The combined effects of the aquatic herbicide fluridone and the lampricide TFM on non-target fish and invertebrates

ABSTRACT Invasive sea lamprey (Petromyzon marinus) and exotic aquatic plants like Eurasian watermilfoil (Myriophyllum spicatum) pose environmental and economic threats to aquatic ecosystems, sometimes within the same water bodies. The lampricide TFM (3-trifluoromethyl-4-nitrophenol) and the aquatic herbicide fluridone [1-methyl-3-phenyl-5-(3-trifluoromethylphenyl)-4(1 H)-pyridinone] are routinely used to control these invasives in the Great Lakes and other water bodies, but the combined toxicity of these two pesticides for non-target aquatic animals has not been reported. Here, we examine the acute toxicity of TFM in combination with fluridone for three non-target species: fathead minnows (Pimephales promelas), brown trout (Salmo trutta), and the freshwater microcrustacean Daphnia magna. The average toxicity of TFM was higher in the presence of fluridone for all three species, although this difference was not statistically significant at all concentrations and time points. Lethal times to 50% were lower for both fish species in the presence of fluridone at the TFM concentrations used in treatments. Brown trout, upon which Maximum Allowable Concentrations of TFM are based, was the least sensitive of the test species to TFM at all time points and concentrations. Although our data indicate that fluridone does increase the toxicity of TFM when close to the maximum concentration recommended by the EPA (12ppm), this increase in toxicity is likely to be within currently accepted limits of non-target mortality in most field application targets (typically <7 ppm). Given the potential presence of other contaminants, however, on-site toxicity tests should still be used whenever possible to evaluate acute mortality of non-target species.


Introduction
Pesticides and herbicides are routinely applied to inland aquatic water bodies to control invasive plants and animals.In general, risk assessments of these chemicals for non-target species, as well as determinations of their toxicity thresholds, are done on a chemical-by-chemical basis.In reality, however, freshwater animals are likely to be exposed to multiple chemicals simultaneously (Chen et al., 2014), and the hazards of these toxins can be greater in combination than in isolation (Cedergreen & Nazir, 2014;Junghans et al., 2006;Walter et al., 2002;Wobeser, 2006): a given chemical below its toxicity threshold concentration can exceed acceptable levels of non-target species mortality when applied in combination with other chemicals (Walter et al., 2002).Although an increasing number of studies have begun to investigate the potential adverse effects of chemical 'cocktails' on non-target wildlife, the joint effects of many contaminants are poorly understood and remain a major challenge for environmental scientists and managers (Chen et al., 2015).
The lampricide TFM (3-trifluoromethyl-4-nitrophenol) is an aquatic pesticide widely used for the control of the sea lamprey (Petromyzon marinus), a species with landlocked populations in the Great Lakes, the Finger Lakes, and Lake Champlain (Wilkie et al., 2019).Sea lamprey are an invasive species in the Great Lakes: it is hypothesized that sea lamprey made their way into Lake Ontario in 1800s through the Erie Canal, and they were first detected in Lake Erie in 1921 and in the upper Great Lakes in the 1930s (Eshenroder, 2014;Gaden et al., 2021;Smith & Tibbles, 1980).Sea lamprey in these landlocked populations parasitize a broad range of hosts and have caused substantial ecological and economic damage, contributing to the collapse of major fish stocks (Marsden & Siefkes, 2019;McDonald & Kolar, 2007).Sea lamprey also have major adverse effects on recreational fisheries in the Finger Lakes and Lake Champlain, although there is debate about their endemicity there (Eshenroder, 2014;Marsden & Siefkes, 2019; U.S. Fish & Wildlife Service; Barber & Steeves, 2004).After screening more than 6000 compounds, TFM was selected for chemical control of lamprey because of its relatively high specificity for sea lamprey larvae and its short duration in the environment, and because it does not bioaccumulate (Hubert, 2003).TFM is applied to tributaries infested with larval sea lamprey on a regular basis as part of lamprey control programs for landlocked populations (Marsden & Siefkes, 2019;McDonald & Kolar, 2007;Sullivan et al., 2021), and its use (e.g.quantity and number of tributaties treated) has increased over the past 20 years (Sullivan et al., 2021).
TFM has been subject to continuing scrutiny by federal, state, and tribal agencies because of concern about its potential lethal and sub-lethal effects on nontarget organisms, including species of conservation concern.Although many non-target species appear to tolerate TFM at the concentrations likely to be experienced in streams (Birceanu et al., 2014;Boogaard et al., 2003Boogaard et al., , 2012;;Middaugh et al., 2014), substantial variation in sensitivity occurs among species and among life-stages within species (Wilkie et al., 2019).Lake sturgeon (Acipenser fulvescens), stonecats (Noturus flavus), mudpuppies (Necturus maculosus), members of the catfish family (Ictaluridae), unionid mussels, and logperch (Percina caprodes) are of particular concern because of their high sensitivity to the lampricide, their ecological role, and their conservation status in some areas (Marsden & Siefkes, 2019;McDonald & Kolar, 2007;Wilkie et al., 2019).Indeed, the lake sturgeon, which is listed as threatened in the Great Lakes basin, is more sensitive to TFM than sea lamprey at some of its life stages (Boogaard et al., 2003).Moreover, adverse effects of TFM may be greater when it is encountered in combination with other toxins: TFM has documented additive and synergistic effects on non-target species when combined with select pesticides, heavy metals, and industrial pollutants (Marking & Bills, 1985).
Other aquatic invasive species, such as Eurasian watermilfoil (Myriophyllum spicatum), an aggressive aquatic plant, co-occur with sea lamprey in its landlocked populations (e.g. the Great Lakes, the Finger Lakes, and Lake Champlain), creating the potential for overlapping chemical control strategies.Eurasian watermilfoil and other aquatic plants (e.g.Hydrilla verticillata) are widely and routinely controlled with the herbicide fluridone (Menninger, 2011;Wagner et al., 2007).At environmentally relevant concentrations, acute toxicity studies indicate that fluridone [1-methyl-3-phenyl-5-(3-trifluoromethylphenyl) -4(1 H)-pyridinone; the active ingredient in SePRO Sonar® Genesis] has low toxicity for most fish and invertebrates (Archambault et al., 2015;Paul et al., 1994).However, it has been linked to elevated mortality and reduced abundance and diversity in water mites (Arrenurus: Megaluracarus; Yi et al., 2011) and sublethal effects (e.g.endocrine disruption) in delta smelt (Hypomesus tranpacificus; Jin et al., 2018) and fathead minnows (Pimephales promelas; Cozzola et al., 2022).Moreover, slow-acting fluridone is typically applied for one to 4 months to effectively control invasive plants (Archambault et al., 2015), sometimes at the whole-lake scale (Wagner et al., 2007) or in specific areas that overlap with TFM treatment (Cornell Cooperative Extension, 2020), prolonging exposure of non-target organisms to fluridone and increasing the likelihood of its co-application with TFM in infested tributaries and inlets.To date, however, the combined effects of fluridone and TFM have not been assessed.
The objective of this study was to evaluate the combined toxicity of TFM and fluridone on three representative species: fathead minnows, brown trout (Salmo trutta), and the freshwater microcrustacean Daphnia magna.Specifically, we tested the hypothesis that fluridone would increase the toxicity of TFM within its recommended application concentration limits.Without such testing, the combined effects of TFM and fluridone are difficult to predict because, like a majority of chemicals (Walter et al., 2002), their modes of action (particularly that of fluridone) are not well understood for non-target species.The lampricide TFM, a halogenated nitrophenol, uncouples mitochondrial oxidative phosphorylation, disrupting the energy metabolism of lamprey larvae and some non-target fishes, although sensitivity varies greatly among species (Wilkie et al., 2019).The herbicide fluridone is a carotenoid synthesis inhibitor in plants (Archambault et al., 2015), but its mode of action in animals has not been determined.Both TFM and fluridone are also potential endocrine disruptors (Hubert, 2003;Jin et al., 2018;Wilkie et al., 2019), although the effects of TFM may be transitory in field applications (Hewitt et al., 1998) and its effect on the reproductive success of non-target species is unclear (Wilkie et al., 2019).Assessing the combined effects of these chemicals will apprise environmental agencies about the potential hazards of simultaneous application across multiple levels of the food chain, informing efforts to minimize non-target impacts in the land-locked sea lamprey populations (Marsden & Siefkes, 2019).

Laboratory assays
Toxicity tests were carried out by the New York State Department of Environmental Conservation (NYSDEC) Aquatic Toxicant Research Unit (Rome, NY) or the Biology Department at Hamilton College (Clinton, NY) and adhered to toxicity test protocols of the United States Environmental Protection Agency (2002).New York State Rome Lab strain domestic brown trout (7.4 ± 2.0 g; 89 ± 7.5 mm) were obtained from the DEC Rome Fish Hatchery and were 7-months old at the time of testing.Juvenile fathead minnows were shipped from Aquatic Biosystems, Inc. (Fort Collins, CO), acclimated for 1 week, and were 35 days old at the start of testing (composite mean weight 0.04 g; 17 ± 1 mm).Daphnia magna neonates were obtained from Aquatic Biosystems, Inc. and were <48-hr old at the start of testing.Animals were not fed during the toxicity tests.They were placed in test beakers immediately following the addition of pesticides to the containers, using spring water from the DEC Rome Fish Hatchery (pH = 8.10, hardness = 132 mg/L CaCO3; dissolved oxygen levels >8.6 mg/L; alkalinity = 114 mg/L CaCO3).The hardness and pH of the water used in the toxicity tests were measured before each test series.Dissolved oxygen (DO) was measured at each fish count interval.The test containers were only aerated for the final 24 hr in the brown trout toxicity tests.No aeration was required in fathead minnow toxicity tests.The DO never dropped below 60% saturation in any of the test containers.
Experimental conditions varied between test species as appropriate for their size and species.For brown trout toxicity tests, 20-L glass containers were filled with 16 L of test solution and held at 13 ± 1°C.For fathead minnow toxicity tests, 2-L glass containers were filled with 1.5 L of test solution and held at 22 ± 1°C.For Daphnia magna toxicity tests, 250-mL beakers were filled with 150 mL of test solution and held at 20 ± 2℃.
Field-grade TFM-HP (High Purity) Sea Lamprey Larvicide (Iofina, Inc.; Batch Number LAM090629A; 36.8%TFM) was obtained from the NYSDEC (Region 5 Bureau of Fisheries in Ray Brook, New York) in conjunction with the Lake Champlain Fish and Wildlife Conservation Office (Essex Junction, Vermont).Preliminary trials were conducted to identify an informative range of target TFM concentrations for each species (e.g.mortality <100% and >0%).Target concentrations (mg/L) of TFM and number of replicates for each species are given in Table 1.Each concentration of TFM was tested with and without 50 μg/L of aquatic herbicide SePRO Sonar® Genesis (active ingredient fluridone; SePRO Corporation).This concentration is within the application concentration rate recommended by SePRO Sonar® Genesis label (30-90 μg/L for single application in lakes, with a maximum of 150 μg/L per annual growth cycle).All herbicides and stock solutions were measured and added using a certified digital pipette (precision = 0.1 µL).
Mortality of test animals was determined by visual inspection and gentle probing at 3, 6, 12, 24, and 48 hr after pesticide application.Lethal concentrations to 50% (LC50) were calculated using the trimmed Spearman-Karber method for each species (Finney, 1978;Hamilton et al., 1977).We calculated the additional lethal concentrations (LC5, LC10, and LC25) using the Probit method.We compared these using the method of Sprague and Fogels (1977) to determine statistical differences.We performed a Dunnett's test using an arcsine square root transformation to compare toxicity of TFM with and without fluridone at all timepoints for each species (Hamilton et al. 1977).

Results
Toxicity of TFM was higher in the presence of fluridone for all three species, time periods, and concentrations, although the difference was only statistically significant at LC50 concentrations (Table 2).The LC50 values indicated that mortality of brown trout was significantly elevated by the addition of fluridone at all time intervals.In contrast, fluridone only elevated mortality significantly for fathead minnows and Daphnia at LC50 concentrations at the early time points (3 hr after application for minnows; 3 and 6 hr after applications for Daphnia).
Brown trout, which is the non-target species upon which the Maximum Allowable Concentrations of TFM are based (Barber & Steeves, 2004), was the least sensitive of the test species to TFM (and combinations of TFM and fluridone) at all time points and LC values (Table 2).At the 12-hr mark, which is the normal application time for TFM (Barber & Steeves, 2004), the LC50 of brown trout was significantly higher than the LC50 of Daphnia, which was significantly higher than fathead minnows (brown trout > Daphnia > fathead minnows).Likewise, the LC25 of brown trout was significantly higher than the LC25 of minnows and Daphnia, and the LC5 values were significantly different among all species (brown trout > fathead minnows > Daphnia).
The addition of fluridone resulted in faster mortality of both fish species, although the effect varied between species and among TFM concentrations (Table 3).For brown trout, lethal times to 50% (TL50) were lower in the presence of fluridone at all TFM concentrations.For fathead minnows, however, the TL50 in the presence of fluridone was only lower at the lowest TFM concentration; TL50 values were similar with and without fluridone at higher concentrations.

Discussion
Since European settlement, the Laurentian Great Lakes have been enormously impacted by the invasion of exotic species (Mills et al., 1993), which have contributed to declines in native biodiversity, changes in food webs, altered nutrient cycling, and the collapse of major fish stocks (Marsden & Siefkes, 2019;Pagnucco et al., 2015).Attempts to control invasive sea lamprey were unsuccessful prior to the implementation of a chemical control program that includes TFM (Barber & Steeves, 2004).At typical levels of acute toxicity, population models suggest that the benefits of reducing predation by sea lamprey outweigh TFM-induced non-target mortality even for the highly sensitive lake sturgeon (Dobiesz et al., 2018).However, simultaneous application of pesticides, like the herbicide fluridone and lampricide TFM, could elevate mortality of non-target organisms above predicted levels.For example, TFM in 'triple combination' with the pesticides malathion and Delnav (the latter of which is no longer sold in the United States) had toxicity levels that were magnified 7.9× above the toxicity expected from the individual chemicals (Marking & Bills, 1985).Our data indicate that the average toxicity of TFM was greater in the presence of fluridone for all three species tested (fathead minnows, brown trout, and the invertebrate Daphnia magna) at all concentrations and time intervals after  application, although the difference was not always statistically significant.Moreover, mortality after TFM application occurred more rapidly in the presence of fluridone for both fish species.
Although our data indicate that fluridone does increase the toxicity of TFM, this increase appears to be generally within accepted limits for non-target mortality.The Maximum Allowable Concentration (MAC) of TFM is based on the LC25 (the concentration that produces 25% mortality) of brown trout after a 9-hr exposure (Barber & Steeves, 2004).Our study indicates that TFM -in combination with fluridone -could exceed this MAC if TFM is applied at the maximum concentration recommended by the EPA (12 ppm; EPA 1999): the LC25 values of TFM for all non-target species after a 12-hr exposure (the normal application time for TMF, although application times can be extended to 24 hr; Barber & Steeves, 2004) were at or below 12 ppm, and these LC25 concentrations were consistently lower in the presence of fluridone.In practice, however, typical target concentrations are generally much lower than this 12 ppm maximum (typically 1 to 6 ppm; personal communication, Dorance Brege, U.S. Fish and Wildlife Service Treatment Supervisor Gaden et al., 2021).Concentrations of TFM within this range are below the estimated LC25 values for all of the three species that we tested after a 12-hr exposure, even in the presence of fluridone.
Minimum lethal concentrations (MLC) of TFM are those that produce at least 99.9% mortality of sea lamprey larvae after a 9-hr exposure (Barber & Steeves, 2004).These minimum lethal concentrations vary with water pH and alkalinity: as pH and alkalinity increase, mortality of both sea lamprey and non-target species declines at a given concentration of TFM, and higher concentrations of TFM are required to achieve the MLC for lamprey (Bills et al., 2003).Based on the chemistry of the stream water used in our study, for example (pH = 8.10, hardness = 132 mg/L CaCO3), the MLC was approximately 3.1 to 3.2 ppm TFM.Stream application concentrations are generally set at 1.0 to 2.0 times the MLC (usually 1.5 MLC; in this case, 4.8 ppm TFM), but not exceeding the MAC (Barber & Steeves, 2004;Marsden & Siefkes, 2019).A target concentration of 4.8 ppm would be well below our estimated MAC: even in the presence of fluridone, this concentration is below our LC5 for brown trout and fathead minnows and below the LC10 for Daphnia magna after a 12-hr exposure.Moreover, MACs are estimated based on LC25 values after a 9-hr exposure, whereas we recorded LC25s at 12 hr; therefore, the difference between the MAC and the target MLCs in field applications is likely to be greater than what we present here.
Our data indicate that simultaneous applications of fluridone and TFM are unlikely to increase acute mortality of non-target species to unaccepted levels at typical field target concentrations (≤7 ppm).We note, however, we did not assess potential sublethal effects of exposure to TFM and fluridone, both of which are potential endocrine disruptors (Hubert, 2003;Jin et al., 2018;Wilkie et al., 2019) that may also affect behaviors such as prey capture ability and locomotion (Cozzola et al., 2022).These sublethal effects are essential to evaluate in aquatic species (Villarroel et al., 2003).We also note the brown trout -the species upon which MAC prediction charts are based -were the least sensitive of the test species to TFM and fluridone.These data align with a comparison of 15 non-target species, indicating that brown trout and other salmonids are relatively tolerant to TFM in comparison with other nontarget species (Boogaard et al., 2003).Therefore, levels of mortality among many non-target species, including Daphnia magna, a keystone species in most freshwater habitats (Altshuler et al., 2011), and lake sturgeon, a species of conservation concern known to be highly sensitive to TFM (Boogaard et al., 2003;Wilkie et al., 2019), are likely to be higher than expected based on MAC prediction values derived from brown trout.For example, another study (performed at similar levels of alkalinity and pH; Boogaard et al., 2003) reported that the 12-hr LC50s of lake sturgeon ranged from 1.47 to 4.5 mg/L of TFM, depending on age class (1.10-2.85 when combined with 1% niclosamide), and is nearly an order of magnitude lower than the 12-hr LC50 of brown trout in our study (14.0 mg/L of TFM) for juvenile age classes of sturgeon.Acute toxicity of TFM in combination with fluridone for lake sturgeon is unknown, but -based on our study -is likely to be elevated, and should be considered in areas where lake sturgeon are present.Managers should consider staggering the timing of TFM and fluridone applications to minimize their overlap, as has been implemented in some programs (Cornell Cooperative Extension, 2020).
Finally, our combined risk assessment does not account for the presence of other contaminants, with which TFM and fluridone could have additive, synergistic, or antagonistic effects on non-target mortality (Marking & Bills, 1985;Panizzi et al., 2017).Large water bodies in populous areas can be challenged by a complex mixture of many chemical contaminants (Gewurtz et al., 2011), some of which could be lethal to non-target species when mixed with small quantities of TFM (Marking & Bills, 1985).To account for existing (often unrecognized) toxic burden in a given system, pretreatment on-site toxicity tests should be used when possible to determine TFM concentrations that do not exceed acceptable levels of acute mortality in non-target species (Barber & Steeves, 2004;Marking & Bills, 1985).

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Andrea Townsend (Associate Professor of Ecology, Hamilton College) examines the responses (behavioral and physiological) of wildlife to anthropogenic disturbances.These include urbanization, emerging infectious disease, and environmental toxins, including the effects of herbicides and pesticides on non-target species as well as the consequences of accidental chemical pollutants.
Marian Burgard examining the effects of fluridone and TFM on Daphnia magna.Eric Paul (Biologist 1 (Ecology), Aquatic Toxicant Research Unit of the New York State Department of Environmental Conservation) studies the non-target toxicity of pesticides to fish and other aquatic organisms.These studies include laboratory testing and testing in field applications in order to better understand the effects on fish and wildlife of New York State.

Table 1 .
Target concentrations (mg/L) of TFM and number of replicates (with and without 50 μg/L fluridone) for each species

Table 2 .
Lethal concentrations (LC; mg/L) of TFM for test species with and without 50 μg/L of fluridone.Asterisks indicate significantly lower LC values in the presence of fluridone (p < 0.05).Confidence intervals (95%) given in parentheses

Table 3 .
Lethal times to 50% (TL50) with TFM concentration for fish, with and without 50 μg/L of fluridone.Range given within parentheses