Impacts of bisphenol A on growth and reproductive traits of submerged macrophyte Vallisneria natans

Bisphenol A (BPA) is considered a contaminant of emerging concern and interferes with the normal activities of living organisms. The toxicity of BPA is evident in animals and terrestrial plants. However, the response of aquatic plants to low BPA concentrations is still unclear. In the present study, effects of varying BPA loadings (targeting at 0.01, 0.1, and 1 mg/L) on the growth and reproductive traits of the dioecious annual submerged macrophyte Vallisneria natans were assessed through a 5-month experiment. The results showed that BPA inhibited the elongation of V. natans leaves but resulted in an increase in leaf number and ramet number under the highest BPA loading treatment (targeting at 1 mg/L). In addition, detectable biochemical changes in the total carbon and soluble sugar contents were found, which both were significantly higher at the highest BPA loading treatment. However, the total biomass did not alter significantly after the BPA treatments, indicating that BPA did not induce direct toxic effects on the growth of V. natans. At the highest BPA loading treatment, female individuals of V. natans allocated less number for ramet than male ones, showing a clear sexual dimorphism. No significant differences between the five treatments were found for the flower or fruit traits, while the germination rate was significantly inhibited for the seeds collected from the highest BPA loading treatment. In conclusion, V. natans tolerated low concentrations of BPA by making a trade-off between ramet (leaf) number and leaf elongation, as well as modulating the total carbon and soluble sugar contents. However, serious consequence of decline in seed viability implied that the impact of BPA on plant reproduction were usually underestimated.


Introduction
Endocrine disrupting chemicals (EDCs), also known as environmental hormones, are exogenous chemicals that interfere with the hormonal regulation of homeostasis, reproduction, development, and behavior in organisms (USEPA 1997). Large quantities of EDCs have been released into environments due to human activities (Casals-Casas and Desvergne 2011). EDCs, such as bisphenol A (2,2-bis(4-hydroxyphenyl) propane, hereafter called BPA), are contaminants with estrogenic activity at very low concentrations and are becoming a major concern for water quality (Zhang et al. 2016).
BPA is formed by combining two phenol molecules bonded through a methyl bridge and two methyl groups (Careghini et al. 2015). Due to its good strength and hardness, thermal stability, and resistance to acids and oils (Ma et al. 2019), BPA has been widely used in the manufacture of various consumer products such as dental sealants, plastic bottles, baby bottles, thermal paper, and canned food (Bhatnagar and Anastopoulos 2017). The global demand for BPA has been reported to be 3.2, 3.9, 5, and 8 million tons in 2003, 2006, respectively (Careghini et al. 2015Hercog et al. 2019). The BPA enters into the aquatic ecosystems by leaching from effluent discharges and seepage from manufactories, which leads to the exposure of aquatic organisms to BPA (Wu and Seebacher 2020). A recent study reported that more than 100 tons of BPA were released into environments worldwide each year (Naveira et al. 2021). Meantime, a low BPA concentration (< 0.063 mg/L) was recorded in surface water in 55 countries (Wu and Seebacher 2020), with relatively high concentrations (< 0.15 mg/L) in industrial wastewater and > 10 mg/L in leachate from hazardous waste landfills (Hing-Biu and Peart 2000; Yamamoto et al. 2001).
Numerous studies have shown that BPA can interact with estrogen receptors, androgen receptors, and thyroid receptors in animals, interfering with endocrine activity, oxidative, and mutagenic potential (Keri et al. 2007;Richter et al. 2007;Tayama et al. 2008). Its toxicity is very pronounced for fish, amphibians, and invertebrates Liu et al. 2020;Petrie et al. 2019). The stress of BPA on plants, however, might be weaker since plants can convert BPA into metabolites, which typically exhibit lower toxicity and estrogenicity (Michałowicz 2014). For terrestrial plants, BPA is only slightly to moderately stressful to plant growth depending on plant species (Alexander et al. 1988;Xiao et al. 2020). The concentrations of BPA at 0.8 mg/L promoted the growth of soybean, whereas extreme high concentrations of BPA (17.2 mg/L) inhibited all determined plant traits . Moreover, the eco-toxicological effects of BPA on aquatic plants were not comprehensive (Xiao et al. 2020). Results of acute toxicity experiments of BPA on floating plant Lemna minor indicated that BPA promoted the plant growth with concentration less than 5 mg/L, but significantly inhibiting the plant growth with the BPA concentration of 50 mg/L (Liang et al. 2022). Meanwhile, studies on another free-floating macrophyte Azolla (Lam.) showed similar results that 20 mg/L BPA significantly inhibited the plant growth and even caused severe leaf damage (Sarkar et al. 2022). However, study on seagrass Cymodocea nodosa reported that BPA with rather low concentrations (0.03-3 μg/L) significantly inhibited the leaf elongation (Malea et al. 2020). How freshwater submerged macrophytes would respond to BPA exposure is still warranted.
Submerged macrophytes are an indispensable part of the freshwater, and the widespread BPA at low concentrations in lakes and rivers were expected to have chemical effects on submerged macrophytes, especially for the establishment of young seedlings. Therefore, we used Vallisneria natans (Lour.) Hara seedlings to conduct a 5-month microcosm study to explore the potential effect of BPA. We hypothesized that (1) low BPA concentrations were expected to stress on growth traits of freshwater macrophyte V. natans similarly with the reported seagrass (H1); (2) the estrogenicity from BPA on animals was not expected for plants; thus, BPA should not have a directly significant effect on the reproductive traits of submerged macrophytes (H2); (3) due to the sexual dimorphism of the chosen species, female and male individuals of V. natans can have different strategies of resource allocation responding to the BPA exposure, female being more sensitive (H3). We believe that our study provides a theoretical basis for understanding the ecotoxic mechanisms of BPA on submerged plants and provides a reference for assessing the potential ecological risk of BPA exposure in aquatic ecosystems.

Plant material and experimental set-ups
We conducted a 5-month microcosm study to analyze the effects of BPA on the growth and reproduction traits of freshwater macrophytes in greenhouse in Wuhan Botanical Garden (114°43′E, 30°55′N), Chinese Academy of Sciences. The study lasted from May to November 2021. The average ambient temperature inside the greenhouse was 29.4 ± 3.1 °C, and the average light intensity on midday was 414 ± 288 μmol photons m −2 s −1 measured with a HOBO (Onset, USA) recorder. The wide-spread dioecious annual submerged macrophyte V. natans was chosen as the indicator species (Chou et al. 2022;Zhou et al. 2016). It is often used as the model aquatic plant in standardized toxicity tests involving environmental stresses and proved to be a sensitive species (Wang et al. 2010;Dou et al. 2013;. Additionally, V. natans is a pioneer species for freshwater ecosystem restoration (Wang et al. 2010), and its decline in natural communities is of great concern (Wang et al. 2015). The experimental setup consisted of 250 plastic cups (top diameter × bottom diameter × height: 7.0 cm × 4.5 cm × 7.8 cm) and 25 plastic buckets (length × width × height: 24 cm × 24 cm × 40 cm, Volume: 20L). In May 2021, we selected the V. natans seedlings with three ca. 4-cm-long leaves and carefully removed the attached algae on the surface of the plant. Each V. natans seedling was planted in the plastic cup as used in Zhi et al. (2018). As the Chinese Sanitary Standard for Drinking Water ( GB5749-2006) stipulates that the content of BPA in tap water should not exceed 0.01 mg/L. BPA was detected in Chinese drinking water with a maximum of 6.5 ng/L (Zhang et al. 2019), so we used tap water to cultivate the experimental material and placed the seedlings in an experimental greenhouse for pre-cultivation. After 1 week (week0), the viable seedlings in plastic cubs were transferred to buckets with 10 cups for each bucket, maintaining a consistent water surface height of 40 cm throughout the experiment (a stable volume of 20 L).

Preparation of BPA Solution and BPA treatments
Since BPA is solid and hardly soluble in water, dimethyl sulfoxide (DMSO) with a mass fraction of 0.005% is used for the preparation of BPA reagents. The stock solution is made at the BPA concentrations of 0.04, 0.4, and 4 mg/ml. A safe BPA concentration limit of 1.5 mg/L for drinking water was established according to the US Environmental Protection Agency (Geens et al. 2011), which is within the concentration range commonly used in previous BPA toxicity studies on plants Seoane et al. 2021). Since BPA degrades relatively quickly in the freshwaters (Cousins et al. 2002), we added a certain volume (5 ml) of stock solution to each bucket weekly for the first three weeks targeting BPA concentrations at 0.01, 0.1 and 1 mg/L for each addition. Five treatments were set: the control without BPA and DMSO addition (CK_W), the second control without BPA but with DMSO addition (CK_D), the three treatments with DMSO, and targeting at 0.01 mg/L BPA (I), 0.1 mg/L BPA (II), and 1 mg/L BPA (III), respectively. Five replicates were set for each treatment. We defined our study as two phases: the first phase (phase I) was from May 27, 2021, to September 2, 2021 (week01-week13), to elucidate the effects of BPA on different morphological characteristics (i.e., growth traits) of plant growth; the second phase (phase II) started on September 3, 2021, when the plants first showed gender differentiation, and lasted till November 1, 2021. During the second phase, we investigated the response of reproductive traits to BPA exposure.

Determination of water chemistry
We measured the total nitrogen (TN), total phosphorus (TP), and phytoplankton chlorophyll a (Chla) contents in the water column at the start and end of the experiment. For the determination of TN and TP, the water samples were first digested with K 2 S 2 O 8 . TN was determined using a spectrophotometric method after the addition of hydrochloric acid, and TP was spectrophotometrically determined as molybdate reactive phosphorus (Ebina et al. 1983). For the determination of Chla, 500 ml of water sample was taken and filtered through a Whatman GF/C filter, then extracted in 95% ethanol and measured with a spectrophotometer (Cao et al. 2014). The results of water chemistry are shown in the Supplementary Information.

Determination of growth traits
To study the effects of low BPA loadings on the growth of submerged macrophytes, the growth traits (leaf number, number of survival individuals, ramet number, and maximum leaf length) were measured weekly in phase I with a minimizing damage to the experimental system. To be specific, the maximum leaf length is the maximum leaf length in each microcosm, not necessarily for a specific individual.
To investigate the effect of BPA on growth traits of mature individuals, five female and five male individuals from each treatment were randomly selected. At the end of the phase I, plastic cups that growing only one individual per cup were brought back to the laboratory for determination (hereafter, the traits were separately analyzed with the treatment Gender). In the laboratory, we carefully rinsed the plants with tap water and a soft toothbrush, and these plants were divided into leaves and roots to determine their leaf morphology (leaf length, leaf width, and leaf area), leaf number, ramet number, and maximum root length. The above-ground biomass, below-ground biomass, and total biomass of each part were measured after drying to constant weight at 80 °C. Dried samples were kept in sealed sterile bags. The projected (single-sided) leaf area was calculated from digital photographs with the AreaAna software (Huazhong University of Science and Technology, China), using a square of known area as a reference (Shao et al. 2017). The dried leaves were ground into powder using a high-throughput tissue grinder (Scientz-48L, Ningbo Scientz Biotechnology, China). The powder was used to determine the total carbon (TC), soluble sugar, and starch contents. The TC content was determined by total organic carbon analyzer (Vario TOC, Elementar, Germany). The soluble sugar and starch contents were determined according to Buysse and Merckx (1993) and McCready et al. (1950). Dried leaf powder of 10 mg was extracted with 4 ml of distilled water in a boiling water bath for 30 min, which was repeated twice. After centrifugation at 4000 g for 10 min, the supernatants were combined and used for the determination of soluble sugars. The residue after extraction was extracted with 2 ml of 9.2 mol/L perchloric acid for 20 min, which was also repeated twice, and the supernatant was collected after centrifugation (4000 g, 10 min) for the determination of starch. The absorbance was measured and calculated at 485 nm and 630 nm, respectively, using an ultraviolet-visible spectrophotometer (TU1810PC, Beijing Purkinje General Instrument Co., Ltd, China).

Determination of reproduction traits
To investigate of the effect of BPA on plant reproduction traits, we firstly recorded the time of the first male spathe and the first female flower blossom in each bucket. Since the number of male and female individuals in each bucket varied, to ensure consistency across treatments, we then randomly collected two male and two female plants to determine the number of fruits, the number of female flowers, and the number of spathes in each bucket. Meanwhile, some individuals were not able to be recognized as female or male, so we transferred the remaining plants to a new growth chamber to identify the gender later. The new chamber was the same size as the experimental bucket and filled only with tap water.
We collected V. natans seeds from each treatment of phase II stage and stored them at 4 °C in the moist and dark place for germination experiments. The seeds were treated by cold-wet stratification (stored at 4 °C with moisture) to relieve seed dormancy (Xiao et al. 2010). The germination trial was conducted from June 17 to July 8, 2022. For each treatment, 250 uniformly sized and full-grained seeds were selected and placed in petri dishes (50 seeds/petri dish, diameter × height: 90 mm × 15 mm) with 1 cm depth distilled water, and then incubated in light incubation room. The average temperature in the light incubation room was set at 28.9 ± 1.4 °C with a 12-h:12-h light-dark cycle (ca. 20 μmol photons m −2 s −1 ). The distilled water was replenished promptly during the experiment, and the water depth of each dish was kept consistent. The number of seed germination was recorded every other day and germination rate was calculated. The experiment was terminated after 21 days (days 01-21) when seeds ceased germinating for 5 days. The germination rate (%) was calculated by the ratio of number of germinated seeds and total number of seeds for testing.

Statistical analysis
To analyze the effect of different BPA loadings treatments during phase I (week01 ~ week13), we used linear mixed models to analyze traits of plant growth (including leaf number, ramet number, and number of survival individuals), with BPA treatments and gender as the fixed factors and time and plant ID as random factors. The maximum leaf length was analyzed with only the BPA treatment as the fixed factors as stated in "Determination of growth traits." A linear model (two-way ANOVA) was used for data sampled from mature individuals of plants, with BPA treatments and gender as the main factor. Growth traits of mature individuals included maximum leaf length, maximum leaf width, leaf area, ramet number, leaf number, maximum root length, above-ground biomass, below-ground biomass, total biomass, and physiological indicators including TC, soluble sugars, and starch.
Finally, to explore the effect of BPA on reproductive traits of V. natans in phase II, we used a linear model to compare the effect of BPA on flowering time, where BPA treatments and gender was the main factor. Also, a linear model (one-way ANOVA) was used to analyze the fruit number, number of female flowers, spathe number, and male to female ratio. Furthermore, a linear model (twoway ANOVA) was used to analyze the compound effects of BPA and gender treatments as the main factor on seed germination rate with BPA and gender treatments. The Tukey method was selected as the post hoc test, and part of these results were shown in the appendix file if not presented below. When needed, the data were log(x + 1) transformed to satisfy the assumptions of the normal distribution. All statistical analyses were performed using the R software (version 4.0.3).

Plant growth in phase I
Plant growth traits on phase I are shown in Table 1. The number of survival individuals was not different between the BPA treatments or between male and female, and both low BPA loadings treatments and gender had no significant effects on plant survival. Among all treatments, the mean maximum leaf length was highest in week05 with 53.2 ± 6.9 cm, 54.3 ± 12.6 cm, 48.5 ± 17.6 cm, 39.4 ± 6.6 cm, and 35.9 ± 9.0 cm in Table 1 Statistical results of plant growth traits in the phase I. BPA treatments are CK_W, CK_D, I II, and III. Post hoc test was conducted for BPA treatments, if needed, and shown in Fig. S1 Significance levels: NS, not significant, ***P < 0.001; **P < 0.01; *P < 0.05. The gender (female or male) of individuals determined for maximum leaf length is not fixed; thus, the two genders have not been separately shown  Fig. 1). The treatments of BPA (I), BPA (II), and BPA (III) were significantly lower than CK_W, decreasing by 20%, 30%, and 34%, respectively. Higher BPA concentrations in BPA (II) and BPA (III) showed significantly higher elongation inhibition than BPA (I), exhibiting a dose effect. Fig. 1 Growth traits of plants in phase I under different BPA treatments, including maximum leaf length (max leaf length, cm), number of leaves (leaf num, ind), and number of ramets (ramet num, ind). The w01-w13 represents week01-week13. The gender (female or male) of individuals determined for maximum leaf length is not fixed; thus, the two genders have not been separately shown Leaf number significantly increased under low BPA loadings treatments. Leaf number in BPA (III) was significantly higher than in CK_W (Fig. S1), while no significant interaction between BPA and gender treatments was found (Table 1; Fig. 1). The most significant increase in average leaf number was in BPA (III), which increased from 3 ind at the beginning of the experiment to 13 ind at the end of the Phase I, which was ca. 40% larger than CK_W.
New ramet has been observed since week07 and the ramet number gradually increased in BPA (III) by the end of the experiment, which was higher than in CK_W by a factor of 9. We found a significant difference of ramet number between the different BPA loadings treatments ( Fig. 1; F = 15.8, p < 0.01), with BPA(III) being significantly higher than BPA(I) (Fig. S1). There was no significant difference between female and male (Table 1). However, a significant interaction between BPA and gender treatments was found (Table 1). Furthermore, ramet number was significantly lower for female individuals than in male ones in BPA(III).

Growth traits of mature individuals by the end of phase I
For mature individuals, maximum leaf length was significantly reduced by BPA loadings treatments, with BPA (II) and BPA (III) being significantly lower than CK_W (Table 2; Fig. 2), with values of 39.3 ± 15.0 cm, 30.9 ± 5.2 cm, 28.24 ± cm, 25.9 ± 5.7 cm, and 26.8 ± 5.7 cm for CK_W, CK_D, BPA (I), BPA (II), and BPA (III), respectively. However, no significant differences between the gender treatments were found. Leaf number, ramet number, and leaf area were significantly affected by BPA loading treatments, which all were higher in BPA(III) than in CK_W (Fig. 2). Aboveground biomass, belowground biomass, and total biomass did not respond significantly to BPA loading treatments. These growth traits, however, were affected by the gender treatments, showing larger values for female individuals than male individuals. Both the BPA loadings and gender treatments did not significantly affect the maximum root length of V. natans.
The TC content in BPA (III) was significantly higher than that in CK_W (Table 2; Fig. 2), being 42.2 ± 2.4% VS 39.3 ± 1.4%, respectively, while no significant difference was found between BPA (I), BPA (II), and CK_W. Furthermore, significant differences in the TC content were found between two genders, with larger values in females than in males. Compared to CK_W, the soluble sugar content of BPA (III) significantly increased, with mean values of 14.7 ± 3.7 mg/g and 20.8 ± 4.9 mg/g for CK_W and BPA (III), respectively. Meanwhile there was no significant difference between CK_W and other two BPA treatments (Table 2; Fig. 2). Similarly with the TC content, females had larger soluble sugar content than males. For starch, no significant differences were found between BPA treatments or between two genders (Table 2; Fig. 2). For all three physiological indicators, there was no difference in the interaction between BPA and gender treatments.

Plant reproduction in phase II
V. natans individuals in our experiment ended with 122 females, 121 males, and 7 deaths. Among all BPA loading treatments, no significant differences were found in initial flowering time, while male individuals flowered significantly earlier than female ones ( Table 2). The spathe number was ca. 8 per individual, and the number of female flowers was ca. 3 per individual, which did not differ between the BPA loading treatments. The fruit number of V. natans averaged within 6 ~ 8 summing all replicates in each treatment and did not differ among the five BPA treatments (Table 2).  Fig. 2 Plant traits of mature individuals by the end of phase I under different BPA treatments, including maximum leaf length (max leaf length, cm), maximum leaf width (max leaf width, mm), number of leaves (leaf num, ind), number of ramets (ramet num, ind), maximum root length (max root length, mm), leaf area (cm 2 ), aboveground biomass (aboveground bio, g), belowground biomass (belowground bio, g), total biomass (total bio, g), total carbon (TC, %), soluble sugar content (soluble sugar, mg/g), and starch content (starch content, mg/g) The seeds of V. natans started to germinate from day03, and the germination rate of seeds gradually stabilized after day05. The germination rate in BPA (III) was significantly lower than in CK_W ( Fig. 3; p < 0.05). By the end of the germination experiment (day21), the germination rate in BPA (III) was only 29.6 ± 6.1% compared to 80.8 ± 5.5% in CK_W.

Discussion
Consistent with our H1, we found that low BPA concentrations affected a subset of growth traits in V. natans, with leaf number, maximum leaf length, and ramet number being sensitive to the highest BPA loading treatment (targeting at 1 mg/L), as evidenced by an inhibition of leaf elongation and a significant increase in both leaf number and ramet number. The highest BPA loading treatment resulted in detectable changes in biochemical variables (TC and soluble sugar) in V. natans; meanwhile, there was the stable total biomass across all five treatments. Regarding our H2, we found no differences in female to male ratio, initial flowering time, female flower number, spathe number, and fruit number of V. natans with respect to the BPA loadings treatments. However, the germination rate of seeds collected from the highest BPA loading treatment was significantly reduced. Finally, for our H3, we found that female individuals had less ramet than male ones at the highest BPA loading treatment, together with different TC and soluble sugar content between two genders, indicating a clear sexual dimorphism.
Previous studies have shown that extremely high concentrations of BPA (> 10 mg/L) inhibited growth traits in soybeans, with root surface area and root length being the most sensitive to BPA in addition to leaf area . This result reflected that for terrestrial plants, the organ that directly exposed to BPA (i.e. root) was relatively more sensitive. Similarly, a study by Malea et al. (2020) on the seagrass seedlings Cymodocea nodosa found that leaf elongation decreased with increasing BPA concentrations (0.03-3 μg/L). Consistently, our study found that leaves of submerged macrophytes that were directly exposed to BPA showed a very sensitive response to BPA treatments, and leaf traits responded significantly to the BPA loading even at 0.01 mg/L. In summary of these results, we discovered a sensitivity rank of leaf traits to BPA stress, which was terrestrial plant (i.e., soybeans) < submerged macrophyte (V. natans) < seagrass (C. nodosa).
Carbon is essential for plant metabolism (Huang et al. 2022), and carbon metabolism provides the main energy and essential nutrients for plants (Paul and Pellny 2003). Soluble sugars (e.g. sucrose, glucose, and fructose) have important functions throughout the plant life cycle and play an active role in plant metabolic regulation and stress tolerance (Smeekens 2000;Rosa et al. 2009). Numerous studies have reported that plants subjected to drought or salt stress have higher soluble sugar content (Foyer et al. 1998;Duby and Sing 1999), which contribute to osmoregulation by elevating the osmotic pressure of cells to maintain normal metabolic activities and play a selfprotective role (Plaxton 1996;Garcia et al. 1997). For example, the soluble sugar content of Lemna minor was significantly higher after the 5 and 20 mg/L BPA treatment (Liang et al. 2022). Similarly, we found that the highest BPA loading treatment induced an increase in carbon demand of V. natans leaves.
Aquatic plants also can reduce the effects of adverse environmental conditions through either avoidance strategies, Fig. 3 Germination rate (%) for the seeds collected from the five treatments by the end of phase II which minimize the stresses encountered, or tolerance strategies, which maximize resistance to damage (Puijalon et al. 2011). For example, Potamogeton alpinus reduces water stress in fast water flows by promoting submerged leaf elongation and producing floating leaves through the avoidance strategy (Robionek et al. 2015). Accordingly, we argued that V. natans adopted avoidance strategies to evade environmental stress through trade-offs in leaf traits, leading to an unaffected total biomass.
BPA has significant estrogenic effects on animals, interferes with normal estrogenic signaling pathways, and has deleterious effects on reproductive development and metabolism (Cano-Nicolau et al. 2016;Cariati et al. 2019;Oehlmann et al. 2000). The results of surveys and experiments on natural populations of V. natans have shown that the female to male ratio is 1:1 (Zhou et al. 2016), similar as in our study. In addition, V. natans did not differ in terms of initial flowering time, number of female flowers, number of spathes, and number of fruits with respect to the BPA exposure. As for the mature individuals, BPA did not play a decisive role in the reproduction of the selected submerged macrophytes, since plant hormones (such as gibberellins and cytokinins) are the key factors that regulate plant reproduction (Khryanin 2002). However, the seed germination was significantly reduced when the parent individuals were exposed to BPA treatments at targeting at 1 mg/L. It is to some extent a generation-delayed inhibition. Similar to our results, study on neonicotinoids, a widely used insecticide, reported that it could significantly disturb the successive reproduction of honey bee colonies without directly acute toxicity on these honey bees (Lu et al. 2012), and this so-called generation-delayed toxicity for bee triggered strict inspection on pesticide usage. Likewise, the phyto-toxicity of BPA in previous studies might be strongly underestimated according to our findings.
Sexual dimorphism is often reported in dioecious flowering plants, where female and male individuals differ between sex characteristics (reproductive organs) and secondary sex characteristics (morphological, physiological, and life history traits) (Greco et al. 2013). Previous studies on the response of V. natans to changes in different water depths have shown that female individuals showed greater investment in leaf length and biomass than male ones and that females were more plastic to environmental stress (Zhou et al. 2019). In our experiments, females had significantly higher TC and soluble sugar contents and biomass than males. However, probably due to the size limitation of our experimental microcosms, there was no difference in leaf length between females and males. Meanwhile, the number of female flowers did not differ between the BPA treatments. However, a significant interaction between the highest BPA loading treatment and the gender treatment for ramet number showed less allocation for vegetative growth (i.e., ramet) in female than male individuals. Female individuals usually bear higher reproductive costs than males due to the need to produce fruits and seeds (Hultine et al. 2016); thus, female individuals were expected to be more sensitive for the allocation of resources in stressed environments than male individuals.

Conclusion
In conclusion, BPA exposure affected some growth and reproductive traits of V. natans. (1) Plant growth traits (leaf number, maximum leaf length, ramet number, total carbon, and starch contents) responded to BPA exposure by applying an avoidance strategy through their own phenotypic plasticity.
(2) Though the macrophyte species can maintain a stable flower and fruit production, serious consequences were found for seed viability under relatively low BPA concentrations, which implied that the impact of BPA on plant reproduction were probably underestimated in previous studies. (3) The chosen species still exhibited sexual dimorphism under the influence of BPA, with female individuals being more sensitive to resource allocation in stressed environments than males. However, our study was conducted in an artificial microcosm. In this regard, more studies in real freshwater are needed in the future to determine their further impact on aquatic ecological security.