Methods

In 1998 and 2000, 2 adult male hawksbill turtles were captured in the set-nets of fishing boats operating at sea and were kept in captivity at the Ocean Expo Park, Motobu-tyo, Okinawa Prefecture, Japan, in an indoor holding tank under Okinawa Prefecture permits (based on the Fishery Act in Japan). The holding tank (5 × 5 × 1 m) was an open-water system and had 4 equal divisions, with the turtles being kept separately in 2 of the divisions. The water temperature of the tank was measured daily and ranged between 20°C and 30°C across a 12-mo period. This water temperature was approximately similar to that experienced by hawksbills under natural conditions, at least by those around Okinawa Island. The light condition of the tank was maintained at 12 hrs light∶12 hrs dark. The turtles were fed at 24-hr intervals with a diet that included fish and squid in quantities equivalent to 1.0%–2.0% of their body weight (BW). The straight carapace length (SCL) and BW of the 2 turtles were as follows: Turtle no. 1, SCL 79.5 cm and BW 54.5 kg; Turtle no. 2, SCL 83.0 cm and BW 71.0 kg.

Semen was collected at 1–2-mo intervals between January 2008 and March 2009 (Turtle no. 1: 9 attempts; Turtle no. 2: 10 attempts). The turtles were removed from the tank and kept stationary with a soft cloth belt on a retention stand (diameter: 500 mm; height: 500 mm), which was assembled from 4 car tires (Fig. 1A). The tail was allowed to hang unsupported beyond the edge of the stand. An electro-transformer for domestic animals (Lane Manufacturing Inc, Lane Pulsator IV-Auto Adjust™, USA) was used, which produces an electric stimulus ranging from 3.0 to 19.5 V (Fig. 1B). The electronic stimulus probe, which was assembled from an acrylic bar, was 700 mm in length and 8 mm in diameter and had 2 stainless steel electrodes (diameter: 1 mm; length: 20 mm) on either side of the tip, covered with acrylic resin, which was made by us (Fig. 1B, C). After measuring the distance from cloacal entry point to the urogenital papilla by endoscopy (Olympus, Japan), the probe, which was washed (lubricated) by normal saline, was inserted 250–350 mm into the cloaca (from which the penis also extends) by leading the probe dorsally around the urogenital papilla, which was then electrostimulated. Over a 5-min period, an electronic stimulus was given for 5 sec followed by a 5-sec rest (1 cycle). A total of 3 cycles (1 set) was performed using 3.0 V in the first cycle, 10.0 V in the second cycle, and 19.0 V in the third cycle, and 3 sets were performed in maximum. The urine of turtles was excreted from the cloaca during the early stages of electronic stimulus and was collected into a vial by suctioning it up with a pipette. When semen was observed on the probe or flowed along the dorsal sulcus to the tip of penis, it was counted as an ejaculation. At this point, we collected all ejaculated semen into a vial by suctioning it up with a pipette within 10 min. The defecation of 2 turtles was not observed during semen collection. After semen collection, the cloaca and penis were washed with normal saline. After a 30-min rest the turtles were returned to the holding tank. The health of both turtles was then assessed by veterinarians and aquarium staff of the Okinawa Churashima Foundation to check that no trauma had been incurred from the electronic stimulus. This study was performed based on the ethical guidelines for animal exhibition and research of JAZA (Japanese Association of Zoos and Aquariums) and was permitted by the board of directors of the Okinawa Churashima Foundation.

View Figure

• Download as Powerpoint
• Download Figure

After semen collection, semen volume (ml) was measured using a 15-ml high-clarity polypropylene conical tube (Becton, Dickinson and Company, USA). Sperm concentration (× 10⁶/ml) was measured using a hematocytometer, Thoma (Sunlead Glass Corp., Japan). Sperm viability (%) was calculated by eosin-nigrosin staining, which differentiates between live and dead sperm. The measurement of sperm concentration and staining has been described in previous studies (Bjorndahl et al. 2003). The pH of semen and urine were measured using a compact pH meter (Horiba Corp., Japan). The sperm motility (%) of normal semen and semen mixed with urine (mixed semen) was measured as previously described (Platz et al. 1978; Tsutsui 2002). The total number of surviving sperm (TNSS) was calculated for each ejaculation based on the results of semen volume, sperm concentration, and viability using the following equation:

Blood samples were collected from the 2 turtles immediately before semen collection. Ten milliliters of blood was sampled from the jugular vein on the left or right side of the neck using a 70-mm, 20-gauge needle (TERUMO Inc., Japan) and a 10-ml syringe (TERUMO Inc.). Blood samples were stored in heparin vacutainers after which plasma was collected by centrifugation (speed: 1500 × g; time: 20 min). The serum concentrations of testosterone were determined by modified double-antibody enzyme immunoassay (EIA) as previously described (Prakash et al. 1987). Serum samples were extracted with a diethylether before assay. The EIA used antibodies of sheep antitestosterone-3-CMO-BSA (GDN250, Colorado State University, Fort Collins, CO, USA; Taya et al. 1985) and horseradish peroxidase conjugated testosterone-3-CMO. Serial dilutions of serum sample from male captive hawksbill turtles resulted in a dose–response curve that was parallel to the standard curve generated with testosterone. The sensitivity of the assay was 2.3 pg per well.

Assumptions for normality were tested using the Shapiro-Wilk test. We used a Wilcoxon signed-ranks test to test for a significant difference in median sperm motility between normal and mixed semen. Statistical significance was assumed as p < 0.05.

Results

Ejaculate containing sperm was obtained from 4 of 9 attempts for Turtle no. 1 and from 10 of 10 attempts for Turtle no. 2. Turtle behavior during the 14 combined ejaculations included the curling of the rear flippers (Fig. 2A), the elongation and erection of the penis (Fig. 2B, C), the formation of the urethral fissure on the penis (Fig. 2D), and the anteroposterior shaking of the neck, which was composed of approximately 5 shakes/sec. In particular, the anteroposterior shaking of the neck was frequently observed just before ejaculation. However, when ejaculation was not observed (in the 5 attempts for Turtle no. 1), the elongation and erection of the penis and the anteroposterior shaking of the neck were not observed.

View Figure

• Download as Powerpoint
• Download Figure

All samples containing semen had a high viscosity. The median of semen volume in the 14 successful attempts was 0.5 ml (interquartile range: 0.2–1.5 ml), the sperm concentration was 325 × 10⁶/ml (100–645 × 10⁶/ml), the semen pH was 7.4 (7.1–7.7), the urine pH was 5.9 (5.9–6.2), and the TNSS was 180.5 × 10⁶ (12.6–750.5 × 10⁶) (Table 1). There was a significant difference in the median sperm motility between normal and mixed semen for Turtle no. 2 (from 2% to 54%; Wilcoxon signed-ranks test, p < 0.05; Fig. 3). Although small sample size prevented us from statistical analysis, a similar pattern was also observed in Turtle no. 1 (from 2.5% to 39.5%; Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Evaluation of semen collected by electro-ejaculation from 2 captive hawksbill turtles (n  =  14 semen samples combined). The total number of surviving sperm (TNSS) was calculated for each ejaculation based on the results of semen volume, sperm concentration, and viability.

View Fullscreen

There was a clear change in the mean water temperature in each month over a 12-mo period, with the temperature ranging from 20.5°C to 29.3°C. The highest temperature was recorded between July and October and the lowest temperatures recorded were between January and March (from winter to early spring, < 22°C, Fig. 4A). When we divided semen collection attempts between low (< 22°C) and high (> 22°C) water temperatures, the TNSS median for Turtle no. 1 was 1035 × 10⁶ (interquartile range: 733.5–1336.5 × 10⁶, n  =  2) and 20.9 × 10⁶ (11.8–30.0 × 10⁶, n  =  2), respectively (Fig. 4B). Correspondingly, the motility of mixed semen was 69.5% (59.3%–79.8%, n  =  2) and 23.5% (20.3%–26.8%, n  =  2) for low and high temperatures, respectively (Fig. 4C). The serum testosterone concentration was 49.0 ng/ml (43.3–53.4 ng/ml, n  =  5) and 11.5 ng/ml (10.2–33.1 ng/ml, n  =  4) for low and high temperatures, respectively (Fig. 4D). In contrast, the TNSS medians for Turtle no. 2 were 795 × 10⁶ (616.9–931.0 × 10⁶, n  =  5) and 3.8 × 10⁶ (1.9–39.0 × 10⁶, n  =  5) for low and high temperatures, respectively (Fig. 4B). In addition, the motility of mixed semen was 79.5% (79.5%–96.0%, n  =  5) and 31.5% (13.0%–41.5%, n  =  5) for low and high temperatures, respectively (Fig. 4C). The serum testosterone concentration was 75.3 ng/ml (73.7–83.7 ng/ml, n  =  5) and 31.5 ng/ml (22.5–31.7 ng/ml, n  =  5), for low and high temperatures, respectively (Fig. 4D).

View Figure

• Download as Powerpoint
• Download Figure

Discussion

This study presents novel information about the semen evaluation of captive male hawksbill sea turtles (Eretmochelys imbricata) based on an extended 15-mo study using the electro-ejaculation technique. We collected information that is useful for developing optimal semen collection techniques for future artificial insemination programs of this and other endangered sea turtle species.

Wood et al. (1982) first studied the use of electro-ejaculation to obtain semen from green sea turtles, followed by Tanasanti et al. (2009) for olive ridley and hawksbill turtles. The former study reported that ejaculate containing semen was obtained from 74.3% of the 74 attempts from green turtles (n  =  28), while the latter study reported a success rate of 75.0% of 8 attempts from olive ridley turtles (n  =  3) and 33.3% of 3 attempts from a single hawksbill turtle. Except for the single hawksbill turtle (Tanasanti et al. 2009), the current study obtained similar results to the previous observations, with a 73.7% success rate of 19 semen collection attempts from 2 individuals. Therefore, we consider this method to be a viable tool for semen collection from hawksbill turtles.

In the semen collection study of green turtles by Wood et al. (1982), the rear flippers were observed to curl in response to the electronic stimulus. This behavior was also observed in the current study (Fig. 3A). We suggest that the curling of the rear flippers is a diagnostic behavior of electro-ejaculation. Furthermore, we frequently observed the anteroposterior shaking of the neck just before ejaculation. Previous studies have not reported the behavior of turtles during ejaculation; however, the observed shaking of the neck in the current study was very similar to the behavior frequently observed in mating captive male hawksbill turtles (I. Kawazu, pers. comm., 2013). Therefore, we suggest that the anteroposterior shaking of the neck is a diagnostic behavior with ejaculation in male hawksbill turtles. During direct observations of mating, it is not possible to observe the moment of ejaculation due to the penis being inserted into the female cloaca. Hence, we suggest that the anteroposterior shaking of the neck presents a viable indicator for ejaculation during mating.

A previous study reported that semen pH was 5.5 for a single hawksbill turtle (Tanasanti et al. 2009). The semen pH for hawksbills in the present study was noticeably higher compared with the previously reported value (median: 7.4; Table 1). Previous studies that have collected semen from numerous animals have suggested that contamination with urine causes the lower semen pH values (Griggers et al. 2001; Miyake 2006). This hypothesis was supported by our results; the pH of the collected urine (median pH: 5.9) was lower compared with that of the semen and was similar to the previously reported values. The pH of semen from domestic mammals and fishes is 6.5–7.7 (Miyake 2006) and 7.5–8.5 (Alavi and Cosson 2005), respectively, which is approximately similar to that obtained in the current study (pH interquartile range: 7.1–7.7; Table 1). Therefore, we suggest that the semen samples collected in the current study were less (or not at all) contaminated with urine compared with the previous studies. This pH difference might also explain the variability in sperm motility between the current study and the previous studies. In the previous study of the single hawksbill (Tanasanti et al. 2009), sperm motility was 60%, whereas the median of sperm motility in the current study rose from 2.5% to 39.5% and from 2% to 50% for Turtles no. 1 and 2, respectively, when normal semen was mixed with urine (Fig. 3). This increase in motility in the presence of urine is extremely interesting, with this study being the first to report this phenomenon.

The sperm of marine and freshwater fish are mainly activated by osmolality that is created between the seminal fluid and surrounding medium (seawater and freshwater), which serve as osmotic and ionic signals (Cosson 2012). When sea turtles mate, the male first mounts the female, then the male curls its tail under the female carapace to come into contact with the female cloaca, and the erected penis is inserted into the female cloaca prior to ejaculation (Miller 1997). Thus, it is unlikely that marine water enters the cloaca of female sea turtles during mating, and the semen of male turtles is not surrounded by water at any point in this process as is observed in fish (Cosson 2012). Therefore, sperm motility after ejaculation might be induced by the presence of urine in the cloaca of the male or female and by the natural secretions in the female cloaca.

Sperm motility of several mammals tends to be compromised by highly anisosmotic solutions such as urine; however, pH is probably not associated with the deleterious effects of urine on semen quality because both substances have a similar pH (Santos et al. 2011). However, the pH value differed between semen and urine in the current study (Table 1). This difference might indicate that the sperm motility of hawksbill turtles is regulated by changes in pH. A similar case in which semen is activated under low pH conditions has been reported for sea urchin, Strongylocentrotus purpuratus (Christen et al. 1982). However, optimum sperm motility in rainbow trout Oncorhynchus mykiss occurs at pH 9.0 (Alavi and Cosson 2005), which is the opposite of our results in which the sperm of hawksbills was activated under low pH conditions (Fig. 3). Gist et al. (2000) reported that the sperm motility of painted turtles, Chrysemys picta, had little effect on spermatozoa at a pH range of 5.9–8.4. The semen collected in the current study was highly viscous, which would be alleviated when mixed with urine. Thus, sperm might also be activated through the relaxation of semen viscosity. Further study on the factors that activate sperm, as conducted on fish (Cosson 2012), would help to clarify the presence of such relationships. In addition, sperm morphology (e.g., the mitochondrial function; Zee et al. 2007) and progressive motility (e.g., swimming speed), which was not examined in the current study, require research to enhance artificial insemination practices. The assimilation of such information on sperm parameters might contribute toward examining the potential of preserving semen over short and long-term (chilling and freezing) periods for the development of artificial insemination techniques similar to that conducted on domestic mammals (Foote 2002).

The serum or plasma testosterone levels of other sea turtles are known to change during the course of the year, termed a seasonal testosterone cycle, in which testosterone peaks during winter and early spring (Licht et al. 1985; Wibbels et al. 1990; Jessop et al. 2004; Rostal et al. 2005; Valente et al. 2011). On the basis of the histological observations of the testis, Wibbels et al. (1990) and Rostal et al. (2005) reported that spermatogenesis is induced by increased testosterone levels in loggerhead (Caretta caretta) and olive ridley turtles. In the current study, the testosterone concentration was high when the water temperature was low at 20°–22°C (between winter and early spring; Fig. 4A, D). During this period, we collected semen samples showing relatively high TNSS and sperm motility from the 2 turtles (Fig. 4B, C). In male sea turtles, spermatogenesis is completed prior to the mating period (Licht et al. 1985; Wibbels et al. 1990; Rostal et al. 1998), with good-quality semen (January–March, between winter and early spring) also being collected just before the mating season in the current study because captive hawksbill turtles mate during April–May (spring) (Kobayashi et al. 2006). Therefore, our results are consistent with the spermatogenic cycle of sea turtles recorded in previous studies that were based on histological and physiological analyses.

The sperm of male sea turtles is stored in the epididymis for 2–3 mo after spermatogenesis (Hamann et al. 2003). However, in the current study semen quantity and quality were low from April onward (Fig. 4B–C), which probably reflects the mating season (Kobayashi et al. 2006). In other words, this observation might be influenced by the frequency of semen collection. For instance, in the current study semen was collected at 1–2-mo intervals. This result indicates that the intervals between semen collections should be carefully selected; for instance, we made 1 to 2 attempts per year from spring onward in the current study. Further study is needed to try to collect semen from nonejaculated hawksbills (at least over a 1-yr period) from spring onward to determine the optimum time for semen collection.

The collection of viable semen is fundamental for successful artificial insemination. The semen collection technique used here contributes significant information about the utility of artificial insemination for endangered sea turtles. In addition, this work provides useful reproductive information about sea turtles, including semen evaluation. For the development of an artificial insemination program for hawksbills, we must also consider certain problems with respect to females. Although ovulation in sea turtles is probably triggered by mating activity (Manire et al. 2008), Kawazu et al. (2014) demonstrated successful artificial ovulation induction by follicle-stimulating hormone (FSH) injection in female hawksbills. However, sea turtle sperm is stored in the sperm storage tubules of the oviduct after mating, at least until late nesting season (Owens 1980; Gist and Jones 1989). Based on this fact, we speculate that sperm should be serviced prior to ovulation induction by FSH injection for successful fertilization (Kawazu et al. 2014). However, we must consider the time required for semen to reach the sperm storage tubules of the oviduct. In conclusion, to further improve the success of artificial insemination in sea turtles, it is important to develop knowledge about the mechanisms of sperm storage and the pattern of sperm usage to fertilize successive clutches of eggs. The acquisition of such information could ultimately contribute toward conserving sea turtles and improving our understanding of male and female sea turtle reproductive processes.