlined in ASTM International (2013). Three of the collected gravid females were determined to have glochidia viability greater than 90%. Glochidia from these females were pooled together in moderately hard water. The viability of the pooled glochidia was 91.0%. The moderately hard water was made according to the US Environmental Protection Agency (2002) method. The physicochemical properties of the moderately hard water are reported in the Supplemental Data, Table S1. The pooled viability of the glochidia was greater than 90% before the start of the test. Viability was determined by taking subsamples of approximately 100 glochidia or more and counting the number of open and closed individuals before and after the addition of a concentrated solution of sodium chloride. Equation 1 was then used to calculate the percentage of viability.

(1)

Pooled glochidia were separated into 250-mL beakers for microcystin-LR concentrations of 2.5, 5, 10, 25, 50, and 100 µg/L, as well as the negative control and media control. Concentrations were made by diluting the culture with moderately hard water. Four replicates were used for each concentration and both controls. Water chemistry parameters were taken at the beginning and end of the 72-h acute test (Supplemental Data, Table S3). Test vessels were kept in a growth chamber at 21 ± 1 °C with a 16:8-h light:dark photoperiod. Viability checks were conducted at the 24, 48, and 72 h by taking a subsample from each beaker and determining the viability as just described. Test vessels were rerandomized every 24 h after observations were completed. The mean viability of glochidia in the negative control and media control remained above 90% at the 24-, 48-, and 72-h sampling points.

Statistical analysis

The geometric means of the measured concentrations of microcystin-LR from day 1 to day 7 were used for the analysis of mussel toxicity during the first 7 d (Supplemental Data, Table S4). For day 14, the geometric means of the concentrations included the measured concentrations on days 1, 7, and 14. For days 21 and 28, the measured concentrations of the previous water change days were included in the calculation of the geometric means (Supplemental Data, Table S4). The measured geometric means and nominal concentrations are compared in the Supplemental Data, Table S5.

A Welch 2-sample t test was used to determine whether there was a significant difference between the controls and media controls. This was done to ensure that the media control was not influencing mussel survival.

Statistical analysis was performed using R Ver 3.5.1. Median lethal concentration (LCx) values of 10, 25, and 50% and their respective standard errors, and 95% confidence intervals were estimated using nonlinear regression. A 4-parameter log–logistic model was fit for a binomial distribution using the drc package in R. Lethal concentrations were based on a mortality response for juveniles and a viability response for glochidia.

To determine the environmental relevance of the lethal concentrations, an environmental exposure distribution (EED) was generated from microcystin-LR monitoring data extracted from Ontario, Alberta, Quebec (Canada), and the Canadian and US Great Lakes (Giani et al. 2005; Chen et al. 2007; Zurawell 2010; Great Lakes Environmental Research Laboratory 2018). These data sources were selected based on public availability and relevance to Canadian freshwater systems. Once collected, data were fit to a log-normal distribution, and the 95th centile was calculated. A second EED was then created using only data extracted from the Great Lakes area (Great Lakes Environmental Research Laboratory 2018).

A Kruskal–Wallis rank sum test was conducted to determine whether there were significant differences between concentration treatment groups and controls. A Dunn's test of multiple comparisons using rank sums was then conducted to determine which pairs of treatment groups were significantly different. All tests were conducted to a 5% significance level (α = 0.05).

Comparison of measured and nominal concentrations

Only microcystin-LR and microcystin-dmLR were detected in the M. aeruginosa. The toxicity of M. aeruginosa to juvenile mussels and glochidia has been expressed as concentration of microcystin-LR, because historically this is how toxicity has been expressed and microcystin-LR was present at a greater concentration than microcystin-dmLR (Supplemental Data, Table S2). In the experiment with juvenile mussels, the measured concentrations of microcystin-LR were less than those of nominal concentrations, and typically decreased over the span of 1 wk between water changes (Supplemental Data, Table S2). The mean percentage differences between the nominal and measured concentrations at the start of a week ranged from 32 to 58% (Supplemental Data, Table S2). The nominal concentrations of microcystin-LR were based on measurements of microcystin-LR in the M. aeruginosa culture before cyanobacteria were isolated for experimentation. The percentage difference between nominal and measured concentrations is likely due to variation in the subsample of M. aeruginosa initially taken for the determination of the nominal concentration for experimentation and the volume of M. aeruginosa that was isolated for use in the experiments. The mean percentage difference between the beginning and end of 1 wk ranged from 17 to 52% for all concentrations (Supplemental Data, Table S2), indicating that microcystin-LR and microcystin-dmLR degraded over each 7-d exposure period before renewal or else the microcystins were taken up and/or removed by the mussels. The measured concentrations were consistent at the beginning of each 7-d period after each renewal, which would indicate that microcystin-LR and microcystin-dmLR were relatively stable during the storage of M. aeruginosa at 4 °C.

In the experiment with mussel glochidia, the percentage difference between nominal and measured concentrations of microcystin-LR ranged from 14.4 to 99.0% (Supplemental Data, Table S3). The majority of the measured concentrations were greater than the nominal concentrations (Supplemental Data, Table S3). As mentioned in the above paragraph, the variation between the nominal and measured concentrations at the beginning of the test was likely due to a difference in the concentration of microcystin-LR between the subsample of the M. aeruginosa culture that was used to determine nominal concentrations and the portion of the culture that was isolated for use in the experiment. This highlights the importance of sampling exposure water at the initiation of the test and not simply relying on the measurement of a stock solution, or in this case, a stock cyanobacteria culture.

Chronic toxicity juvenile test

There was no significant difference in survival between negative controls and media controls at 7, 14, 21, or 28 d of exposure (p > 0.05). This finding indicates that the media used to culture the microcystin did not have a significant effect on the mussels alone. There was no mortality in negative controls and media controls after the 28-d test.

During the juvenile mussel chronic experiment, microcystin-LR was observed to have a significant effect on the mortality of juvenile mussels after 21 and 28 d of exposure relative to the control treatment (p < 0.05). On days 7 and 14, no significant effect was observed (p = 0.57). The LC10 values for 21 and 28 d of exposure were determined to be 0.45 and 1.4 µg/L, respectively (Table 1). Interestingly, the LC10 value for 28 d of exposure to microcystin was a value similar to the Health Canada acceptable drinking water concentration of 1.5 µg/L, meaning that toxicity was at similar concentrations to acceptable drinking water guidelines for 10% of juvenile fatmuckets tested (Kotak and Zurawell, 2007). The no-observable-effect concentrations were determined through Dunn's test of multiple comparisons to be 1.6 µg/L for 21 d of exposure and 1.5 µg/L for 28 d of exposure. When higher proportions of mortality were examined, the LC25 and LC50 values were determined to be 7.6 and 42 µg/L for 21 d, and 0.97 and 2.1 µg/L for 28 d (Table 1). At 28 d of exposure, complete mortality was observed at the 2 highest concentrations of microcystins.

Freshwater and marine mussels have previously been found to accumulate high concentrations of microcystins in their tissues while demonstrating few negative health impacts (Chen and Xie 2007; Vareli et al. 2012). The ability of mussels to metabolize microcystins varies between species. Burmester et al. (2012) compared the responses of the invasive zebra mussel (D. polymorpha) and the native Swollen River mussel (U. tumidus) when exposed to 10 and 50 µg/L of microcystin-LR in a 7-d exposure study. They found that the invasive species had a stronger detoxification response over the 7-d period (Burmester et al. 2012). However, most detoxification abilities decreased over time and at higher concentrations (Kim et al. 2018).

Previous studies have confirmed that the toxicity of microcystins to freshwater mussels and fish was both concentration and time dependent (Burmester et al. 2012; Kim et al. 2018). Wu et al. (2017) demonstrated the time dependency of the effect of microcystins in a study that examined the cumulative impact of toxic M. aeruginosa and hypoxia on triangle sail mussels (Hyriopsis cumingii) over a period of 21 d. They found that the accumulation of microcystins in the exposed mussels increased over time (Wu et al. 2017). On day 7 of their experiment, triangle sail mussels were observed to have damage in their gills, digestive gland, and stomach (Wu et al. 2017). These effects were amplified as length of exposure increased to 21 d (Wu et al. 2017). Similarly, a study examining the toxicity of microcystins to common carp (Cyprinus carpio) found time-dependent damage in the gills and livers at 10 µg/L after 14 d of exposure (Jiang et al. 2011). Clearwater et al. (2012) observed that 92% of juvenile New Zealand freshwater mussels (Echyridella menziesii) survived after exposure to 5300 µg/L of microcystins for 96 h. Although this concentration was much higher than those used in the present study, it is important to note that Clearwater et al. (2012) quantified total microcystins, not solely microcystin-LR. In the present study, microcystin-LR was the dominant microcystin, but microcystin-dmLR was also present and might have contributed to some level of observed toxicity. Microcystin-dmLR (also referred to as [D-Asp³]-microcystin-LR) has been observed to be approximately 4 times more toxic than microcystin-LR (Shimizu et al. 2014). Although statistical analysis was completed based on measurements of microcystin-LR, microcystin-dmLR was also measured and is reported in Supplemental Data, Tables S2 and S3. The observation of time-dependent toxicity reported in these studies corresponds with what was observed in our study, because mortality in juvenile mussels did not occur until after 21 and 28 d of exposure.

The 28-d LC25 and LC50 values were below concentrations reported in aquatic systems throughout North America. In Mexico's Valle de Bravo Reservoir, microcystin-LR concentrations were reported as ranging from 0.25 to 5.56 µg/L (Alillo-Schanez et al. 2014). In Vancouver Lake (WA, USA), concentrations varied in 2009 and 2010 from 12 to 15 µg/L, well above the calculated 28-d LC50 value of 6.1 µg/L reported in the present study (Table 1; Lee et al. 2015). Within Canada in 2010 and 2011, Quebec's Missisquoi Bay had concentrations ranging from 0.01 to 24 µg/L. Considerably higher concentrations have been observed in western Canada in the past. In 1993, Driedmeat Lake (Alberta, Canada) had peak concentrations of 6400 µg/L of microcystin-LR (Kotak and Zurawell 2007). In 2003, concentrations of 400 µg/L were seen in Hamilton Harbour (ON, Canada; Kotak and Zurawell 2007).

When environmental monitoring data from Alberta, Quebec, and the Great Lakes were examined, it was evident that reported concentrations greater than 100 µg/L were not the most prominent exposure concentrations (Giani et al. 2005; Kotak and Zurawell 2007; Chen et al. 2007; Zurawell 2010; Great Lakes Environmental Research Laboratory 2018). The 95th centile of the EED indicates that 95% of reported environmental concentrations were less than 4.9 µg/L for the Canadian data (Figures 1 and 2) and 3.5 µg/L for the Great Lakes data (Figures 3 and 4). The 28-d LC50 value of juvenile mussels in the present study was greater than the 95th centile of the EED (Figure 2). However, the 21- and 28-d LC10 and 28-d LC25 values were below the 95th centile of the Canadian EED (Figures 1 and 2). This trend is further emphasized by the percentage of the EED that exceeded the LCx values (Table 1). Furthermore, the half-life of microcystin was 90 to 120 d/m depth of the water column, although it can degrade more quickly in shallow water that is exposed to sunlight (Walker and Steinberg, 2000). More frequent monitoring could better characterize the persistence of microcystins and how they may fluctuate throughout the seasons. Because microcystins could be more persistent than the length of the present 28-d chronic study, future experiments should consider extending exposure periods to better characterize the effects of this toxin (Walker and Steinberg 2000; Schmidt et al. 2014). Future studies should consider the fluctuations throughout the seasons and the type of water body because the present EED does not represent exposure under all seasonal conditions, or varying water depths.

Acute toxicity glochidia test

There was no significant difference in the viability of glochidia between controls and media controls at 24, 48, or 72 h of microcystin exposure (p > 0.05). Therefore, the BG-11 media alone had no significant influence on the viability of the glochidia (Supplemental Data, Table S3).

In the acute glochidia test, the effects of the concentrations of microcystin-LR on glochidia viability was not significant after periods of 24, 48, and 72 h (p > 0.05). A decrease in viability was observed for concentrations above 25 µg/L after 48 h of exposure, but there was no statistically significant relationship between viability and concentration. The effect concentrations for the glochidia test were calculated, but these values were not significant (Table 2). The large amount of variability between replicates at the higher nominal concentrations likely contributed to a lack of significant trends.