Introduction
Reproductive cycles in vertebrates are regulated by the hypothalamus. The secretion of gonadotropin releasing hormone (GnRH) stimulates the pituitary gland to release the gonadotropins luteinizing hormone (LH) and follicle stimulating hormone (FSH) (Licht, 1980; Kang et al., 2020). Gonadotropins (GTH) are composed of a common α-subunit, and a unique β-subunit with a biological role specific to each hormone (Aizen et al., 2007a; Aizen et al., 2007b). Both gonadotropins regulate gonadal functions and thus are essential for reproduction. In mammals, FSH and LH have a distinct effect on endocrine regulation traits. In females, LH stimulates estrogen and progesterone production in the ovary. FSH is responsible for the development of the ovarian follicle, and stimulates secretion of estrogen from the follicle (Kluge et al., 2012).
Green turtles (Chelonia Mydas) have been subject to a long, and worldwide, history of human exploitation, resulting in endangerment of the global green turtle`s population (Seminoff et al., 2015). Since the green turtles exhibit high natal homing behavior and due to their long and complex life cycle, human intervention is needed for to increase the green turtle population of nesting females. The Israel National Sea Turtle Rescue Center established a captive breeding stock in 2002 to provide this needed help. To rehabilitate the Mediterranean green turtle population, a deeper understanding and prediction of reproductive stages are essential. Although the reproductive ecology and endocrinology of green turtles have been the subject of numerous studies, little is known regarding their reproductive biology.
There is little information on the cycling and function of FSH in green sea turtles, and the physiological significance in this species is not yet fully understood. Several studies in female reptiles have shown a stimulation of follicular development following injection of mammalian FSH (Kawazu et al., 2014; Tripathy and Rai, 2017; Kang et al., 2020). Previous studies by Licht, 1980 and Wibbels et al., 1992 regarding FSH function in green sea turtle supported these results. An increase in FSH concertation at the time of ovulation in green sea turtles from April to August has also been observed (Licht et al., 1979; Wibbels et al., 1992).
In males, FSH was shown to stimulate testicular growth, and increased production of an androgen-binding protein by Sertoli cells that are necessary for sustaining the maturing sperm cell. Furthermore, FSH also stimulates spermatogonial mitosis (Licht et al., 1979; Ng et al., 1986; Bousfield et al., 2007). In males, an increase in FSH concertation was concurrent with an increase in testosterone, reaching maximal concentration after summer testicular recrudescence (Licht et al., 1985; Mendonca and Licht, 1986).
Gonadotropins regulate second messenger pathways leading to the production and secretion of sex steroids (Aizen et al., 2007a; Aizen et al., 2007b). There are three main sex hormones that affect behavior and reproduction in vertebrates: progesterone, estradiol, and testosterone. In females, progesterone is a potential ovulation inhibitor; testosterone affects mating receptivity towards males, and estradiol, is the primary stimulus for vitellogenin (Licht et al., 1979). In males, testosterone is significantly correlated with testicular growth and spermiation, and estradiol is essential for spermatogenesis (Mahmoud and Licht, 1997).
The development of tools such ELISA or Radioimmunoassay (RIA) for the measurement of gonadotropins until now has required purified protein in large amounts and specific antibodies (Aizen et al., 2007b). Obtaining high quality and quantity of native gonadotropins requires purifying GTHs from pituitary glands, which would involve the killing of a significant number of animals. This is very problematic, especially for the green sea turtle that is an endangered species (Seminoff et al., 2015).
Recombinant FSH production has a distinct advantage since production is not dependent on pituitary availability (Kasuto and Levavi-Sivan, 2005). The yeast host system is one of the most powerful tools for expressing recombinant proteins. Methylotrophic yeast species such as Pichia pastoris have been developed for this purpose and are very popular due to their low cost (Karbalaei et al., 2020). This yeast line can produce high yields of recombinant proteins with glycosylation (Balamurugan et al., 2007). The advantages of this expression system include appropriate folding in the endoplasmic reticulum, and secretion of recombinant proteins to the medium, making the harvesting easier, and avoiding yeast cell protein contamination in the purification process (Li et al., 2007; Yang et al., 2013). Furthermore, they have high growth rate, coupled with ease of high cell-density fermentation, high levels of productivity in an almost protein-free medium, ease of genetic manipulation of well characterized yeast expression vectors as in Escherichia coli. (Li et al., 2007). The time needed to assess recombinant gene expression with the P. pastoris is approximately 3-4 weeks. This time frame includes transformation of the yeast, screening the transformants for integration, and an expression time-course analysis (Aizen et al., 2007a; Aizen, 2011). P. pastoris culture can reach 120 g/L of dry cell weight density (Aizen, 2011).
There are many immunoassay platforms available to quantify protein concentrations in biological fluids. ELISAs are preferred in many cases due to their sensitivity, specificity, accuracy, and ability to tolerate harsh buffers or pre-treatments (Molés et al., 2012). Both indirect and sandwich ELISAs allow detection of antigen or antibody at very low concentrations hence their specificity. The main advantage of competitive ELISA is that the method detects differences in antigen-antibody complex with high sensitivity, even when the specific detecting antibody is present in relatively small amounts. Competitive ELISA is less sensitive to sample dilution and sample matrix effects than the sandwich ELISA, and there is less variability between duplicate samples and between assays (Barlow et al., 1986). Taken together one should consider using a sandwich ELISA. The use of two specific antibodies, one against the beta subunit and the other against the alpha subunit allows us to specifically detect each glycoprotein hormone, depending on the anti-beta used, in its natural quaternary structure. (i.e. associated with alpha and excluding free beta detection). With regards to the alpha subunit in glycoproteins, it holds some challenges, as it is not very immunogenic and highly conserved among glycoproteins among different species and the same species (identical alpha for FSH, LH and TSH).
RIA methods were widely used for the measurement of gonadotrophin concentrations based on the principle of competitive binding. In this method, an unlabeled antigen competes with a radiolabeled antigen for binding to an antibody with the appropriate specificity (Webster et al., 1990). RIA is often based on a polyclonal antibody and may show cross-reactivity to substances other than the test analyte. ELISA, on the other hand, is based on two antibodies, which ensures high specificity for the analyte. In addition, the ELISA method can utilize antibody quantities that are higher than those allowed by RIA’s radioactive tracer (Wewer Albrechtsen et al., 2014), and ELISA exhibits lower coefficient of variation (CV) than RIA (Warnken et al., 2016).
ELISA also offers excellent performance with small sample volumes, and the procedure takes about half the amount of time as the RIA assay. In addition, RIA uses radiolabeled reagents that require a license for storage and special waste disposal. The ELISA method avoids these constraints of using radioactive chemicals (Webster et al., 1990; Wewer Albrechtsen et al., 2014).
In recent years the use of analytical methods to determine concentrations of gonadotropins such as LC-MS/MS and LC-MRM/MS were implemented due to the limited availability of high-quality antibodies, and ability to distinguish between different FSH glycovariants and showed comparable sensitivity to ELISA assays, with advantage of being faster and more selective (Melchiorre et al., 2021). The majority of assays LC-MS is currently at least complementary if not superior to immunoassay and that the benefits LC-MS offers will see its increased use over immunoassay, especially as the technology advances but we are not there yet (Cross and Hornshaw, 2016).
The reproductive biology of sea turtles has been studied both in the wild and in captivity; most of the research is based on nesting females, due to the ease of sampling during the nesting season. Only a few studies on hormonal cycles of adult males have been investigated in the wild (Wibbels et al., 1990; Tim et al., 2004; Blanvillain et al., 2008) and others have been studied in captivity (Licht et al., 1979; Licht et al., 1985; Rostal et al., 1998). To study GTH biological function we need assays for measuring GTH concentrations, and in order to achieve that we need to produce properly folded and biologically active recombinant GTHs.
In this study, we used the yeast Pichia pastoris expression system, since this system provided sufficient quantities of recombinant protein (Daly and Hearn, 2005) for the development and validation of an ELISA. We report on the production of a recombinant green sea turtle FSH (rcmFSH) and the development of a competitive ELISA for measuring green sea turtle FSH using the recombinant protein. For the first time, we were able to directly measure FSH concentrations in plasma samples, by comparing plasma FSH concentrations with steroids concentrations. We observed that FSH concentrations increased during the prior and during the nesting season from January to June (nesting season is April to August) along with progesterone concentrations. We were able to identify a pattern whereby FSH concentrations rise followed by rising estradiol concentrations. This helps us predict the reproductive stages of the turtles in the breeding facility, which can enhance successful breeding in captivity. Our main goal is to monitor annual concentrations of sex steroids and gonadotropins to characterize green turtle hormonal profiles. A better understanding of the green sea turtle reproductive biology may boost future efforts on population rehabilitation and conservation.
Materials and methods
Research population and site
The Sea Turtle Rescue Center under was established by the Nature & Parks Authority of Israel is located in Alexander River National Park (Coordinates: 32.4016667, 34.8666667)
The reproduction facility features two 450 cubic meters tanks and an artificial nesting beach. Water temperature is kept at 20-22°C, and periodically altered. The turtle’s diet is composed of romaine lettuce (2% BW) and frozen fish (Herring, Flounder and Salmon, 0.2% BW). All turtles are tagged with metal plates on either their front or back limb. The research population consists of 12 females (weight: 83.58 ± 22.74 Kg; CCL: 82.81 ± 7.52 cm; CCW: 75.31 ± 7.44 cm) and 11 males (weight: 66.16 ± 17.89 Kg; CCL: 76.09 ± 8.16 cm; CCW: 69.5 ± 7.2 cm), from the captive breeding stock in Mikhmoret (
Table S1
). 6 of the 12 females nested in 2 nesting seasons: 474, 495, 729 and 760 in 2019; 173 and 1055 in 2020.
Blood sample collection
Blood sampling was performed approximately every two months in the Israel National Sea Turtle Rescue Center. Five ml of blood were taken from the dorsal vein of the turtle's neck using a sterile 21-gauge 8.9 cm needle. Blood was moved into a biobased blood collection tube that contains an anticoagulant agent heparin sodium. The blood samples were stored at 4°C during the collection (no more than one hour). Once collection ended, hematocrit count of each turtle was first examined to obtain information on their health. We centrifuged two capillaries for each sample at 17800 RCF for 12 min and calculated the average of the two capillaries. For blood fractionation, we centrifuged the 5 ml tubes at 3200 RCF for 15 min. Each tube contained approximately 3 ml of plasma. We divided each individual plasma sample into three different 1.5 ml Eppendorf tubes for hormone measurement, blood biochemistry and backup. The plasma was then stored at -80°C.
Steroid extraction and measurement
For steroid extraction we used a 500 µl plasma sample and added 5 ml diethyl ether to glass tubes and mixed 3 times for 15 seconds with vortex. We transferred the glass tubes to -80°C for 45 minutes, the upper phase in the glass tube was transferred to a new glass tube and the diethyl ether was evaporated by heating the tubes in a 55°C water bath. Extraction products were dissolved in 500 µl Steroid Assay Buffer (SAB). The concentrations of sex steroids (estradiol, progesterone and testosterone) were measured by commercial Cayman ELISA kits ®. (501890, 582701 and 582601 respectively). Concentration detection levels were 6.6-4,000, 7.8-1,000 and 3.9-500 pg/ml for estradiol, progesterone and testosterone respectively, with sensitivity range of 6-15 pg/ml. In order to assess the concentration range of each steroid in the research population, preliminary tests were conducted. For each steroid assay at least three dilutions were tested, with samples from both sexes and different sampling dates to assess the annual concentration variation (data not shown).
Hormone production and purification
Recombinant FSH was produced using the methylotrophic yeast P. pastoris, as a biologically active, single-chain polypeptide according to Aizen et al. (2007a) protocol, green sea turtle (Chelonia mydas) FSHB (GenBank Accession No. XM_007067910.1) and green sea turtle (Chelonia mydas) CGA (GenBank Accession No. XM_007052575.1) subunits coding gene sequences were translated to amino acid sequences using the ExPASy Proteomics Server (http://ca.expasy.org/tools/dna.html). The presence of signal peptide cleavage site was predicted by SignalP 4.1 Server (Nielsen, 2017). The protein coding gene sequences were used to design two expression vectors: FSH β-subunit peptide chain for future antibody production, and FSHβα peptide chain for standard curve. FSHβα vector was constructed with a flexible linker allowing a smooth rotation of the two subunits to reach the natural structure of the entire gonadotropin. To allow flexibility to the construct tethered polypeptide, we used 2xGS-6His-2xGS between the β-subunits and α-subunits. Both vectors were transfected into chemically competent E. coli, JM109 strain via heat shock based on manufacturer’s protocol (Promega). Selection and expression were generally performed according to (Aizen et al., 2007a). Recombinant proteins were purified using nickel–nitrilotriacetic acid–agarose (Ni-NTA; Qiagen). As a negative control, P. pastoris was transformed with an expression vector that did not contain the cmFSH cDNA, and fractions were prepared in the same manner. The eluted proteins were concentrated by lyophilization, and protein concentration was measured using Bradford reagent. The eluted proteins were stored at -80°C until use. Recombinant proteins were deglycosylated using PNGaseF (Peptide N-glycosidase F) and analyzed by Western blot with anti-HIS staining.
Antisera
The antiserum against green sea turtle FSH β-subunit was produced by a commercial company ADAR Biotech Ltd. (Rehovot, Israel). The antiserum was raised in two rabbits with approximately 40 ml of sera harvested from each rabbit. Sera antibodies titers against the antigen were determined by direct ELISA by ADAR Biotech Ltd. resulting in high titer of antibodies against the FSH β-subunit. Final bleeding was performed, and the antiserum was then aliquoted, and stored at -80°C.
Characterization of FSH by western blot analysis
FSH protein samples were electrophoresed on 16% SDS-polyacrylamide unstained gels with 6-8% stacking gel. We added Laemmli SDS reducing buffer (Bio-Rad, Hercules, CA) with 2-mercaptoethanol concentrated at 4X to the protein samples for SDS-PAGE under reducing conditions. Samples where heated in a Thermomixer comfort (Eppendorf AG, Hamburg, Germany) for 10 min at 100°C. The electrophoresis was carried in TGS (Tris-Glycine-SDS) running buffer (Bio-Rad, Hercules, CA) at 110V for one hour. The SDS PAGE gels were then washed once with DD H2O. Gels were blotted to nitrocellulose membranes (BioRad) using Bio-Rad Trans-Blot® Turbo TM Blotting System. Membranes were blocked with Anti-His HRP blocking (Qiagen, Limburg, Netherlands) or 5% low-fat milk powder in PBS. Bound antibodies were detected with either penta-His antibody (Qiagen, Limburg, Netherlands) or with turtle FSHβ antibody. When using anti-His antibody, the membranes were treated with penta-His-HRP antibody (1:2000 dilution) overnight at 4°C. When using anti-FSHβ, the membranes were incubated in PBS plus 1% low-fat milk powder with the antibodies (diluted 1:5000) overnight at 4°C and then with goat-antirabbit horseradish peroxidase conjugate (GAR-HRP) (diluted 1:5000) for 1h at room temperature (RT). After washing, all membranes were treated with enhanced chemiluminescence reagent, EZ-ECL kit (Biological Industries, Beit Haemek, Israel) to reveal immunoreactive bands.
Development of ELISA for FSH
Competitive ELISA for green sea turtle FSH was developed using FSHβ-subunit primary antibody, recombinant FSHβ-subunit for coating ELISA microplates, and recombinant single chain βαFSH for the standard curve. The protocol was based on (Aizen et al., 2007b).
Coating
We coated ELISA 96-well microplates (MaxiSorp™ Nunc-Immuno, Thermo Scientific™, Denmark) with 100 µl/well of recombinant green sea turtle FSHβ diluted 1.5 ng/10ml (0.015 ng/well) in a solution of carbonate buffer (50 mM sodium carbonate (NaHCO3), pH 9.6). The plates were incubated in complete darkness overnight at RT without shaking. For non-specific binding, two wells (A12 and B12) were not coated with the FSHβ solution.
Washing and blocking
The next day, coated microplates were washed (3 times X 5 min) with 200 µl/well of PBST buffer (10 mM Na2PO4, 2 mM KH2PO4 (pH 7.4), 140 mM NaCl, 3 mM KCl, and 0.05% Tween 20). The microplates were blocked with 200 µl/well of 1% bovine serum albumin (Sigma, Rehovot, Israel) in PBST buffer for 1 h in complete darkness at RT without shaking to reduce background.
Pre-incubation of standards and samples
Standard was diluted in 0.1% BSA in PBST with 10% of normal goat serum (NGS, Biological Industries). Unknown plasma samples were diluted (1:15 for males, 1:15 for females and 1:30 for nested females) with 0.1% BSA in PBST, each sample was diluted in a volume of 300 µl allowing duplicate analysis. The unknown samples and FSHβα standards were pre-incubated in 96-well microplates (Sarstedt AG & Co., Nümbrecht, Germany) with the primary antibody against FSHβ in final dilution of 1:5000 dissolved in 0.1% BSA in PBST in complete darkness overnight at RT without shaking.
Samples and standards dispensed
After the pre-incubation of standard and unknown samples with the primary antibody, each sample was dispensed into the coated and blocked microplates (100 µl/well) and incubated in complete darkness overnight at RT without shaking.
Incubation with secondary antibodies
The next day the plates were washed (3 X 5 min) with PBST. 100 µl/well of GAR-HRP diluted 1:5000 in PBST 0.1% BSA buffer were added and incubated for 2 h in complete darkness at RT without shaking to detect the formed antigen–antibody complexes.
Color development
Before color development the plates were washed again (3 times X 5 min) with PBST. The visualization of enzyme complexes was executed by addition of 100 ml/well of ABTS® 2-Component Microwell Peroxidase Substrate (KPL, Zotal, Israel). The reaction was carried out in complete darkness at RT for 20 min. Absorbance was read at 405 nm, using a Synergy HTX Multi-Mode Reader (BioTek, USA).
Validation of the ELISA for green sea turtle FSH
The ELISA was validated for cmFSH determinations in plasma samples of Chelonia mydas and was tested for cross-reactivity with additional turtle species (loggerhead, hawksbill and african softshell turtle). To conduct the parallelism analysis, whole blood samples were collected from sexually mature turtles held in captivity at the NPA Center. Displacement curves for plasma samples were obtained by serial dilutions of the plasma in the ELISA buffer (PBST buffer containing 10% NGS) and compared with the ELISA standard curve. The ELISA was also validated for FSH in other turtle species. Plasma samples were diluted 1:2. All samples were used in the ELISA at a total volume of 100μl solution per well. Intra assay coefficient of variation (CV) was determined by assaying 24 replicates of green sea turtle samples on the same assay plate. Inter assay variation was determined by assaying the same sample 9 times on different assay plates.
Statistical analysis
All data are presented as means ± SEM (Standard Error of the Mean). To test parallelism between displacement curves, we used linear regression in Graph-Pad Prism 9.0 software (GraphPad, San Diego, CA) and compared slopes. The significance in differences of hormone level means between groups was determined by one-way analysis of variance (ANOVA) followed by Newman–Keuls test. To test for parallelism between various regressions lines, we performed logit transformation, where logit (B/Bo) = log [r/(1-r)], and r = B/B0, B represents the binding at each point, and B0 represents the maximum binding. We used the linear regression parameter in Graph-Pad Prism 9.0 software to compare the slopes.
Discussion
Herein we describe for the first time the development of an ELISA for FSH in green sea turtles, with recombinant gonadotropin we produced using the yeast P. pastoris expression system. The ELISA was also found to be applicable for other sea turtle species. We also observed plasma steroid concentration during the reproductive cycles that revealed a positive correlation between FSH and estradiol concentrations in females; estradiol concentrations were increased immediately after FSH elevation. In addition, nested females presented an increase in FSH concentrations prior to progesterone elevation in January to April, slightly before egg laying.
We must emphasize that the turtles in this research have been in the breeding facility since hatching. Our results describe patterns of FSH and steroids concentrations for turtles that were kept in captivity their entire life. With that we believe that our findings can help to better understand what is occurring during the time periods before nesting and after nesting especially for turtles that will not nest every year.
Measuring hormone concentrations in turtles is challenging on many levels. Taking blood from females arriving on the beaches to lay eggs may interfere with their normal behavior during this critical period. Finding males on the beach is rare, and so measuring their hormone concentrations may not be feasible (Goldberg et al., 2013). When injured turtles arrive at the Sea Turtle Rescue Center, they are usually young, and distinguishing between males and females is not possible, Owens et al. (1985) showed that injecting mammalian FSH in immature turtles increased the testosterone concentration in the blood and was helpful to determine if the turtles are truly males, but this method has not been tested since then (Owens and Morris, 1985). Measuring testosterone concentrations in young turtles does not correlates with their gender, especially in young age.
Therefore, studying the breeding stock at Mikhmoret was important for achieving a hormonal profile of the green sea turtles. In this research we developed and validated a competitive ELISA for measuring FSH concentrations in plasma samples from green sea turtle based on the approach of production recombinant FSH as a single chain molecule using the P. pastoris expression system. Earlier research using the P. pastoris expression system has shown that the main advantage of this system is the ability to produce a recombinant protein that is similar to the natural hormone and can elicit an In-vitro and In-vivo biological response (Kasuto and Levavi-Sivan, 2005). In addition, the protein can be efficiently purified from the culture medium (Aizen et al., 2007a; Aizen et al., 2007b). The use of a recombinant gonadotropin can increase its immunogenicity, as well as the specificity of the generated antibodies (Daly and Hearn, 2005).
Generally, immunoassays that are developed to determine GTHs concentrations have been based on native GTHs that were purified from pituitary glands (Molés et al., 2012). Purification of native FSH is a very demanding process, and to purify a sufficient amount of GTHs we would need a large number of pituitary glands, which would require killing a large number of turtles. Recombinant proteins can be produced continuously, independent of turtle pituitary gland availability. Additionally, recombinant FSH subunits can ensure there is no cross-reactivity with similar GTHs that share the same α-subunit (Aizen, 2011). The open reading frame for FSH α-subunit is 120 amino acids, and for FSH β-subunit is 138 amino acids. The subunits were cloned and validated for their predicted protein sequences from NCBI. The sequences were compared to different turtle species (The common box turtle ; T. carolina, Painted turtle; C. picta, Chinese pond turtle; C. reevesii, and Chinese softshell turtle ; P. sinensis) and were found to be practically identical. In different studies on the FSH molecular structure resemblance in FSH α-subunit between vertebrates was found (Chien et al., 2005; Shen et al., 2006). To date, an ELISA for the measurement of FSH in sea turtles or in any other reptile has not been developed.
Different studies for development of ELISA to measure FSH were done in several fish and mammalian species: tilapia, rat, mice, flatfish, yellowtail kingfish, Japanese rice fish, Japanese eel and more (Pappa et al., 1999; Shimizu et al., 2012; Chauvigné et al., 2015; Burow et al., 2019; Kazeto et al., 2019; Nocillado et al., 2019; Pappas, 2019; Ongaro et al., 2021). FSH in green sea turtles was last measured in the early 90's (Wibbels et al., 1990) using radioimmunoassay (RIA), which is based on the principle of all immunoassays: a known amount of antigen is labeled with a radioactive element and mixed with a known amount of antibody directed against that antigen (Webster et al., 1990).
The RIA method is problematic because it uses radiolabeled reagents that require a special license for storage and waste disposal (Sancho et al., 2013). In addition, ELISA is more accurate and reproducible at lower concentrations than RIA (Webster et al., 1990).
LC-MS/MS and LC-MRM/MS techniques present the potential for surpassing the accuracy of RIA and ELISA, exhibiting heightened sensitivity towards extremely low concentrations. It's crucial to consider the requirements for regular biomolecule quantitation from a broad range of fields. While immunoassays have been instrumental, they might need to evolve to match the escalating throughput demands. As long as the goals involve reduced per-sample expenses and enhanced precision, the prospects of LC and LC-MS appear promising as a prospective solution.
The FSH concentrations measured in our ELISA were 17-650 ng/ml for females and 24-710 ng/ml for males. The results for FSH level ranges in females do not concur with studies by P. Licht et al., 1979 (0.3-10.6 ng/ml) and Wibbels et al., 1992 (0.1-2 ng/ml). In these studies green sea turtle males did not reveal measurable concentrations of FSH, but observed females FSH concentrations of 3-10 ng/ml in the snapping turtle (Chelydra serpentina) (Mahmoud and Licht, 1997), 2-16 ng/ml in the musk turtle (Sternotherus odoratus) (Mendonca and Licht, 1986) and 6-26 ng/ml in the painted turtle (Chrysemys picta) (Licht et al., 1985).
Our results are higher than those reported in these previous studies. The RIA method was used in all these studies. The antisera used in these studies were not specific enough for FSH and failed to maintain selective neutralization between the gonadotropins (Licht et al., 1977). Licht and Wibbles have conducted measurements on FSH levels in females during the nesting period (20 to 50 hours after each turtle had nested). Our measurements encompass both the months preceding nesting and the period following nesting completion (every two months). It appears that FSH, as previously observed in reptiles, plays a role in elevating the levels of estradiol and testosterone and vitellogenin (Smelker et al., 2014; Bruno et al., 2021). This can explain the high concentrations of FSH observed months prior to nesting. In contrast, non-nesting females exhibited consistently lower FSH concentrations compared to the nesting females.
The ELISA developed in this study showed sensitive characteristics compared to other GTH immunoassays. High precision represented by low intra- assay and inter-assay coefficients of variation. The CV values are similar to those from immunoassays for measuring FSH reported in other species, with a range of 5–8% and 7–15% for intra-assay and inter-assay (Aizen et al., 2007b; Molés et al., 2012; Caron et al., 2018; Nocillado et al., 2019). These findings indicate that recombinant FSH was produced as expected, and that the antiserum correctly recognized the target FSH in the plasma (Kazeto et al., 2019).
The sensitivity of the ELISA was 0.13 ng/ml for measuring FSH concentrations in plasma which was similar or more sensitive in comparison to the sensitivities reported for measuring FSH in fish ELISA’s; 0.24 pg/ml for tilapia (Aizen et al., 2007b), 78 pg/ml for yellowtail kingfish (Nocillado et al., 2019) and 10 pg/ml for flatfish (Chauvigné et al., 2015). The differences in assay sensitivity may be affected by the purification level of the recombinant antigens and antisera (Nocillado et al., 2019).
To increase the specify limitations of the ELISA presented herein, one should consider developing a sandwich ELISA that will permit us to detect each glycoprotein hormone with 2 antibodies, one for the beta subunit and the other for the alpha specific epitope even though the immunogenicity of the alpha subunit is not very high.
Validation of the assay was demonstrated by parallelism between serial dilutions of standard curve and serial dilutions of displacement curves of green sea turtle FSH in plasma. These results indicate that the structure of the rcmFSH we produced is immunologically identical to the native FSH of green sea turtle (Aizen et al., 2007b; Nocillado et al., 2019). We also tested parallelism between serial dilutions of standard curve and serial dilutions of displacement curves for potential measurements of gonadotropins in other turtle species. Parallelism in tested plasma samples of three other species was observed: loggerhead sea turtle, hawksbill sea turtle and African softshell turtle. As seen in several studies on fish species, when parallelism is observed between the standard curve and several dilutions of the same sample it indicates similar protein structure between the different species (Aizen et al., 2007a; Aizen et al., 2007b; Aizen, 2011) Therefore, the structure of green sea turtle FSH is similar to the FSH of the other three turtle species. One should keep in mind that the developed ELISA assay presented herein cannot distinguish between free beta and beta associated with alpha. The two subunits of the gonadotropins are co-translated and they non-covalently combine in the endoplasmic reticulum of gonadotrope (FSH, LH), thyrotroph (TSH), or trophoblast (CG) cells (Cahoreau et al., 2015). Probably, if free FSHβ is present in the blood it can be a result of FSH dissociation.
This discovery will allow us the use of this assay for measuring FSH in these species and possibly more turtle species. In mammals, FSH in females controls the menstrual cycle and stimulates growth of follicles in the ovaries, whereas males FSH is an important hormone regulating spermatogenesis (Licht et al., 1979; Sairam and Krishnamurthy, 2001; Kluge et al., 2012). It is estimated that the mechanism is similar in most vertebrates, but the specific functions of FSH in reptiles are not fully understood (Tripathy and Rai, 2017; Kang et al., 2020). It was suggested that FSH has a distinct function in the stimulation of estradiol during the inter-nesting period that stimulates follicular development in females, and spermatogenesis in males (Wibbels et al., 1992; Mills et al., 2008; Kawazu, 2015; Tripathy and Rai, 2017).
After the development and calibration of the ELISA, we used it to measure concentrations of FSH from turtles in the Mikhmoret breeding facility using plasma samples taken since 2018. Our results show that successful nesting females that laid eggs (2019 and 2020) had an elevation in FSH concentrations during and before nesting season (January- August) compared to non-nesting females that presented relatively low FSH concentrations. FSH concentrations in non-nesting females were highest from April 2021 to September 2021. The females did not lay eggs in that year even though mating attempts were observed. We assumed that the females that haven’t nested yet didn’t reach sexual maturity and therefore the pituitary continues to secrete FSH at high concentrations in order to enter the puberty phase as in the case of other vertebrates.
High FSH concentrations increase the secretion of progesterone due to the ruptured follicle that forms a corpus luteum and produces high concentrations of progesterone (Wibbels et al., 1992; Al-Habsi et al., 2006). In female green sea turtles the high progesterone concentrations were reported slightly before nesting (Al-Habsi et al., 2006; Ikonomopoulou et al., 2014), associated with early stages of ovulation that follow the formation of new active corpora luteum (Licht et al., 1979; Licht, 1980; Al-Habsi et al., 2006) In the loggerhead turtle on the other hand high concentrations were also recorded in April prior to nesting similar to our results in the green sea turtle, but much higher concentrations were recorded when female was sampled 3 days after egg laying for 5 consecutive egg laying sessions with concentrations between 12-15ng/ml (Kakizoe et al., 2010). Additionally, increased progesterone concentrations stimulates albumin that is needed for the egg production (Owens and Morris, 1985). As shown in our results, in nested females, and relatively low in non-nested females.
Nested females showed elevated testosterone concentrations six months prior to mating (November) until the concentrations peaked three months prior to nesting in April, and gradually decreased to lower concentrations at the beginning of the nesting season (June). It has been suggested that a rise in testosterone concentrations can act as a breeding behavior initiator, and a possible trigger to male-attractive pheromones (Licht et al., 1979; Owens and Morris, 1985). Licht et al. (1979) conducted research with captive green turtles in the Cayman farm and reported that testosterone elevation was two months prior to mating until they peaked at the beginning of mating activity (mean >700 pg/ml) followed by testosterone level decrease during nesting activity. loggerhead turtle on the other hand showed testosterone concentrations elevation from February to May (Kakizoe et al., 2010) different from the green turtle.
Our results show a consistent rise in estradiol concentrations as FSH concentrations rises in both males and females. In females, we observed that estradiol concentration start to rise in November during follicular growth, declining towards July, concurring with observations from a study by Blanvillain et al. (2011) for green sea turtles and loggerhead turtles (Kakizoe et al., 2010). High concentration of estradiol were observed in April slightly before the females nested and concentration remain at the same concentrations until they decreased in September. The secretion of estradiol may serve both behavioral and physiological functions; it might increase the drive of migration to nesting beaches, regulate ovarian follicle maturation (Al-Habsi et al., 2006) and stimulate vitellogenesis (Kawazu et al., 2016).
In males, we observed a decrease in estradiol concentrations during August and September - consistent with the study by G. D. Blanvillain et al. (2011), in which the FSH concentrations dropped after the nesting season along with the drive to seek a mating partner. The gradual increase in testosterone concentrations in males during the several months before the onset of mating activity, August to November in our case, was expected as described earlier by Licht (Licht et al., 1979),
Overall FSH role in the green sea turtle was found to be correlated with the release of ovarian and testicular steroids. Further development of GTH ELISA for LH and TSH will contribute to a better understanding of the endocrine controls of the reproductive system in green sea turtle and other related species as well, clarifying the functional differentiation between these GTHs and their effects on other sex hormones. Future work should be directed towards the development and production of recombinant LH and TSH of the green sea turtle, this will help to better test and validate the assay specificity compared to other turtles’ glycoproteins hormones. This research has the potential to encourage the development of methods for increasing the fertility and reproduction of the turtles in the Mikhmoret breeding facility.