Evolutionary change in metabolic rate of Daphnia pulicaria following invasion by the predator Bythotrephes longimanus

Abstract Metabolic rate is a trait that may evolve in response to the direct and indirect effects of predator‐induced mortality. Predators may indirectly alter selection by lowering prey densities and increasing resource availability or by intensifying resource limitation through changes in prey behavior (e.g., use of less productive areas). In the current study, we quantify the evolution of metabolic rate in the zooplankton Daphnia pulicaria following an invasive event by the predator Bythotrephes longimanus in Lake Mendota, Wisconsin, US. This invasion has been shown to dramatically impact D. pulicaria, causing a ~60% decline in their biomass. Using a resurrection ecology approach, we compared the metabolic rate of D. pulicaria clones originating prior to the Bythotrephes invasion with that of clones having evolved in the presence of Bythotrephes. We observed a 7.4% reduction in metabolic rate among post‐invasive clones compared to pre‐invasive clones and discuss the potential roles of direct and indirect selection in driving this change.

Specifically, the invasive predators may reduce their prey abundance to such an extent that the prey is no longer resource limited (Walsh et al., 2016;White et al., 2006). An additional indirect effect of predators is that they can influence prey behavior, which may, in turn, further modify energy acquisition. For example, the presence of predators may cause prey to choose low-quality habitats resulting in increased (rather than decreased) resource limitation under increased predator abundance (Lima & Dill, 1990). However, it remains largely unknown how invasive predators influence phenotypic evolution through such indirect effects, adding a hitherto unconsidered dimension to the nexus of evolutionary responses caused by invasive predators.
Given this potential for predators to not only exert direct selection, but to also alter the resource availability for the prey, traits that are involved in the conversion of energy may be of particular interest in this context. Metabolic rate (MR) is one such trait. It is the measure of rate for energy-producing enzymatic reactions in a living being and thus has been connected to the rate of all biological activities (Brown et al., 2004). MR has been linked with key life-history traits, such as growth, reproductive output, and survival (Pettersen et al., 2018).
Furthermore, the broad-sense evolvability in MR is similar in magnitude to other physiological and life-history traits, pointing toward its evolutionary potential (Einum et al., 2019;Hansen et al., 2011).
Mechanistically, a relationship between resource availability and MR can be hypothesized in terms of the "increased-intake hypothesis" (Nilsson, 2002). A high resource availability allows a high intake, but with associated metabolic costs of maintaining high capacity in organs required for processing food. Conversely, if resource availability is low, paying this cost will not be beneficial and selection may favor reduced metabolic investment into such organs, and hence reduced MR. Indeed, studies have suggested that a relatively low MR might be favored in resource-limited environments (reviewed by Burton et al., 2011). More recently, a study comparing guppy populations experiencing different predatory pressures shows a higher MR in guppies originating from high predation environment (Auer et al., 2018). However, in another comparative study of guppy populations, Handelsman et al. (2013) reported the opposite pattern. Finally, theoretical work indicates that the relationship between resource availability and optimal MR may be complex and that the optimal MR will be relatively higher at intermediate resource availability (Einum, 2014). Collectively, these studies suggest that the response of prey MR to increased predation pressure may be challenging to predict and that more empirical studies are required.
The freshwater zooplankton Daphnia has been used as a model organism for evolutionary research due to its substantially investigated life cycle and easy rearing under laboratory conditions (Lampert, 2006;Miner et al., 2012). In addition, Daphnia's ecology is well studied, as it is a key link in freshwater food webs, prey for fish, and other zooplanktivorous organisms, and highly efficient predator for phytoplankton and bacteria. It reproduces by parthenogenesis (i.e., clonally) under favorable conditions. This makes it a strong candidate for ecological and evolutionary studies involving changes in trophic control and related consequences (Stollewerk, 2010;Strecker & Arnott, 2008). Furthermore, Daphnia has been proven to be an ideal organism to use in so-called "resurrection ecology" experiments. Under certain ecological conditions, Daphnia produces resting eggs as a result of sexual reproduction (also referred to as ephippia), some of which sink to the bottom of the water body and become embedded in the sediment where they can remain viable for years thereafter. The resting eggs within the ephippia can be hatched in the laboratory, and the resulting genotypes (each resting egg gives rise to a genetically unique individual, hereafter referred to as clones) kept in a common environment thus provide the means to observe phenotypic evolution among genotypes from vastly different time periods (Frisch et al., 2014;Landy et al., 2020). Robison et al. (2018) reported predator-induced phenotypic plasticity in MR in Daphnia, but beyond demonstration that it is a heritable trait (Einum et al., 2019), there is a lack of studies regarding the potential for evolutionary responses in MR. In the current study, we explore how MR evolves in Daphnia pulicaria from Lake Mendota, Wisconsin, USA. This lake has been invaded by the zooplanktivorous Bythotrephes longimanus, a zooplankton species native to Northern Europe (Hasnain & Arnott, 2019;Strecker et al., 2006). The main ecological impact caused by this invasion became evident in 2009, when a sharp decrease in the native population of D. pulicaria was observed (Walsh et al., 2016). Although possibly present prior to 2009, no Bythotrephes impacts on Daphnia had been detected previously.
Invasion by Bythotrephes caused a rapid and substantial (~60%) decline in the biomass of D. pulicaria. This reduction in population density was accompanied by a one-meter decrease in Secchi depth, a measurement indicating a substantial increase in surface water phytoplankton abundance (Walsh et al., 2016). Since phytoplankton is a key food source for Daphnia, one might expect that this leads to an increased resource availability, and with corresponding effects on the evolution of MR. However, this is complicated by the observation that the presence of Bythothrephes results in increased use of deeper, colder, and less productive parts of lakes among their zooplankton prey (Pangle et al., 2007), which may counteract the effect of increased food abundance in the surface layers. This, in addition to any direct selective pressures the predator may pose on this trait, makes it challenging to predict if and how MR may have responded to the invasion. Using a resurrection ecology approach, we compared the MR of D. pulicaria clones originating from prior to the Bythotrephes invasion with that of clones having evolved in the presence of Bythotrephes.

| Metabolic rate measurement
We used the rate of oxygen consumption, i.e., respiration rate, to estimate MR (Barry & Syal, 2013;Fuhrman et al., 1961). Measurements were conducted during January-June 2020 on animals that were fed and allowed to swim freely to approximate MR under field conditions.
Oxygen consumption was measured at 17°C using a 24-channel fluorescence-based respirometry system (SDR SensorDish Reader, PreSens) placed in an incubator (Memmert Peltier-cooled incubator IPP). The oxygen consumption rate was measured using a 24-well microplate (capacity 200 µl, Loligo Systems) equipped with planar oxygen sensors. The wells were pre-filled with airsaturated ADaM before the transfer of individual Daphnia and were then sealed with an adhesive PCR film (Thermo Scientific). The effects of changing temperature on metabolic rate were minimized using a water bath at 17°C during the transfer of individuals into the wells and the sealing process. Oxygen concentration readings in the wells were taken every 3 min in darkness for 1.5-2 h using SDR version 38 software (PreSens). After the oxygen consumption measurement, all individuals were photographed under a stereomicroscope (Leica Microsystems GmbH) and measured digitally for body length (mm), from the tip of head crest to the point of tail spine attachment to the carapace, using IMAGEJ version 1.49v software (National Institutes of Health).
The experiment was conducted using only female individuals (Daphnia predominantly produce daughters by parthenogenesis under favorable conditions). To standardize feeding status, all individuals to be measured were kept in isolation in 50ml ADaM for 48 h prior to an experimental run, during which time they were fed twice.
A total of 19 runs with 22 individuals (one in each well) per run were conducted, resulting in 418 measurements. However, 62 of these were excluded from statistical analyses due to individuals being injured during the process, or air bubbles being observed within the wells following measurements. Furthermore, to exclude data points where estimated oxygen consumption had high uncertainty, we only included data that had a strong linear decline in oxygen content during measurements. Thus, for each individual, we extracted the r 2 values from the fitted linear regressions between time since the start of the experiment and measured oxygen concentration, and excluded individuals (n = 8) for which r 2 < .95. In each run, two "blank" wells containing only ADaM were used as controls to estimate the level of microbial respiration, and the mean of this was subtracted from the observed oxygen consumption. Based on the volume of the wells, we

| Statistics
We used linear mixed-effect modeling to model log e -transformed MR as a response to log e body length and clone type (pre-and postinvasion) as fixed effects, and with run and clone line as random effects. Four different models with different fixed terms but the same random effects were fitted. The first model included the additive effects of body length and clone type. The second model contained both body length and clone type as well as their interaction. The third and fourth models used only body length and clone type as fixed effects, respectively.
Models were compared using Akaike Information Criterion (AIC) after fitting with maximum likelihood, while parameter estimates of the two best models were obtained by refitting the models with restricted maximum likelihood. The two models with the highest support were inspected for linearity, homoscedasticity, and normal distribution of residuals. All statistical analyses were done using R v. 4.0.2 (R Core Team, 2020), and linear mixed-effect models were fitted using the lmer function in the lme4 package (Bates et al., 2015).

| RE SULTS
The two models that contained effects of both body length and clone type received considerably stronger support than those containing only one of these terms (Table 1). Support for these two top models was similar in magnitude. For the model with only additive effects, MR increased with body length, and post-invasion clones had a lower MR than pre-invasion clones (Figure 1a). For the model that included the interaction, the interaction term was relatively weak, with the 95% CI including zero (Table 2), and the predicted MR for post-invasion clones was lower than that for the preinvasion clones across the whole range of body lengths (Figure 1b).
For the model with only additive effects, the predicted MR of an individual of the mean body length (2.2 mm) was 9.01 × 10 −5 mg/h and 8.34 × 10 −5 mg/h for pre-and post-invasive clones, respectively.
This corresponds to a reduction of 7.4% in MR of the post-invasive clones compared to the pre-invasive clones.

| DISCUSS ION
In this study, we tested if the invasion of a non-native predator is associated with an evolutionary change in the metabolic rate (MR) of its prey species. We used Daphnia pulicaria clones from Lake Mendota, USA, as the study prey species since this lake has been documented to have an established population of the non-native predator Bythotrephes longimanus that have caused large cascading effects on the pelagic ecosystem of this lake (Walsh et al., 2016).
A resurrection ecology approach enabled us to rear pre-and postinvasion clones in the lab and to compare their MR under highly controlled experimental conditions. Our results demonstrate an evolutionary change in MR among post-invasive clones, supporting the view that the predation regime influences the optimality of this Note: Measurement run and clone line were included as random intercept in all models.
Abbreviations: K, number of parameters (K-2) in model; ΔAIC C , difference in AIC C values between the given model and best-fitting model among the candidates; w i , AIC C weight; acc w i , sum of AIC C weights.
TA B L E 1 AICc comparisons of candidate models explaining variation in metabolic rate in Daphnia pulicaria as a function of body length and clone type (pre-and post-invasion by the predatory Bythotrephes longimanus)  (Pangle & Peacor, 2009;Strecker & Arnott, 2008). Under DVM, individuals move to deeper waters at dawn and return toward the surface at night, and this behavior has a genetic component (Hasnain & Arnott, 2019;Rozenberg et al., 2015). One might speculate that this behavior is genetically correlated with MR. However, if this is the case, it may seem most likely that these traits would be positively correlated. Selection for DVM in post-invasive clones would then be expected to result in a correlated increase rather than a decrease in MR. Direct selection by the predator may therefore be less likely to drive the observed evolutionary decline in MR.

F I G U R E 1 Relationship between log metabolic rate (MR) and log body length for
An alternative mechanism behind the observed evolutionary change is that even if the invasive predator has caused an increase in productivity of surface waters, the Daphnia do not benefit from this due to the shift in their vertical distribution associated with DVM.
Among genotypes that undertake such habitat shifts and successfully avoid predation, a new selective regime will be experienced, as they spend considerable amounts of time in the deeper water layers which receive less light and are less productive in terms of phytoplankton abundance. Under such an "ecology of fear" scenario (Zanette & Clinchy, 2019), the novel predator may in fact not cause a release of food restriction. Rather, the behavioral escape by the prey may create a stronger food limitation which may in turn select for a reduced MR (Burton et al., 2011). However, even this mechanism may not be straightforward to predict. First, the deeper and less productive layers of lakes are also colder, which reduces the MR (Gillooly et al., 2001) and hence the metabolic demand for resources. Second, the Daphnia may compensate for the reduced food intake during daytime by increasing their food intake during night when they move up into the productive surface layers (but see Lampert, 1989 for a critical review of such compensation in Daphnia). To evaluate the likelihood of such a mechanism, further efforts should be made to understand how Bythotrephes shape the seasonal dynamics of energetic constraints in D. pulicaria, for example, through studies of size-specific fecundity (Lampert, 1978) or lipid contents (Tessier & Goulden, 1982).
Although we focus on the potential role of changing predatory patterns, Lake Mendota has for more than a century been heavily influenced by human activity, including wastewater discharge and agricultural and urban runoff leading to pronounced eutrophication (Lathrop, 2007). Thus, we cannot rule out that changes in the selective regimes associated with changing environmental conditions aside from the change in predation level can have caused the observed evolutionary changes. However, these other influences have a long history that predates the time of origin for the clones used in the current study , and thus it is not clear that these would result in a consistent temporal change in selection regimes.
Data on nutrient inputs and concentrations in this lake have been highly variable among years without showing clear temporal trends at least since the 1970s (Lathrop & Carpenter, 2014).
The majority of previous studies on the role of predators in shaping MR of their prey have focused on phenotypically plastic responses. These studies have shown that the extent and direction of change in metabolic rate in response to predators are highly species-specific, and even context specific. For example, tadpoles may increase their MR during short-term exposure to a predator, whereas the long-term response is a reduction in MR, presumably to compensate for the behavioral reduction in feeding activity (Barry & Syal, 2013;Steiner & Buskirk, 2009 Note: "Additive" denotes the best model, whereas "Interaction" denotes the second-best model in Table 1.
TA B L E 2 Summary of the best-fitting linear mixed-effect model estimating the effects of body length and clone type (pre-and post-invasion by the predatory Bythotrephes longimanus) on metabolic rate in D pulicaria some studies (Beckerman et al., 2007), but not in others (Pauwels et al., 2010;Stibor & Machacek, 1998). Lobsters decrease their MR when exposed to predators (Briceño et al., 2018). If such variable plastic responses are adaptive, it is perhaps not surprising that similar differences exist in terms of evolutionary responses across species and systems. Strikingly, even separate populations within a single species (guppies) show opposite evolutionary trajectories of MR in response to a change in predation regime (Handelsman et al., 2013vs. Auer et al., 2018. The current study adds to the limited knowledge on evolutionary responses of MR to a change in predator exposure. We tentatively conclude that the response observed in the present study is not a response to selective predation. Rather, it may be an outcome of a change in the resource limitation. However, we emphasize that there is a need for more studies across populations and species, in combination with controlled experimental studies, to obtain a more mechanistic and better predictive understanding.

ACK N OWLED G M ENTS
This work recieved funding from the National Science Foundation (IOS 1651613) and the Research Council of Norway (223257/F50 and 244046). We thank Line E. Sundt-Hansen for comments on an earlier draft.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to draw graphs and tables in the study is available on