THE JOURNAL OF EXPERIMENTAL ZOOLOGY 260:130-134 (1991)
RAP1D COM MUNICAT1ON
Synergism Between Temperature and Estradiol: A Common Pathway in Turtle Sex Determination? THANE WIBBELS, J.J. BULL, AND DAVID CREWS Institute of Reproductive Biology, Department of Zoology, University of Texas at Austin, Austin, Texas 78712 ABSTRACT In many reptiles, the temperature at which the eggs are incubated determines the sex of the hatchlings. Administration of estradiol will counteract the masculinizing effects of a male-producing temperature, resulting in female hatchlings. To address whether temperature and estrogen are biologically equivalent, two experiments were conducted with the red-eared slider turtle, Trachemys scripta. In the first experiment, varying dosages of estrogen were administered at Stage 17 (the middle of the temperature-sensitive window) to eggs maintained a t two temperatures, 26°C (which normally produces all males) and 28.2"C (which produces mostly males but lies at the threshold of the transition from male- to female-producing temperatures). Results indicate that estrogen and temperature exert a synergistic effect on sex determination. In the second experiment, estrogen was administered at different stages of embryonic development. The results indicate an estrogen-sensitive period ranging from Stage 14 through Stage 21, a period similar to the temperature-sensitive period for this species. The results of these experiments are consistent with the hypothesis that temperature and estradiol act in a common pathway in temperature-dependent sex determination.
A variety and abundance of reptiles exhibit temperature-dependent sex determination (TSD) in which the incubation temperature of the embryo determines whether it hatches as male or female (Bull, '80; Raynaud and Pieau, '85; Ewert and Nelson, '91). Application of the steroid hormone estradiol causes the embryos of many reptiles to develop as female (Pieau, '74; Gutzke and Bull, '86; Bull et al., '88; Crews et al., '89, '91); some estradiol-sensitive species have TSD, but some have genotypic sex determination (GSD). As yet, there is virtually no evidence as to whether endogenous estradiol might be involved in the natural process of sex determination. Although any demonstration that endogenous estradiol is indeed involved awaits the elucidation of molecular details of the sex determination pathway, one might attempt to demonstrate that estradiol is not involved in the natural process by revealing that estradiol-induced female determination has different characteristics than does normal female determination. The present study explores two avenues for comparing the effects of temperature and estradiol on female determination in a turtle with temperature-dependent sex determination. One set of experiments investigates whether tempera@ 1991 WILEY-LISS, INC.
ture and estradiol exhibit synergism on sex determination: if the two factors operate independently of one another (i.e., fail to exhibit synergism), then i t would seem likely that they operate in different pathways. The other set of experiments identifies the estradiol-sensitive period of female determination for comparison with the temperature-sensitive period. Both results are consistent with the hypothesis that temperature and estradiol act through the same biochemical pathway.
MATERIALS AND METHODS Animals
Eggs of red-eared slider turtles, Trachemys scriptu, were obtained from a commercial supplier and placed under constant temperature incubation within 7 days of oviposition. Details of our procedures for egg handling and incubation (Bull et al., 'go), as well as the source of eggs, have been reported previously (Crews et al., '91). Synergism A group of approximately 400 eggs was divided between incubation temperatures of 26"and Received January 22, 1991; revision accepted May 14, 1991.
TEMPERATURE AND ESTRADIOL IN TURTLE SEX DETERMINATION
28.2"C within 1 4 days of oviposition. At embryonic Stage 17, estradiol-17P was applied to the upper surface of eggs in the following dosages: 0 pg (controls),0.01, 0.1, 0.5, 1.0, or 10.0 pg per egg (dissolved in 5 pl of ethanol) (Crews et al., '91). Eggs were maintained at their pre-application temperature until hatching, and the gonadal sex of turtles was determined macroscopically within 6 weeks of hatching (as described by Crews et al., '91). Estrogen-sensitiue period Within 7 days of oviposition, approximately 450 eggs were placed at 26°C for the duration of development. Estradiol was applied t o the upper surface of eggs (either 1.0 or 10.0 pg, dissolved in 5 p1 ethanol) at specific embryonic stages, and the eggs were allowed t o develop to hatching at this temperature (approximately 20-25 eggdembryonic stageldosage). Additionally, ethanol only (5 p1) was applied to eggs in control groups (n = 10 eggdgroup) at Stages 15, 17, and 20. Sex was diagnosed as above.
131
PE,PA be the increases in proportion female due to estradiol and temperature increment, respectively (measured at 26°C). Our null hypothesis (independence of temperature increment and estradiol) is
Calculation of these quantities from the experimental design, however, requires two more equations:
The experimental design provides direct estiso, all varimates of P,, PT+E,PT + A,and P T + A + E ables in (2) and (3) may be estimated. From (21, a measure of synergism (S) is
Statistics of synergism The question of interest is whether the propor- where the existence of synergism between A and tion female is increased under the joint combina- E implies S > 0, and independence implies S = 0. Although the requisite quantities may be caltion of estradiol and warmer temperature more than if the two effects operate independently. As culated from the data, the observations are samthis question is intrinsically quantitative, we re- ples and do not necessarily reflect the true values quire a mathematical model to describe these ef- of the quantities we seek (henceforth, carats ( -1 fects. Write Pi as the proportion female due to will be used t o designate sample values, whereas "treatment" z. The null model we will use is the absence of a carat will designate true values). We thus require a sampling model to evaluate the Pi+j= Pl + (1 - P,)Pj; (1) statistical significance of any observed S > 0. Our sampling model was numerical, along the followthat is, if we have two effects, i and j , whose sepa- ing lines: nl binomial variables were drawnwith rate effects are Pi and Pj, the combined effect of mean P,, n2 with mean P T + En3 , with mean P T+A , both P(i+j,is the proportion female expected as if n4 with meanPT+A+E (nl,.n2,n3,n4 being the samthey acted independently. ple sizes from the respective treatments). A value The design compares the effects of a given dos- of S was calculated from this numerical trial, and age of estradiol at 26°C versus 28.2"C. We thus this process was repeated 10,000 times. The probregard 26°C as the baseline of the experiment, ability that the confidence interval for Pi included with two treatments that each increase the pro- zero was taken as the fraction of trials for which portion female, estradiol and an increase in tem- the numerical S was less than or equal to zero. perature (an increment of 2.2"C). (To incorporate unisexual samples, Pi= 0, 1, the Let mean for the binomial distribution was chosen to be positive, as if the sample size had been ni + 1 and the n + 1st individual had been of the missp, be the proportion female at 26°C; ing sex. For 26"C, however, we let P = 0.005 in PT+, be the proportion female at 28.2"C; these trials because we have not observed a fePT+E be the proportion female at 26°C along with estradiol; male develop at 26°C in well over 200 hatchlings. PT+,+E be the proportion female at 28.2"C These procedures necessarily bias our test in faalong with estradiol; vor of the null hypothesis.)
T. WIBBELS ET AL.
132
=
1.0 p g Estradiol-17p m l O . 0 p g Estradiolk17p 100
90
y
2
80
70 60
b
50
W
40
2
CTRL
001
0 10
050
100
1000
ESTRADIOL- 17p DOSAGE(pg) Fig. 1. Effect of varying dosages of estradiol administered during incubation a t two different temperatures in the redeared slider turtle (Truchemys scrzptu). Shown is percent female hatchlings. Ratios of numbers of females to numbers of males for treatments (left to right): 0:24, 1:30, 0:24, 2:23, 4:36, 21:8, 5:19, 24:6, 17:7, 24:0, 22:0, 23:O.
RESULTS Sgnergism The combination of estradiol and incubation at 28.2”C produced more females (at most dosages) than expected from each effect independently, indicating significant synergism (Fig. 1). Specifically for dosages 0.1, 0.5, and 1.0 pg of estradiol, the observed magnitude of synergism was significantly in excess of zero (0.1 pg: S = 0.6, the confidence interval around this observation includes S = 0 with probability P < 0.0001; 0.5 pg: S = 0.57, S = 0 is included with probability P < 0.0001; 1.0 pg: 8 = 0.24, S = 0 is included with probability P < 0.005). The observed synergism was not significantly in excess of zero at the 0.01 pg dose, and the all-female sex ratios obtained at 10.0 pg precludes any determination of synergism for that dosage. Overall, therefore, there is highly significant evidence of synergism. Estrogen-sensitive period The proportions of females obtained for estradiol applications at different stages are shown separately for the 1.0 and 10.0 pg dosages (Fig. 2). Embryos exhibited a narrower period of “sensitivity” t o the lower dosage, ranging from Stages 14-20, as opposed t o Stages 11-21 at the higher dosage. All controls were male (n = 30).
DISCUSSION These results reveal that estradiol brings about female sex determination in ways similar to the effects of female-producing temperatures. First
u e: 30 w a 20 10
0 1 1 12 13 1 4 15 16 17 18 19 2 0 21
22
EMBRYONIC STAGE Fig. 2. Effect of two dosages of estradiol administered at different embryonic stages in the red-eared slider turtle (Truchemys scrzpta). Eggs were incubated at a male-producing temperature (26°C).Ratios of numbers of females to numbers of males for treatments (left to right): 18:0, 0:19,20:0, 0:15, 13:0, 8:11, 18:0, 7:9, 18:0, 11:5, 20:0, 8:11, 14:0, 1:16, 14:4, 0:7, 0:19, 0:22.
consider the synergism experiment. We know that embryos develop over approximately a 10°C range of incubation temperatures, yet (in the absence of exogenous estradiol) only males develop from 23°C to 27-28”C, and only females develop above 30°C. Thus the sex ratio changes dramatically over a short range of incubation temperatures, beginning at about 28°C. One way to describe this phenomenon is that a small increment in incubation temperature has no impact on the proportion female at low or high temperatures, but it has a major impact at 28°C. The motivation for the synergism experiment was that if estradiol is working via the same pathway as temperature, then small dosages of exogenous estradiol should have a larger effect on sex determination at 28°C than at lower temperatures. The results were consistent with this hypothesis and inconsistent with the hypothesis that the effect of estradiol was independent of incubation temperature. Although synergism between temperature and estradiol is expected if they both operate in the same pathway, it does not follow that synergism implies a common pathway. Two alternative hypotheses consistent with these results that are also consistent with estradiol operating outside of the normal sex determining pathway are that (1) estradiol is simply more biologically active at higher temperatures, and (2) the physiological processes by which estradiol acts (e.g., outside of
TEMPERATURE AND ESTRADIOL IN TURTLE SEX DETERMINATION
the normal sex determining pathway) are themselves enhanced at higher temperatures, rendering a given dosage of estradiol more effective. Hypothesis (1)is easily tested using species in which females develop at cool temperatures, males at warmer temperatures. Such a study is possible in reptiles such as the leopard gecko, Eublepharis mucularius, a species in which eggs incubated at 32.5"C normally result in mostly male offspring (Bull, '80; Bull et al., '88; Gutzke and Crews, '88). Administration of exogenous estrogen to leopard gecko eggs incubated at 325°C results in all female hatchlings (Bull et al., '88) that can develop into fertile adults (D. Crews and A. Tousignant, unpublished data). Hypotheses of type (2), however, can only be tested by learning the molecular details of normal sex determination (and estradiol effects). The second parallel between the effects of warm temperature and estradiol is revealed by a similarity of sensitive periods during embryogenesis; a similar parallel is already evident for the snapping turtle, Chelydra serpentina (Yntema, '79; Gutzke and Chymiy, '88). Prior work on red-eared sliders has indicated that temperature-sensitivity of sex determination occurs at least from Stage 15 through Stage 21 (the range of the sensitive period detected depends on the magnitude of incubation temperatures used) (Wibbels et al., '91). The present results suggest that the developmental window of estradiol sensitivity for this species is similar, lying within Stages 14-21. However, such comparisons are necessarily superficial, because we do not know the period of exposure and activity for a given application of estradiol the way that we know the period of exposure for a controlled temperature. A previous study indicates that some estradiol rapidly enters the egg after being applied t o the shell, and exogenous estradiol is still detectable in the embryo at least nine days later (Crews et al., '91). Thus, exogenous estradiol has a long "half-life" after being applied t o the eggshell, which could be due to low levels of estradiol metabolism in the egg and/ or some estradiol initially being sequestered in the eggshell and then gradually entering the egg. Comparison of the 10 c1.g and 1.0 pg results suggest that such prolonged half-life may be important, as the lower dosage reveals a much later onset of the sensitive period. From these considerations, we suggest that the onset of the estradiolsensitive period is no earlier than Stage 14 (the earliest stage at which female determination was obtained with the 1.0 pg dosage) and that the end
133
of the period is at Stage 21 or later (the last stage at which female determination was obtained with the 10.0 pg dosage). The present results motivate extension of this work t o other species. In particular, soft-shell turtles, Trionyx spinifera, lack temperature effects on sex determination yet exhibit estradiolinduced female determination (Bull et al., '88). Whether exogenous estrogens can result in a permanent sex reversal in reptiles with GSD mechanisms remains to be demonstrated. The existence of synergism between incubation temperature and estradiol in such a species would lend support t o the view that similar molecular/physiological processes underlie sex determination in species with TSD and those lacking it. Estradiol also provides a tool for understanding the period of gonadal commitment in such species, as no other means of manipulating sex determination during development has been available for such species.
ACKNOWLEDGMENTS We thank A. Alexander, F. Chan, S. Bell, and P. Kellogg for technical assistance. This research was supported by NIH grant HD24976 and NIMH Research Scientist Award 00135 to D.C., NIH NRSA training grant HD-07264 to the Institute of Reproductive Biology, and by NIH NRSA fellowship HD-07319-01A1 to T.W. LITERATURE CITED Bull, J.J. (1980) Sex determination in reptiles. Q. Rev. Biol., 55:3-21. Bull, J.J., T. Wibbels, and D. Crews (1990) Sex determining potencies vary among female incubation temperatures in a turtle. J . Exp. Zool., 256:339-341. Bull, J.J., W.H.N. Gutzke, and D. Crews (1988) Sex reversal by estradiol in three reptilian orders. Gen. Comp. Endocrinol., 70:425-428. Crews, D., J.J. Bull, and T. Wibbels (1991) Estrogen and sex reversal in turtles: A dose-dependent phenomenon. Gen. Comp. Endocrinol., 81:357-364. Crews, D., T. Wibbels, and W.H.N. Gutzke (1989) Action of sex steroid hormones on temperature-induced sex determination in the snapping turtle (Chelydru serpentinu). Gen. Comp. Endocrinol., 75:159-166. Ewert, M.A., and C.E. Nelson (1991) Sex determination in turtles: Diverse patterns and some possible adaptive values. Copeia, 199150-69. Gutzke, W.H.N., and D.B. Chymiy (1988) Sensitive periods during embryogeney for hormonally induced sex determination in turtles. Gen. Comp. Endocr., 71:265-261. Gutzke, W.H.N., and D. Crews (1988) Incubation temperature determines adult sexuality in a reptile. Nature, 332:832-834. Gutzke, W.H.N., and J.J. Bull (1986) Steroid hormones reverse sex in turtles. Gen. Comp. Endocr., 64:368-372. Pieau, C. (1974) Differenciation du sexe en fonction de la
134
T. WIBBELS ET AL.
temperature chez les embryons d’Emys orbicularis L. (Chelonien). Effets des hormones sexuelles. Ann. Embryol. Morphol., 7:365-394. Raynaud, A., and C. Pieau (1985) Embryonic development of the genital system. In: Biology of the Reptilia. Development B. C. Gans and F. Billett, eds. John Wiley & Sons, New York, Vol. 15, pp. 149-300.
Wibbels, T.R., J.J. Bull, and D. Crews (in press) Chronology of temperature-dependent sex determination. J. Exp. Zool. Yntema, C.L. (1979) Temperature levels and periods of sex determination during incubation of eggs of Chetydru serpentina. J. Morphol., 159:1’7-28.
RAP1D COM MUNICAT1ON
Synergism Between Temperature and Estradiol: A Common Pathway in Turtle Sex Determination? THANE WIBBELS, J.J. BULL, AND DAVID CREWS Institute of Reproductive Biology, Department of Zoology, University of Texas at Austin, Austin, Texas 78712 ABSTRACT In many reptiles, the temperature at which the eggs are incubated determines the sex of the hatchlings. Administration of estradiol will counteract the masculinizing effects of a male-producing temperature, resulting in female hatchlings. To address whether temperature and estrogen are biologically equivalent, two experiments were conducted with the red-eared slider turtle, Trachemys scripta. In the first experiment, varying dosages of estrogen were administered at Stage 17 (the middle of the temperature-sensitive window) to eggs maintained a t two temperatures, 26°C (which normally produces all males) and 28.2"C (which produces mostly males but lies at the threshold of the transition from male- to female-producing temperatures). Results indicate that estrogen and temperature exert a synergistic effect on sex determination. In the second experiment, estrogen was administered at different stages of embryonic development. The results indicate an estrogen-sensitive period ranging from Stage 14 through Stage 21, a period similar to the temperature-sensitive period for this species. The results of these experiments are consistent with the hypothesis that temperature and estradiol act in a common pathway in temperature-dependent sex determination.
A variety and abundance of reptiles exhibit temperature-dependent sex determination (TSD) in which the incubation temperature of the embryo determines whether it hatches as male or female (Bull, '80; Raynaud and Pieau, '85; Ewert and Nelson, '91). Application of the steroid hormone estradiol causes the embryos of many reptiles to develop as female (Pieau, '74; Gutzke and Bull, '86; Bull et al., '88; Crews et al., '89, '91); some estradiol-sensitive species have TSD, but some have genotypic sex determination (GSD). As yet, there is virtually no evidence as to whether endogenous estradiol might be involved in the natural process of sex determination. Although any demonstration that endogenous estradiol is indeed involved awaits the elucidation of molecular details of the sex determination pathway, one might attempt to demonstrate that estradiol is not involved in the natural process by revealing that estradiol-induced female determination has different characteristics than does normal female determination. The present study explores two avenues for comparing the effects of temperature and estradiol on female determination in a turtle with temperature-dependent sex determination. One set of experiments investigates whether tempera@ 1991 WILEY-LISS, INC.
ture and estradiol exhibit synergism on sex determination: if the two factors operate independently of one another (i.e., fail to exhibit synergism), then i t would seem likely that they operate in different pathways. The other set of experiments identifies the estradiol-sensitive period of female determination for comparison with the temperature-sensitive period. Both results are consistent with the hypothesis that temperature and estradiol act through the same biochemical pathway.
MATERIALS AND METHODS Animals
Eggs of red-eared slider turtles, Trachemys scriptu, were obtained from a commercial supplier and placed under constant temperature incubation within 7 days of oviposition. Details of our procedures for egg handling and incubation (Bull et al., 'go), as well as the source of eggs, have been reported previously (Crews et al., '91). Synergism A group of approximately 400 eggs was divided between incubation temperatures of 26"and Received January 22, 1991; revision accepted May 14, 1991.
TEMPERATURE AND ESTRADIOL IN TURTLE SEX DETERMINATION
28.2"C within 1 4 days of oviposition. At embryonic Stage 17, estradiol-17P was applied to the upper surface of eggs in the following dosages: 0 pg (controls),0.01, 0.1, 0.5, 1.0, or 10.0 pg per egg (dissolved in 5 pl of ethanol) (Crews et al., '91). Eggs were maintained at their pre-application temperature until hatching, and the gonadal sex of turtles was determined macroscopically within 6 weeks of hatching (as described by Crews et al., '91). Estrogen-sensitiue period Within 7 days of oviposition, approximately 450 eggs were placed at 26°C for the duration of development. Estradiol was applied t o the upper surface of eggs (either 1.0 or 10.0 pg, dissolved in 5 p1 ethanol) at specific embryonic stages, and the eggs were allowed t o develop to hatching at this temperature (approximately 20-25 eggdembryonic stageldosage). Additionally, ethanol only (5 p1) was applied to eggs in control groups (n = 10 eggdgroup) at Stages 15, 17, and 20. Sex was diagnosed as above.
131
PE,PA be the increases in proportion female due to estradiol and temperature increment, respectively (measured at 26°C). Our null hypothesis (independence of temperature increment and estradiol) is
Calculation of these quantities from the experimental design, however, requires two more equations:
The experimental design provides direct estiso, all varimates of P,, PT+E,PT + A,and P T + A + E ables in (2) and (3) may be estimated. From (21, a measure of synergism (S) is
Statistics of synergism The question of interest is whether the propor- where the existence of synergism between A and tion female is increased under the joint combina- E implies S > 0, and independence implies S = 0. Although the requisite quantities may be caltion of estradiol and warmer temperature more than if the two effects operate independently. As culated from the data, the observations are samthis question is intrinsically quantitative, we re- ples and do not necessarily reflect the true values quire a mathematical model to describe these ef- of the quantities we seek (henceforth, carats ( -1 fects. Write Pi as the proportion female due to will be used t o designate sample values, whereas "treatment" z. The null model we will use is the absence of a carat will designate true values). We thus require a sampling model to evaluate the Pi+j= Pl + (1 - P,)Pj; (1) statistical significance of any observed S > 0. Our sampling model was numerical, along the followthat is, if we have two effects, i and j , whose sepa- ing lines: nl binomial variables were drawnwith rate effects are Pi and Pj, the combined effect of mean P,, n2 with mean P T + En3 , with mean P T+A , both P(i+j,is the proportion female expected as if n4 with meanPT+A+E (nl,.n2,n3,n4 being the samthey acted independently. ple sizes from the respective treatments). A value The design compares the effects of a given dos- of S was calculated from this numerical trial, and age of estradiol at 26°C versus 28.2"C. We thus this process was repeated 10,000 times. The probregard 26°C as the baseline of the experiment, ability that the confidence interval for Pi included with two treatments that each increase the pro- zero was taken as the fraction of trials for which portion female, estradiol and an increase in tem- the numerical S was less than or equal to zero. perature (an increment of 2.2"C). (To incorporate unisexual samples, Pi= 0, 1, the Let mean for the binomial distribution was chosen to be positive, as if the sample size had been ni + 1 and the n + 1st individual had been of the missp, be the proportion female at 26°C; ing sex. For 26"C, however, we let P = 0.005 in PT+, be the proportion female at 28.2"C; these trials because we have not observed a fePT+E be the proportion female at 26°C along with estradiol; male develop at 26°C in well over 200 hatchlings. PT+,+E be the proportion female at 28.2"C These procedures necessarily bias our test in faalong with estradiol; vor of the null hypothesis.)
T. WIBBELS ET AL.
132
=
1.0 p g Estradiol-17p m l O . 0 p g Estradiolk17p 100
90
y
2
80
70 60
b
50
W
40
2
CTRL
001
0 10
050
100
1000
ESTRADIOL- 17p DOSAGE(pg) Fig. 1. Effect of varying dosages of estradiol administered during incubation a t two different temperatures in the redeared slider turtle (Truchemys scrzptu). Shown is percent female hatchlings. Ratios of numbers of females to numbers of males for treatments (left to right): 0:24, 1:30, 0:24, 2:23, 4:36, 21:8, 5:19, 24:6, 17:7, 24:0, 22:0, 23:O.
RESULTS Sgnergism The combination of estradiol and incubation at 28.2”C produced more females (at most dosages) than expected from each effect independently, indicating significant synergism (Fig. 1). Specifically for dosages 0.1, 0.5, and 1.0 pg of estradiol, the observed magnitude of synergism was significantly in excess of zero (0.1 pg: S = 0.6, the confidence interval around this observation includes S = 0 with probability P < 0.0001; 0.5 pg: S = 0.57, S = 0 is included with probability P < 0.0001; 1.0 pg: 8 = 0.24, S = 0 is included with probability P < 0.005). The observed synergism was not significantly in excess of zero at the 0.01 pg dose, and the all-female sex ratios obtained at 10.0 pg precludes any determination of synergism for that dosage. Overall, therefore, there is highly significant evidence of synergism. Estrogen-sensitive period The proportions of females obtained for estradiol applications at different stages are shown separately for the 1.0 and 10.0 pg dosages (Fig. 2). Embryos exhibited a narrower period of “sensitivity” t o the lower dosage, ranging from Stages 14-20, as opposed t o Stages 11-21 at the higher dosage. All controls were male (n = 30).
DISCUSSION These results reveal that estradiol brings about female sex determination in ways similar to the effects of female-producing temperatures. First
u e: 30 w a 20 10
0 1 1 12 13 1 4 15 16 17 18 19 2 0 21
22
EMBRYONIC STAGE Fig. 2. Effect of two dosages of estradiol administered at different embryonic stages in the red-eared slider turtle (Truchemys scrzpta). Eggs were incubated at a male-producing temperature (26°C).Ratios of numbers of females to numbers of males for treatments (left to right): 18:0, 0:19,20:0, 0:15, 13:0, 8:11, 18:0, 7:9, 18:0, 11:5, 20:0, 8:11, 14:0, 1:16, 14:4, 0:7, 0:19, 0:22.
consider the synergism experiment. We know that embryos develop over approximately a 10°C range of incubation temperatures, yet (in the absence of exogenous estradiol) only males develop from 23°C to 27-28”C, and only females develop above 30°C. Thus the sex ratio changes dramatically over a short range of incubation temperatures, beginning at about 28°C. One way to describe this phenomenon is that a small increment in incubation temperature has no impact on the proportion female at low or high temperatures, but it has a major impact at 28°C. The motivation for the synergism experiment was that if estradiol is working via the same pathway as temperature, then small dosages of exogenous estradiol should have a larger effect on sex determination at 28°C than at lower temperatures. The results were consistent with this hypothesis and inconsistent with the hypothesis that the effect of estradiol was independent of incubation temperature. Although synergism between temperature and estradiol is expected if they both operate in the same pathway, it does not follow that synergism implies a common pathway. Two alternative hypotheses consistent with these results that are also consistent with estradiol operating outside of the normal sex determining pathway are that (1) estradiol is simply more biologically active at higher temperatures, and (2) the physiological processes by which estradiol acts (e.g., outside of
TEMPERATURE AND ESTRADIOL IN TURTLE SEX DETERMINATION
the normal sex determining pathway) are themselves enhanced at higher temperatures, rendering a given dosage of estradiol more effective. Hypothesis (1)is easily tested using species in which females develop at cool temperatures, males at warmer temperatures. Such a study is possible in reptiles such as the leopard gecko, Eublepharis mucularius, a species in which eggs incubated at 32.5"C normally result in mostly male offspring (Bull, '80; Bull et al., '88; Gutzke and Crews, '88). Administration of exogenous estrogen to leopard gecko eggs incubated at 325°C results in all female hatchlings (Bull et al., '88) that can develop into fertile adults (D. Crews and A. Tousignant, unpublished data). Hypotheses of type (2), however, can only be tested by learning the molecular details of normal sex determination (and estradiol effects). The second parallel between the effects of warm temperature and estradiol is revealed by a similarity of sensitive periods during embryogenesis; a similar parallel is already evident for the snapping turtle, Chelydra serpentina (Yntema, '79; Gutzke and Chymiy, '88). Prior work on red-eared sliders has indicated that temperature-sensitivity of sex determination occurs at least from Stage 15 through Stage 21 (the range of the sensitive period detected depends on the magnitude of incubation temperatures used) (Wibbels et al., '91). The present results suggest that the developmental window of estradiol sensitivity for this species is similar, lying within Stages 14-21. However, such comparisons are necessarily superficial, because we do not know the period of exposure and activity for a given application of estradiol the way that we know the period of exposure for a controlled temperature. A previous study indicates that some estradiol rapidly enters the egg after being applied t o the shell, and exogenous estradiol is still detectable in the embryo at least nine days later (Crews et al., '91). Thus, exogenous estradiol has a long "half-life" after being applied t o the eggshell, which could be due to low levels of estradiol metabolism in the egg and/ or some estradiol initially being sequestered in the eggshell and then gradually entering the egg. Comparison of the 10 c1.g and 1.0 pg results suggest that such prolonged half-life may be important, as the lower dosage reveals a much later onset of the sensitive period. From these considerations, we suggest that the onset of the estradiolsensitive period is no earlier than Stage 14 (the earliest stage at which female determination was obtained with the 1.0 pg dosage) and that the end
133
of the period is at Stage 21 or later (the last stage at which female determination was obtained with the 10.0 pg dosage). The present results motivate extension of this work t o other species. In particular, soft-shell turtles, Trionyx spinifera, lack temperature effects on sex determination yet exhibit estradiolinduced female determination (Bull et al., '88). Whether exogenous estrogens can result in a permanent sex reversal in reptiles with GSD mechanisms remains to be demonstrated. The existence of synergism between incubation temperature and estradiol in such a species would lend support t o the view that similar molecular/physiological processes underlie sex determination in species with TSD and those lacking it. Estradiol also provides a tool for understanding the period of gonadal commitment in such species, as no other means of manipulating sex determination during development has been available for such species.
ACKNOWLEDGMENTS We thank A. Alexander, F. Chan, S. Bell, and P. Kellogg for technical assistance. This research was supported by NIH grant HD24976 and NIMH Research Scientist Award 00135 to D.C., NIH NRSA training grant HD-07264 to the Institute of Reproductive Biology, and by NIH NRSA fellowship HD-07319-01A1 to T.W. LITERATURE CITED Bull, J.J. (1980) Sex determination in reptiles. Q. Rev. Biol., 55:3-21. Bull, J.J., T. Wibbels, and D. Crews (1990) Sex determining potencies vary among female incubation temperatures in a turtle. J . Exp. Zool., 256:339-341. Bull, J.J., W.H.N. Gutzke, and D. Crews (1988) Sex reversal by estradiol in three reptilian orders. Gen. Comp. Endocrinol., 70:425-428. Crews, D., J.J. Bull, and T. Wibbels (1991) Estrogen and sex reversal in turtles: A dose-dependent phenomenon. Gen. Comp. Endocrinol., 81:357-364. Crews, D., T. Wibbels, and W.H.N. Gutzke (1989) Action of sex steroid hormones on temperature-induced sex determination in the snapping turtle (Chelydru serpentinu). Gen. Comp. Endocrinol., 75:159-166. Ewert, M.A., and C.E. Nelson (1991) Sex determination in turtles: Diverse patterns and some possible adaptive values. Copeia, 199150-69. Gutzke, W.H.N., and D.B. Chymiy (1988) Sensitive periods during embryogeney for hormonally induced sex determination in turtles. Gen. Comp. Endocr., 71:265-261. Gutzke, W.H.N., and D. Crews (1988) Incubation temperature determines adult sexuality in a reptile. Nature, 332:832-834. Gutzke, W.H.N., and J.J. Bull (1986) Steroid hormones reverse sex in turtles. Gen. Comp. Endocr., 64:368-372. Pieau, C. (1974) Differenciation du sexe en fonction de la
134
T. WIBBELS ET AL.
temperature chez les embryons d’Emys orbicularis L. (Chelonien). Effets des hormones sexuelles. Ann. Embryol. Morphol., 7:365-394. Raynaud, A., and C. Pieau (1985) Embryonic development of the genital system. In: Biology of the Reptilia. Development B. C. Gans and F. Billett, eds. John Wiley & Sons, New York, Vol. 15, pp. 149-300.
Wibbels, T.R., J.J. Bull, and D. Crews (in press) Chronology of temperature-dependent sex determination. J. Exp. Zool. Yntema, C.L. (1979) Temperature levels and periods of sex determination during incubation of eggs of Chetydru serpentina. J. Morphol., 159:1’7-28.
