Atrazine induces complete feminization and chemical castration in male African clawed frogs (Xenopus laevis)

All of the control animals reared to sexual maturity (n = 40) were males, on the basis of external morphology, whereas only 90% of the atrazine-treated animals (36 of 40) appeared male at sexual maturity (on the basis of the presence of keratinized nuptial pads on the forearms and the absence of cloacal labia). The other 10% of atrazine-exposed animals (n = 4) lacked visible nuptial pads on the forearms and had protruding cloacal labia, typical of females (Fig. 1). Upon dissection of two of the apparent females and laparotomy in another two, we confirmed that animals with cloacal labia were indeed females from the present study, on the basis of the presence of ovaries (Fig. 1F). To date, two atrazine-induced females have been maintained, mated with control males (Fig. 1G), and produced viable eggs (Fig. 1H). The resulting larvae were all male when raised to metamorphosis and sampled (n = 100), confirming that atrazine-induced females were, in fact, chromosomal males. Furthermore, atrazine-induced females lacked the DM-W further confirming that these atrazine-induced females were indeed chromosomal males (Fig. 2). These ZZ females expressed gonadal aromatase, as did true ZW females (n = 4, from our stock colony), but ZZ males (n = 8, control or treated) did not (Fig. 2).

For methods regarding generation of sex-reversed (F1) males and F2-ZZ larvae, as well as animal husbandry, see SI Materials and Methods. ZZ larvae (F2) were reared in atrazine (2.5 ppb dissolved in ethanol) from hatching, through metamorphosis [Nieuwkoop and Faber (NF) stage 66] and throughout postmetamorphic life for comparison with control (ethanol-treated) animals. Total ethanol concentration was 0.003%. Atrazine levels and the absence of atrazine in control tanks were monitored by ELISA (Abraxis BioScience).

These analyses were conducted on sexually mature animals (2 or 3 years after metamorphosis, as indicated). For larynges, the proportion of the dilator larynges that extended below the thiohyral was determined (Fig. 3 F–H). For nuptial pads, histologic sections (8 μm) were cut through the geometric center of the nuptial pad. The maximum cross-sectional area of breeding glands was determined and compared with the maximum cross-sectional area of mucous and serous glands. For testis, the stages of spermatogenesis from five random testicular tubules and from the largest tubule from each cross-section were analyzed, as well as the proportion of testicular tubules from the largest cross-sections with and without spermatozoa bundles. For other measurements and corrections conducted, see SI Materials and Methods.

Genomic DNA was isolated from toe tips prepared by tissue lysis and proteinase k protein digestion. The ZW genotype was determined by multiplex PCR amplification (37 cycles) of DM-W (W specific) (44). The four animals determined to be female, on the basis of external morphology, were analyzed for comparison with four males. For conditions and primers, see SI Materials and Methods.

RT-PCR for cyp19 aromatase was conducted using RNA extracted from gonads of the two atrazine-treated females that were euthanized, along with four control males and four stock ZW females as controls. For conditions and primers, see SI Materials and Methods.

To compare the ability of control and atrazine-exposed males to attract females and achieve amplexus, we conducted the following behavioral studies. Males and females were marked for identification with unique single black or white stitches placed (without anesthesia) in the dorsal skin using silk thread. Stock ZW females maintained for this purpose (SI Materials and Methods) were then injected with hCG (1,000 IU) at 1500 hours. Four stock females (ZW), four control males (no hCG injection), and four atrazine-exposed males (no hCG injection) were all placed in a circular pool (diameter = 168 cm, height = 41 cm) filled with 264 L of fresh dechloraminated water. Animals were left overnight. At 0600 hours the next day (1 h before lights on), amplectant pairs were observed and animals identified under red light. Pairs and single males were removed from the pool and blood samples immediately taken from all males via cardiac puncture without anesthesia, as described previously (10). Sampling was alternated between controls and atrazine-treated males and the time of capture and time of blood draw recorded for each. Blood plasma was collected after centrifugation at 209 × g at 4 °C and stored at −20 °C. Plasma testosterone was extracted and measured by RIA as described by Hayes et al. (10). These behavioral trials were replicated five times, but data were obtained for only four owing to a failure in lighting during one trial. In all cases, all control and atrazine-treated males were virgins and had never been exposed to females. Stock females were also virgins and had never been exposed to males. In each replicate, control and atrazine-treated males were matched for size (snout–vent length and body weight) so that there were no significant differences (ANOVA: P > 0.05) in male size between groups and size had no effect on mating success (SI Materials and Methods). To examine the frequency of successful copulations, a G test (52) was used with replicate (i.e., trial) nested within treatment. For testosterone analysis, an ANOVA was conducted to examine differences in testosterone levels between control and atrazine-treated males (SYSTAT; SPSS). Data were also examined using a Kruskal-Wallis test.

To examine fertility, two studies were conducted. In the first, conducted on December 8, 2007, nine control and nine atrazine-treated males (both without hCG injections) were paired with stock ZW females (hCG-injected). Females (from the same San Diego colony) were injected and paired with individual males in aquaria (46 × 25 × 20 cm) with 10 L of fresh 10% Holtfreter’s solution and left overnight at 22 °C. At 900 hours, eggs were collected. Eggs were aerated and allowed to develop for 72 h, after which time they were fixed in Bouin’s fixative for 48 h and then preserved in 70% ethanol (two changes over 48 h). Fertility was determined by counting the number of undeveloped eggs and the number of developed embryos (NF stages 14–34). Fertility data (proportions) were arcsine transformed before ANOVA. In addition, mate choice data were subjected to analysis by G test. All males and females used in this study were virgins and had no previous exposure to the opposite sex. The animals used in this study were distinct from those used in the breeding studies above. This study was repeated on December 12, 2008 with five control and five atrazine-exposed virgin males. In both cases, the control males and the atrazine-treated males were tested in separate rooms, so that vocalizations from one group did not affect the results in the other. The animals in both of these studies were the same ones used for the histologic analyses and morphometric studies described above.