Temperature-Dependent Sex Determination in Reptiles: Proximate Mechanisms, Ultimate Outcomes, and Practical Applications

DEVELOPMENTAL GENETICS 15297312 (1994)

DAVID CREWS, JUDITH M. BERGERON, JAMES J. BULL, DEBORAH FLORES, ALAN TOUSIGNANT, JAMES K. SKIPPER, AND THANE WIBBELS Institute of Reproductive Bio1og.v and the Department of Zoology, University of Texas at Austin, Austin

ABSTRACT In many egg-laying reptiles, INTRODUCTION the incubation temperature of the egg determines Much progress has been made in our understanding the sex of the offspring, a process known as tem- of sex determination and differentiation in a variety of perature-dependent sex determination (TSD). In organisms. Indeed, until about 25 years ago, it was TSD sex determination is an “all or none” process assumed that all amniote vertebrates (mammals, and intersexes are rarely formed. How is the exter- birds, and reptiles) shared similar genotypic sex-deter- nal signal of temperature transduced into a genetic mining (GSD) mechanisms. Then in 1966 Madeline signal that determines gonadal sex and channels Charnier discovered that in the egg-laying lizard Ag- sexual development? Studies with the red-eared arna agama the incubation temperature of the egg de- slider turtle have focused on the physiological, bio- termines the sex of the hatchling. This process of tem- chemical, and molecular cascades initiated by the perature-dependent sex determination (‘ED) has now temperature signal. Both male and female devcl- been demonstrated in many turtles, some lizards, and opment are active processes-rather than the or- all crocodilians [Bull, 1980; Ewert and Nelson, 1991; ganized/defadt system characteristic of verte- Janzen and Paukstis, 19911. brates with genotypic sex determination-that In mammals, birds, and many other gonochoristic require simultaneous activation and suppression of vertebrates (separate sexes in separate individuals), testis- aqd ovary-determining cascades for normal gonadal sex is determined at fertilization by specific sex determination. It appears that temperature ac- chromosomes and results in a 1:1 sex ratio. In mam- complishes this end by acting on genes encoding mals, a gene on the Y chromosome (designated Sry) for steroidogenic enzymes and steroid hormone re- channels gonadal development to result in testes and ceptors and modifying the endocrine microenvi- the individual to develop a male-typical phenotype; the ronment in the embryo. The temperature experi- absence of this gene results in the formation of ovaries enced in development also has long-term and a female-typical phenotype (Fig. 1, top). Thus, the functional outcomes in addition to sex determina- Sry gene serves to switch the developmental cascade tion. Research with the leopard gecko indicates leading to maleness. However, Sry is part of a large that incubation temperature as well as steroid hor- family of genetic transcripts, and comparative studies mones serve as organizers in shaping the adult in a variety of species have not found other Sry-like phenotype, with temperature modulating sex hor- genes to be sex-linked in nonrnammals [Tiersch et al., mone action in sexual differentiation. Finally, prac- 1991, 19921. tical applications of this research have emerged for Unlike GSD, where it is not possible to deviate the the conservation and restoration of endangered primary sex ratio from unity [Beamer and Whitten, egg-laying reptiles as well as the embryonic devel- 19911, species with TSD do not inherit sex-specific opment of reptiles as biomarkers to monitor the chromosomes from their parents, and gonadal sex is estrogenic effects of common environmental contaminants. G 1554 Wiley-Liss, Inc.

Key words: Sex determination, sexual differenti- Received for publication October 25, 1993; accepted January 12, 1994. ation, reptiles, temperature-dependent sex determination, behavior, steroidogenic enzymes, aro- Address reprint requests to Dr. David Crcws, Department of Zoology, University of Texas at Austin, Austin, TX 78712. matase, reductase, estrogen, androgen, steroid hormone receptors Dr. Wibbels is now at the Department of Biology, University of Ala- bama at Birmingham, Birmingham, AL 35294.

0 1994 WILEY-LISS, INC. 298 CREWS ET AL. GENOTYPIC SEX DETERMINATION Fertilization

Gonad Sexual Gonad Determining +Hormones + Differentiation -aFormation Genes of Phcnotypc

TEMPERATURE-DEPENDENT SEX DETERMINATION Fertilization Temperature

Gonad Ihn/.yme.! Gonad Sexual 4-b ~Orm~~s+ Determining + + Hormones + Differentiation Receptors Gcnes of Phenotype

Fig. 1. In the current model of vertebrate sex determination and not irrevocably set by the genetic composition inherited at fertiliza- sexual differentiation (top), gonadal sex is fixed at fertilization by tion, but rather depends ultimately on which genes encoding for ste- specific chromosomes, a process known as genotypic sex determina- roidogenic enzymes and hormone receptor arc activated during devel- tion. Only after the gonad is formed do hormones begin to exert an opment by temperature. Incubation temperature modifies the activity influence, modifying specific structures that eventually will differ as well as the temporal and spatial sequence of enzymes and hormone between the sexes. Research on reptiles with temperature-dependent receptors such that sex-specific hormone milieus, created in the uro- sex determination indicates that sex determination in these species is genital system of the developing embryo, determine gonad type. fundamentally different in at least one way (bottom). Gonadal sex is not determined at fertilization. In TSD each individual threatened and endangered species and monitoring en- has equal potential to become male or female and there vironmental contamination. is little, or no, genetic predisposition for specific responses to incubation temperature. Rather, incubation temperature serves as a switch that initiates the cas- PROXIMATE MECHANISMS OF cade that leads to the development of testes and sup- TEMPERATURE-DEPENDENT presses the cascade that leads to the development of SEX DETERMINATION ovaries (and vice versa) (Fig. 1, bottom). How is the physical stimulus of temperature trans- Reptiles with TSD may offer model systems to better duced into a physiological signal that ultimately acts understand the events that comprise sex determination on a molecular switch to determine sex? We have used and sexual differentiation in amniote vertebrates. the red-eared slider (Trachemys scriptu) to address this Mammals and birds share their most recent common question. In this species, an incubation temperature of ancestry with reptiles. It has been suggested that en- 26°C produces all males, 31°C produces all females, and vironmental sex determination may have been the pre- 29.2”C is the threshold temperature that produces ap- cursor to GSD [Ohno, 1967; Janzen and Paukstis, proximately a 1:l sex ratio (Fig. 2). The following four 19911. Before such promise can be realized, basic infor- points have been established in this and other species: mation both on how sex is determined as well as how sexuality develops in TSD species is required. Our pur- 1. Sex determination is sensitive to both the dura- pose here is to summarize a set of studies conducted in tion and magnitude of incubation temperature [Wib- this laboratory on two species, the red-eared slider and bels et al., 1991al. the leopard gecko. An obvious starting point is to de- 2. Sex determination remains labile to temperature termine how TSD is achieved. We consider the physi- changes through the early stages of gonadal differen- ological, cellular, and molecular mechanisms of TSD. A tiation and initial sex-specific changes in the gonads second issue concerns the functional outcomes of TSD are reversible [Wibbels et al., 1991a,b]. The tempera- on characteristics other than gonad phenotype. We ture-sensitive period corresponds to the middle of em- demonstrate how the temperature experienced during bryogenesis ( = sex determination period), beginning embryonic development modifies the morphology, prior to the differentiation of the gonads into recogniz- physiology, and behavior of the adult. Finally, we pro- able ovaries or testes but ending well before hatching vide new information on some practical applications of [Wibbels et al., 1991al. this research such as benefiting the conservation of 3. Temperature exerts an “all-or-none” effect on the MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 299

'0 %%$@, %%$@, Estradiol Testosterone < ,& D i h ydrote s t 0st e r on e

Fig. 3. Biosynthesis of testosterone to dihydrotestosterone (via reductase) and estradiol (via arornatase).

of other androgens. We assessed the activity of 3a-, 3p-, and 17P-HSDH enzymes before, during, and after the temperature-sensitive period in the red-eared slider turtle. HSDH reaction product was observed in the liver, adrenal gland, and kidney before and during the 0 temperature-sensitive window at either incubation 26 27 28 29 30 31 32 temperature and in the genital ridge during the last Temperature ("C) trimester of development [Thomas et al., 19921. Similar findings have been reported for the green sea turtle Fig. 2. Relationship between incubation temperature and sex ratio (Chelonia mydas) [Merchant-Larios et al., 19891 and in the red-eared slider turtle. Depicted is the mean sex ratio obtained the European pond turtle (Emys orbicularis) [Pieau, over a 4-year period. Total number of individuals followed by the 19731. number of replicate experiments are shown in parentheses; standard error indicated by vertical bar. Stipple indicates transitional temper- Testosterone (T) is the precursor of both dihydrotes- ature. tosterone (DHT) and estradiol (E2), being converted by reductase to DHT and by aromatase to E2 (Fig. 3). The ontogenetic pattern in aromatase activity parallels ovarian or testicular nature of the gonad; intersex in- that of HSDH enzymes. In the European pond turtle, dividuals are rarely produced (<,001 incidence) [Crews aromatase activity is greatest in embryos incubating at et al., 19911. a female-producing incubation temperature compared 4. During the window of temperature sensitivity, to embryos incubating at a male-producing tempera- the effect of temperature appears to be cumulative; the ture; indeed, there appears to be a positive feedback longer an embryo is maintained at a particular tem- between E2 and aromatase at a female-producing temperature, the more extreme the new temperature must perature [see review by Pieau et al., 19941. Further- be to override the effects of the original temperature more, initial studies show aromatase levels to be [Wibbels et al., 1991bl. higher before, compared to after, the temperature-sensitive period in the kidney-adrenal-gonadal complex Our studies explore one general model suggested of the red-eared slider [J. Bergeron, A. Wozniak, J. originally by Claude Pieau: incubation temperature Hutchison, and D. Crews, unpublished data]. No infor- determines sex by altering the hormone environment mation is available on the ontogenetic pattern of re- of the sexually indifferent embryo, channeling devel- ductase activity. opment into a male or a female direction. Specifically, If incubation temperature modifies the steroid mi- incubation temperature acts on the expression of genes crocnvironment in the embryonic turtle, which in turn encoding for steroidogenic enzymes and steroid hor- determines gonadal sex, then the nature and pattern of mone receptors, which together guide the differentia- steroid hormone secretion may differ with incubation tion of the embryonic gonad into testes or ovaries (Fig. temperature. Dorizzi et al. 119911 found higher estro- 1, bottom). What is the evidence supporting this hy- gen content in gonads of European pond turtle embryos pothesis? In oviparous vertebrates, the yolk is a rich thaL were incubating at a female-producing incubation repository of hormones and their precursors [Bern, temperature compared to embryos incubating at a 19901. These precursor molecules are modified by ste- male-producing temperature during the temperature- roidogenic enzymes in the biosynthesis of androgens sensitive window. In the red-eared slider turtle, the and estrogens. If incubation temperature modulates steroid content in the urogenital tissues (kidney-adre- the nature, quantity, and activity of steroidogenic en- nal-gonad) of embryos incubated at male- and female- zymes in the embryo, these enzymes could act on ste- producing temperatures differs [White and Thomas, roid precursors to produce temperature-specific hor- 1992a1. This finding suggests that the onset of ste- monal milieus in the embryo. Hydroxysteroid roidogenesis occurs only after the onset of the temper- dehydrogenase (HSDH) enzymes are involved in tes- ature-sensitive period and not before. Consistent with tosterone and estrogen biosynthesis (36-HSDH and this interpretation is the finding that in in vitro exper- 176-HSDH) and 3a-HSDH is involved in the synthesis iments, ovine follicle-stimulating hormone (FSH) fails 300 CREWS ET AL. to stimulate steroidogenesis in isolated gonads until Orthogonal Model after the temperature-sensitive period but does elicit Mascb,ln,zatlon + - &masculinization +,A\ + significant responses in the kidney and adrenal before I__\ or during the temperature-sensitive period [White and defeminizatlon- + Feminization + Thomas, 1992131. I--\

Unidimensional Model FUNCTIONAL OUTCOMES OF A

1’- - - ~

TEMPERATURE-DEPENDENT Masculinization ( ~ - ~~ J 3 1 Feminization SEX DETERMINATION - Since sex determination in TSD depends fundamen- Fig. 4. Two models of sexual differentiation in vertebrates. (Top) tally on incubation temperature, is the process of sex- Orthogonal model that posits that the dimensions of masculinity and ual differentiation similarly affected? Before this ques- femininity are separate and independent. This model applies to species having genotypic sex determination. (Bottom) Unidimensional tion can be answered, it is necessary first to review model which posits that masculinity and femininity are at opposite briefly our understanding of sexual differentiation in ends of the same continuum and hence functionally linked. This vertebrates. It has been established that in mammals, model may apply to species with temperature-dependent sex determi- androgenic steroid hormones along with other protein nation. hormones (e.g., miillerian inhibiting substance) produced by the embryonic testes, cause the development of male-typical morphological, physiological, and be- feminization, or feminization and corresponding de- havioral traits; if testes are not formed, a female phe- masculinization. For such a model to have heuristic notype results. Thus, in various mammals, perinatal value in understanding sex differences in TSD species, castration of genetic males results in adults having two important issues must be addressed. First, are in- female-typical morphology, physiology, and behavior, dividuals from temperatures that produce both males whereas perinatal androgen treatment of genetic fe- and females sexually equivalent to those individuals males produces adults with a masculinized phenotype. from incubation temperatures that produce all females The modern perspective of sexual differentiation was or all males? That is, are all females (or males) similar developed from work on species exhibiting genotypic regardless of incubation temperature? Second, is the sex determination (GSD). Initially the embryo is bipo- temperature that produces one sex more potent than tential, having the properties of both sexes. During de- the temperature that produces the opposite sex? In velopment of the male the wolffian ducts develop into other words, could a certain temperature serve as an the epididymides and the vasa deferentia and the miil- organizer in TSD species in a fashion similar to the role lerian ducts degenerate; in the case of female develop- of steroid hormones in GSD species? ment, the mhllerian ducts become the oviducts and Because it takes 5-7 years for the red-eared slider uterus and the wolffian ducts degenerate. An analogy turtle to reach reproductive maturity, we have em- often is made between the development of the urogen- ployed the leopard gecko (Eublepharis rnauularius) as ital duct system and the brain circuitry underlying sex- an animal model for the study of the functional out- ual behavior. That is, there exists in each individual comes of TSD. This lizard species is of moderate size dual neural substrates underlying sexual behavior, one (40-90 a), reaches sexual maturity at approximately subserving male-typical behaviors such as mounting 45 weeks of age, and is fully grown at 65 weeks of age. and intromission behavior and one subserving female- Incubation of leopard gecko eggs at 26°C produces only typical behaviors such as sexual receptivity. Like the female hatchlings, 30°C produces a female-biased sex duct systems, the development of these structures is ratio (80:20), and 325°C produces a male-biased (20:80) differentially affected during development by gonadal sex ratio. An incubation temperature of 35°C (at or hormones. Thus, in mammals, the genetic male is said near the lethal maximum) again produces only females to be masculinized and defeminized and the genetic [Viets et al., 19931. By studying animals that differ female as being feminized and demasculinized (Fig. 4, only in (1) incubation temperature experienced, as well top). This has been termed the orthogonal model of as (2) their prenatal, (3) and/or postnatal hormonal ma- sexual differentiation. nipulation, it is possible to determine the extent to Given the “all-or-none’’ nature of TSD, it is evident which differences between and within the sexes are due that in such species maleness does not develop inde- to incubation temperature and to what extent they are pendently of femaleness (and vice versa) and that we due to the individual’s gonadal sex. For example, fe- must envision another model of sexual differentiation. males determined by incubation temperature can be Originally, sexual differentiation, or masculinity and compared to females determined by estrogen at male- femininity, were viewed as representing ends of a sin- producing temperatures. gle continuum (Fig. 4, bottom); in such a unidimen- Adult leopard geckos exhibit marked sexual dimor- sional model individual differences would arise from phisms in morphology, physiology, and behavior. For the degree of masculinization and corresponding de- example, specialized secretory pores anterior to the clo- MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 301 aca are open in males; in females from low incubation midpoint in the temperature-sensitive period in this temperatures they are closed, whereas in females from species [Wibbels et ul., 1991aI. Eggs are then random- higher, male-biased incubation temperatures, they are ized into control and experimental treatments. In all open. The onset of sexual maturity and plasma hor- experiments, eggs in control groups at each incubation mone levels in leopard geckos varies in a similar fash- temperature receive a single treatment consisting of 5 ion as a function of the incubation temperature of the p.1 of 95% ethanol; inhibitor and/or ligand is also dis- egg. Females from a male-biased incubation tempera- solved in 5 ~1 of 95% ethanol. After receiving treat- ture take longer to reach sexual maturity [Tousignant ments, all eggs are returned to their respective incu- and Crews, 1994al and have significantly lower circu- bators until they hatch. Otherwise methods are as lating concentrations of E2 and significantly higher T described in the text and figure legends or in the ref- [Gutzke and Crews, 19881 levels. Males from a female- erences therein. biased incubation temperature have significantly Turtles are euthanized within two weeks after hatch- higher levels of E2 [Tousignant and Crews, 1994al as ing. Gonadal sex and developmental status of the Mul- adults than do males from a male-biased incubation lerian ducts are assessed by examination under a dis- temperature. section microscope. The gonads of hatchling red-eared This spectrum of effects evident in females from slider turtles are relatively well differentiated and, male-biased incubation temperatures is reminiscent of with rare exception, appear distinctly testicular or ova- the well-known masculinizing effects following admin- rian when viewed under a dissection microscope istration of exogenous androgens on neonatal female [Crews et al., 1991; Wibbels et al., 1991al. Ovaries are mammals and the more recently discovered intrauter- long and flat whereas testes are shorter, more round, ine position effect in polytocous mammals. In mam- and have visible seminiferous tubules [Crews et al., mals, the embryonic environment influences behav- 19911. The developmental status of both the cranial ioral expression in adults. Sociosexual behavior of and the caudal halves of the mullerian ducts is exam- adults varies according to intrauterine positioning of ined and scored as either absent, regressed but visible, male and female fetuses and is a consequence of expo- normal (as in a typical female hatchling), or hypertro- sure to the hormones from fetal neighbors. For exam- phied. A phallus is also noted if present. In the initial ple, in rats, mice, and gerbils, females situated next to studies all gonads were processed for histological ex- males in utero (2M) have a masculinized adult pheno- amination [Crews et al., 1991; Wibbels et al., 1991 a]. In type, compared to females situated next to females all instances, the histological assessment of sex has (OM) [vom Saal, 19811. Furthermore, 2M females ex- coincided with macroscopic assignment of sex. In sub- hibit a higher intensity of aggression toward female sequent studies, the gonads of three to five individuals stimulus animals and a decrease in attractivity to male from each experimental group are processed for histo- stimulus animals, compared to OM females. In the ger- logical examination of the gonad to confirm sex assign- bil, intrauterine position affects the age of maturity of ment. females; OM females mature early whereas 2M females mature late [Clark et aZ., 19881. Early-maturing (OM) RESULTS AND DISCUSSION females have lower androgen levels and higher estradiol levels than those of late-maturing (2M) females Proximate Mechanisms of TSD [Clark et al., 19911. Could an incubation temperature Is there evidence that steroid hormones are the that produces predominantly males act in an analogous physiological equivalent of temperature in the sex fashion by altering the neuroendocrinology of the fe- determination cascade? We have examined the possi- males born at these temperatures? ble role of steroid hormones in both female and male sex determination, although most of our work to date has centered on estrogen and female sex determina- MATERIALS AND METHODS tion. Freshly laid turtle eggs are obtained commercially Estrogen-sensitivity is correlated with tempera- (Robert Kliebert, Hammond, LA). After transport to ture-sensitivity. Administration of exogenous estro- our laboratory, eggs are held at room temperature un- gen will override the effect of a male-producing incu- til viability is established by candling. They are then bation temperature so that all of the hatchlings will be placed in containers with moistened vermiculite (ver- female [Crews et al., 1989, 19911. The period of sensi- miculite:water, 1:l) and the containers placed in tivity to E2 coincides with the mid-trimester window of reach-in incubators (Precision) programmed to provide temperature-sensitivity; only in this period are incipi- a constant temperature. Embryonic development is ent males induced to become female [Wibbels et al., monitored by candling eggs and by dissecting 2-4 eggs 1991al. Also, the feminizing ability of E2 varies with approximately twice a week to verify specific develop- incubation temperature. Incubation temperature and ment stages, based on criteria described by Yntema E2 synergize, with E2 exerting a significantly more 119681. In studies involving experimental treatment, potent effect as the threshold temperature is ap- eggs are incubated until stage 17, the approximate proached (Fig. 5). Finally, the morphological changes 302 CREWS ET AL.

100

I t .40 a, aa, .20

LO Temperature ("C)

Fig. 5. Interaction between incubation temperature and exogenous Fig. 6. Administration of the aromatase inhibitor Ciba Geigy estradiol in determining female development in the red-eared slider 16949A induces male development. In all studies eggs were spotted turtle. Plotted is the female-inducing ability of exogenous estrogen with the chemical in alcohol or with alcohol only. Sample sizes aver- (see Wibbels et al., 1991, for mathematical formula). In all studies, age 25 individuals, ranging from 21 to 28; total number of individuals eggs were spotted with hormone in alcohol or with alcohol only. Sam- depicted = 623. ple sizes average 26 individuals, ranging from 22 to 40; total number of individuals depicted = 706. Not all cells filled because certain groups were not tested. Cells that appear to counter general trend are within experimental error. sponse is temperature sensitive, with lower dosages required as the threshold incubation temperature is approached. Similarly, administration of T and Ciba induced by exogenous estrogen administration are Geigy 16949A to eggs incubating at 31°C results in identical to those induced by incubation temperature 100% of the hatchlings being male. [Wibbels et al., 19931. Estrogen is concentrated in specific tissues and Estrogen response is specific to estrogen-like cells. Estrogen target areas during the temperature- compounds. If E can act as a physiological parallel to sensitive period have been identified using autoradiog- female-producing temperature in TSD, does it do so by raphy [Gahr et al., 19921. Hyperfilm autoradiography a hormone receptor mechanism and is it specific to es- shows uptake primarily in the liver, the kidney-adre- trogen? We have evaluated the effect of a variety of nal-gonad area, and the bony structures before, dur- steroid hormones and their agonists and antagonists at ing, and after the period of sex determination in both male- and female-producing temperatures. Em- embryos incubated at female- or male-producing tem- bryos incubated at a male-producing temperature are peratures. Emulsion autoradiography reveals that dur- feminized by the estrogen agonists diethylstilbestrol ing and after the temperature-sensitive window, la- (DES) and R2858, but not by the androgen agonist beled cells were found primarily in the liver, kidney, R1881 [Wibbels and Crews, 19921. Interestingly, treat- adrenal, and oviduct. Labeled cells in the gonad are ment of eggs incubating at a male-producing tempera- distributed evenly throughout the medulla and the cor- ture with T also feminizes approximately one-half the tex. individuals, presumably through the aromatization of Distribution and quantity of estrogen-respon- T to E2; at a female-producing temperature, compara- sive target cells are known. It is assumed, but re- ble dosages of T or E have no discernible effect. Exog- mains to be proved, that the estrogen-concentrating enous progesterone, corticosterone, and DHT do not cells identified by autoradiography contain estrogen feminize embryos. These results suggest that steroid- receptor (ER) and not simply estrogen-binding protein. induced feminization is mediated via an estrogen-spe- Our first effort at documenting this was an instructive cific receptor. failure. In ten turtle species tested, spanning six fam- Aromatase inhibitors will block female develop- ilies, there was no immunoreactivity to mammalian ment. The biosynthesis of E2 depends on the aromati- monoclonal antibodies (e.g., H222Spy and H226Spy), zation of T to E (Fig. 3). Administration of either of two although these same antibodies did react with embry- aromatase inhibitors (Ciba Geigy CGS 16949A or CGS onic and adult tissue from various lizards and the al- 20267) to eggs incubating at female-producing temper- ligator [M. Gahr and D. Crews, unpublished data]. atures will cause male development (Fig. 6). This re- Since we were unable to find an available antibody MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 303

DNA Binding Hinge Hornnone Binding Domain Region Domain

Chicken Ill IWIIIII II 111

Rat III~IIIII II II Ill1

Human fl1111 I I I I Ill

Consensus - M - K - X R Q 2 - - - ~ TURTLE rMlKhKRQReeqd ,;;;;; ITL Chic ken 250 q:G e M K q K R Q R e e q d Rat 261 GGrMl KhKRQRddle Human 256 RKDRRGGrMlKhKRQRddge

Consensus -R------R --__-- SP TURTLE aR-tagpstemRsphpltSP ChLcken 270 ~RngeasstelRaptlwtSF Rat 281 gRnemgts gdmRd arilwp S P Human 276 gRgevgsagdmRaanlwpSP

Consensus L--K--KKNS-ALSL,TA-QM 'TURTLE LviKhsKKNSpALSLTAeQM Chicken 290 LvvKhnKKNSpALSLTAeQM Rat 301 LviKhtKKNSpALSLTAdQM Human 295 LniKrsKKNSlALSLTAdQM

Fig. 7. Schematic of turtle estrogen receptor (tER) gene. For these tER with subclones indicating regions sequenced to date below, The studies, mRNA was isolated from oviduct target tissue of an adult hatch marks in the middle portion are a graphic representation of red-eared slider. With PCR primers designed to the most conserved regions where tER differs from chicken, rat, or human ER. The lower regions of the DNA- and hormone-binding domains of ER gene, DNA portion presents the deduced amino acid sequence for tER compared fragments were amplified from the cDNA prepared from this mRNA. to the corresponding regions of published sequences for ER of other These products (three independent clones each of which spans 0.9 kb) species. Boxed segment of the sequence signifies the hinge region. were subcloned and, based on the nucleic acid composition, the amino Dashes in the consensus sequence indicate nonconserved amino acids. acid sequence determined. Top portion shows functional domains of

that recognized turtle estrogen receptor (tER), an al- Second, it reveals great differences within the hinge ternative strategy was undertaken involving the region between the two domains; only 45% similarity cloning and sequencing of tER (Fig. 7). Comparison of exists. the sequence of the tER is interesting in two regards. Using nucleotide probes, we have localized and quan- First, it affirms the high degree (>go%) of similarity tified tER-mRNA during the critical periods of sex de- with the published sequences of the DNA- and steroid- termination. In situ hybridization located cells contain- binding domains of the human, rat, and chicken ER. ing tER in the liver, adrenal, kidney, oviduct, and the 304 CREWS ET AL.

Fig. 8. In situ hybridization of ER-mRNA in the developing gonad clone MGPCR3C2, as shown in Figure 7. Tissues are 20 micron cryo- in embryos incubating at a male- and a female-producing tempera- sections, fixed immediately after sectioning. Hybridization was per- ture. Light-field micrography (upper left) of testis and corresponding formed in 50% formamide, 10% dextran sulfate, 0.3M NaCI, 10 mM dark-field image indicating gene transcript (upper right). Light-field Tris-Cl IpH 8.0), 1 mM EDTA, and 1 x Denhardt's solution at 45°C. micrography (lower left1 of ovary and corresponding dark-field image Slides were washed at 50°C in sodium citrate solutions of increasing indicating gene transcript (lower right). "S-labeled antisense ER ri- stringency, prior to a 2-week exposure using emulsion. boprobe was generated using T7 polymerase transcription of cDNA

gonad (Fig. 8). We hypothesize that the mechanism of determination, and the relationship between incuba- action of temperature in TSD includes the regulation of tion temperature and DHT. the genes coding for steroid hormone receptors. Prelim- Reductase inhibitors will block male develop- inary studies using the ribonuclease protection assay ment. The fact that exogenous DHT will induce male (RPA) to quantify tER-mRNA reveal that tER-mRNA development when administered to eggs incubating at varies in abundance with both developmental stage a threshold temperature suggests that male develop- and incubation temperature LJ . Bergeron, T. Wibbels, ment is mediated by an androgen receptor. Testoster- and D. Crews, unpublished data]. one is reduced to DHT by the enzyme reductase (Fig. 31, Androgens play a role in male sex determina- and administration of the reductase inhibitor 4MA to tion. Although exogenous E2 overcomes the effects of a incubating eggs prevents male development at both male-producing temperature and induces female devel- threshold or male-biased incubation temperatures opment, neither exogenous T nor DHT can induce male (Fig. 11). Comparable results are obtained with development at an incubation temperature that pro- MK906. Administration of T and reductase inhibitor to duces all females. However, if eggs incubating at a eggs incubating at a male-producing temperature re- threshold temperature receive exogenous DHT, most or sults in the production of female hatchlings. all of the offspring will be male [Wibbels et al., 19921 Feedback model for hormone action in sex (Fig. 9). Significantly, simultaneous administration of determination in TSD. Taken together, our results DHT and E to eggs incubating at a threshold temper- suggest that female and male sex determination are ature will cause hatchlings to have ovotestes (Fig. 10); separate processes that are differentially affected by such individuals are never observed normally in the incubation temperature, rather than by the organiza- red-eared slider. We presently are using the same tionidefault system characteristic of genotypic sex strategy employed in the research with estrogen to de- determination. Thus, females may result from the termine: under what conditions DHT is effective, when activation of an ovary determining cascade and DHT exerts its action, the hormone specificity of male inhibition of a testis determining cascade; conversely, MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 305

100

Frl d 70 c 60

WE u!z c4w 20 10

DHT 62 T TREATMENT Fig. 9. The role of estrogen in stimulating female development and androgen in stimulating male development in eggs incubated at lhe threshold temperature (292°C) in the red-eared slider. C, ethanol Control; DHT, dihydrotestosterone; E2, estradiol; T, testosterone. Sample sizes range from 20-30 per group. Dashed line represents 50:50 sex ratio.

Fig. 10. The ovotestis that results from simultaneous administration of estrogen and androgen on eggs incubated at an intermediate temperature. Note the thick membrane separating the cortical and medullary components. males may result from the activation of a testis hypothetical model (Figure 12, top panel) postulates determining cascade and inhibition of an ovary that at a female-producing temperature the gene(s) determining cascade (Fig. 12). Central to the steroid encoding for aromatase are activated and the geneb) hormone-mediation hypothesis (Crews et al., 1989, encoding for reductase remain at constitutive levels 1991; Dorizzi et al., 1991; Pieau et al., 1994) is the role (or perhaps suppressed via negative feedback control of T as a precursor molecule destined for conversion to of E2), resulting in increased E2 production. For DHT (via reductase) or E2 (via aromatasej. One example, Claude Pieau and colleagues (reviewed in 306 CREWS ET AL.

DHT and E2 to specific, high-affinity, intranuclear receptor proteins (AR and ER, respectively). which in turn would activate the putative receptors such that the hormone-hormone receptor complex would bind to hormone response elements on the DNA. The consequence of such events would be a stimulation of the transcription of genes associated with the sex-determining cascade of one sex, and an inhibition of the expression of genes associated with the sex-determining cascade of the opposite sex. Such a model would account for the following facts: (1) The effect of temperature or exogenous steroids is all-or-none; that is, individuals are either male or female. (2) Incubation at a threshold temperature results in a 50:50 sex ratio, rather than intersexes being formed. (3) Intersexes can be formed experimentally by the simultaneous administration of DHT and E2 to eggs incubating at a threshold incubation temperature or by high dosages of T at female-biased incubation temperatures. (4) Exogenous E2 will overcome the ef- Fig. 11. Administration of the reductase inhibitor 4-MA induces fects of male-producing temperatures, and there is a female development. In all studies, eggs were spotted with the chem- correlation between estrogen-sensitivity and tempera- ical in alcohol or with alcohol only. Sample sizes average 25 individuals, ranging from 21-30; total number of individuals depicted = 632; ture-sensitivity. (5) Exogenous DHT cannot overcome 100-pg dosage at 26°C not filled because group not tested. the effects of a female-producing temperature; although exogenous DHT will induce male development in eggs incubating at a threshold incubation temperature, there is no apparent dynamic between incubation Pieau et al., 1994) have documented a positive temperature and sensitivity to DHT. (6) Steroid-in- feedback relationship between E2 secretion and aro- duced gonadal feminization is specific to ER-responsive matase production at a female-producing temperature estrogens whereas gonadal masculinization appears in the European pond turtle. On the other hand, at a to be specific to AR-responsive androgens; and (7) Ad- male-producing temperature the gene(s1 encoding for ministration of an aromatase inhibitor will block fe- reductase may be enhanced and the gene(s) encoding male development and induce male development, for aromatase inhibited, resulting in increased DHT whereas administration of a reductase inhibitor will production. As yet unproven, a positive feedback block male development and induce female develop- relationship similar to that between aromatase and E2 ment, in eggs incubating at intermediate incubation may exist between DHT and reductase. Alternatively, temperatures. the finding of a lack of a dose-response in the effect of the reductase inhibitor and the absence of a dynamic between DHT sensitivity and incubation temperature Functional Outcomes of TSD as occurs between E2 sensitivity and incubation Incubation temperature has a permanent and cumu- temperature indicates that the genejs) for reductase lative effect on the morphology, physiology, and behav- are constitutively expressed at the various incubation ior of the individual. temperatures. Thus, incubation temperature may act Incubation temperature and gonadal hormones only indirectly via the testis determining gene(s1 have separate roles in sexual differentiation. It having a negative feedback on the regulation of must be appreciated that in species with TSD, in- aromatase gene expression, leading to suppression of cubation temperature and gonadal sex co-vary. Thus, E2 (Figure 12, bottom panel). For example, Haqq et al. any difference between individuals could be due to the (1993) demonstrated recently in the rat that the incubation temperature of the egg, the gonadal sex of putative male-determining factor (SRY) may control the individual, or both factors combined. If the con- male development through regulation of aromatase tribution of each is to be assessed, they must be disso- and MIS genes. ciated. Two approaches can be used to assess the rela- We propose further that incubation temperature active contribution of each in the process of sexual tivates the gene(s) encoding for steroid hormone recep- differentiation: (1) comparing individuals of the same tors (e.g., male-producing temperature upregulating sex from different incubation temperatures, and (2) AR and female-producing temperature upregulating hormonally manipulating embryos to reverse gonad ER). Whatever the actual enzymatic regulation, the phenotype. Thus far the data indicate that while go- resulting hormonal milieu would lead to the binding of nadal sex is important in shaping sex differences, it is MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 307

Fve fe,

Female- aromatase -@ f E2 -e 0 ODC -@ Ovary producing Tsp temperature (0 reductase -@ no A in DHT -d 0 TDC

Male- 0 aromatase no A in E2 -b-/ 0 ODC producing Tsp{ - temperature Q reductase -b f DHT _c)Q TDC -@ Testes

I Key: Tsp=Temperature-sensitive promoter derepression, or stimulation ODC-Ovary determining cascade 0 repression, or no activation TDC-Testis determining cascade

positive feedback

Female- -+ @ ODC -b Ovary producing temperature reductase -+ no A in DHT ,-d 0 TDC

Male- aromatase W no A in E2 -4 0 ODC producing --\- -- 0 temperature --- -- reductase -+ f Dm A\@TDC Testes 6 -(SRY) -

Fig. 12. Two hypothetical scenarios of temperature-dependent sex of steroidogenic enzyme activity and by testis determining factors (an determination in the red-eared slider (Trdchem.ys scripfa).The tem- SRY-like gene product) acting on aromatase production (Bottom perature experienced during the middle third of incubation initiates panel). In both models, incubation temperature is hypothesized to the cellular and molecular cascades that result in male or female activate genes encoding for steroid hormone receptors (e.g.,male-pro- offspring. Temperature may act via a temperature-sensitive promoter ducing temperature upregulating androgen receptor and female-pro- directly on genes encoding for steroidogenic enzymes, resulting in ducing temperature upregulating estrogen receptor) in turn contrib- changes in the steroid hormone milieu (Top panel). Alternatively, uting to the activation of ovary and testis determining cascades. temperature may act indirectly via steroid hormone feedback control secondary in importance to incubation temperature. Crews, 1994al (Fig. 13A). If the ovaries are removed This is seen particularly in the growth rate and behav- from females from low incubation temperatures on the ior of adults. day of hatching, they grow as males from a higher in- In the leopard gecko, the male is the larger sex. How- cubation temperature (Fig. 13B). Ovariectomy in ever, females from a male-biased incubation tempera- adulthood does not cause a similar weight gain, sug- ture grow faster and larger than do females from a gesting a developmental effect only. Finally, when the female-biased incubation temperature, indeed growing effect of gonad type and incubation temperature is ex- as rapidly and as large as males from lower, female- perimentally dissociated using estrogen treatment of biased incubation temperatures [Tousignant and eggs at a male-biased incubation temperature, these 308 CREWS ET AL. estrogen-determined females grow as would control, temperature-determined males (Fig. 13C). Incubation temperature also has a profound effect on TEMPLRATURL AND the frequency of species-typical aggressive behavior. In GOUADALSFX the initial stages of aggression the animal raises and arches its body, standing on its toes and waving the tail slowly. This is followed by a lateral orientation of the animal's body toward the stimulus animal and a slow approach. If the stimulus animal does not flee, then it -G Male 32 5" (n=10) t Mde 30" (n=5) will be attacked. Submissive behavior consists initially -w Female 32 5' (n=8) of the animal flattening its body so that the ventrum is +Female 30" (n=251 pressed to the substrate; at this initial stage the tail tip is twitched. If the aggressive animal approaches, the ec R mrmor GOYADEC~IYJW submissive animal will thrash its tail rapidly and then 0'4 TEMPFRATLIRE flee. 70 MODULATION 1 T, Aggression appears to be a male-typical trait in the 60 T leopard gecko. Females usually show little or no ag- 50 gression in response to males, whereas males will pos- ture and often attack another male as he approaches; 40 males rarely attack females. Females from female-bi- 30 ased incubation temperatures are less likely to be ag- 20 gressive toward male stimulus animals compared to females from a male-biased incubation temperature 10 [Flores et al., 19941 (Fig. 14). Hormone-determined fe- 0 males from a male-biased temperature are as aggres- C EFFECT OF IN( LIBATION sive as temperature-determined females from a male- TEMWRATLRE VS biased temperature. This finding suggests that the aggressive behavior in a female leopard gecko is less affected by ovarian hormones than by incubation temperature. In a sexual encounter, the male will slowly approach the female, touching the substrate or licking the air with his tongue. There is an attractivity pheromone contained in the skin of females that elicits the characteristic tail vibration of males where the tail is vi- brated rapidly, creating a buzzing sound [Mason and 0 5 15 25 55 65 Gutzke, 19901; females have never been observed to Age (Wee&) exhibit this tail vibration behavior, even if treated with exogenous androgen. Thus, attractivity is a fe- Fig. 13. The relative influence of incubation temperature and male-typical trait measured by a sexually active male's gonadal sex on body growth in the leopard gecko. Illustrated is the courtship behavior toward a stimulus animal. Females average body weight (iSE)of individuals from different incubation from a male-biased incubation temperature are less at- temperatures or hormonal manipulations. In the leopard gecko, 26°C produccs only females, 30°C produces a female-biased (80201 sex tractive than are females from female-biased incuba- ratio, and 325°C produces a male-biased (20:80)sex ratio. A Males tion temperatures [Flores et al., 19941 (Fig. 14). Hor- and females incubated at either 30.0" or 32.5"C. B: Animals mone-determined females from a male-biased incubated at either 26" or 32.5"C and then receiving a sham operation incubation temperature are both attractive and aggres- or surgical castration on the day of hatching. C: Eggs incubated at 26" or 32.5"C and receiving estrogen or ethanol at the mid-trimester sive. of development. Each individual was weighed weekly from hatching until adulthood; only the data for 5-week intervals are presented Practical Applications here. After hatching, all animals were raised in isolation while The preservation and restoration of endangered or exposed to a 14:10-hr:30:18"C daily photothermal regimen and fed a standard diet. In B the castrated male sample size initially was threatened species is becoming a more pressing prob- believed to be nine, but laparoscopy and RIA for androgens in the lem every day. This is particularly the case with ovip- circulation revealed that in 7!9 instances the castrations were arous reptiles, many of which are classified as threat- incomplete. Interestingly, only two female individuals were found to ened or endangered. This situation is due in large part have a partial ovariectomy as adults and in one of these females; the to the exploitation of reptiles as an important protein records indicate that one gonad was lost in the body cavity after detachment; this ovary has attached to the liver and yolks follicles resource (meat and eggs) in Third World countries. In but does not ovulate them. This finding is consistent with the addition, the leather and shell byproducts of many rep- literature indicating that an intact neural connection to the gonad is tiles represent substantial commercial opportunities. necessary for ovulation. MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 309 these negative features.1 It provides an easy and inex- loo 1 pensive solution to a pressing problem. Estrogen-spot- o Aggression ting of eggs has no mortality associated with the ap- 80 - plication of hormone and is absolutely effective in producing only female offspring [Crews et al., 19911; tn C the efficacy of the estrogen-spotting method for ensur- 60 - c ing the production of females now has been demon- C 2- strated in a variety of threatened or endangered spe- 2 40- cies: Mugger’s crocodile, Olive Ridley sea turtle, the * 0 freshwater Cagle’s map turtle, and the New Cale- donian gecko. Further, long-term studies with the leop- 20 ard gecko indicate that the fecundity of such estrogen- determined females is indistinguishable from that of temperature-determined females [D. Crews, unpub- 0 I-.[ lished datal. We have also established a long-term 26 28 29 30 31.5 32.5 34 35 breeding pond to monitor the egg-laying of estrogen- Incubation Temperature (“C) determined red-eared slider turtles once reproductive age is reached. Fig. 14. Effect of incubation temperature on aggression, a male- The potential impact of the estrogen-spotting tech- typical trait, and attractivity, a female-typical trait,, in female leopard nique can be seen in the following example. Let us geckos. Higher incubation temperatures result in a greater number of assume that 10 breeding females exist and that each females exhibiting aggression and decreasing levels of attractivity. female will produce 30 eggs each year. Let us assume Measure of aggression (0)was the female’s response to a male stimulus. Attractivity (0)was measured by the male’s interest in courting also that the young become sexually mature in their the female. Dashed line represents the sex ratio resulting from dif- third year. Finally, we will assume that a 1:l sex ratio ferent incubation temperatures. occurs in unmanipulated normal animals in the popu- lation. In all examples no mortality is considered and each female produced has equal fecundity. [These as- Unfortunately, black market trade of wild-caught spec- sumptions are clearly unrealistic, but any decreases in imens has further aggravated this problem. production will be equivalent in both the normal and Because incubation temperature determines the sex the manipulated animals.] With estrogen-spotting to ratio at birth, understanding TSD has profound conse- ensure 100% females, the number of females will in- quences for breeding programs. More than a decade ago crease exponentially. Thus, estrogen-treatment will re- Mrosovsky and Yntema [19801 and Morreale et al. sult in 10,200 females being produced over a 4-year 119821 suggested that then present practices of exca- period, compared to the 2,700 females produced nor- vating, moving, and artificially incubating turtle eggs mally. At the end of seven years this difference be- could result in the production of mostly, if not all, male comes 633,100 versus 56,150! individuals. This fate has been realized [Shaver et al., The timing and amount of estrogen applied to the 1988; Taubes, 1992; Behler, 19931. reptile egg are critical. At the appropriate time and in Incubating eggs at female-producing temperatures the proper dosage, estrogen can override the effects of would seem to be the most effective method for gener- temperature and result in healthy female hatchlings ating female offspring. This would be true if (1) the that grow normally and lay eggs as adults. However, if dynamics of the temperature-sex ratio interaction applied too early in development or in too great an have been established empirically for the species in amount, the result is increased embryonic mortality question, and (2) there is ready access to controlled and, in those animals that survive to hatch, stunted environment chambers. Often times we are dealing growth with reduced reproduction or, in extreme in- with so few eggs that experimental validation is im- stances, intersexuality with consequent failure to possible. This can result in a frustrating “hit and miss” breed as adults LTousigant and Crews, 1994b; Gross strategy. Further, for many countries and even entre- and Guillette, 19941. Thus, the phenomenon of TSD preneurs in the United States the cost of precision con- also makes clear another practical application, namely trolled environment chambers necessary to determine the use of turtle embryos as a bioassay for environmen- the pattern of TSD of a particular species is prohibi- tal quality. In temperate and tropical zones, reptiles tive. Even if incubators are affordable, nesting sites are often in remote and inaccessible areas, making controlled temperature studies even more costly. Further- ‘The estrogen-spotting procedure is patented (“A Method for Prefer- more, in these areas, power failures are common, with ential Production of Female Turtles, Lizards, and Crocodiles,” U.S. the result that valuable eggs die. Patent 5,201,280) with all rights assigned to Reptile Conservation International,Inc., a tax exempt, nonprofit organization devoted to the The technique of spotting eggs with estrogen to in- conservation of threatened and endangered reptiles; RCI makes this duce female development potentially circumvents technology available free of charge to established conservation groups. 310 CREWS ET AL.

TABLE 1. Effects of Some PCB Compounds on Sex Determination*

% hatchlings % hatchlings with female gonads with oviducts Compound (low dose1high dose! (low doselhigh dose) A 2',5',-Dichloro-3-biphenylol 010 0114 B 2,2',4-Trichlorodiphenyl ether 710 2110 C 2,2 ' ,5' -Trichloro-4-biphenylol 010 710 D 2,3,4'-Trichlorobiphenyl 710 010 E 2,3',5-Trichlorobiphenyl 018 718 F 2',4',6'-Trichloro-4-biphenylol Oil00 OIlOO G 2',3',4',5'-Tetrachloro-4-biphenylol 4150 817 1 H 2,4,4',5-Tetrachlorodiphenylether 010 010 J 2,4,4',6-Tetrachlorobiphenyl 710 017 K 2,4,4",6-Tetrachloro-p-terphenyl 010 017 L 2',3',4',5,5'-Pentachloro-2-biphenylol 010 010 EtOH Control 0 0 E2 Control 100 100 *Eleven compounds were applied to eggs incubated at 27.8"C in two doses per compound. Doses were A, B, C, F, G, H, J = 10 kg, 100 kg; D, L = 5 pg:, 50 bg; E = 25 kg, 250 pg; K = 3.35 pg, 33.5 pg. The E2 Control consisted of 10 pg E2. commonly occur at the boundaries of human habita- phenyls [Korach et al., 1988; McKinney et al., 19901. tion andlor in areas of environmental catastrophes As metabolites of other PCBs, hydroxylated PCBs such and hence could serve as an early warning signal of as F and G may exist in steady-state concentrations in environmental contamination [Gross and Guillette, aquatic environments, potentially exposing wildlife to 19941. This is possible because the commercial value their effects by direct contact or through the food chain of some turtle and crocodilian species is so great that [McKinney et al., 19901. eggs are available in large numbers. This presents Because purified PCB compounds are rarely found in an opportunity to develop a nonmammalian labora- the environment, we decided in the second series of tory model system for testing the developmental ef- experiments to look at combinations of the same PCBs. fects of xenobiotic compounds found in the environ- All eggs were incubated at 27.8"C and received a low ment. (10 pg), medium (100 kg), or high (145-190 pg) dosage There is abundant evidence that manmade products of compounds. Some eggs received a cocktail of all can become environmental toxins [Colborn and Clem- PCBs, except for the two that caused sex reversal (F ent, 19921. For example, some of the effects of toxins, and G). Others were exposed to combined hydroxybi- such as polychlorinated biphenyls (PCBs), can have phenyls, again excluding F and G. Lastly, some eggs estrogenic effects [Korach et al., 19881 and induce were treated with combined non-hydroxylated PCBs. reproductive anomalies in adult female animals in In all three conditions, there was no evidence of sex nature [Colborn and Clement, 19921. The fact that in reversal. the red-eared slider turtle species sex determination Since we knew compounds F and G showed estro- can be influenced by minute amounts of exogenous genic activity at the slightly higher temperature, we estrogenic ligands provides a sensitive indicator of decided to try these two compounds at a temperature potential contamination by environmental estrogens. that produces 100% males (26°C). Both compounds Using the all-or-nothing nature of the response of showed significant sex reversal at this temperature. red-eared slider turtle embryos to exogenous estrogen, When combined, F and G synergized, resulting in a we recently assayed 11 common PCB's [Bergeron et al., significant increase in ovarian development at a dose of 19941. Table 1 demonstrates that only two of the 10 pg or <1 ppm, whereas F alone and G alone re- compounds tested, 2',4',6'-trichloro-4-biphenylol (F) quired at least a 10-fold higher dose to show sex rever- and 2',3',4',5'-tetrachloro-4-biphenylol(GI, were found sal. Exogenous E2 produces similar results at a dose of to have estrogenic activity as indicated by the 0.5 kg, or 0.04 ppm [Wibbels et aE., 19911. production of female hatchlings from eggs incubated at Such experiments demonstrate that it is possible to a male-producing temperature. In these instances, use the eggs of reptiles with TSD as biomarkers of en- only the high dosage produced females complete with vironmental contamination. Indeed, recent studies fully developed oviducts. The former compound with the alligator have involved assessment of the de- showed 100% sex reversal at 100 pg or just below 9 gree of actual xenobiotic contamination (via physical ppm. In tests using mouse tissue, these same two and/or biochemical methods) followed by the applica- compounds show an appreciable affinity for ER, due in tion of the amount of contamination to incubating eggs part to their conformational properties as hydroxybi- under controlled laboratory conditions; results of such MECHANISMS, OUTCOMES, AND APPLICATIONS OF TSD 311 studies are similar to those for the red-eared slider REFERENCES [Gross and Guillette, 19941. Beamer WG, Whitten WK (1991): Do histocompability genes influence sex ratio (% males)? Keprod Fertil Dev 3:267-276. CONCLUSIONS Behler JI, (1993): SSP Reportc on the Chinese alligator (Alligator sinensis). AAZPA Commun 42:3-4. Sex determination is the product of coordinated gene Bergeron JM, Crews D, MacLachan dA (1994):PCBs as environmen- expression. The ability to manipulate sex in TSD per- tal estrogens: Turtle sex determination as a biomarker of environ- mits a degree of unparalled control, enabling study of mental contamination. Environ Health Perspect (submitted). Bern HA (1990): The “new” endocrinology: Its scope and its impact. the normal pattern of gene expression not possible with Am Zoo1 302377-885. other amniote vertebrate species having genotypic sex Hull

Ohno S (1967): “Sex Chromosomes and Sex-linked Genes.” Berlin: at different embryonic stages on sex determination, growth, and Springer Verlag. mortality in the leopard gecko (Eub/eph,arismncularius). J Exp Zoo1 Pieau C (1973): Variation de l’activite enzymatique rl5-3P-hydroxy- 268:17-21. st6roide dkshydrogknasique dans les glandes genitales d’embryons Viets BE, Tousignant A, Ewert MA, Nelson CE, Crews D 11993): dEmys orbicularis L. (Chcilonienr, en fonction de la temperature Temperature-dependent sex determination in the leopard gecko, d’incuhation. CR Hebd Sen Acad Sci Paris Ser D 276:197-200. Eublepharzs nzacularius. J Exp Zool 265:679-683. Pieau, C., Girondot, M., Desvages, G., Dorrizzi. M.. Richard-Mercier, vom Saal F (1981): Variation in phenotype due to random intrauter- N., Zaborski, P. (1994): Environmental control of gonadal differen- ine position of male and female fetuses in rodents. J Reprod Fertil tiation. In Short RV, Balaban E (eds): “The Differences Between the 62:633-650. Sexes.” Cambridge: Cambridge University Press, pp 433-448. K’hite KB, Thomas P (1992ai: Adrenal-kidney and gonadal ste- Shaver DJ, Owens DW, Chaney AH, Caillouet CW Jr, Rurchfield P, roidogenesis during sexual differentiation of a reptile with temper- Marquez MR (1988): Styrofoam box and beach temperatures in re- ature-dependent sex determination. Gen Comp Endocr 88:lO-19. lation to incubation and sex ratios of Kemp’s ridley sea turtles. In White RB, Thomas P (1992b): Stimulation of in vitro steroidogenesis Schroeder BA (compiler): “The Eighth Annual Workshop on Sea by pituitary hormones in a turtle (Truchernys scriptaj within the Turtle Conservation and Biology.” Fort Fisher, North Carolina: Na- temperature-sensitive period for sex determination. Biol Reprod 47: tional Oceanic and Atmospheric Administration, pp 103-108. 952-959. Taubes G (1992): A dubious battle to save the Kemp’s Ridley sea Wibbels T: Bull JJ, Crews L) (1991a):Chronology and morphology of turtle. Science 256:614-616. temperature-dependent sex determination. J Exp Zoo1 260:371- Thomas EO, Licht P, Wibbels T, Crews 11 (1992): Development of 381. hydroxysteroid dehydrogenase activity in sexually undifferentiated Wibbels T?Bull JJ; Crews D (1991b):Synergism between temperature and differentiated embryos of the turtle Trachemys (Pseudemys) and estradiol: A common pathway in turtle sex determination? J scripta. Biol Reprod 46:140-145. Exp Zool 260:130-134. ‘l’iersch TR, Mitchell MJ, Wachtel SS (1991):Studies on the phyloge- Wibbels T, Bull JJ, Crews D (1992):Hormone-induced sex determi- netic conservation of the SRY gene. Hum Genet 87:571-573. nation in an amniotic vertebrate. J Exp Zool 262454-457. Tiersch TR, Simco BA, Davis KB, Wachtel SS (1992): Molecular genetics of sex determination in channel catfish: studies on SRY, Wihhels T, Crews D (1992):Specificity of steroid hormone-induced sex ZFY, Bkm and human telomeric repeats. Biol Reprod 47:185-192. determination in a turtle. J Endocr 133:121-129. Tousignant A, Crews 1) (1994a):Incubation temperature and gonadal Wibbels T, Gideon P, Bull JJ, Crews D (19931: Estrogen- and temper- sex affect growth and physiology in the leopard gecko (Eubleppharzs ature-induced medullary regression during gonadal differentiation ntncularius), a lizard with temperature-dependent scx determina- in a turtle. Differentiation 53:149-154. tion. J Morph (Submitted.1 Yntema CL (196tlj:A series of stages in the embryonic development of Tousignant A, Crews D (199413):Effect of exogenous estradiol applied Che1,ydra serpentina. J Morph 125:219-252.