Data analysis was done using RStudio version 2022.07.2 working on R version 4.2.2 (2022-10-31). Furthermore packages: Tidyverse (version 1.3.2), DADA2 (version 1.26.0), Vegan (version 2.6-4), phyloseq (version 1.42.0), and stats (version 4.2.2) were used for all data processing, statistical analysis, and graphing. Statistical significance was measured using two-way ANOVA and multiple pairwise comparisons were coupled with Tukey’s post hoc. Community relative abundances were compared using ADONIS and was corrected using the Benjamini Hochberg protocol.

For the batch culture trial, a microcystin toxin-producing Planktothrix agardhii strain (strain designation 1,031) was used, and infected with chytrid isolate C2. A water pump was used to simulate natural water movement. The water circulation batch culture infection prevalence was 0–1.5% while the stagnant batch culture ranging from 1.20% to 83% (Figure 2A). Sporangia were counted as a proxy for chytrid abundances which were significantly different between the two cultures (p = 1.03E − 02; Figure 2B). There was no significant difference between the top and the bottom of the batch culture (Supplementary Figure 2). Infection prevalence were significantly different between the stagnant and the water circulation (p = 1.15E − 06; Supplementary Table 1). Because a microcystin producing P. agardhii strain was used, the dissolved fraction of toxin was tested to assess if chytrid infection increased toxin release in large cultures. Dissolved microcystin concentrations were consistently below 1 μg L⁻¹ and concentrations between the water circulation and stagnant batch cultures were significantly different (p = 2.87E − 03; Supplementary Table 1). Looking at microcystin per filament of Planktothrix revealed differences between the two treatments (p = 1.05E − 08; Supplementary Table 1) and throughout the experiment (p = 2.47E − 06).

To better understand the environment in which chytrid infections would occur, physiochemical data was collected within each mesocosm tube as well as in the cove waters for a period of 8 weeks, from 27 August 2021 to 22 October 2021. During this time, water temperatures ranged from 14.24°C to 27.83°C and had averages of 20.9°C ± 4.21°C, 20.7°C ± 4.24°C, and 20.7°C ± 4.18°C for control, stagnant, and circulation treatments, respectively (Figure 3A). Furthermore, over the 8-week period, temperatures were between 20°C and 22°C for about 25% of the time. The pH ranged 8.52–10.77 with stagnant water having the greatest fluctuations during the experiment (Figure 3B). Conductivity ranged 295.00–491.10 μS cm⁻¹, again with the stagnant water mesocosms having the greatest range (Figure 3C). Lastly, Secchi disk measurements as a proxy for light attenuation had an average of 21 cm for all mesocosms during the experiment and ranged from 10 to 62 cm with the highest depth values taken in the water circulation mesocosms. During peak chlorophyll-a measurements, the stagnant water mesocosms developed a film of cyanobacteria, which decreased Secchi disk measurements for these samples (Supplementary Figure 4).

Nutrient samples were obtained throughout the experiment at the control site and in the mesocosms to track nutrient availability to the community. Nutrient concentrations during the experimental period were 4.09–33.05 μmol L⁻¹ for total phosphorus and 138.24–694.16 μmol L⁻¹ for total nitrogen (Figures 4H,F). Treatments had large differences between the replicate mesocosms causing high standard deviation (Supplementary Table 1). Interestingly, stagnant mesocosms had large spikes in ammonium, dissolved reactive phosphorus and total kjeldahl nitrogen (Figures 4A,B,E) while the water circulation treatment had large spikes of nitrate and nitrite early in the experiment (Figures 4C,D). These spikes could be caused by changes in pH and changes in dissolved oxygen (Supplementary Figure 5). Additionally, the spikes in the water circulation mesocosms could be caused by the circulation which could allow for nutrient resuspension. Total phosphorus and total nitrogen were significantly different between the treatments (p = 6.40E − 06 and p = 2.92E − 02 respectively). There were some differences between other individual nutrients during the experiment, see Supplementary Table 2. Furthermore, the spikes seen at date 24 September 2021, and leading up to that time correspond to a period when dissolved oxygen was low in both treatments with an average of 2.54 ± 1.36 mg L⁻¹ and 1.60 ± 0.36 mg L⁻¹ for stagnant and water circulation, respectively, on 24 September 2021 (Supplementary Table 3).

Chlorophyll-a concentrations have been used as a proxy of productivity within a water body. Extractive chlorophyll concentration ranged from 24.77 to 896.5 μg L⁻¹ and had an average of 356.6 ± 62.23 μg L⁻¹, 376.5 ± 236.8 μg L⁻¹, and 347.2 ± 304.9 μg/L L⁻¹ for control, stagnant water, and water circulation, respectively (Figure 5). This high standard deviation is due to a large amount of variability between the six mesocosms and there was a cold spell where air temperatures dropped from 28.3°C to 13.4°C during the day and at night it went from 13.9°C to 7.8°C which caused the mesocosms to drop from an average of 24.53°C to 14.58°C across both mesocosm treatments and the control (Figure 3A). Stagnant mesocosm concentrations were significantly higher than the water circulation (p = 4.57E − 09) and the control treatment (p = 3.76E − 02) but water circulation was not significantly different from the control (p = 6.33E − 02; Supplementary Table 4). There was a noticeable decline in chlorophyll during the beginning part of the experiment, with an increase on 01 October 2021, when the mesocosms were reset, and a subsequent decline for the rest of the experiment until the mesocosms were removed on 22 October 2021.

Microcystin toxin concentrations were measured during the experiment to look for the potential increase in the dissolved fraction as a response to increasing cell breakage from chytrid infections. The World Health Organization (WHO) has set a recreation exposure guideline to 10 and 1.5 μg L⁻¹ for drinking water. Both particulate and dissolved were measured by mixing the three mesocosms for each treatment into one sample for analysis. Dissolved concentration averaged 0.58 ± 0.32 μg L⁻¹ and particulate concentrations averaged 4.87 ± 2.86 μg L⁻¹. All samples from the particulate fraction were diluted 10-fold (Figure 6). Two-way ANOVA resulted in significant differences between all treatments and dates in particulate fraction (p = 7.79E − 08 and p = 5.90E − 06, respectively). The dissolved fraction was not significant between treatments and dates (p = 3.98E − 01 and p = 3.97E − 01, respectively). While particulate toxins were significantly different total toxins (both particulate and dissolved) was not significantly different between treatments and dates (Supplementary Table 5).

Infection prevalence was quantified using microscopy for the duration of the experiment. Infections reached a max of 4.12% infection with an average of 0.27% ± 0.76% over all samples and dates (Figure 7). There were infections in 25% of the samples examined (n = 112), with more infections occurring in the stagnant samples vs. the other two treatment types. There were trends within the data that suggest that 9.1–9.4 and 10.3–10.6 pH, 19°C–23°C, and 316–412 μS cm⁻¹ were ideal for infections but due to the low number of samples with infections we could not statistically show significance. Furthermore, there was a significant difference between infections in the three treatment groups (p = 8.85E − 07, Supplementary Table 6) with the stagnant treatment being different from both the control (p = 1.05E − 03) and water circulation treatment (p = 1.72E − 06; Supplementary Table 7). There was no statistical difference in infections between the control and the water circulation treatment (p = 9.94E − 01). Temperatures between 19°C and 23°C had the highest percent of the counted filaments infected (Figure 7) and infections within this range were significantly higher than outside of the range (p = 1.52E − 05; Supplementary Table 8). Samples were taken at both top and bottom of each mesocosm and the control with no apparent difference between the two depths, as was seen as the in vitro experiments.

In vitro batch cultures were utilized to help predict the dynamics of in situ mesocosm experiments by controlling for all factors except for water agitation. Agitation using a water pump maintained similar levels of biomass between the stagnant and water circulation carboy and indicated that water movement could significantly reduce chytrid infections on P. agardhii (Figure 2). Additionally, Planktothrix growth under moving water and stagnant water conditions are different. In batch culture water movement increased the growth rate of Planktothrix. This could be due to the introduction of CO₂ from the movement on the surface layer of water or from the lack of parasite infectability. This confirmed previous work that suggested water agitation would reduce chytrid infections in the wild, either by preventing zoospores from finding their host through signal disruption or by preventing swarming and attachment (McKindles et al., 2021a,b).

The presence of microcystin, a toxin of great concern, was also evaluated, as it was previously suggested that chytrid infection could cause the cells to release high concentrations of microcystin into the dissolved fraction. However, the batch culture experiment showed that even when the chytrid infection prevalence reached 40%–60%, there was almost no detectable dissolved microcystin presence (0.06 to 0.6 μg L⁻¹). Therefore, chytrid infections may not have a significant impact on microcystin production by Planktothrix, which support data in another recent study (Agha et al., 2022).

The main goal of this work was to build upon previous studies looking at the environmental effects on chytrid infection on Planktothrix agardhii within the context of a real environment. The mesocosms were accessing an environment without amendment. Most studies have been conducted in laboratory-controlled conditions, which are important to get an estimation of what is possible, but in situ studies like the one conducted here can confirm lab observations and identify other potential factors that may constrain or maximize chytrid pathogenesis. Our 16S and 18S community analysis showed a population shift following installation of the mesocosms. GLSM has a dominant Planktothrix bloom through many months of the year (Steffen et al., 2014), meaning it would be an ideal study location.

Physiochemical measurements during the sampling period indicated that both temperature and conductivity were well suited for chytrid pathogenesis as found in previous studies (McKindles et al., 2021b). In lab studies using local to Ohio isolates of both host and parasite indicated that infections were most prominent at 0.226–0.426 mS cm⁻¹ (McKindles et al., 2021b), which was consistent with the conductivity range of our study site at GLSM (0.246–0.491 mS cm⁻¹; Figure 3). Temperature was also indicated to have a significant effect on infection prevalence, with an optimal infection temperature of 21°C (McKindles et al., 2021b). Similarly, our mesocosm had the highest precent of infections between 19°C and 22°C (Figure 7). During the sampling period, the mesocosm were exposed to a wide range of water temperatures, from 14°C to 28°C (Figure 3). Infections could occasionally be identified at these extreme temperatures, but they were infrequent and not at a high prevalence (Figure 7), further suggesting the significance of thermal refuges for chytrid infections. Previous work looking at infection dynamics of Chytridiomycota and Planktothrix species has identified these thermal refuges in which the host is able to grow at temperatures which are limiting to the parasite. In controlled laboratory studies, Rohrlack et al. (2015) identified a lower thermal refuge for Planktothrix rubescens at temperatures below 11.6°C, while McKindles et al. (2021b) identified an upper thermal refuge for Planktothrix agardhii at temperatures above 27.1°C. To date many of the experiments looking into chytrid-host interactions have been laboratory based and have reported infection rates that are not seen in the environment. More real-world experimentation like what is being done here should be done to elucidate what factors are preventing chytrid pathogenesis in the environment.

Nutrients in Grand Lake Saint Marys have been an ongoing topic of concern (Jacquemin et al., 2018). During the 8 weeks that the mesocosms were in total nitrogen in the control ranged from 138.24–694.16 μmol L⁻¹ (Figure 4F) and the total phosphorus ranged from 4.091–33.05 μmol L⁻¹ (Figure 4H). In the water circulation mesocosms, there was a spike of nitrate (Figures 4C,D) causing a high TN:TP ratio which could have allowed for Planktotrhix to be less competitive (Figure 4G). These spikes on 20 September 2021 and leading up to that time correspond to a period when dissolved oxygen was low in both treatments (Supplementary Figure 5). One possible explanation for the increase could be the decreasing temperature or the large rain event that dropped nearly 12.7 cm of rain, reducing phytoplankton production and thus relieving the use of these nutrients causing the spike. Interestingly, stagnant mesocosms had large spikes in ammonium, dissolved reactive phosphorus and total kjeldahl nitrogen (Figures 4A,B,E) this was also during the time when temperatures were low and right after the rain event. Ammonium plays a key role in sustaining Planktothrix blooms (Hampel et al., 2019), despite this, chlorophyll-a declined on 24 September 2021 and did not fully recover.

By reducing the water flow and creating a stagnant environment with the untreated mesocosm, both the bioactivity and cyanobacterial toxin production increased. Measurements of chlorophyll-a were used a proxy for bioactivity for each treatment, with significantly higher values in the stagnant water mesocosms compared to the water outside of the mesocosms, which was higher still than the water circulation mesocosm (Figure 5). The increased chlorophyll-a measurements for the stagnant water mesocosms were due to an increase in abundance of cyanobacteria, particularly Planktothrix species (Figure 8; Supplementary Figure 6). Due to Planktothrix being a fast migrating buoyant cyanobacteria, it is able to take advantage of stable, stagnant water column by adjusting its depth for optimal light conditions (Carey et al., 2012). This increase was also reflected in the increases in microcystin total toxin concentrations (Figure 6). In the batch culture experiment, the stagnant water treatment had a reduced Planktothrix carrying capacity, whereas in the mesocosms, Planktothrix was more prevalent in the stagnant condition than the water circulation treatments despite having more parasite pressures. This is likely because in the environment the Planktothrix community is more diverse than a single strain that is known to be susceptible to infection like in the batch culture. In other studies where they look at mixed populations of both host and parasite they find that mixed cultures leads to a decline in infection prevalence (McKindles et al., 2023). While McKindles et al. utilized smaller benchtop experiments more lab experiments with mixed cultures would likely be necessary to see if non-susceptible strains of Planktothrix present in the stagnant water of the batch culture could then thrive as seen in the environmental mesocosms. While the cyanobacterial population dominated in the stagnant water mesocosms, it was replaced by a Chlorophyta population in the water circulation mesocosm (Figure 8; Supplementary Figure 11), perhaps indicating other environmental factors are involved population shifts besides parasitic pressure.

Within the stagnant water mesocosms, Planktothrix proliferated and so did the organisms that could lead to their demise. 18S reads indicated that the stagnant mesocosm was dominated by Ciliophora, Dinoflagellata, and Fungi. Ciliates are an important group of microorganisms in freshwater ecosystems, playing key roles in controlling algal populations and nutrient cycling and can therefore play an important role in mitigating the negative impacts of algal blooms. Many of the taxa within the Ciliophora and Fungi division were ones that have been reported to consume (Filip et al., 2012; Combes et al., 2013; Farthing et al., 2021) or parasitize Planktothrix species (Kagami et al., 2007; Agha et al., 2016; Frenken et al., 2017; McKindles et al., 2021a,b). Combined with the measures of diversity provided by the 16S and 18S data, this study reveals the importance of water movement in structuring the phytoplankton population. Overall, water movement suppresses Planktothrix dominance, despite the reduction of chytrid parasitism by coexisting Rhizophydiales. Combined with the fact that infections were rarest at 23°C and 25°C, the effects of climate change and lake warming may further reduce chytrid success. Consequently, the role of chytrid pathogenesis has limited impact in this system, despite the potential importance of trophic upgrading to grazers in the food web (Gerphagnon et al., 2019).