Sex, Drugs, and Rotifers: Endocrine Disrupting Water Contaminants and Rotifer Reproductive Cycling

Charlotte Hovland1

Advised by Kristin Gribble2

1The Biological Sciences Collegiate Division, University of Chicago,

Chicago, IL 60637 USA

2Josephine Bay Paul Center. Marine Biological Laboratory, Woods Hole, MA 02543 USA

Abstract Monogonot rotifers are near-ubiquitous invertebrate zooplankton, distinguished by their cyclically parthanogenetic reproductive activity. This reproductive cycling is regulated by steroid hormones, possibly including estrogen. In this study, I observed the effects of two estrogen agonists and known environmental pollutants, 4-nonylphenol and 17α-ethynylestradiol, on rotifer population growth rates, sex ratios, and egg quality. Neither nonylphenol or ethynylestradiol, nor both in conjunction had any effect on population growth rates and they had only mixed effects on sex ratios. However, the estrogen agonists did have a marked impact on egg quality. The estrogen agonists’ effects on egg development were sex specific, with male eggs showing greater susceptibility. Nonylphenol and ethynylestradiol influenced egg quality at far lower concentrations in combination than were sufficient to produce results in tests of individual compounds. This suggests interactions between the two endocrine-disrupting compounds that may be cause for environmental concern.

Keywords: Rotifers, endocrine disruption, mixis

Introduction Rotifers (phylum Rotifera) are zooplanktonic invertebrates. Rotifers are “ubiquitous,” appearing globally in freshwater and marine systems, and even in moist soils (Fontaneto and De

Smet, 2015). Although they are small (generally < 1mm in length), rotifers are of great ecological importance. As low-level consumers, rotifers take in energy and nutrients from phytoplankton, detritus, and single-celled bacteria and protozoa (Fontaneto and De Smet, 2015).

In turn, rotifers are consumed by macro-organisms, including large zooplankton, benthic worms, and larval fish. Rotifers package energy and nutrients from the smallest scale organisms in their ecosystems and allow their export to larger animals, bridging the gap between the micro and macroscopic worlds within aquatic ecosystems (Fontaneto and De Smet, 2015; Huang et al.,

1 2012). Monogonota is the most specious rotifer sub-group, and includes my study organism,

Brachionus manjavacas.

Monogonot rotifers are cyclically parthenogenetic, and males and females are highly sexually dimorphic, with females composing most of the population. Under ideal conditions, asexual (amictic) females produce eggs by mitosis, resulting in the parthanogenetic development of clonal, diploid daughters (Radix et al., 2001). However, at high populations densities and in response to changes in photoperiod, sexual (mictic) reproduction is triggered, and a portion of the eggs released by the amictic females develop into mictic females, which produce haploid ova by meiosis (Snell, 2011). If these eggs go unfertilized, they develop into small, haploid males, which are then available to mate with the mictic females. Once mating occurs, the zygote develops into a ‘resting egg,’ an embryo surrounded by a thick shell. The resting eggs settle out of the water column and enter a period of dormancy in the sediments below (Fontaneto and De

Smet, 2015). Once conditions are again favorable, the resting eggs emerge and complete their development into amictic females.

Mictic reproduction is density dependent and regulates rotifer population dynamics. The production of resting eggs removes rotifers from active circulation in preparation for winter, during population booms, or when their aquatic habitat shrinks, concentrating the rotifers in the remaining water and increasing their population densities. This allows rotifer populations to endure dramatic fluctuations in habitat quality. Mixis is thought to be under the control of a pheromone, released into water by female rotifers (Snell, 2006). As in bacterial quorum sensing, the strength of the signal is related to the density of rotifers releasing the signaling molecule, so that the rotifers will only transition to mixis when certain density thresholds are surpassed (Snell,

2 2011; Stout et al., 2010). Once the mixis signal has been triggered, the rotifers’ reproductive response is under hormonal control (Snell, 2011; Stout, 2010).

Steroid hormones in particular are increasingly well-supported candidates for the primary regulators of rotifer reproduction. Snell et al. (2006) observed similarities between fragments of a purported mixis signaling protein and a protein that induces steroidogenesis in humans, while

Stout et al. (2010) revealed the presence of progesterone receptors in male and female rotifers.

Recently, Jones et al. (2017) discovered an estrogen-like receptor in Brachionus rotifers. This estrogen-like receptor is so highly conserved with respect to mammalian estrogen receptors that it is able to bind human estradiol as a ligand. Several teams of researchers have shown that rotifers respond when their environment is dosed with human steroid hormones, including progesterone, estrogen, and testosterone (Snell and DesRosiers, 2008; Gardallo et al., 1997;

Preston et al., 2000). This has led to concerns that rotifer reproductive cycling may be disrupted by water contaminants that are androgen and estrogen mimics or antagonists. These contaminants include β –estradiol, ethynylestradiol, nonylphenol and nonylphenol-ethoxylate, dioxin, and endosulfan (Depledge and Billinghurst, 1999; Roefer, 2000; Swartz et al., 2006; and

Zhang, 2015). In isolation, ethynylestradiol and nonylphenol have been shown to reduce the number of females in rotifer populations as well as the proportion of mictic females (Radix,

2001). Endocrine agonists and antagonists have previously been shown to cause changes to the sex ratios of fish populations and to hinder sexual development and reproduction in several vertebrate groups (Depledge and Billinghurst, 1999). Diverse marine invertebrates, including mollusks and arthropods, have been found to be sensitive to endocrine disruption (Depledge and

Billinghurst, 1999). Endocrine disruption is of particular concern in the effort to monitor and regulate water pollution, as pollutants may produce reproduction-disrupting endocrine effects

3 even at low concentrations that do not result in outright toxicity (Depledge and Billinghurst,

1999).

Little is yet known about the potential synergistic, additive, or antagonistic effects of multiple endocrine-disrupting contaminants—despite the fact that most contaminated water sources contain multiple contaminants, and contaminants cannot be expected to appear in isolation in ecologically relevant environments (Depledge and Billinghurst, 1999; Standley et al.,

2008). When Thorpe et al. studied the combined effects of nonylphenol and ethynylestradiol in developing rainbow trout (Oncorhynchus mykiss), they found an additive response on increasing vitellogenin concentrations in the trouts’ blood plasma (2001). Given that steroid hormone receptors appear to be widely conserved in bilatarians, it is of interest to see if additive effects of endocrine-disrupting contamination are also observed in invertebrate members of impacted ecosystems.

In this study, I investigated the effects of ethynylestradiol, an artificial estrogen manufactured as an oral contraceptive and present in wastewater, and nonylphenol, an estrogen- mimicking contaminant released from plastics and used as an industrial surfactant (Swartz et al.,

2006; Soares, 2008), on the growth rate and mictic/amictic ratios of rotifer populations, and on the size of their eggs. I examined the effects of these contaminants in and above environmental concentrations, as well as their potential effects in combination.

Given that monogonot rotifers appear to have active estrogen-like receptors localized to their reproductive structures (Jones et al., 2017), I expected that their reproductive cycles would be influenced by exposure to the estrogen agonists ethynylestradiol and nonylphenol. Based on the work of Preston et al. (2000) and Radix et al., (2001), if the compounds were to have any effect, I would have expected them to lower population growth rates, however, I was uncertain as

4 to whether any effect would be seen at environmental-level concentrations. As the estrogen hormone signaling pathway is deeply conserved between rotifers and vertebrates (Jones et al.,

2017), I expected that, as in the case of the rainbow trout, a combination of contaminants would have a greater effect than would each contaminant alone (Thorpe et al., 2001).

Methods

I reared Brachionus manjavacas rotifers under eleven chemical treatments: a control without added contaminants; a control to which only ethanol was added; three treatments to which 4-nonylphenol (Standard grade, Sigma-Aldrich, Milwaukee USA) was added dissolved in ethanol to concentrations of 1 ug/L (0.0045 uM), 5 ug/L (0.0227 uM), and 50 ug/L (0.2273 uM); three treatments to which 17α-ethynylestradiol (> 98%, Sigma-Aldrich, Milwaukee USA) was added to concentrations of 1 ng/L (3.378x10-6 uM), 5 ng/L (1.689x10-5 uM), and 50 ng/L

(1.689x10-4 uM); and three additional treatments with both 4-nonylphenol and 17α- ethynylestradiol in combination. In the treatments with both 4-nonylphenol and 17α- ethynylestradiol, one treatment contained 4-nonylphenol at a concentration of 1 ug/L and 17α- ethynylestradiol at a concentration of 1 ng/L, while the next contained 5 ug/L 4-nonylphenol and

5 ng/L 17α-ethynylestradiol, and the final treatment contained 50 ug/L 4-nonylphenol and 50 ng/L 17α-ethynylestradiol. Each treatment was carried out in three replicates.

I prepared each replicate in a 50 mL glass vial. Initially, I filled each replicate with 20 mL from a single, mixed solution of Tetraselmis suecica, Instant Ocean, and B. manjavacas. The initial concentration of T. suecica was ~6x105 cells per millilitre, while the initial concentration of rotifers was ~5 individuals per millilitre. The rotifers were sourced from mixed-age laboratory cultures maintained by Kristin Gribble at the MBL. I spiked each vial individually with the appropriate volume of contaminant solution. Given that the greatest added volume of

5 contaminant was 200 uL, only 1% of the 20 mL volume of the total solution, it was not necessary to add any additional volume to the control vials.

I incubated the control vials and the vials treated with 17α-ethynylestradiol for 7 days, but the ethanol controls and the vials treated with 4-nonylphenol and with both contaminants were only incubated for 6 days. After the control vials and 17α-ethynylestradiol replicates had incubated for 5 days, and the ethanol controls, 4-nonylphenol replicates, and combined replicates had incubated for 4 days, I replenished each vial with 6 additional millilitres of algae solution. I also added contaminant solution to the vials so that the increased volume of algae solution would not result in more dilute contaminants. All incubation took place at 21°C and on a 12 hour light- dark cycle.

I took 1 mL subsamples from each replicate during each day of their incubation period. I observed these live subsamples using a dissecting microscope, noting the number of active male rotifers. I then fixed each subsample with 20 uL of Lugol’s solution. Using a dissecting microscope, I counted the rotifers and characterized them as mictic females, amictic females, nonovigerous females, or males. I also counted the free eggs in each subsample, and determined whether they were diploid female eggs, haploid male eggs, or resting cysts.

After the seventh day of incubation for the control and ethynylestradiol vials and the sixth day of incubation for the ethanol control, nonylphenol, and combined treatments, I transferred 2 mL subsamples from each replicate into fresh vials, each containing 18 mL of algae solution and the appropriate concentration of contaminants. I allowed these vials to incubate undisturbed for 5 days. On the fifth day, I took 1 mL subsamples from each vial, fixed them with Lugol’s solution, and counted and categorized the mictic, amictic, and nonovigerous females, males, and eggs. I then replenished the vials with 6 mL of algae solution, added the appropriate amounts of

6 contaminants to maintain their concentrations, and allowed the vials to incubate overnight. The next day, I took 7 mL subsamples from each vial, vortexed each subsample for a minute to detach as many eggs as possible from the female rotifers, and treated the subsamples with

Lugol’s solution. I let the treated subsamples settle overnight and then removed 5 mL of the supernatant. I prepared microscope slides using 30 uL of the material remaining at the bottom of each subsample. I viewed these slides using a light microscope under a 40x objective lens. I photographed the first 8-12 female and male eggs I saw from each sample. If there were not at least 8 eggs of each type present on a single slide, I prepared a second slide from the same sample. Resting cysts were rare, and I took photographs of each that I observed.

I processed the resulting photos (Figure 1) using Fiji (Schindelin et al., 2012), and used a known scale to measure the area, perimeter, feret diameter (maximum caliper diameter), and roundness of each egg, where:

퐴푟푒푎 푅표푢푛푑푛푒푠푠 = 4 ∗ 휋 (푀푎푗표푟 푎푥푖푠)2

To test whether the contaminants resulted in significant effects, I used a Student’s t-test to compare the mean proportions of mictic, amictic, nonovigerous, and male rotifers in treated populations to those in control populations and to compare the average dimensions of the eggs between the treated populations and control populations. I compared the populations that had been treated with ethynylestradiol only to the control populations. The nonylphenol treated populations and the combined contaminant treated populations all contained ethanol as a solvent, and thus were compared to the ethanol control populations. Growth rates were determined by plotting the numbers of adult female rotifers and fitting an exponential curve to the data.

Results

7 Time Series

No general toxicity seems to have been caused by any of the compounds used at any of the test concentrations. All of the rotifer populations increased exponentially over time. Their rate of growth, in terms of the number of adult females, was not impacted by any of the contaminant treatments (Figure 2), nor was the time it took for the female population to double

(Figure 3).

End Point Population Mixis Ratios

No significant differences were observed between the total numbers of either rotifers or eggs in between any treatments and their respective controls. However, the proportions of mictic

(Figure 4), and eggless females did vary somewhat between treatments, as did the ratios of male and female eggs (Figures 5-6).

Here, it should be noted that the proportions of male and female eggs differed significantly between the control replicates made with only Instant Ocean and the control replicates to which ethanol was added. In the following analyses, all treatments containing nonylphenol dissolved in ethanol have been compared only to the ethanol control, while treatments containing only ethynylestradiol have been compared only to the control without ethanol. Nevertheless, the presence of ethanol may represent a confounding factor in my examination of estrogen-mimicking compounds.

The 1 ug/L nonylphenol treatment resulted in a lower proportion of mictic females and a higher proportion of nonovigerous females, with the proportion of mictic females out of total females dropping from an average of 5.3% (S.E. = 0.41%) to an average of 2.3% (S.E. = 0.63%)

8 (P = 0.0155) (Figure 4). The rotifers reared in 1 ug/L nonylphenol also produced a lower proportion of male eggs and a correspondingly greater proportion of female eggs (Figures 5-6).

In the ethanol control, male eggs averaged 43% of the total egg yield (S.E. = 4.1%), while female eggs made up 57% of the total eggs (S.E. = 4.1 %). In the 1 ug/L nonylphenol treatment, male eggs composed only 27% of total eggs and female eggs made up the remaining 73% (S.E. =

3.2%). However, no significant deviations from the control were observed for the 5 ug/L or 50 ug/L nonylphenol treatments (Figure 5).

Mictic females made up a smaller proportion of total adult females in the 1 ng/L ethynylestradiol and 1 ug/L nonylphenol combined treatment than in the ethanol control (5.3% vs. 3.3%) (P = 0.0156). A similar, but more dramatic effect was observed when both compounds were present at 50 ng/L and 50 ug/L, respectively (P = 0.0067). When both compounds were present at their highest concentration, on average mictic females only made up 1.2% of the female population (S.E. = 0.64%). In this treatment, the proportion of female eggs out of total eggs also increased, from an average of 57% to an average of 81% (S.E. = 4.9%) (P = 0.0184).

Ethynylestradiol alone did not appear to have any significant effects on the sex ratios of either adult rotifer populations or their eggs.

Resting cysts appeared so infrequently that it was impractical to subject their counts to statistical testing.

Egg Qualities

Area: After 5 days of undisturbed incubation (making 11/12 days of total contaminant exposure), all treatments had plentiful male and female eggs, but there were not enough resting cysts to produce sample sizes sufficient for further analysis.

9 The area of female eggs ranged from 8069.456 um2 to 20014.049 um2 with an overall average of 14813.35 um2 (S.E. = 128.08 um2). A significant difference between the mean area of contaminant-exposed eggs and that of the control (P = 0.0051) was found only when the rotifers were incubated with both ethynylestradiol and nonylphenol at the highest tested concentrations

(Figure 7). The mean area of the eggs from the ethanol control group was 15253.97 um2 with a standard error of 443.04 um2, while the mean area of female eggs developed in 50 ng/L ethynylestradiol and 50 ug/L nonylphenol was 13483.55 um2 with a standard error of 415.83 um2. On average, the affected eggs were 11% smaller by area than the control eggs.

Male eggs were smaller than female eggs, with an overall average area of only 5909.15 um2 (S.E. = 42.71 um2). Significant differences were observed between the mean areas of the controls, the 50 ug/L nonylphenol treatment, the 50 ng/L ethynylestradiol treatment, and all treatments using both contaminants (Figure 8). On average, the eggs developed in 50 ng/L ethynylestradiol were 4.4% larger by area than those of the control, while the male eggs developed in 50 ug/L nonylphenol and in both nonylphenol and ethynylestradiol were smaller, on average, than those in the ethanol control treatment (Table 1). Although the combination of compounds produced statistically significant effects at lower concentrations than nonylphenol alone, at the concentration when both produced an effect (50 ug/L nonylphenol and a combination of 50 ug/L nonylphenol and 50 ng/L ethynylestradiol), there was not a significant difference in the effect’s magnitude. The mean area of male eggs treated with both ethynylestradiol and nonylphenol was not significantly different from those treated with nonylphenol alone (P = 0.9028).

Perimeter: Neither ethylestradiol or nonylphenol had a significant effect on female egg perimeter as individual compounds. When the compounds were applied in combination, results

10 were mixed. The mean perimeter of the female control eggs was significantly greater (P < 0.05) that that of the female eggs produced in 1 ng/L ethynylestradiol and 1 ug/L nonylphenol or that of the female eggs produced in the highest concentration of both compounds (Figure 9).

However, the intermediate treatment, with 5 ng/L ethynylestradiol and 5 ug/L nonylphenol, showed no significant deviation from the control (P = 0.1004).

The mean perimeters of male eggs (Figure 10) roughly corresponded to the mean areas.

Whether measured by area or perimeter, male egg size was affected by the combined contaminants at lower concentrations than affected the female eggs. All cultures treated with both compounds had significantly smaller mean male egg perimeters compared to controls. Male eggs from the 50 ug/L nonylphenol treatment had, on average, 4.5% smaller perimeters than male eggs from the ethanol control, and the difference between the 50 ug/L nonylphenol and ethanol control treatments was found to be significant (P = 0.003). None of the ethynylestradiol treatments were found to have any effect on male egg perimeter.

Feret Diameter: Under no circumstances did the average feret diameters of the female eggs differ significantly from that of their relevant controls (Figure 11).

In contrast, the mean diameter of male eggs from the 50 ug/L nonylphenol treatment was significantly smaller than that of the ethanol control (P = 0.0020). All treatments using both contaminants also resulted in male eggs with small mean feret diameters compared to the ethanol control (Figure 12).

Roundness: Out of a maximum of 1, the average female overall had a roundness of

0.8081 (S.E. = 0.0037). The average egg developed in 50 ng/L ethynylestradiol was significantly rounder than the average control (0.8669 as compared to 0.7872; P = 0.00004). The average female egg developed in 1 ng/L ethynylestradiol and 1 ug nonylphenol was less round than the

11 average control (P = 0.0258), however no other treatments had any significant effects on female egg roundness, including higher concentrations of both contaminants (Figure 13).

Male eggs that developed in 50 ng/L ethynylestradiol were, on average, significantly rounder than control male eggs (0.8371 as compared to 0.7997, P = 0.0348). Male eggs that developed in 50 ug/L nonylphenol were significantly rounder than control eggs (P = 0.0046).

Male eggs from the low concentration treatment with both compounds were rounder than the control, as were male eggs from the highest concentration treatment, but the intermediate concentration did not show significant effects on male egg shape (P = 0.0778) (Figure 14).

Discussion

Neither nonylphenol nor ethynyestradiol in isolation, nor both in conjunction, had a significant effect on the growth rate or doubling time of B. plicatilis populations. These findings are consistent with those of Radix et al. (2001), who found that ethynylestradiol only decreased rates of population increase at concentrations of 1.72 uM or higher, while nonylphenol only decreased the rate of population increase at concentrations higher than 0.59 uM. The maximum concentrations used in the present study, selected to approximate the concentrations found in

Cape Cod wastewater (Swartz et al., 2006; Rudel et al., 1998; Bhandari, 2015) were 0.2273 uM nonylphenol and 1.689x10-4 uM ethynylestradiol. The concentrations of nonylphenol and ethynylestradiol typically seen in coastal waters (4.5x10-4 uM to 0.381 uM nonylphenol; 5x10-6 uM ethynylestradiol) are not sufficient to significantly impact the growth rates of local rotifer populations (Swartz et al., 2006; Bhandari, 2015). However, nonylphenol and ethynylestradiol are only two of many estrogen agonists that have been detected entering the ocean, and the effects of many compounds in conjunction, although present at low concentrations, remain worthy of study and concern.

12 Exposure to estrogen agonists did not have any significant effect on the total number of females or on the proportion of amictic females present. Treatment with nonylphenol or with both nonylphenol and ethynylestradiol did result in a significantly lowered ratio of mictic females to total females in three cases, but these effects did not follow a clear pattern with respect to concentration. While somewhat perplexing, this is consistent with the 2001 findings of

Radix et al., who report, “it was difficult to establish a clear dose-response relationship” between exposure to nonylphenol and changing ratios of mictic and amictic females. Previous researchers have speculated that sexual reproductive processes in rotifers are more likely to be affected by endocrine disruption than are asexual reproductive processes, perhaps because of the greater complexity of endocrine signaling involved in carrying out sexual reproduction (Preston et al.,

2000). This study, which saw some effect of estrogen agonists on mictic female proportions but none on the proportions of amictic females, could lend support to this hypothesis.

Preston et al. found alterations to the proportion of fertilized mictic females (mictic females bearing resting cysts) to be a reliable indicator of endocrine disruption, with the proportion of fertilized mictic females declining when exposed to 50 ug/L nonylphenol (2000).

However, in my experiment, which tested 50 ug/L nonylphenol, fertilized mictic females were so rarely seen in any of the treatments that no useful comparisons between treatments could be made. This scarcity of fertilized mictic females is likely a result of the short time period over which the experiment was run. Each population was allowed to develop for only two generations. It’s possible that a longer incubation time would have allowed more males to accumulate, resulting in higher rates of fertilization and a greater accumulation of resting eggs.

While the total numbers of eggs did not vary between treatments, and the proportions of male and female eggs did not vary with any clear pattern, several trends emerged when I

13 compared the shapes and sizes of eggs extracted from the different treatments. In general, male eggs were much more susceptible to effects of the contaminants than were female eggs. Male eggs responded to single contaminants when female eggs did not, and consistently responded to lower concentrations of the combined contaminants. Preston et al. (2000) hypothesized that certain “aspects of the life history of male rotifers” could cause them to be more vulnerable to endocrine disruption. Although the mechanisms that would cause such a disparity are not well characterized, they could stem from differences in the biochemical pathways used in the development of males and females. This sex specific susceptibility to estrogen agonists appears to extend to eggs, and indicates that nonylphenol and ethynylestradiol are causing endocrine effects, not merely general toxicity, which would be expected to impact all egg morphologies equally.

Additionally, effects on egg size and shape occurred at concentrations of the combined compounds that were insufficient to produce significant change when each compound was tested individually. For all measures of egg size, nonylphenol alone had a significant effect on male egg size only at its highest concentration. When both compounds were present, significant effects were observed at all concentrations. The interactions between the two compounds do not appear to be simply additive: in the case of male egg area, high concentrations of nonylphenol alone resulted in a slightly smaller average egg, while high concentrations of ethynylestradiol resulted in a slightly larger average egg. If the effects of the two compounds were simply added together, one might expect to see the two compounds counteract each other. Instead, even the lowest concentration of both contaminants resulted in male eggs that were, on average, smaller than the control, an effect which was not achieved by similar concentrations of nonylphenol or ethynylestradiol alone. Additionally, at concentrations where both an individual compound and

14 the combination of compounds had an effect, the magnitude of the effect was not significantly greater under the combined treatment. In order to more definitely characterize the interaction between nonylphenol and ethynylestradiol, a wider range of concentrations should be tested in order to establish curve by which responses to various doses of individual and combined compounds could be compared. This question is of great interest going forward, as real environments are usually polluted with trace amounts of many different contaminants, rather than high concentrations of a only a few compounds. If the interaction of multiple endocrine disruptors allows them to have greater effects at lowered concentrations, then even trace amounts of endocrine agonists may be cause for grave environmental concern.

Acknowledgments

This work would not have been possible without Kristin Gribble’s mentorship, Emily

Corey’s aid and patience, and the generosity of the Gribble Lab in sharing their knowledge and resources with me.

References

Bhandari, R.K., S. L. Deem, D.K. Holliday, C.M. Jandegian, C.D. Kassotis, S.C. Nagel, D.E. Tillitt, F.S. vom Saal, C.S. Rosenfeld. 2015. Effects of the environmental estrogenic contaminants bisphenol A and 17α-ethinyl estradiol on sexual development and adult behavior in aquatic wildlife species. General and Comparative Endocrinology 214: 195- 219.

Depledge, M.H. and Z. Billinghurst. 1999. Ecological Significance of Endocrine Disruption in Marine Invertebrates. Marine Pollution Bulletin 39: 32-38.

Fontaneto, D. & W. De Smet 2015. Rotifera. 217-300.

Gallardo, G. W., A. Hagiwara, Y. Tomita, K. Soyano, and T.W. Snell. 1997. Effect of some vertebrate and invertebrate hormones on the population growth, mictic female production, and body size of the marine rotifer Brachionus plicatilis Müller. Hydrobiologia 358:113-120.

Huang, L., Y. Xi, C. Zha, L. Zhao, and X. Wen. 2012. Effects of dieldrin and 17β-estradiol on life history characteristics of freshwater rotifer Brachionus calyciflorus Pallas. Journal of Freshwater Ecology 27(3): 381-392.

Jones, B.L., C. Walker, B. Azizi, L. Tolbert, L.D. Williams, and T. W. Snell. 2017. Conservation of estrogen receptor function in invertebrate reproduction. BioMed Central Evolutionary Biology 17:65.

Preston, B.L., T.W Snell, T.L Robertson, and B.J. Dingmann. 2000. Use of Freshwater Rotifer Brachionus calyciflorus in Screening Assay for Potential Endocrine Disruptors. Environmental Toxicology and Chemistry 19(12): 2923-2928.

Radix, P., G. Severin, K.W. Schramm, A. Kettrup. 2001. Reproduction disturbances of Brachionus calyciflorus (rotifer) for the screening of environmental endocrine disrupters. Chemosphere 47:1097-1101.

Roefer, P., S. Snyder, R.E. Zegers, D.J. Rexing, and J.L. Fronk. 2000 Endocrine-disrupting chemicals in a source water. American Water Works Association Journal 92(8): 52-58.

Rudel, R.A., S. J. Melly, P.W. Geno, G. Sun, and J.G. Brody. 1998. Identification of Alkylphenols anad Other Estrogenic Phenolic Compounds in Wastewater, Septage, and Groundwater on Cape Cod, Massachusetts. Environmental Science and Technology 32: 861-869.

Schindelin, J.; Arganda-Carreras, I. & Frise, E. et al. 2012. Fiji: an open-source platform for biological-image analysis. Nature methods 9(7): 676-682

Snell, T.W., J. Kubanek, W. Carter, A.B. Payne, J. Kim, M.K. Hicks, and C-P Stelzer. 2006. A protein signal triggers sexual reproduction in Brachionus plicatilis (Rotifera). Marine Biology 10.1007

Snell, T.W. 2011. A review of the molecular mechanisms of monogonot rotifer reproduction. Hydrobiologia 662:89-97.

Soares, A., B. Guieysse, B. Jefferson, E. Cartmell, and J.N. Lester. 2008. Nonylphenol in the environment: A critical review on occurrence, fate, toxicity and treatment in wastewaters. Environment International 34:1033-1049.

Standley, L.J., R.A. Rudel, C.H. Swartz, K.R. Attfield, J. Christian, M. Erickson, and J.G. Brody. 2008. Wastewater-contaminated groundwater as a source of endogenous hormones and pharmaceuticals to surface water ecosystems. Environmental Toxicology and Chemistry 27(12): 2457-2468.

Stout, E.P., J.J. La Clair, T.W. Snell, T.L. Shearer, and J. Kubanek. 2010. Conservation of progesterone hormone function in invertebrate reproduction. PNAS 107(26): 11859- 11864.

Swartz, C.H., S. Reddy, M.J. Benotti, H. Yin, L.B. Barber, B.J. Brownawell, and R.A. Rudel. 2006. Steroid Estrogens, Nonylphenol Ethoxylate Metabolites, and Other Wastewater Contaminants in a Groundwater Affected by a Residential Septic System on Cape Cod, MA. Environ. Sci. Technol. 40: 4894-4902.

17 Thorpe, K.L., T.H. Hutchinson, M.J. Hetheridge, M. Scholze, J.P. Sumpter, and C.R. Tyler. 2001. Assessing the Biological Potency of Binary Mixtures of Environmental Estrogens using Vitellogenin Induction in Juvenile Rainbow Trout (Oncorhynchus mykiss). Environ. Sci. Technol. 35:2476-2481.

Zhang, Y.Z. 2015. “Occurrence and Reduction of Pharmaceuticals and Personal Care Products (PPCPs) in Groundwater and Wastewater of Cape Cod, MA.” SES MBL

18 Tables

Table 1: Statistically significant differences in haploid (male) egg area

Figures

Figure 1: Representative images of different egg morphologies

Figure 2: Comparison of population growth rate across treatments

Figure 3: Comparison of population doubling time across treatments

Figure 4: Percentage of mictic females out of adult females

Figure 5: Percentage of male eggs out of total eggs

Figure 6: Percentage of female eggs out of total eggs

Figure 7: Comparison of diploid (female) egg areas

Figure 8: Comparison of haploid (male) egg areas

Figure 9: Comparison of diploid (female) egg perimeters

Figure 10: Comparison of haploid (male) egg perimeters

Figure 11: Comparison of diploid (female) eggs – maximum diameter

Figure 12: Comparison of haploid (male) eggs – maximum diameter

Figure 13: Comparison of roundness among diploid (female) eggs

Figure 14: Comparison of roundness among haploid (male) eggs

19 Table 1: Results of Student’s t-test comparing the mean areas of eggs between treatments. The null hypothesis was that the mean areas would be the same between contaminant treatments and the relevant controls.

A. Male Eggs Treatment Mean Area (um2) S.E. (um2) P-value Control 6158.16 100.55 N/A Ethanol Control 6057.90 119.63 N/A P-value relative to Control 17EE 1 5854.38 147.66 0.0945 17EE 5 6076.66 248.60 0.7629 17EE 50 6428.84 83.19 0.0426 P-value relative to Ethanol Control 4N 1 6022.17 115.23 0.8305 4N 5 6040.91 63.07 0.9006 4N 50 5575.70 104.95 0.0037 Both 1 5522.81 117.33 0.0023 Both 5 5638.03 139.32 0.0259 Both 50 5559.88 154.30 0.0136

B. Female Eggs

Treatment Mean Area (um2) S.E. (um2) P-value Control 15544.63 365.33 N/A Ethanol Control 15253.97 443.04 N/A P-value relative to Control 17EE 1 16018.22 277.61 0.3067 17EE 5 15111.66 434.60 0.4444 17EE 50 15626.45 348.13 0.8718 P-value relative to Ethanol Control 4N 1 14363.94 443.63 0.1612 4N 5 15269.80 400.68 0.9788 4N 50 14269.99 496.94 0.1456 Both 1 14359.03 370.75 0.1271 Both 5 14188.79 390.61 0.0745 Both 50 13483.55 415.83 0.0051

A. B.

Figure 1: Typical female (A.) and male (B.) eggs, and resting cyst (C.) Stained with Lugol’s solution and captured at 40x magnification.

21 A.

Figure 2: Population growth, in terms of total adult females. Growth rates were unchanged between controls and treatments. Error bars represent standard error.

Figure 3: Doubling time of female rotifer population. In no case did the average doubling time of any treatment differ significantly from its relevant control. (Respective P values, 0.131, 0.977, 0.346, 0.256, 0.349, 0.434. All are > 0.05.)

Figure 4: Mictic females made up 1.2 to 5.3 percent of adult females. Treatment with 1 ug/L nonylphenol, with 1 ng/L ethynylestradiol and 1 ug/L nonylphenol, or with 50 ng/L ethynylestradiol and 50 ug/L nonylphenol all depressed the proportion of mictic females. Error bars show standard error. * denotes significance (P < 0.5)

Figure 5: Male eggs made up 23-44% of the total eggs observed. The proportion of male eggs differed from the relevant control only in the 1 ug/L nonylphenol treatment (4N 1). Error bars represent standard error. * denotes significance (P < 0.5)

Figure 6: Female eggs made up 55 to 88% of the total eggs observed. Error bars represent standard error. * denotes significance (P < 0.5)

Figure 7: The mean areas of diploid, female eggs collected after 11/12 total days of incubation. Error bars show standard error. A significant difference from the control (P = 0.0051) was found only when the rotifers were incubated with both 50 ng/L ethynylestradiol and 50 ug/L nonylphenol (Both 50). * denotes significance (P < 0.5)

Figure 8: The mean areas of haploid, male eggs collected after 11/12 total days of incubation. Error bars show standard error. Significant deviations from the control (P < 0.05) were found in the 50 ng/L ethynylestradiol treatment (17EE 50), the 50 ug/L nonylphenol treatment (4N 50), and in all treatments with both contaminants (Both 1, Both 5, and Both 50). * denotes significance (P < 0.5)

Figure 9: The mean perimeters of diploid, female eggs. Error bars show standard error. The mean perimeter of the control eggs was significantly greater than that of the eggs developed in 1 ng/L ethynylestradiol and 1 ug/L nonylphenol (Both 1) or in 50 ng/L ethynylestradiol and 50 ug/L nonylphenol (Both 50), but was not statistically significantly different from the intermediate Both 5 treatment. * denotes significance (P < 0.5)

Figure 10: Mean perimeters of haploid, male eggs. Error bars represent standard error. Eggs from the 50 ug/L nonylphenol treatment (4N 50), and from all the cultures treated with both compounds (Both 1, Both 5, Both 50), had significantly smaller mean perimeters compared to the controls. * denotes significance (P < 0.5)

Figure 11: Feret diameters (diameter at widest point) of diploid, female eggs. Error bars display standard error. Under no circumstances did the diameters deviate significantly from the control values.

Figure 12: Feret diameters (diameter at widest point) of haploid, male eggs. Error bars show standard error. The mean diameter of the eggs from the 50 ug/L nonylphenol treatment (4N 50) was significantly smaller than that of the control. This was also the case for all treatments using both contaminants (Both 1, Both 5, Both 50). * denotes significance (P < 0.5)

Figure 13: Roundness of diploid, female eggs. Error bars show standard error. The average egg developed in 50 ng/L ethynylestradiol (17EE 50) was significantly more circular than the average control (P = 0.00004). The average egg developed in 1 ng/L ethynylestradiol and 1 ug nonylphenol was less round than the average control. * denotes significance (P < 0.5)

Figure 14: Roundness of haploid, male eggs. Error bars show standard error. The average egg developed in 50 ng/L ethynylestradiol (17EE 50) was slightly more round than the average control (P = 0.0348), as was the average male egg developed in 50 ug/L nonylphenol (4N 50) (P = 0.0046). The average eggs developed in the lowest and highest concentrations of both compounds (Both 1 and Both 50), were significantly rounder than the average control, but this was not the case for the intermediate concentration (Both 5). * denotes significance (P < 0.5)