Mesocosm experiments validate induction of Daphnia vertical migration by the fish‐derived kairomone 5α‐cyprinol sulfate

The fish‐derived bile salt 5α‐cyprinol sulfate (CPS) has been identified as a kairomone inducing the predator avoidance behavior “diel vertical migration” (DVM) in Daphnia magna in response to fish. However, conclusions about the ecological significance of CPS have been derived from laboratory experiments only. Using a mesocosm approach, we investigate whether the role of CPS as a kairomone can be confirmed in the field. We demonstrate that CPS induces downward migration during daytime in a field‐derived Daphnia community mainly consisting of Daphnia longispina and Daphnia cucullata within the experiment. In the study lake, the actual population of D. longispina shows a similar pattern of DVM, while concomitant quantification of CPS by HPLC‐MS confirms that CPS in situ concentrations are sufficiently high for induction of daytime downward migration of D. longispina in this oligo‐mesotrophic lake. Together, these observations infer that CPS is a significant kairomone‐inducing DVM‐like behavior in Daphnia.

confirms that CPS in situ concentrations are sufficiently high for induction of daytime downward migration of D. longispina in this oligo-mesotrophic lake.Together, these observations infer that CPS is a significant kairomone-inducing DVM-like behavior in Daphnia.
Infochemicals that mediate the perception of predators by prey and which impart a benefit to the receiving organism while not being beneficial for the producer are termed kairomones (Pohnert et al. 2007).Such chemical signaling has been well-established for inducible defenses in freshwater systems, in particular within predator-prey systems by planktivorous fish and prey organisms of the genus Daphnia.In response to kairomones from fish, Daphnia spp.deploy several inducible defenses against zooplanktivores including life history changes (e.g., earlier reproduction, smaller size at first reproduction; von Elert and Stibor 2006;Hahn and von Elert 2020), morphological defensive structures (e.g., helmets; Laforsch and Tollrian 2004;Hahn and von Elert 2022), and diel vertical migration (DVM; Loose 1993a).
DVM, in response to fish, describes the daytime migration of zooplankton into the dark hypolimnion of a waterbody, which can serve as a refuge from visually oriented predators (Ringelberg et al. 1991).During nighttime, zooplankton migrate back into the epilimnion.The metabolic costs associated with DVM (Loose and Dawidowicz 1994) are compensated by reduced predation losses (Lampert 1993).DVM contributes substantially to vertical nutrient transport (Haupt et al. 2009(Haupt et al. , 2010) ) and to the biological carbon pump in marine systems (Brierley 2014), while also affecting the foraging of planktivorous fish and distribution of phytoplankton biomass (Reichwaldt and Stibor 2005;Haupt et al. 2009).
Recently, the first fish-derived kairomone that induces DVM in Daphnia magna has been identified (Hahn et al. 2019).The only other identified kairomone perceived by Daphnia reported are fatty acids conjugated to the amino group of glutamine that are released by larvae of the insect Chaoborus (Weiss et al. 2018).In both cases, kairomone identification was based on small-scale laboratory bioassays.For identification of the fish kairomone, the DVM of a D. magna clone was investigated in tubes of 1-m length in which the temperature stratification of a lake was simulated while the length of the water column and the temperature-, light-, and food regime diverged from natural conditions (von Elert and Loose 1996).Using a clone of D. magna, the DVM-inducing kairomone was found to be the bile salt 5αcyprinol sulfate (CPS), a compound exuded by various fish species (Hahn et al. 2019).D. magna mainly occurs in ponds where it may coexist with fish; however, DVM traditionally is described in lakes with smaller-sized Daphnia species (Larsson and Dodson 1993).Hence, it remains to be demonstrated that CPS also induces DVM in these smallersized Daphnia species.In addition, unfortunately, little is known about the concentration and effect of CPS in nature.
Here, we conducted a field mesocosm experiment in which we tracked the effect of purified CPS on the vertical distribution of a field-derived Daphnia community.We tested two different CPS concentrations to estimate CPS threshold concentrations for DVM induction and quantified CPS concentrations within the mesocosms over time to estimate their chemical stability.Subsequently, we measured vertical daytime and nighttime distributions of Daphnia longispina in the experimental lake and quantified concomitant CPS concentrations.

Methods
To test the effects of CPS addition on DVM, we performed a factorial mesocosm experiment during the summer of 2019 in the dimictic oligo-mesotrophic Lake Klostersee.This lake contains Daphnia cucullata and D. longispina as the major Daphnia species (Seitz 1984;Ili c et al. 2021) and has zooplanktivorous cyprinid fish (for details, see Supporting Information S1.1).In summer 2022, we determined vertical profiles of light, temperature, chlorophyll a (Chl a), as well as the daytime and nighttime vertical distribution of D. longispina and CPS.To ensure that the chlorophyll profile from 2022 is representative, such data was also collected during 2023.See Ahlers et al. (2023) for all raw data used in this study.

Mesocosm experiment
In May 2019, lake plankton (< 250 μm) was enclosed in nine 12-m long bottom-sealed mesocosms (Supporting Information Fig. S1).Thereafter, zooplankton was collected in Lake Klostersee with a plankton net (100-μm mesh) from 12 to 0 m and inserted into the mesocosms.The experiment was started 33 d after the establishment of the fish-free mesocosms to ensure degradation of any initially present fish kairomone.Then, on 11 June 2019, the mesocosms were amended with either CPS (1x or 10x concentrated) or control extract to initiate the experiment (Supporting Information Fig. S1; for details, see Supporting Information S1.2).Samples for quantification of CPS in the mesocosms were taken at 0-2 and 9-11 m 48, 72, and 120 h after CPS addition (Supporting Information Fig. S1, for details, see Supporting Information S1.2).
Four vertical layers of the mesocosms (0-3, 3-6, 6-9, and 9-12 m) were sampled with 20-cm diameter closing plankton net (type Apstein, Pokorny-Netze GmbH; 140 μm mesh size).To avoid disturbance of vertical gradients, sampling was initiated at the shallowest layer.Each of the nine mesocosms was sampled on 13 June 2019 during daytime (48 h after CPS) and nighttime (60 h after CPS) and on 14 June 2019 during daytime (72 h after CPS, see Supporting Information Fig. S1).Zooplankton samples were fixed in 4% (v : v) of formaldehyde and 4% (w : v) saccharose; a dissecting microscope was used for taxa identification and counting.Treatment effects on vertical distributions were tested by PERMANOVAs, and effects on the share of the Daphnia community in different depth layers were tested by one-way ANOVA and posthoc tests (for details, see Supporting Information S1.5).For analysis of nutritional and physical parameters, see Supporting Information S1.6, S1.7.

Vertical profiles in Lake Klostersee
On midday and midnight of 17 August 2022, 18 August 2022, and 19 August 2022, the layers 0-2, 2-4, 4-6, 6-8, 8-10, and 10-12 m were sampled in duplicate in the center of the lake, zooplankton was fixed, and D. longispina was counted as indicated above.Deeper strata were not sampled because of muddy sediments.The entire sample was counted for each duplicate before pooling data.A Cochran-Mantel-Haenszel test (software R, version 3.6.2) was performed to test for differences in the vertical distribution of D. longispina between daytime and nighttime.On the same dates samples, for CPS were taken (for details, see Supporting Information S1.8).
Vertical profiles of temperature were measured on 13 August 2015, 31 August 2016, 04 September 2019, 16 August 2022, and 12 August 2023.A vertical profile of light was measured on 17 August 2022 (see Supporting Information S1.7).R-package ggplot2 was used to fit an exponential model to the vertical profiles of light intensity and to apply geomline to connect data points of the vertical temperature profile.
On 17 August 2022, a 2-m long tube sampler was used to collect a vertical profile of Chl a.Samples of 2 L were stored at À20 C.After thawing, the samples were filtered onto GFF filters and Chl a concentrations were determined spectrophotometrically (Tilzer 1983).On 12 August 2023, a vertical profile of Chl a was obtained using a Hydrolab HL4 OTT hydroMet multiprobe.

Mesocosm study in 2019
In an earlier mesocosm study, the Daphnia community had been given several days to adapt to changes in fish abundance before vertical profiles of Daphnia were sampled (Loose 1993a,b).Accordingly, on 13 June 2019, we investigated daytime depth distributions at 48 and 72 h after kairomone addition, while the nighttime profile was measured after 60 h (Supporting Information Fig. S1).We found an effect of CPS on the daytime depth distribution of Daphnia spp.48 h after CPS addition (PERMANOVA results, R 2 = 0.56, p < 0.01; Fig. 1A; Supporting Information Table S5).In the control treatment, most Daphnia resided in the layer between 3 and 6 m of depth ($ 68% of the community; Fig. 1A).In the 1x CPS mesocosms, Daphnia was located mainly (42%) in the layer between 0 and 3 m depth (Fig. 1A), while after addition of 10x CPS the most favored layer was the one between 6 and 9 m ($ 52% of the community).Accordingly, a significantly larger proportion of the Daphnia community resided in the bottom part of the mesocosms (6-12 m depth) during daytime (Supporting Information Fig. S5A; Supporting Information Table S6).The same treatmentdependent trends of vertical daytime distribution were observed for individual species, D. cucullata and D. longispina (Supporting Information Fig. S5C,E).The effect of CPS on the vertical daytime distribution was significant for D. longispina (PERMANOVA: R 2 = 0.63, p < 0.05; Fig. 1E; Supporting Information Table S5) but not for D. cucullata (PERMANOVA: R 2 = 0.43, p < 0.1; Fig. 1C; Supporting Information Table S5).As expected for animals that perform DVM, CPS did not affect the vertical distribution of Daphnia at night (60 h; Fig. 1B,D,F; Supporting Information Fig. S5B,D,F; Supporting Information Table S7).As anticipated, in situ concentrations of CPS in the mesocosms were highest in the 10x CPS treatment and lowest in the control 48 h after its addition.Concentrations were higher in the epilimnion than in the hypolimnion and decreased over time (Table 1; Supporting Information Fig. S6).Seventy-two hours after CPS addition, there was no effect of CPS on the daytime distribution of Daphnia (Supporting Information Fig. S7): neither in terms of the vertical distribution of individuals nor the share of the population that resided in the lower half of the mesocosms (Supporting Information Tables S8, S9).
Initial nutrient concentrations in the mesocosms (Supporting Information Table S10) did not differ among treatments Depicted are the relative distributions of Daphnia populations in percent (mean AE SE, n = 3) over the depth ranges "0-3 m," "3-6 m," "6-9 m," and "9-12 m." Data are depicted for the total Daphnia community (A, B), for Daphnia cucullata (C, D), and for Daphnia longispina (E, F) in response to different CPS concentrations and a control.PERMANOVA results are obtained after 100,000 permutations, testing for the effect of the treatment on distribution patterns.(Supporting Information Table S11); likewise, temperature profiles (Fig. 2C-E) were not distinguishable 48 h after CPS addition (permanova, 1000 permutations, F model,2 = 0.743, p = 0.782).
Although light intensities seemed to diverge between treatments mainly at the water surface (Supporting Information Fig. S8), mean coefficients for light attenuation ranged from À0.149 to À0.167 m À1 across treatments and were not significantly different among treatments (Fig. 2A; one-way ANOVA, F 2 = 0.802, p = 0.49).Accordingly, the mean depth of the euphotic zone (1% I 0 ) ranged from 10.3 to 11.4 m across treatments and was not significantly different (Fig. 2B; Kruskal-Wallis one-way ANOVA on ranks, H 2 = 2.489, p = 0.339).
On 17 August 2022, the coefficient for light attenuation (À0.234 m À1 ) was within the 95% confidence interval (CI) of the respective coefficients in the mesocosms; accordingly, the depth of the euphotic zone (1% I 0 , 9.1 m) fell within the 95% CI of that in the mesocosms.The vertical profile of Chl a revealed a deep chlorophyll maximum (DCM) of 20.4 μg L À1 at 12 m during August 2022 (Fig. 3D) and at 11 m during 2023 (Supporting Information Fig. S10), suggesting that DCMs are typical for this lake.
During 2022, at the same time as the DVM profiling, a temperature profile was collected.At each depth, the 2022 temperatures fell within the 95% CI of water temperatures from in total five different years spanning from 2015 to 2022 (Supporting Information Fig. S11).

Discussion
In a small-scale bioassay setup, CPS has been shown to induce "normal" DVM in D. magna (Hahn et al. 2019).Hence, we expected similar effects in the field mesocosm experiment in 2019.Under these near-natural conditions, CPS induced a deeper daytime residence in the Daphnia community and in D. longispina.Such pronounced DVM behavior in response to fish has been demonstrated previously for D. longispina (Stich and Lampert 1981;King and Miracle 1995;Makino et al. 1996).Similarly, the limited response of D. cucullata is also consistent with previous research that fish prefer larger prey over D. cucullata (Laforsch and Tollrian 2004), and that the smaller daphnid does not undergo DVM.
In accordance with the normal DVM pattern, the effect of CPS on vertical distribution in the mesocosms during 2019 was absent during nighttime, although there was a trend that animals preferred the upper layer during nighttime.A similar pattern with a more pronounced absence of upward migration at night was also observed for D. longispina in Lake Klostersee 3 years later in 2022.This absence of a strong upward migration at night can be explained by several factors.First, food quantity and quality may not have been optimal in the epilimnion of Lake Klostersee, while upward migration of zooplankton during the night can increase development due to higher epilimnetic temperatures (Williamson et al. 2011).This factor may not override the effects of optimum food conditions.A DCM is commonly present in oligo-mesotrophic lakes in summer (Williamson et al. 2011), like in Lake Klostersee, possibly resulting in elevated food quality of DCMs relative to epilimnetic seston, as seen elsewhere (Rothhaupt 1991;Williamson et al. 1996;DeMott et al. 2004).We hypothesize that food-mediated constraints might have contributed to the absence of a clear upward migration at night in the lake.Pronounced DCMs may result from long stable stratification of mesocosms water columns similar to the situation in the lake, as suggested by common light and temperature profiles in mesocosms and lake.Second, meteorological conditions may have limited DVM at night: light not only serves as a proximate cue for DVM (Ringelberg 1999), but it also reduces the amplitude of upward migration under full moon conditions (Gliwicz 1986;Dodson 1990).Here, the nighttime sampling of zooplankton took place under a clear night sky 4 days before (mesocosm) and 5 days after (lake sampling) full moon.We infer that moonlight may have dampened the upward migration of Daphnia at night both in mesocosms during 2019 and Lake Klostersee during 2022.
Comparison among treatments of the mesocosm experiment during 2019 helped identify a threshold for CPS effects on Daphnia DVM.Forty-eight hours after CPS addition to mesocosms, there was no effect of CPS on vertical Daphnia distribution in the low concentration trials (0.4-2 nmol CPS L À1 ), whereas Daphnia migrated in the high CPS treatment (6.7-33 nmol CPS L À1 ).This pattern indicates that more than 2 nmol L À1 are required for DVM induction in this lake and that the CPS threshold concentration for induction of DVM is reached or even exceeded in the presence of 33 nmol CPS L À1 in the respective mesocosms.Furthermore, due to high rates of CPS loss from the mesocosms, it is possible that migration may have been triggered by higher CPS concentrations that had been present earlier in the experiment.Interestingly, in Lake Klostersee during 2022, the pattern for normal DVM coincided with ≥ 20 nmol CPS L À1 in the water column, values which lay within the range of CPS concentrations that altered Daphnia distribution in mesocosms.In conclusion, CPS concentrations in the lake are most likely sufficiently high to cause deeper daytime residence of D. longispina.
These evidence that the DVM of D. longispina in the lake can be explained by in situ CPS concentrations was obtained in 2022.The fact that stratification regimes are consistent from 2015 to 2023 shows that the stratification in 2022 was typical for Lake Klostersee and suggests that the finding that DVM in the lake coincides with sufficiently high CPS concentrations is a representative result.
DVM has been induced in a D. magna clone by 100 pmol CPS L À1 under laboratory conditions (Hahn et al. 2019), whereas 2 nmol CPS L À1 did not affect the daytime position of Daphnia spp. in mesocosms.The higher threshold in the mesocosms may result from (i) insufficient CPS degradation in control mesocosms so that higher CPS concentrations may have been necessary to affect daytime distributions in comparison to controls; (ii) species-specific CPS thresholds with the smaller-sized D. longispina in mesocosms and in Lake

Fig. 2 .
Fig. 2. Depth profiles of temperature and attenuation coefficients and depths of euphotic zone in experimental mesocosms 48 h after addition of 5αcyprinol sulfate (CPS) during 2019.Boxplots of attenuation coefficients of light (A) and depths of euphotic zones (= 1% I 0 ), (B) represent three biological replicates per treatment.Depth-dependent values of temperature are depicted for three biological replicates of control (C), 1x CPS-treated (D), and 10x CPS-treated (E) mesocosms; the magenta line represents smoothed conditional means of the vertical temperature profiles across all replicates of the respective treatment.

Fig. 3 .
Fig. 3. Vertical profiles of Lake Klostersee during 2022.Distribution of Daphnia longispina during daytime (A) and nighttime (B).Depicted are relative distributions (mean AE SE, n = 3) over the depth ranges "0-2 m," "2-4 m," "4-6 m," "6-8 m," and "8-10 m." Data represent samplings from three consecutive dates.(C) Vertical profile of light intensity and temperature on 1 st day of sampling.The green line represents the fit of an exponential model to the vertical profile of light intensity; the magenta line connects the temperature data points.(D) Vertical profile of chlorophyll a on 1 st day of sampling; the black line connects the chlorophyll data points.

Table 1 .
Concentrations of 5α-cyprinol sulfate (CPS) in the experimental mesocosms 48 h after CPS addition during 2019.Concentrations represent the means of two technical replicates.