Disentangling the role of light and nutrient limitation on bacterivory by mixotrophic nanoflagellates

Many phytoplankton taxa function on multiple trophic levels by combining photosynthesis and ingestion of bacteria, termed mixotrophy. Despite the recognition of mixotrophy as a universal functional trait, we have yet to fully resolve how environmental conditions influence community grazing rates in situ. A microcosm study was used to assess bacterivory by mixotrophic nanoflagellates following nutrient enrichment and light attenuation in a temperate lake. We found contrasting results based on assessment of mixotroph abundance or bacterivory. Despite an interactive effect of nutrient enrichment and light attenuation on mixotroph abundance, significant differences within light treatments were observed only after enrichment with P or N + P. The greatest abundance of mixotrophs across treatments occurred under co‐nutrient enrichment with full exposure to irradiance. However, bacterivory by mixotrophic nanoflagellates was greatest under shaded conditions after either N or P enrichment. We suggest that PAR availability dampened the stimulatory effect of nutrient limitation, and bacterivory supplemented a suboptimal photosynthetic environment. In a saturating light regime, the mixotrophic community was less driven to ingest bacteria because photosynthesis was able to satisfy energetic demands. These findings quantify community bacterivory in response to environmental drivers that may characterize future ecosystem conditions and highlight the importance of considering grazing rates in conjunction with abundance of mixotrophic protists.

aquatic food webs that integrates mixotrophy is of vital importance given the potential for mixotrophic activity to enhance trophic transfer. However, there are significant gaps in our knowledge regarding in situ mixotroph activity under a variety of environmental conditions, especially in inland waters.
Whereas grazing by heterotrophic forms is typically constrained by prey abundance, activity of mixotrophs can be regulated by a variety of abiotic factors such as temperature, light availability, and dissolved nutrient concentration . Mixotrophy is theorized to be advantageous under sufficient irradiance when dissolved nutrients are limiting (Edwards, 2019). Studies support the success of mixotrophic nutrition in oligotrophic environments where dissolved nutrients are limiting (Unrein et al., 2013). However, an increase in bacterivory by mixotrophs has been demonstrated under photosynthetically active radiation (PAR) limitation as a mechanism to supplement reduced carbon fixation (Wilken et al., 2020). It is well known that there exists species-specific variability in the relative reliance on either photosynthesis or bacterivory, but there is no consensus for how mixotrophs may respond to light in combination with stoichiometric imbalances. Changes in global climate are likely to be accompanied by increases in surface water temperature that may strengthen and extend the duration of thermal stratification in freshwater ecosystems. Resultant modifications to water column structure include restriction in the upward flow of nutrients that may be favorable to mixotrophic protists. Additionally, inflows of terrestrially derived organic matter are predicted to increase in response to more frequent and severe precipitation events and can significantly attenuate PAR in freshwater lakes.
Due to the wide range of strategies amongst mixotrophic protists, especially in response to environmental conditions, the understanding of activity of mixotrophs under climate change scenarios is a critical challenge (Gonzalez-Olalla et al., 2019). It is necessary to conduct in situ experiments that explore the responses of different metabolic processes under multiple stressors that are emerging in aquatic systems. The goal of this field study was to assess bacterivory by mixotrophic nanoflagellates (MNAN) under modification of the light environment (differential shading of solar radiation) with nutrient enrichment (+N, +P, +NP). We hypothesized that rates of bacterivory by MNAN would be greatest under shaded conditions with single-nutrient enrichment in which increased bacterivory would be driven by compensation for decreased photosynthetic carbon fixation plus a stoichiometric imbalance.
The study was performed at Lake Lacawac, located on the Pocono Plateau of northeastern Pennsylvania, U.S.A. (41°22′55.2″ N, 75°17′31.2″ W). It is a mesotrophic and moderately dystrophic, 21-ha freshwater lake. Long-term observations provide evidence that the lake has experienced significant declines in water clarity Pilla et al., 2018). Previous work has also indicated seasonal dominance and importance of mixotrophic nanoflagellates in spring and mid-summer, as well as under ice (Berninger et al., 1992;Princiotta & Sanders, 2017). Subsurface water samples (0.5 m depth) from the lake were collected in 25-L plastic containers in July 2017 before immediate transport to the on-site laboratory. Water samples were filtered through 100 μm mesh to remove zooplankton grazers and distributed into 32 pre-rinsed UV-transparent Bitran bags (30.5 × 30.5 cm). A total of eight bags were supplemented with two nitrogen species (hereafter +N treatment) with a solution of 168 μg · L −1 NO 3 and 168 μg · L −1 NH 4 (24 μM N total). For phosphorus enrichment (hereafter +P treatment), eight additional bags received a solution of 31 μg · L −1 NaH 2 PO 4 (1 μM P total). An equal number of bags received an addition of both macronutrients at the same concentrations (+NP treatment, n = 8), and a final set did not receive any nutrient addition (C, control treatment, n = 8). Bags were randomly fastened to two racks constructed from ¾ inch PVC tubing, supported by 0.09 mm plastic sheeting in such a way that half of each nutrient treatment (n = 4 replicate bags) were covered by multiple layers of neutral density mesh that blocked 75% of incident solar radiation, whereas the remaining bags received ambient solar radiation. Microcosm set-up was incubated and undisturbed for 7 d in the center of the lake.
Grazing rates by MNAN were determined both before (in ambient, non-nutrient amended lake water) and after the 7-d experimental period via short-term feeding experiments. Subsamples were removed from experimental bags, transferred into 15.24 × 15.24 cm Bitran bags, and incubated at ambient temperature and light conditions in the on-site laboratory. Fluorescent polycarbonate microspheres (0.6 μm diameter, Polysciences, Inc.) were added to each replicate bag at a final concentration of ~5 × 10 5 microspheres · mL −1 , and the resultant ratio of natural bacteria to fluorescent tracers ranged from 1.7 to 5.7. Subsamples from each bag were fixed after 20 min (Sherr & Sherr, 1993). A single sample from each treatment was taken to account for coincidental, nonspecific association events between MNAN and prey surrogates (hereafter T0). Samples were fixed sequentially with Lugol's iodine, Na 2 S 2 O 3 , and formalin to prevent egestion of microspheres (Sherr et al., 1987).
Determinations of MNAN abundance and grazing were prepared from fixed samples that were concentrated by ~50% via settling in darkness for a minimum of 24 h. A volume of each concentrated sample was filtered onto either a 0.2 or 0.8 μm polycarbonate filter and mounted onto a slide with mounting media containing DAPI stain (Vector Laboratories). Enumeration of MNAN, fluorescent microspheres, and natural bacteria in the water, as well as ingestion of microspheres was determined using epifluorescence microscopy on an Olympus BX41 microscope at 1000×. A minimum of 100 nanoflagellate cells were first visualized using a DAPI filter set. MNAN were identified as such based on the presence of chlorophyll autofluorescence plus the presence of at least one microsphere within the cell boundary. Coincidental ingestion was subtracted from rate of tracer uptake (T0) to account for false-positive accounts of grazing. Bacterial ingestion rate was calculated by multiplying microsphere ingestion rate, corrected by coincidental ingestion (T0), to the ratio of natural bacteria to fluorescent microspheres.
Ambient radiation data were collected with a LI-190SA quantum sensor (LI-COR, Inc.) coupled to a CR-10 data logger (Campbell Scientific) that averages incident PAR over 15-min intervals (Hargreaves, 2019). Temperature conditions during the incubation were derived from the average of two thermistor temperature sensors (Campbell Scientific, type 107B) deployed from a raft located at the lake's center at 0.1 and 0.5 m below the lake surface. Chlorophyll a and total macronutrient concentrations (TN, TP) were determined at the beginning of the experiment, from subsamples of the initial water collection (n = 4 for chl a, n = 4 for TN, and n = 4 for TP), and at the end of the 7-d incubation period from each replicate bag. To determine concentration of chl a, 50-mL aliquots were filtered onto a GF/F filter and stored frozen until analysis on a Turner Designs 10 AU Fluorometer. Pigment was extracted in acetone/ methanol solution for 48 h (Pechar, 1987). Additional replicates of whole subsurface water were transferred to pre-rinsed bottles, acidified with H 2 SO 4 , and stored for macronutrient measurement by spectrophotometric technique. Total phosphorus was determined following digestion with potassium persulfate solution by molybdenum blue method. Total nitrogen was determined by the cadmium reduction method after persulfate digestion.
Statistical analyses were run using R statistical software (V 3.6.2, packages multcomp and car). Nonparametric statistics were completed in JMP. A two-way analysis of variance was used each for MNAN ingestion rate and abundance against treatments (nutrient enrichment and light availability). Normality and homogeneity of variance were confirmed with a Shapiro-Wilk and Levene's test, respectively. Ingestion rates and MNAN abundances were square-roottransformed in order to meet statistical assumptions. If the effect of either treatment was significant, subsequent post-hoc LSD HSD tests were used to determine which treatments were significantly different from one another based on a probability value of p < 0.05. Nonparametric methods (Kruskal-Wallis test with posthoc Steel-Dwass) were used to assess the effect of nutrient enrichment and light availability on chl a concentration, as normality and homoscedasticity were not meet, even after transformation of the data.
Daily total photosynthetically active radiation (PAR, 400-700 nm) ranged from 27 (day 6) to 58 (day 3) mol · m −2 . Water temperature was 24°C (±0.21) during the experimental period. Average ambient TN and TP in Lake Lacawac, as measured prior to experimentation, were 192 (±9) and 13 (±1) μg · L −1 , respectively, with an N:P ratio of 15. After the 7-d experimental period, control bags that did not receive nutrient additions had N:P ratios of 17, and experimental bags that received single-nutrient additions were limited by the opposing macronutrient (N:P +N = 57; N:P +P = 7). Those that received both nitrogen and phosphorus (+NP) had an N:P ratio of 18. There was a significant effect of nutrient enrichment (χ 2 = 19.3, df = 3, p = 0.0002), but not light availability (χ 2 = 2.5, df = 1, p = 0.11) on chl a. At the end of the incubation period, chl a values in experimental bags ranged from 2.9 to 15.3 μg · L −1 ; highest and lowest chl a concentrations occurred after addition of +NP in shaded bags and after addition of +P in bags fully exposed to solar irradiation, respectively (Figure 1).
Initial MNAN abundance measured prior to experimental manipulation was 381 (±35) cells · mL −1 , representing 43% of the PNAN community. MNAN abundance increased or was similar relative to ambient conditions in all treatments, with final abundances in experimental treatments ranging from 412 to 1044 cells · mL −1 (Figure 2), representing an average of 48% of the PNAN community. Greatest MNAN abundance was observed in the treatment exposed to full solar radiation with co-nutrient enrichment (+NP). Relative to the appropriate controls, MNAN abundance as a percentage of the total PNAN community increased F I G U R E 1 Chlorophyll a concentrations (μg · L −1 ) in experimental bags measured after experimental period. Control refers to bags that did not receive nutrient enrichment. Letters above nutrient treatments represent results of post-hoc Steel Dwass test.
with P-addition/shaded conditions (58%) and + NP/ full exposure to solar radiation (53%). In the control (no nutrient enrichment) and + N treatments, MNAN abundances were similar between shaded and exposed bags (Table 1). However, under +P and +NP enrichment, significant, yet opposing, differences in MNAN abundance were observed between light treatments.
Hourly ingestion rate by MNAN in Lake Lacawac just prior to beginning the incubation was 15 (±5.1 SE) bacteria · cell −1 · h −1 . After the 7-d experimental period, MNAN ingestion rates ranged from 4 to 25 bacteria · cell −1 · h −1 (Figure 3). Maximum MNAN ingestion rates were recorded in the treatments that received either N or P enrichment with concurrent shading. Similar to MNAN abundance, there was an effect of nutrients, but not light, on bacteria grazing rates (Table 2), and the interaction between light and nutrients was significant. Whereas under shaded conditions, grazing rates were highest after N or P enrichment, in full exposure to solar radiation there were no significant differences in MNAN grazing across nutrient treatments, including the control.
This short-term microcosm experiment provides in situ evidence that an interaction between nutrient enrichment and light availability influences abundance and bacterivory by mixotrophic nanoflagellates. Differences between these response variables emphasize the need to directly measure grazing rates in conjunction with abundance of mixotrophs. Peak bacterivory by mixotrophic nanoflagellates was observed under N-and P-limitation with shading, supporting previous work demonstrating that grazing may be used to account for reductions of photosynthetic efficiency in a suboptimal light environment. In this scenario, grazing could overcome limitation by dissolved N or P. However, the experimental addition of N and P together (+NP) in full solar radiation led to the greatest mixotroph abundance but not the greatest ingestion rates, which suggests that photosynthetic, not phagotrophic, growth likely supported the increase in MNAN abundance observed with co-nutrient enrichment. As evidenced by an increase in chlorophyll, the phytoplankton community in general also responded to co-nutrient enrichment, and in the treatment and under co-nutrient enrichment in full solar radiation, MNAN represented 53% of the photosynthetic nanoflagellate community.
MNAN abundance was nominally elevated in the +P shaded treatment and ingestion rate in this treatment was significantly greater than that of other shaded treatments, except for +N. Under P enrichment with concurrent shading, it is possible that MNAN switched from photosynthetic to heterotrophic growth where ingestion could satisfy nitrogen limitation and potentially supplement carbon from photosynthetic growth, a response noted for some mixotrophs (Carpenter et al., 2018;Johnson, 2015). An alternate explanation for the increased abundance in this treatment (+P, shaded) is that placement of the microcosm bags just below the water surface exposed the plankton in the exposed bags to inhibiting levels of solar radiation. Since nitrogen limitation, which is to be expected in the phosphorus alone supplementation, is known to increase photoinhibition in phytoplankton (Berges et al., 1996;Li et al., 2021;Salomon et al., 2013), this could also explain why MNAN abundance and grazing were reduced in the full sunlight +P treatment. In this scenario, photoinhibition would be reduced in the shaded +P treatment, and particulate N, in the form of ingested bacteria, could overcome the limitation by dissolved N. The additional dissolved nitrogen in the +NP treatment may have reduced the potential for photoinhibition noted in the full sunlight +P treatment, allowing for faster growth, even with reduced ingestion, by the MNAN in the +NP treatment. We also cannot disregard that there was also likely species-specific variability in MNAN response to light/nutrient conditions that would have been captured by an analysis of the community F I G U R E 2 Abundance (cells · mL −1 ) of MNAN within experimental microcosms. Dotted line indicates ambient abundance of MNAN as measured prior to experimental manipulation. Letters indicate results of LSD test on square-roottransformed data. Bacterivory by phagotrophic protists is hypothesized to be critical for transfer of energy and organic matter through aquatic food webs, and the major impact that mixotrophs can have on bacteria make them an important part of this narrative (Gerea et al., 2019;Urabe et al., 2000). However, bacterivory as a response to environmental conditions is not universal, and variability exists along the gradient of nutritional strategies (Wilken et al., 2020). Differing reliance on photosynthesis versus phagotrophy are likely to dictate the biochemical composition of a mixotroph and, therefore, its importance as a food source. Regardless, recognition of the widespread occurrence of mixotrophy is relatively recent and only beginning to be included in large-scale modeling efforts (e.g., Edwards, 2019;Leles et al., 2018), and work that explores abiotic drivers of bacterivory in situ can be helpful to parameterize such models. This study demonstrates some of the complexity necessary to evaluate mixotrophy and provides further insight into the impact of light/nutrients on MNAN dynamics. Without any indication of bacterivory, mixotroph abundance would suggest that they had a much larger impact on the bacterial community in higher light with a nutrient-replete (+NP) environment than under the nitrogen or phosphorus limitation (+N, +P) and reduced light. However, the combined abundance and feeding data suggest a more similar, or potentially an even greater, impact of mixotroph bacterivory in the +P treatment, especially under PAR limitation. Assessment of grazing rates along with abundances are imperative to fully characterize mixotroph dynamics and their role in the food web.

ACK NOWLEDG M ENT S
Field support provided by Nicole Berry and Caleigh Wildenstein. Dr. Bruce Hargreaves maintains environmental monitoring network at Lacawac Sanctuary Field Station, which provided PAR data. This work was supported by Fulbright funding from The Ministerio de Educación de la Nación Argentina to MSV (Supervisor: CEW). This is contribution no. 199 of Estacion de Fotobiología Playa Unión.