Methods: Study Area and Sample Collection

This study was carried out in wet evergreen and moist deciduous forests of the Anamalai Tiger Reserve in Western Ghats, India during 2007–2009. The elevations in the study sites ranged from 500 to 850 m above mean sea level. The annual rainfall ranged from 1500 to 2200 mm with 2 distinct monsoons, southwest (June–August) and northeast (September–November). These species are poorly detected and remain in crypsis, buried under litter debris in the forest floor; therefore, systematic collection of fecal samples was not feasible. Fresh fecal samples from wild chelonians using methods described in Deepak et al. (2009) and Deepak and Vasudevan (2012). Freshly collected fecal samples were dried in an oven at 65°C, pulverized, and sent to the lab (Centre for Cellular and Molecular Biology) within 1 mo for further hormone extraction and assays (Umapathy et al. 2013). Sixty-nine fecal samples (V. silvatica male = 29, female = 20; I. travancorica male = 8, female = 12) were collected between 2007 and 2009.

Methods: Fecal Steroid Extraction

Fecal steroid metabolites were extracted as described previously (Schwarzenberger et al. 1991). Briefly 0.2 g of dried feces was taken in a test tube with 2 ml of 90% methanol and vortexed at high speed for 30 min. The samples were then centrifuged 500 × g for 20 min. The supernatant was recovered and stored at −70°C until it was assayed. Extraction efficiency of fecal steroid was estimated by adding a known amount of ³H-labelled progesterone tracer in fecal samples before extraction.

Methods: High-Performance Liquid Chromatography (HPLC)

To evaluate the applicability of the assay for use with chelonian fecal extracts, HPLC tests were carried out (Shimadzu CTO-10AS). Steroid metabolites were separated and identified on a steroid-specific reverse-phase C-18 column (Waters column, Symmetry C-18, 4.6 × 20 mm, 3.5 µm, Intelligent Speed [IS] column) using a gradient of 20%–64% acetonitrile (ACN):water (H₂O) at a flow rate of 1 ml/min; run time 8 min. Prior to HPLC run, pooled fecal sample was extracted and passed through a Sep-Pak C-18 column (Sep-Pak plus C-18 Cartridges; Waters, Milford, MA) and eluted with 3 ml of absolute methanol (100%) as described by Weingrill et al. (2004). The supernatant was dried under a flow of nitrogen and resuspended with 100 μl of methanol. Reference steroid standards (Steraloids, USA) and the fecal hormone samples were eluted and monitored from 190 to 400 nm. The eluent fractions (250 µl) were collected every 15 sec, vacuum-dried, resuspended in Enzyme Immuno Assay (EIA) buffer and assayed in duplicate to determine immunoreactivity by using the corresponding enzyme-linked immunosorbent assay (ELISA). The HPLC elution profiles showed that the extracted samples were identical to that of standard testosterone, estradiol, cortisol, and 5α-pregnan. Most of the profiles have showed only 1 immunoreactive peak for the cane turtle (Fig. 1) and Travancore tortoise (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Methods: Enzyme Immunoassays

Fecal progesterone metabolite (progestagen) was measured using one of the major progesterone metabolites—5α-pregnan-3α-ol-20-one. Previously an EIA against 5α-pregnan-3α-ol-20-one was developed and standardized in a wide range of animals using fecal samples (Umapathy et al. 2013). Cross-reactivity of the 5α-pregnane antiserum in the assay was as follows: 5α-pregnane-3α-ol-20-one 100%, 5α-pregnane-3β-ol-20-one 34%, 4-pregnane-3, 20-dione 27%, 5β-pregnan-3α-ol-20-one 8%, and others < 1% (Umapathy et al. 2013). Fecal estradiol metabolite concentration was measured using estradiol polyclonal antibody (R0008). Estradiol antibody and horseradish peroxidase (HRP) conjugate were obtained from C. Munro (University of California, Davis). Cross-reactivity of estradiol antiserum in the assay was as follows: estradiol 100%, estrone 3%, and others < 1%. Fecal cortisol was measured by using cortisol polyclonal antibody (R4866), cortisol antiserum, and HRP conjugates that were obtained from C. Munro (University of California, Davis). Cross-reactivity of this antiserum was as follows: cortisol 100%, prednisolone 9.9%, prednisone 6.3%, cortisone 5%, and others < 1% with corticosterone, desoxycorticosterone, 21-desoxycortisone, testosterone, androstenedione, androsterone, and 11-desoxycortisol (Young et al. 2004; Kumar et al. 2014). Fecal testosterone metabolites were measured by using testosterone polyclonal antibody (R156/7). Testosterone antiserum and HRP conjugate were obtained from C. Munro (University of California, Davis). Cross-reactivity of this antiserum was as follows: testosterone 100%, 5α-dihydrotestosterone 57.4%, < 0.3% with androstenedione, and < 0.1% with androsterone, dihydroepiandrosterone, β-estradiol, and progesterone (Kumar et al. 2014).

Methods: EIA Procedure

Direct competitive EIAs were performed by following a previously reported procedure (Munro and Lasley 1988; Young et al. 2004). Antibody was coated onto a 96-well plate (Nunc-Immuno maxisorp; Fisher Scientific International, Inc, Hampton, NH) before diluting with a coating buffer (0.05 M sodium bicarbonate buffer, pH 9.6), incubated overnight at 4°C and washed 4 times with washing buffer (0.15 M NaCl and 0.05% Tween 20; Sigma-Aldrich, India). Fecal extracts (50 µl) or Standards diluted in assay buffer (0.1 M phosphate-buffered saline, pH 7, containing 0.1% bovine serum albumin) were added to the wells, followed by 50 µl of conjugated HRP steroids. After 2 hrs of incubation at room temperature, the plate was washed and 50 µl of TMB (tetramethyl benzidine)/H₂O₂ (Genei, Bangalore, India) substrate was added and kept in the dark for 10–15 min. The reaction was stopped by addition of 50 µl of 1 N HCl and absorbance read at 450 nm in the ELISA reader (Thermo Multiskan Spectrum Plate Reader, version 2.4.2; Thermo Scientific, Finland).

On account of small sample size (V. silvatica male = 29, female = 20; I. travancorica male = 8, female = 12), values of individual fecal-steroid concentration were pooled with referenced to sex, months of collection, breeding season, and nonbreeding season based on available literatures (Vasudevan et al. 2010; Deepak and Vasudevan 2012). Data are reported as the mean ±standard error of mean (SEM). Difference between mean hormone metabolite concentrations was analyzed using nonparametric Mann-Whitney U-test. Correlation analysis was carried out using Spearman rank correlation between values of fecal cortisol and progesterone metabolites. All statistical analysis and figures were made using the SPSS (ver. 17.0) statistical package for windows.

Results: Extraction of Steroid Metabolites

Recovery of known amount of unlabeled steroids ranged from 83.47% to 92.3% for 5α-pregnan-3α-ol-20-one, testosterone, estradiol, and cortisol in fecal extracts by EIAs for both the species. Similarly the extraction efficiency was between 80% and 85%.

Results: Validation of EIAs

All the EIAs were validated by demonstrating parallelism between the serial dilution of pooled fecal extracts (endogenous antigen) and respective standards (exogenous antigen) with no difference in slopes (p > 0.05). Assay sensitivity was calculated at 90% binding for each EIAs. The assay sensitivity of antibodies were 23.4, 39, 39, and 120 pg/ml for testosterone, estradiol, cortisol, and 5α-pregnan-3α-ol-20-one EIAs, respectively. The intra- and interassay coefficients of variation (CV) were 5.2% and 8.6% (n = 10), 3.98% and 5.27% (n = 10), 6.83% and 8.54% (n = 10), and 6.2% and 10.2% (n = 10) for estradiol, cortisol, testosterone, and 5α-pregnan-3α-ol-20-one, respectively.

Vijayachelys silvatica (Female): Monthly mean of fecal testosterone ranged from 29 ng/g (October) to 158 ng/g (March), and it elevated before the onset of the mating season (Fig. 3a). Monthly mean fecal estrogen concentrations ranged from 28 ng/g in March to 62 ng/g in June (Fig. 3b). Overall, fecal estrogen was higher and more variable during the breeding season than during the nonbreeding season, but no significant difference was observed between the 2 seasons (p > 0.05), most likely because of small sample size in the nonbreeding season (Fig. 4b). Increased fecal cortisol concentration was observed in the month of July (Figs. 3c and 4c). A significant and positive correlation was observed between fecal progestagen and cortisol (r_s = 0.59, p = 0.006). Monthly mean progestagen concentrations ranged from 75 to −385 ng/g. The lowest concentration was in June and the highest was in November. Overall, fecal progestagen concentrations were higher in the breeding season than nonbreeding season and it also varied during breeding season (Fig. 3d).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Vijayachelys silvatica (Male): Monthly mean fecal testosterone concentrations ranged from 42 ng/g (December) to 131 ng/g (August; Fig. 3a). Overall the fecal testosterone concentrations were higher and variable during the onset of breeding (June) and peak breeding season (August) than during nonbreeding season (December). Further, it declined toward the end of breeding season (Fig. 4a). Monthly mean fecal cortisol concentration ranged from 0.24 to −7.40 ng/g (Fig. 3c). The lowest concentration was observed in the month of December (nonbreeding season) and the highest was observed in July (breeding period; Fig. 4c). Some males also showed high cortisol concentrations during the mating season. Overall, mean monthly fecal cortisol concentrations were elevated during June and July (i.e., mating season for this species).

Indotestudo travancorica: Fecal testosterone in both sexes was higher during the breeding season than during the nonbreeding season (Fig. 4e). Fecal estrogen concentration in females ranged from 37 to 102 ng/g between June and December (Fig. 3f). The values fluctuated widely and there was no significant difference in fecal estrogen concentration in females between breeding and nonbreeding seasons (p > 0.05). Fecal cortisol did not show any pattern between the sexes except an increased monthly mean during October (Fig. 4g). Fecal progestagen concentration was higher during nonbreeding season than during breeding season; however, the difference was not significant because of small sample size (p > 0.05; Fig. 4h).

Discussion

This is the first attempt to study the hormone profiles of V. silvatica and I. travancorica using noninvasive steroid hormone analysis from fecal samples collected from their forest habitats. Overall, V. silvatica had higher fecal-steroid concentration during breeding season than during nonbreeding season. Similarly, I. travancorica showed elevated fecal steroids during breeding season. Increased testosterone concentrations prior to the commencement of breeding and during the breeding season has been reported in many higher vertebrates (Sadlier 1969) and reptiles (Rostal et al. 1998) because it helps in preparing the male for successful mating and reproductive behavior and females for vitellogenesis, ovulation, and nesting. In male Galápagos tortoises (Chelonoidis nigra), plasma testosterone and corticosterone increased for a few months before the onset of the mating season and peak levels were attained when the highest number of copulations occurred (Rostal et al. 1998). In the present study, male testosterone levels increased during and prior to the breeding season in V. silvatica. Females of V. silvatica also attained higher testosterone levels during the breeding season.

Estradiol stimulates vitellogenesis, final maturation of ovarian follicles prior to their release, and recrudescence of the ovary after courtship and mating in reptiles (Ho 1987; Rostal et al. 1994; Lance et al. 1995). Estradiol levels were elevated during the premating period when vitellogenesis occurred in Galápagos tortoises (Rostal et al. 1998). In the present study, we only detected estradiol in females and found large variation in fecal estradiol levels during the breeding season. With limited samples used in the present study, it is difficult to explain variation in levels of estradiol observed in the species.

Progesterone in chelonians increases during the preovulatory period or after ovulation and declines following egg-shelling (Licht 1982). In oviparous vertebrates, progesterone has a nutritive function in stimulating production of the egg-white protein, avidin (O'Malley 1967). In gopher tortoises (Gopherus polyphemus), plasma progesterone did not show significant variation with season, but females that had shelled eggs showed slightly higher progesterone than those females without eggs (Ott et al. 2000). Huot-Daubremont et al. (2003) reported elevated plasma progesterone during the periovulatory period in Hermann's tortoise (Testudo hermanni hermanni). In the present study, only females had detectable levels of progesterone and, in particular, V. silvatica fecal progesterone concentration increased during the breeding season and then declined soon after, implying a similar role for the hormone in this species, although we did not know the exact reproductive state of the females.

In many animals, including reptiles, stress hormones (Glucocorticoids) tend to increase during breeding season for the generation of energy (Romero 2002; Moore and Jessop 2003). However, stress levels vary among individuals depending on age, sex, reproductive state, body condition, social status, and external conditions such as temperature, rainfall, food availability, and habitat condition (Dunlap and Schall 1995; Moore et al. 2001). The measurements of stress levels in the present study showed elevated and variable cortisol concentrations 1 mo before commencement of mating in the male V. silvatica. Elevated cortisol concentrations in males during breeding might be due to aggressive interactions between males (Deepak and Vasudevan 2013). The females also had high cortisol concentration during the entire breeding season. Female turtles were not often involved in aggressive encounters; therefore, elevated cortisol concentration might play a role in the generation of energy for egg production and nesting (Moore and Jessop 2003). Elevated levels of corticosterone during breeding has been reported in Galápagos marine iguanas (Amblyrhynchus cristatus; Rubenstein and Wikelski 2005), leatherback sea turtles (Dermochelys coriacea; Rostal et al. 2001), and several other reptiles species (Romero 2002).

Fecal hormone concentrations in I. travancorica suggest that breeding is timed immediately after the monsoon in line with expectation. Delayed oviposition could expose the eggs to low temperatures that prevail in the region. We hypothesize, based on evidence from progesterone concentrations, that late clutches probably overwinter and progressively undergo development during the onset of summer.

The present study demonstrates that fecal hormones could be used for assessing reproductive function and stress status in endangered forest-dwelling chelonian species. Noninvasive fecal-steroid hormone monitoring is a promising tool for chelonian in situ and ex situ conservation programs.